ABSTRACT Title of Document: MODEL BASED SYSTEMS ENGINEERING APPROACH FOR COLLABORATIVE REQUIREMENTS IN COOLING WATER SYSTEM DESIGN Binyam Abeye, Master of Science, 2013 Directed By: Professor John Baras Department of Electrical Engineering and ISR Evaluation of the manufacturing process industry confirms that there is still manual exchange of product data between design and procurement engineers and equipment suppliers. Manual data exchange incurs human error, increases the cost, and takes more time. Also manual data exchange prevents designers from automatically evaluating a larger pool of suppliers and verifying supplier requirements. This thesis proposes to develop a collaborative requirements framework using a Model Based System Engineering approach to representing, communicating, and verifying requirements. Collaborative requirements entail that equipment data and process system requirements are shared in a common way to encourage automated of equipment tradeoff and requirement traceability. The collaborative requirement framework includes SysML to represent the multiple views of requirements, Multilevel Flow Model functional diagrams to depict the high level qualitative functionality, and lastly an optimization tool to verify requirements. Overall, this thesis shows the benefits of using the collaborative requirements framework automating data exchange between design engineers and equipment suppliers. MODEL-BASED SYSTEMS ENGINEERING APPROACH FOR COLLABORATIVE REQUIREMENTS IN COOLING WATER SYSTEM DESIGN By Binyam Girma Abeye Thesis submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Masters of Science 2014 Advisory Committee: Professor John Baras, Chair Professor Mark Austin Professor Linda Schmidt ? Copyright by Binyam Abeye 2014 ii Acknowledgements At the beginning of this all was my dear Professor Baras. I am thankful for risk he took in giving me this opportunity and his patience with me through the ups and down. Though the time was short, he has taught life lessons that greatly appreciate. Next I would like to extend a sincere word of gratitude to the very knowledgeable researchers I work alongside with at the National Institute of Standards and Technology (NIST). Mark Palmer was a tremendous help in getting me up to speed with my project work, both on this thesis and NIST work. Additionally, the information he was able to attain through teleconferences and meetings with industry gave me the opportunity to even write this thesis. Without his persistence and reminders I don?t know what I would have done. Peter Denno was a highly great aide in guiding my technical research and the approaches I should take for my thesis. The many one-on-one discussions helped me gain clarity and direction with what I wanted to do for my research. Edward Barkmeyer was the one that brought everything all together. His wisdom was by far one of the most valuable aspects of working at NIST that I will take with. Over the time I have worked at NIST I have probably talked with him the most and he is never bashful to let me know the truth about anything. His candor was priceless and impacted me in many ways. Also I would like to acknowledge my system engineering professor, Professor Mark Austin. He has aided me through most of my systems engineering course and has taught me many lessons. His openness and ability to drop whatever he is doing to listen to my concerns was priceless. I don?t know if this thesis would be as it is without his guidance and multiple inquisitions about ?how is it going?? Also I would like to show iii great appreciation for the guidance and leniency shown to me by Professor Linda Schmidt. The independent study I worked with her has enlighten me about much of the topics that my thesis is involved with. I thank her for her continued support and willingness to sit on my committee. Last, but most definitely not least, I would like to thank my family and friends. My family has bear with me much and this very well may be their thesis, which is the reason I am dedicating this thesis to them. Their unconditional love was a strong foundation that would be impossible for me to fathom what I would do without it. My friends both inside and outside the office were great in keeping me sane and giving me advice in all aspects of life. I am blessed to have these people in my life that have supported me in a vast amount of ways. I thank you all from the bottom of my heart. Finally, this research was supported by the NIST research grant 70NANB11H148 under the Collaborative Requirements Engineering (CRE) project. iv TABLE OF CONTENTS Acknowledgements ..................................................................................................... ii TABLE OF CONTENTS .......................................................................................... iv List of Tables .............................................................................................................. vi List of Figures ............................................................................................................ vii Chapter 1: Introduction ............................................................................................. 1 Problem Statement .................................................................................................... 1 Current Trends .......................................................................................................... 2 Proposed Methodology ............................................................................................. 3 Thesis Overview ....................................................................................................... 7 Chapter 2: Prior Related Work ................................................................................. 8 Resource Description Framework (RDF) for Component Selection ........................ 8 Product Data Sheet Ontology.................................................................................. 10 Integrated Product and Process Design ................................................................... 11 Chapter 3: Closed Loop, Heat Transfer, Liquid Circulating System (CHL)...... 12 Introduction ............................................................................................................. 12 CHL Description ..................................................................................................... 13 CHL Requirements ................................................................................................. 17 Chapter 4: Systems Modeling Language (SysML) for CHL ................................ 21 Introduction ............................................................................................................. 21 Use Case Diagrams ................................................................................................. 23 Requirement Diagrams ........................................................................................... 27 Activity Diagrams ................................................................................................... 31 Block Diagrams ...................................................................................................... 34 Internal Block Diagrams ......................................................................................... 39 Parametric Diagrams ............................................................................................... 40 Chapter 5: Functional Modeling with MFM .......................................................... 42 Introduction ............................................................................................................. 42 Implementation for Thesis ...................................................................................... 43 CHL MFM Model ................................................................................................... 44 Functional Reasoning.............................................................................................. 46 Chapter 6: Formulating the Optimization Problem .............................................. 47 Purpose .................................................................................................................... 47 Optimization Tool ................................................................................................... 48 Problem Formulation .............................................................................................. 50 Objective Function .................................................................................................. 51 Constraints .............................................................................................................. 51 Chapter 7: Optimization Results and Trade-off .................................................... 53 Analysis of High Impact Parameters ...................................................................... 53 Sensitivity Analysis with Pump Efficiencies .......................................................... 66 CHL Trade Off and Traceability............................................................................. 69 Negotiation aided by Optimization ......................................................................... 75 Chapter 8: Conclusion and Future Work ............................................................... 83 Conclusion .............................................................................................................. 83 Future Work ............................................................................................................ 83 v Appendices A: CPLEX Code ................................................................................... 84 Appendices B: Component Engineering Data ........................................................ 87 Appendices C: SysML Diagrams ............................................................................. 96 Appendices D: Tabular Requirements.................................................................. 103 Appendices E: CHL MFM Model ......................................................................... 109 References ................................................................................................................ 110 vi List of Tables Table 1 Instances of RFQ generated in Excel ............................................................. 38 Table 2 Parameters for Analysis ................................................................................. 53 Table 3 Flow rate Margin analysis (16 in) .................................................................. 54 Table 4 Flow rate margin analysis (18 in) .................................................................. 57 Table 5 Max Power analysis (18 inch) ....................................................................... 59 Table 6 Max Power analysis (16 in) ........................................................................... 60 Table 7 Pressure Margin Analysis (16 in) .................................................................. 62 Table 8 Pressure Margin Analysis (18 in) .................................................................. 64 Table 9 System Configuration Choices (16 and 18 inch) ........................................... 70 Table 10 CHL Negotiation Objectives ....................................................................... 76 Table 11 Reliability (Specific speed) Objective Results ............................................ 79 Table 12 Cost (Capacity Factor) Objective Results .................................................... 79 Table 13 Performance (Efficiency) Objective Results ............................................... 80 Table 14 Objective Values for 16 inch Connection .................................................... 80 Table 15 Objective Values for 18 inch Connection .................................................... 80 vii List of Figures Figure 1 MBSE Approach for Process Plant Design .................................................... 6 Figure 2 Connection Relation created by Inferences .................................................... 9 Figure 3 Compatibility Relation created by Inference .................................................. 9 Figure 4 IPPD Architecture ........................................................................................ 11 Figure 5 PFD of CHL System ..................................................................................... 16 Figure 6 System Power and Heat Load Requirements from Mitsubishi .................... 18 Figure 7 Component Requirements From AP1000 DCD ........................................... 18 Figure 8 Design Basis Requirements .......................................................................... 20 Figure 9 SysML Diagrams .......................................................................................... 21 Figure 10 Pathways from Goals and Scenarios to Structure and Behavior of System 22 Figure 11 Development of System Specifications ...................................................... 22 Figure 12 CHL Automation Use Case Diagram ......................................................... 24 Figure 13 CHL Service Use Case Diagram ................................................................ 24 Figure 14 CHL System Requirements ........................................................................ 28 Figure 15 Surge Tank Requirements .......................................................................... 29 Figure 16 Control Valve Requirements ...................................................................... 30 Figure 17 Activity Diagram of Heat Transfer Process ............................................... 32 Figure 18 Activity Diagram with actions Allocated to CHL Structure ...................... 33 Figure 19 Product Data Sheet Ontology UML Model ................................................ 35 Figure 20 Block Definition Diagram of CHL ............................................................. 36 Figure 21 RFQ data as Instances in BDD ................................................................... 37 Figure 22 Internal Block Diagram of CHL ................................................................. 40 Figure 23 Parameteric Diagram of Plate Heat Exchanger Constraint ........................ 41 Figure 24 MFM Functional Model Symbols .............................................................. 42 Figure 25 MFM Model Example: Water Mill ............................................................ 43 Figure 26 MFM diagram MagicDraw Implementation .............................................. 44 Figure 27 CHL Heat Transfer MFM Model ............................................................... 45 Figure 28 CHL MFM and SysML Relationships ....................................................... 46 Figure 29 CPLEX Input and Output Data................................................................... 49 Figure 30 Cost vs Flow rate Margin (16 in) ............................................................... 56 Figure 31 Cost vs Flow rate Margin (18 in) ............................................................... 58 Figure 32 Cost vs Max power (18 in) ......................................................................... 60 Figure 33 Cost vs Max power (16 in) ......................................................................... 61 Figure 34 Cost vs Pressure Margin (16 in) ................................................................. 63 Figure 35 Cost vs Pressure Margin (18 in) ................................................................. 65 Figure 36 Pressure Margin Sensitivity Analysis (16 in) ............................................. 66 Figure 37 Pressure Margin Sensitivity Analysis (18 in) ............................................. 67 Figure 38 Flow rate Sensitivity Analysis (18 in) ........................................................ 68 Figure 39 Flow rate Sensitivity Analysis (16 in) ........................................................ 69 Figure 40 Pressure Margin vs Pump Efficiency (16 inch) .......................................... 70 Figure 41 Pressure Margin vs Flow Margin (16 inch) ................................................ 71 Figure 42 Flow Margin vs Pump Efficiency (18 inch) ............................................... 72 Figure 43 Pressure Margin vs Pump Efficiency (18 inch) .......................................... 73 Figure 44 CPLEX Relaxation Suggestion .................................................................. 74 viii Figure 45 Sample Pump Characteristic Curve ............................................................ 77 Figure 46 Specific Speed vs Efficiency for 16 inch Connection ................................ 80 Figure 47 Capacity Factor vs Efficiency for 16 inch Connection .............................. 81 Figure 48 Specific Speed vs Capacity Factor for 16 inch Connection ....................... 81 Figure 49 Cost vs Efficiency Negotiation Limit ......................................................... 82 Figure 50 Specific Speed vs Efficiency Negotiation Limit ........................................ 82 1 Chapter 1: Introduction Problem Statement The aspiration for the future in manufacturing is automatic access of supplier data for manufacturing design engineers easily evaluate and determine the best suppliers for their system components. Additionally the designer?s manufacturing requirements will trace to the specific attributes of the supplier equipment in an easy automated way. Overall this automated process of building manufacturing systems will lead faster, cheaper, and with less probability of errors manufactured systems. Today the exchange of manufacturing equipment data and system requirements is a manual process where both the design engineers and equipment suppliers must manually input system requirements and equipment data into their own data management systems to evaluate information. This type of data exchange is costly not only in time and money of the design engineers and suppliers, but also in quality and performance of manufacturing systems, which affects all users of the manufacturing system. Therefore this thesis will propose a method for the representation, communication, and verification of requirements to aid the data exchange process between the design engineers and equipment suppliers. The method will include system engineering principles and optimization techniques. Specifically system engineering principles deal with integrating all the disciplines in the development process from the concept to operation and it considers both the business and technical needs of all customers and their goals [1]. 2 Current Trends In the manufacturing industry there is a big push for smart manufacturing. Smart manufacturing is the application of information technology into all aspects of the manufacturing process and products, which can fundamentally change how products are invented, manufactured, shipped and sold [2]. Introduced in the late 1990s, smart manufacturing is now reemerging as the solution to data management and enhancing manufacturing operations because of the new technological innovations with software management tools. Companies such as IBM [3] and Siemens [4] are using smart manufacturing principles in software to increase productivity and efficiency. One of the major software solutions for smart manufacturing is Product Lifecycle Management (PLM). PLM software evaluates the business processes that govern a product from the beginning to the final stages of a product?s life cycle to produce the best possible value for the business of the enterprise, customer, and other involved partners [5]. Some examples of successful use of PLM software (e.g. Siemens PLM NX) include the collaboration between NASA and JPL to design and simulate the latest Mars rover Curiosity [6]. Such cases show that PLM can be beneficial to the design of products, but there are also some caveats to their usage. First, PLM software conflicts with the processes set in place by manufacturing companies. Usually, one-off software solutions are created by manufacturing company engineers to support their version control, partner collaboration, change approval management, and other applications. With PLM all those custom functions become obsolete [7]. As a result, PLM limits the business and engineering 3 capabilities of the manufacturing company. Secondly, PLM struggles with dealing with domain-specific knowledge (information specifically important to the manufacturing company). Differing perspectives on the product domain lead to poor verification of data. As a consequence information flow is poorly linked between the design engineers and equipment suppliers. This problem is embodied by companies like Bis-sell Homecare, who have a tremendous amount of domain-specific knowledge and struggles to represent that information in PLM software. Instead companies like Bis-sell have resorted to knowledge-based engineering (techniques that capture decision-making knowledge and also offer a medium for exploiting efficient strategies used by experts [8]). Currently Bis-sell has expressed interest into system engineering techniques to strengthen their knowledge-based engineering [9]. Proposed Methodology This thesis shows how Model Based System Engineering (MBSE), functional modeling, and optimization tools can aid in traceability, communication, and verification analysis of system and component requirements. By using MBSE, functional modeling, and optimization tool requirements (both qualitative and quantitative) can be verified in a way that the current PLM systems are unable to do (mainly in the information flow and tracing of that flow). Additionally this method will allow for requirement and equipment data exchange between different suppliers and customers in the business enterprise by the use of data models (represented using MBSE). As a result, product data and their associated constraints are communicated automatically between multiple participants, spanning across the lifecycle of a project and allowing for better reasoning on requirements. 4 The first part of the framework is the system models, created by MBSE principles. MBSE is the formalized application of modeling to support system requirements, design, analysis, verification and validation activities beginning in the conceptual design phase and continuing throughout development and later life cycle phases [10]. This modeling formalism is used because it allows for the representation of system structure and behavior, as well as allow for the representation of textual and quantitative requirements in an integrated manner. As a result, MBSE allows for requirement management, ensuring the organization of requirements documents. Specifically within requirement management MBSE allows for tracing, prioritizing, change management, and communicating requirements. The MBSE language used is Systems Modeling language (SysML) because it is an industry-standard, providing good visual modeling to support system engineering [10]. Functional modeling is used because of its ability to represent a products or subsystem?s overall function with respect to a formal function representation [11]. This allows for a higher abstraction for representing how functions are related. One type of functional modeling language is Multilevel Flow Modeling (MFM). MFM was designed for industrial process functional modeling and allows for the representation of how functions satisfy high level requirements (labeled as goals within MFM). Therefore, MFM is highly useful because of its ability to represent qualitative requirements and how they relate to requirements in a formal way (that fosters to reasoning). This thesis will focus on using MFM to perform functional modeling. 5 Lastly, an optimization tool is used because of its ability to verify requirements and determine the best system designs. Along with verification, such tools also allow for greater understanding as to how requirements are effect certain low level behavior and structure. These attributes are highly desirable in this framework because they quantify the impact of requirements and how they relate to all parts of the system. Also, this functionality allows for deeper understanding into how the system can be improved by altering equipment specifications (low level structure), which enable negotiation. The optimization tool used in this thesis is IBM ILOG CPLEX Optimization Solution because of its strong mathematical programming solver, which is capable of high order mixed integer programming. Figure 1 shows the steps this thesis will follow to trace from system requirements to conceptual design of a water cooling system. The process begins by collecting requirements for the water cooling system from various design and procurement engineers. Then the equipment specifications, process specifications (qualitative requirements), and operational specifications are derived. Finally the equipment requirements are represented in SysML, process requirements are represented in MFM diagrams, and the operational requirements are represented in the optimization tool. Once modeled the requirements from each part of the framework are linked with respect to their shared requirements. This thesis will apply this step by step approach for a water cooling system. 6 Figure 1 MBSE Approach for Process Plant Design 7 Thesis Overview This thesis will demonstrate the collaborative requirements framework on a small process plant subsystem known as the Closed Loop, Heat Transfer, Liquid Circulating (CHL) system. Specifically, this framework will examine the process of representing, communicating, and verifying requirement during the final design and procurement phases of the CHL system lifecycle. In Chapter 2, prior related research is compared to the concepts in the thesis. Chapter 3 describes the CHL system requirements (equipment, process, and operation) and the relationship of requirements. Chapter 4 summarizes SysML and how the CHL system was modeled in SysML. Chapter 5 introduces functional modeling with the MFM language and the software implementation to support the language. Chapter 6 defines how the optimization problem is formulated with respect to the operational requirements (represented as constraints) using CPLEX. Chapter 7 describes the results of using this framework for collaborative requirements. The results include the optimization results and the methods used for integrating the models. Discussion, evaluation and conclusion are in Chapter 8. 8 Chapter 2: Prior Related Work Resource Description Framework (RDF) for Component Selection RDF is a model for data exchange on the Web, but can be extended to show directed and labeled graph models. At the core of the models are triples, which are the linking structure of RDF. Triples represent the relation between two entities as ?? where the ?Subject? and ?Object? represent the entities and the ?Predicate? represents the relation [12]. The two entities represent nodes in the graph and the relation is the edge between the entities. Previous work focused on RDF- based component selection. The project used RDF because it allowed for automated component and system requirement checking. Using RDF triples, plant equipment (pumps, heat exchanger, valves, and surge tank) were related to their attributes (pressure, flow rate, cost, etc.). This type of triple represented the product model for equipment. Next, triples were generated using inferences, which were based on component interface requirements. Inferences would check whether two equipment could be connected (e.g. If node is a pump and another node is a valve the inference generates a connection relation between the two nodes) and compatible (whether they could operate together based on engineering specifications). Figure 2 and Figure 3 show the results from the inferences. Lastly, a tradeoff analysis was conducted to determine the best configurations based on cost, reliability, and functionality. For the thesis, the inference requirements of this work were used in the development of requirements used in the thesis. Also the idea of RDF was tested for requirement checking. Still RDF for the system component selection is still limited to evaluating component to component requirements and not system to component 9 requirements (e.g. the power required for the pump based on all the components selected in the system). For this reason, RDF will only be explored for simple requirement checking. Also, the graphs would grow exponentially large if the attributes and component connections were managed in this way, making this method difficult to scale up. Figure 2 Connection Relation created by Inferences Figure 3 Compatibility Relation created by Inference 10 Product Data Sheet Ontology Work conducted at the National Institute of Standards and Technology (NIST) focused on developing a Product Data Sheet Ontology (PDSO) for collaborative requirements. The reason for a PDSO was to push for automated data exchange. Currently, data in product data sheets are not computer interpretable, which prevents automated exchange. Ontologies provide meaning to the data sheet elements so that a computer can interpret and use the data for exchange. In order to develop a good ontology, a common dictionary of terms must be shared among all users of the ontology. Therefore, the PDSO mapped common data sheet terminology to standard- based terminology (ISO15926 part 4) and definitions. This ensured a common definition of data sheet terminology. PDSO ontologies were generated from the Unified Modeling Language (UML) models of a general data sheet and three common process components (centrifugal pump, valve, and pressure transmitter). This research uses the concept of modeling component data in a similar way to map terminology to standards, but modeling is in SysML. 11 Integrated Product and Process Design Another motivation for using MBSE for collaborative requirements was the University of Maryland project on Integrated product and process design (IPPD). The IPPD is a decision making tool that aides the process for selecting components for the construction of a microwave modules. The tool optimized the component selection by reducing the cost, improving quality, and gaining leverage in time to market the product. To optimize the component selection, the tool used a multi-objective optimization model that selected the components and processes for a conceptual design that were Pareto optimal according to the previous metrics described. Overall, the tool improved the coordination and communication of requirements between the process design and product design by using a common interface [13]. Similar to the IPPD tool, this thesis aims to use a common interface (SysML) to coordinate and communicate requirements between the engineering design and supplier specifications. The thesis also used aspects of the IPPD architecture (in Figure 4) as guidance for incorporating the optimization. Figure 4 IPPD Architecture 12 Chapter 3: Closed Loop, Heat Transfer, Liquid Circulating System (CHL) Introduction The CHL system is a class of process cooling water system that focuses on temperature reduction of process fluid. The CHL system was develop through the Collaborative Requirements Engineering (CRE) project at the National Institute of Standards and Technology (NIST) [14]. The project involved working closely with representatives of the power and chemical process industries to identify a type of system common to many types of facilities and plants. The fruit of those discussions with industry was the CHL System. This thesis will use the CHL system because it is of the information provided by the project and the collaboration with industry. This collaboration from different industries permitted the comparing of multiple forms of information representation and determining the management challenges in requirements engineering. 13 CHL Description A process flow diagram (PFD) shows the interconnection of components in the closed loop, heat transfer, and liquid circulating system (CHL) and the main equipment that will be focused on for this thesis (see Figure 5). As well as the piping, the main system component at will be examined are the surge tank (pressure vessel), centrifugal pump, control valve, and plate heat exchanger. The goal of the CHL system is to remove heat from certain process fluids at a specific mass flowrate and heat load with recirculated cooling water within a closed- loop system. This goal is achieved by the centrifugal pump and plate heat exchanger. At start, the system is fully filled with water and a pump forces the flow of water by increasing the pressure of the fluid at the pump outlet. This pressure difference across the pump causes the water to flow through the pipes at a certain flow rate that is maintained throughout the system. The specific flowrate for the system is constant to allow for stable operation of the plate heat exchanger and other equipment. The plate heat exchanger inputs the cooling fluid at a certain temperature and flow rate to reduce the temperature of the process fluid that is also entering the heat exchanger. Entering through different ports and flowing through different chambers, the cooling fluid and process fluid exchange heat through the thin metal plates inside the plate heat exchanger. Afterward the cooling water exits the heat exchanger to be feed back to the inlet of the centrifugal pump and the process fluid is output to an external process system. 14 In addition to the centrifugal pump and plate heat exchanger, safety equipment is also used to support the main function. Safety equipment helps control and handles deviations in system pressure and temperature. Safety equipment include the surge tank, control valve, instruments, check valves, gate valves, and flow and temperature elements. This thesis will only focus on the surge tank and control valve in terms of safety equipment. The surge tank provides the necessary pressure of the inlet of the centrifugal pump and also aids in temperature fluctuations in the system by changing the cooling fluid volume. The water level in the tank determines the outlet pressure of the tank. Therefore, changes in the water level result in changes to the outlet pressure. The outlet pressure serves the centrifugal pump operation. The centrifugal pump needs a certain inlet pressure to operate safely. In addition, the surge tank serves the system operation. When the system pressure surpasses certain limits of a level the surge tank will intake more cooling water, resulting in the water level in the tank increasing to accommodate for the system?s over-pressurization. Similarly, when the fluid temperature in the feedback is too high the surge tank will intake the fluid, resulting in a water level rise. The reason this happens is because the temperature raises the pressure of the fluid. Control valves are also included in the CHL system. The control valve maintains the flow rate of the cooling water in the system. In the CHL system they are located at the outlet of the plate heat exchanger and at the outlet of the refrigeration system .For this thesis we will only focus on control valves that proceed after the heat exchangers. They are used in situations when the cooling water flow 15 rate or pressure rises or fall outside normal operation levels. The control valve reacts by either shrinking or widening its aperture to stabilize the cooling water?s flow rate or pressure. Also the control valve is dependent upon instrumentation to react to system flow rate and pressure changes. Since instrumentation is not considered in this thesis, the main focus on the control valve will be on sizing it for the system minimum and maximum pressure and flow rate levels and not reaction time and other control aspects. Some of the parameters that would be focused on include the pressure drop and the maximum flow rate allowance. 16 Figure 5 PFD of CHL System 17 CHL Requirements The main sources of requirement information on the CHL system came from nuclear power industry, data sheet industry standards, and the chemical process plant industry. Each industry provided a different perspective on the CHL system and contributed their own requirements problems with respect to the representation, communication, and verification of requirements. From the nuclear power industry, the CHL system is closely related to the component cooling water systems (CCWS), a common non-safety subsystem in a nuclear plant. Several CCWS control and requirement documentation were used for developing requirements for the CHL system. These requirements on components provided the key metrics that CHL equipment designers would need from component suppliers. Additionally, the DCDs also provided system requirements that showed how system specifications changed with respect to different scenarios of the system. From a greater standpoint, this information provided insight into what specifications were most important for communication with suppliers. An example of system and component requirements is shown below in Figure 6 [16] and Figure 7 [17]. 18 Figure 6 System Power and Heat Load Requirements from Mitsubishi Figure 7 Component Requirements From AP1000 DCD 19 Industry data sheet standards also provided a variety of requirement specifications with respect to the standards domain. Specifically these requirements focused on CHL components. Of all the components, the centrifugal pump and heat exchanger were well represented in terms of standards. For the centrifugal pump ASME B73.1, ANSI/API 610, and ISO 15926 were incorporated to the component requirements. For the heat exchanger the ISO 15926 and private industry data sheets were used. For the control valve and surge tank the ISO 15926 and handbook data sheets were used. These requirements, as a whole, showed how the component requirements for the CHL were commonly represented for design and communication to suppliers. In terms of the system requirements, the chemical plant industry provided project documentation, which gave insight into main requirements needed for specific aspects of design. Additionally, process simulation tools, such as CHEMCAD and AFT Fathom, provided clarity into how component requirements were verified. Overall collection of these system requirements provided an understanding of what CHL system requirements are most important for verification. Another aspect that is important to the CHL system requirements is traceability. Most of the provided information involved specifications, irrespective of their development. Figure 8 shows the requirements taxonomy for the CHL system and how requirements for one component feed into the other components [17]. This is very important because it provides for traceability and requirement verification. These requirements will be reexamined in the modeling section to show how requirements are represented in this manner. 20 Figure 8 Design Basis Requirements 21 Chapter 4: Systems Modeling Language (SysML) for CHL Introduction To apply MBSE principles to the CHL system this thesis has proposed to use OMG Systems Modeling Language (SysML). SysML is the main language for implementing MBSE. It is a general-purpose graphical modeling language that supports the analysis, specification, design, verification, and validation of complex systems [18]. Figure 9 (below) represents the main diagrams supported by the SysML language [18]. The diagrams represent the behavior, requirements, or structure of a system. Primarily the models of most importance for the CHL are the activity, use case, block definition, internal block, parametric, and requirement diagrams for the CHL system. Figure 9 SysML Diagrams While it is a visual modeling language that provides a metamodel for semantics (rules governing the creation and the structure of SysML models) and notation (representation of meaning, graphical or textual) it is not a methodology or 22 tool [19]. Since SysML is methodology independent, there is freedom to use the SysML language as fitting for the system in design. From coursework at the University of Maryland a set methodology is proposed that is shown in and [20]. These methods are used in developing the diagrams. Figure 10 Pathways from Goals and Scenarios to Structure and Behavior of System Figure 11 Development of System Specifications 23 Use Case Diagrams Use cases describe the functionality of a system in terms of how it is used to achieve the goals of its various users. They are also used to capture system requirements in terms of system uses. Use cases can be further elaborated with detailed descriptions of their behavior, using activities, interactions, or state machines [21]. Use case diagram visually show the relations between use cases and actors with respect to the system boundary. For the collaborative requirement framework use cases serve as a beneficial method to representing functional capabilities in a visual format. Additionally, this use case representation allows for building relationships between system behavior and requirements for the system (see requirement section for more). To show the benefits CHL use cases were developed. Using the functional descriptions from the nuclear power design control documents for a component cooling water system (CCW) two use case diagrams were developed for the CHL system (see Figure 12). This first use case diagram shows how the CHL system interacts with other mechanical systems for the purpose of automated operation. As shown there are three primary use cases, which include Monitor Flowrate, Monitor Process Fluid Heat Removal, and Monitor Surge Tank Fluid Level. These use cases depict the ways that the user will use the system, which the CHL system must accommodate for. The second use case diagram (see Figure 13) focuses on the interaction the process fluid, refrigeration system, and the CHL system. 24 Figure 12 CHL Automation Use Case Diagram Figure 13 CHL Service Use Case Diagram 25 To further elaborate on the use case diagrams, each use case can be described in detail through use case scenario descriptions. Elaborating on use cases is necessary for the collaborative requirements framework to show the fine details of a process plants behavior. Below is an example of a scenario for the ?Remove heat from Process Fluid? in the second use case (Figure 13). Use Case 1: Remove heat from Process Fluid ? Actors: Process Fluid System, Refrigeration System ? Preconditions: 1. CHL pump must be operating at steady state 2. All equipment is working error free ? Basic Flow of events: 1. The Refrigeration system decreases the temperature of the cooling fluid to 41 deg F. 2. Cooling fluid enters the heat exchanger at 6500 gpm and 41 deg F. 3. Process fluid enters the heat exchanger traveling at 3000 gpm flow rate and 90 deg F. 4. Heat gets transferred within the heat exchanger from the process fluid to the cooling fluid. 5. Cooling fluid exits the heat exchanger at 70 deg F and the process fluid exits the heat exchanger at 70 deg F. ? Alternative Flow 4: 26 4a. Process fluid exits the heat exchanger at undesirable temperature. 1. The cooling fluid flow rate is increased to increase heat transfer. a. Performed by increasing the power to the pump or opening the valve downstream to increase flowrate. 2. The cooling fluid temperature out of the Refrigeration is decreased to encourage more heat transfer. ? Post Condition: 1. Cooling fluid is feedback into the CHL system. 2. Process fluid is returned to the Process Fluid System. Overall use case diagrams and use case descriptions serve as a first step in defining the system behavior and developing behavioral requirements. Unfortunately there is no method for currently validating or reasoning on these use cases, which would benefit in the automated aspect of the collaborative requirement framework. This is the reason another functional modeling tool is also used along with the use case diagram (describe later in MFM section). Otherwise use cases still serve an important purpose in their relationship to requirements and requirement diagrams. 27 Requirement Diagrams Once finished collecting all the user requirements from the use cases, requirement diagrams can be developed to show how requirements are related. There are several requirement relationships that will be used for describing the CHL requirements. First relationship is containment. A containment relationship shows the decomposition of requirements, showing the high level requirement and all the sub requirements that are included within it. The second relationship used is the derived requirement relationship. This relationship shows how a general requirement can related with a more detailed requirement based on calculations or other forms of justification. The third relationship used is the verify relationship. The verify relationship connects a requirement with the method with which the requirement would be evaluated on the system. Most of the verify relationships used in the CHL requirements will connect requirements to constraint blocks (one way of verification). The last relationship used is the satisfy relationship. This relationship shows what block or component in the system the requirement will be associated with (what structural or behavior aspect of the system must ?satisfy? this requirement). In Figure 14 the requirements diagram of the CHL system is shown. The diagram shows how from one high level requirement there were many sub requirements that were contained within it (a containment requirement used). Requirements can also be viewed in a tabular format that is in Appendices D: Tabular Requirements. 28 Figure 14 CHL System Requirements 29 As stated earlier, by using the verify relationship requirements can be linked to the verification method used. SysML can represent verification methods such as inspection, analysis, demonstration, and test. For the CHL all the components have engineering equations associated with them, so analysis used as the verification method. One form of analysis is through constraints (bounded equations). Below in Figure 15 is an example of a Surge Tank requirement and its verification method (a constraint called ?SurgeTankSizing?). Figure 15 Surge Tank Requirements Lastly, requirements allow for referencing to the source where the requirement was taken from. For example, in Figure 16 the requirement titled ?ValveDifferentialPressure? is sourced from a software tool (AFT Fathom). This allows for requirements that were once separated to be joined together, without losing their original source. Sourcing can also be seen in ?ValveFlowrate? and ?ValveMassFlowrate? requirements. 30 Figure 16 Control Valve Requirements The remaining requirements for the CHL system are located at Appendices C: SysML Diagrams and Appendices D: Tabular Requirements. 31 Activity Diagrams The main diagram used to describe activity in a system is the activity diagram. These diagrams define the actions in the system that are required to achieve a certain functionality (determined through use cases) along with the flow of input/output and control between the actions [21]. Describing the CHL system in this manner allows for a strictly functional view of the system without any allocation to components or structure of the system. Since the CHL system is already provided (the structure of the system) this activity diagram is not used for design, but for requirement tracing, since requirements can be satisfied by both structure and behavior. In Figure 17 an activity diagram shows how the different actions feed into one (with object flows) and the sequence of actions that are taken (the control flow of the system). Also, activity diagram mirror functional block diagrams that are used in the Product Process Design, which gives credibility to using activity diagram to represent the CHL system?s behavior. 32 Figure 17 Activity Diagram of Heat Transfer Process 33 Another highly beneficial aspect of behavioral modeling with activity diagrams is the ability to allocate actions to the structure. During the design stage this allows for a better understanding of the requirements imposed on the structure. This allows for traceability from the requirements gather in the use cases to the activities that achieve the function of the use cases to the structure that embody the behavior. Below in Figure 18 shows the allocation of actions to component in the CHL system structure. Figure 18 Activity Diagram with actions Allocated to CHL Structure 34 Block Diagrams The way SysML models structure is through blocks, a modular unit of structure that can represent a system, component, item that flows through a system, conceptual entity, or other logical abstraction [21]. This flexibility allows the blocks to represent manufacturing component models. These models can represent what designers use to specify the component to satisfy the system functionality and also used to generate documents to send to vendors as RFQs. From the PDSO work, the distinction between the designed component, the product model, and actual component (physical component) is the way they are referenced (tag numbers, part number, and serial numbers), but they are required to be the same in terms of engineering parameters. Therefore a model that can relate design components to product models (from the vendor) and check for their alignment would build toward collaborative requirements. Below in Figure 19 shows the connection between these representations of the component and their attributes. 35 Figure 19 Product Data Sheet Ontology UML Model The system architecture for the CHL system is show in a block definition diagram (BDD), which shows all the models of the components in the CHL system. Each block contains the attributes associated with that component as well as the components constraints, operations, and associated requirements which it satisfies. Below in Figure 20 the BDD is shown. 36 Figure 20 Block Definition Diagram of CHL 37 Another interesting aspect of the blocks is that instances of them can be created. Like in java with classes and objects, blocks are the template for what is contained in a component model for design and RFQ, but the instances are the actual specification with values supplied for the attributes. This allows for RFQ information to entered into the instances of the components and sent out to multiple suppliers. Below in an example of RFQ information for the components and fluids in the CHL system are shown. The MagicDraw tool used for building the SysML models also allow for the generation of excel files, so the RFQ data can be exported to excel to allow for communication of requirements. shows the output from the exported excel. Figure 21 RFQ data as Instances in BDD 38 Table 1 Instances of RFQ generated in Excel # Name Slot 1 ProcurementInstances 2 cyclopentane boilingPoint = "121" density = "46.88" maxTemperature = "90" minTemperature = "7" prandtlNumber = 4.29 specificGravity = 0.74 specificHeat = "1.1217" thermalConductivity = 0.025 viscosity = "0.438" 3 phxRFQ inletTempCold = "41" inletTempHot = "90" outletTempCold = "70" outletTempHot = "70" connectionDiameter = "16" allowablePressureDrop = "70" heatLoad = "87017715.12" massFlowrateCold = "3003614.48" massFlowrateHot = "1128210.12" 4 pipeRFQ length = "200" maxHeadLoss = "500.0" nominalSize = "16" 5 pumpRFQ connectionDiameter = "16" designVolumetricFlowrate = "6500" differentialHead = "700" ratedEfficiency = 0.65 6 tankRFQ designHead = "200" designStress = "19580" nominalDiameter = "60" 7 valveRFQ designVolumetricFlowrate = "6500" maxDifferentialHead = "40" connectionDiameter = "16" 39 8 water boilingPoint = "100" Density = "62.36" prandtlNumber = 0.0 specificGravity = 1.0 specificHeat = "1" thermalConductivity = 0.3351 minTemperature = "41" maxTemperature = "70" viscosity = "0.89" Internal Block Diagrams In addition to a BDD there is also a internal block diagram (IBD) that shows how the component in the CHL system are connected together. This is similar to the process flow diagram (PFD) that was first shown to describe the system. Another industry diagram also specializes in describing the connection between component and enumerating the requirements for each component (and the system) on the diagram. This diagram is known as the Piping and Instrumentation Diagram (P&ID). Since the industry has a vast amount of knowledge in these diagrams (PFDs and P&ID) the IBD should be used for small scale examination of the flow between components. Figure 22 shows the IBD of the CHL system (excluding the piping and control valve). 40 Figure 22 Internal Block Diagram of CHL Parametric Diagrams Apart from just showing attributes and connection of components there is also the ability to show the constraints on the attributes of the components. Constraints are added to the models by the use of constraint blocks and the parameteric diagram. Parameteric diagrams allow for specialization of blocks (parts) to be constrained by constraint blocks. The constraint blocks consist of equations and parameters. The parameters of the constraint are associated with the attributes of the component associated with the constraint. This way the actual physical and behavioral constraints the component truly has can be modeled and tied directly to the block (through the part). Below in Figure 23 a parameteric diagram is shown for plate heat exchanger and it constraint on its heat load. The parameteric diagrams for the Centrifugal Pump, Valve, Pipe, and cost analysis are shown in Appendices C: SysML Diagrams. 41 Figure 23 Parameteric Diagram of Plate Heat Exchanger Constraint 42 Chapter 5: Functional Modeling with MFM Introduction Over several decades, researchers from the Technical University of Denmark have created a modeling language for industrial process plants. The purpose of the modeling language was to represent functional behavior of the industrial process with respect to the goals of the system by using means-end and whole-part relations. This functional modeling allows for qualitative reasoning, which reasons about knowledge of physical phenomena and systems that cannot be done by quantitative methods [22]. This capability makes MFM beneficial for communicating requirements that are quantitative. Therefore, this thesis will apply MFM to connect requirements that are qualitatively based. Currently MFM has no dedicated software implementation, so this thesis will develop a software implementation of the model. Below in Figure 24 is a legend of symbols to represent MFM models. Also in Figure 25 there is an example MFM diagram [22]. Figure 24 MFM Functional Model Symbols 43 Figure 25 MFM Model Example: Water Mill Implementation for Thesis The MFM language is developed as a UML profile within the MagicDraw software. The profile consists of the different function types, relations, function structures, and goals. By creating this profile a domain specific language (DSL) is created. Following the creation of the profile, customizations or rules were applied to the elements (connection rules between functions). Afterward a custom diagram was created for the MFM language, where MFM diagrams can be created via the MagicDraw software interface. Figure 26 shows an example of an MFM diagram created in the MagicDraw interface. 44 Figure 26 MFM diagram MagicDraw Implementation CHL MFM Model To demonstrate the value of MFM modeling, models of the CHL functionality were developed using the MFM language. As a result of developing the models a greater understanding of the means-to-end relationships were developed. These types of relationships guide in the requirement traceability, since there is an understanding on how functions are related. Above in Figure 27 a part of the MFM diagram for the CHL system (whole diagram available in Appendices E: CHL MFM Model) shows how there is a heat balance between the cold cooling fluid and hot process fluid. Another beneficial aspect of the MFM diagrams is that the model elements are linkable to the SysML components. Figure 28 (below) shows how the surge tank storage functionality relates to the structural representation of the surge tank in SysML. Also the requirements for the pump and valve are related to their functional 45 representations in the MFM model. This capability makes these MFM models highly useful and allow for traceability between the two models. Figure 27 CHL Heat Transfer MFM Model 46 Figure 28 CHL MFM and SysML Relationships Functional Reasoning Another benefit to using the MFM language is the ability to perform reasoning on the models. The main type of reasoning that is performed on the model is cause- effect reasoning. For the MFM model, the focus of this reasoning is on the goal- function and function-function patterns [22]. Therefore this reason is ideal for determining how changes in one component requirements or their functionality will affect all the other system functionality downstream and the overall goal of the system. From that perspective MFM models aid in both the representation and verification of system functionality and requirements in the collaborative requirements framework. 47 Chapter 6: Formulating the Optimization Problem Purpose Within the process plant industry tremendous amount of work has been conducted in designing the best process, most suitable plant structure, and optimal parameters of the process. The only area that has not had much attention involves the best design of the plant equipment [23]. This area is difficult to address because it is highly interconnected with the other areas of design. For that reason, the optimization of this thesis will focus on the optimization on the plant equipment design based on the requirements from the process that focus on normal operation and the equipment requirements. In addition to design, optimization can also aid in the understanding of requirements and how they affect the component selection process. This aspect is extremely important when trying to negotiate requirements between equipment designers/procurement engineers and the equipment/pipe suppliers. Therefore this thesis will also use optimization to determine the best group of equipment from a list of suppliers that will satisfy the individual equipment requirements as well as the process requirements. This is beneficial because it allows the equipment designers/procurement engineers to grasp what needs to be changed in the suppliers? equipment specifications to achieve their process requirements. Also, the requirements can be traced to the equipment specifications that have the most impact. All of these concepts aid the equipment designers/procurement ability to negotiate with the equipment suppliers and have more insight into how much more optimal the 48 process plant can be with the available suppliers. In order to perform the design and selection optimization a software package was used name IBM ILOG CPLEX. Optimization Tool The optimization software package used for the project was IBM ILOG CPLEX Optimization studio. The CPLEX optimizer can solve integer programs, very large linear programming problems using either primal or dual variants of the simplex method, quadratic programs, and convex quadratically constrained problems. This thesis uses a powerful mathematical programming engine (CPLEX engine) and a constraint programming engine (CP engine). The CPLEX engine can solve linear, mixed-integer, quadratic, and quadratically constrained programs. The CP engine can solve models with complex combinatorial constraints and uses powerful constraint- propagation and branch-and-bound techniques. Additionally, the CPLEX Optimization Studio provides the Optimization Programming Language (OPL) for modeling the constraint problem. Two of the main features used from OPL were the interface to Excel that allowed for the input and output of data to Excel spreadsheets and the easy to use OPL script that can run optimization programs multiple times with changing constraint bounds and store results in text files for later analysis [24]. Below is a model of how CPLEX was generally used in the thesis work (see Figure 29). 49 Figure 29 CPLEX Input and Output Data 50 Problem Formulation In order to evaluate the components selected we have to use a constraint programming language. Each component has their own physical and functional constraints, but the main benefit of constraint programming is the ability to evaluate component selections with respect to system constraints that depend on the connection of all component (or interaction between groups of components). System constraints mirror the high level requirements. Therefore, these high level requirements can be immediately validated in the component selection using the constraint programming. Additionally, this shows if there is a viable collection of components (from supplier data) that can satisfy the system requirements. Otherwise, if there isn?t a group of components that satisfy the system constraints a recommendation can be provided as to what needs to change in order to get viable system. Recommendations can range from changing one parameter of one supplier data to changing multiple parameters for multiple components. These recommendations may also give more information as to whether or not the system constraints should be loosened or can be made stricter. To begin formulating the component selection process with constraint processing several aspects must be made clear. First the objective function (the goal constraint programming is to satisfy the maximization, minimization, or equality of a specific equation related to the constraint variables) must be determined. For this component selection problem the objective is to minimize total cost (the sum of the cost for each component). Second, component constraints that only involve a single component must be defined. Third, the constraint that involves multiple components 51 must be described. Lastly, the system constraints must be represented in terms of some set of the components. Each of these steps involves understanding the characteristic functionality of the component. Once the functionality of each component is determined, then their interaction (how their functionality serve other components and aid in their functionality) can be described in constraints. The combination of these interactions then yields a system functionality that can be controlled with system constraints. Objective Function Minimize Cost: i=type of components j=number of vendors x is Boolean to determine whether the component which component from 20 vendors is selected C is a cost matrix that has the cost for each of the individual components Constraints Subject to: 20 1 (i) 1j j x = =? Ensures that there is only one component picked out of the 20 vendors, i=the component type 20 1 (i)(i,y) ( )D jj j x SV y= ?? D = Component engineering data from vendor i= the component type y= engineering parameter index (number of parameters vary for each component vary) SV= a vector system variable that constrain the component selection ( 5 20 1 1 min *ij ij i j C x = = ?? 52 These types of constraints include the max flow rate for the pump and max power constraints on the pump. 20 1 (i)(i,y) ( )D jj j x SV y= ?? (Same as above) These types of constraints include surge tank supply head minimum, heat exchanger and pump efficiency. 20 1 (i)(i,y) ( )D jj j x SV y= =? (Same as above) These types of constraints include the pipe, heat exchanger, valve, and pump connection constraint (has to all be the same size). 20 20 1 2 1 1 ( ) ( )1 2( ,y) ( ,y) ( )D Dj jj jj jx x SV yi ii i= =+ ?? ? These constraint include more than one component (component interaction) and the SV represents a system variable margin. Constraints on the pipe, valve, and hx with respect to the allowable pump flow rate margin would fit this constraint, as well as the pump supply head margin equation (involve all components). Also the pumps supply pressure from the surge tank constraint is modeled in the same fashion. 53 Chapter 7: Optimization Results and Trade-off Analysis of High Impact Parameters After executing the optimizer a list of high impact system and component parameters were determined. The reason they are high impact is because they interconnect each component or greatly affect the functionality of another component or the overall system. The high impact system and component parameters are listed below in Table 2. Table 2 Parameters for Analysis High Impact System/Component Parameter Components Involved Pressure Margin Centrifugal Pump, Plate Heat Exchanger, Control Valve, Surge Tank, and Pipe Flow rate Margin Centrifugal Pump, Plate Heat Exchanger, Control Valve, and Pipe Centrifugal Pump Power Centrifugal Pump Using these high impact parameters, the design options were evaluated. This results in a range of viable system options that had to be evaluated. This range of options allow for a tradeoff of system components and component arrangements to take place. In addition to cost, component efficiencies, power, and system volumetric flow rate can be evaluated. The flow rate margin, pressure margin, and power were all compared with respect to the objective (to minimize cost). 54 The constraint on volumetric flow rate started with the system requirement that the flow rate must be at least up to the design parameter (6500 gpm). From there the constraint was applied on the other equipment. The valve, heat exchanger, and piping all have max flow rate tolerances that need to be consider. So in order to select a pump, there must first be a check over the design space to find if there is a heat exchanger, valve, and pipe that can withstand that specific flow rate. Margin is added to this selection process to show how close the pumps provided flow rate is to the other equipment?s rated flow rate. Ideally the margin should be minimized to only compensate for variations in the system operation, such as switches in operation mode or to allow for extra time to react to system safety problems (such as a leak or broken equipment). The optimal selection of components for their specific flow rate margin is shown below in Table 3, Table 4, Figure 30, and Figure 30. Table 3 Flow rate Margin analysis (16 in) 16 inch connection Hx Pu V St Pi Obj cMarg 1 1 1 1 3 794000 1000 1 2 1 1 3 797185 990 1 2 1 1 3 797185 980 1 2 1 1 3 797185 970 1 2 1 1 3 797185 960 1 2 1 1 3 797185 950 1 2 1 1 3 797185 940 1 2 1 1 3 797185 930 1 2 1 1 3 797185 920 1 2 1 1 3 797185 910 1 2 1 1 3 797185 900 1 3 1 1 3 799308 890 1 3 1 1 3 799308 880 1 3 1 1 3 799308 870 1 3 1 1 3 799308 860 1 3 1 1 3 799308 850 55 1 3 1 1 3 799308 840 1 3 1 1 3 799308 830 1 3 1 1 3 799308 820 1 3 1 1 3 799308 810 1 3 1 1 3 799308 800 2 4 1 1 3 805748 790 2 4 1 1 3 805748 780 2 4 1 1 3 805748 770 2 4 1 1 3 805748 760 2 4 1 1 3 805748 750 3 4 1 1 3 810772 740 3 4 1 1 3 810772 730 3 4 1 1 3 810772 720 3 4 1 1 3 810772 710 3 4 1 1 3 810772 700 6 5 1 1 3 827084 690 6 5 1 1 3 827084 680 6 5 1 1 3 827084 670 6 5 1 1 3 827084 660 6 5 1 1 3 827084 650 7 5 1 1 3 832108 640 7 5 1 1 3 832108 630 7 5 1 1 3 832108 620 7 5 1 1 3 832108 610 7 5 1 1 3 832108 600 10 6 1 1 3 848243 590 10 6 1 1 3 848243 580 10 6 1 1 3 848243 570 10 6 1 1 3 848243 560 10 6 1 1 3 848243 550 56 Figure 30 Cost vs Flow rate Margin (16 in) Results for the component with 16 inch connections show that as the volumetric flow rate margin increases, the cost of the system decreases. This is not a surprise, since over sizing the components allows for the selecting of cheaper components. The flow rate margin decreases all the way to 550 gpm. It is also interesting to point out that the valve, surge tank, and pipe remain constant for all the system configurations. 400 500 600 700 800 900 1000 1100 780000 800000 820000 840000 860000 Flo w ra te M ar gin (g pm ) Cost ($) Cost vs Flow rate Margin (16 in connection) hx1,pu1,v1,st1,pi3 hx1,pu2,v1,st1,pi3 hx1,pu3,v1,st1,pi3 hx2,pu4,v1,st1,pi3 hx3,pu4,v1,st1,pi3 hx6,pu5,v1,st1,pi3 hx7,pu5,v1,st1,pi3 hx10,pu6,v1,st1,pi3 57 Table 4 Flow rate margin analysis (18 in) 18 inch connection Hx Pu V St Pi Obj cMarg 11 11 11 1 8 874995 1000 12 11 11 1 8 880020 990 12 11 11 1 8 880020 980 12 11 11 1 8 880020 970 12 11 11 1 8 880020 960 12 11 11 1 8 880020 950 13 11 11 1 8 885044 940 13 11 11 1 8 885044 930 13 11 11 1 8 885044 920 13 11 11 1 8 885044 910 13 11 11 1 8 885044 900 14 11 11 1 8 890068 890 14 11 11 1 8 890068 880 14 11 11 1 8 890068 870 14 11 11 1 8 890068 860 14 11 11 1 8 890068 850 15 11 11 1 8 895092 840 15 11 11 1 8 895092 830 15 11 11 1 8 895092 820 15 11 11 1 8 895092 810 15 11 11 1 8 895092 800 16 11 11 1 8 900117 790 16 11 11 1 8 900117 780 16 11 11 1 8 900117 770 16 11 11 1 8 900117 760 16 11 11 1 8 900117 750 17 11 11 1 8 905141 740 17 11 11 1 8 905141 730 17 11 11 1 8 905141 720 17 11 11 1 8 905141 710 17 11 11 1 8 905141 700 18 11 11 1 8 910165 690 18 11 11 1 8 910165 680 18 11 11 1 8 910165 670 18 11 11 1 8 910165 660 18 11 11 1 8 910165 650 19 11 11 1 8 915190 640 19 11 11 1 8 915190 630 19 11 11 1 8 915190 620 58 19 11 11 1 8 915190 610 19 11 11 1 8 915190 600 20 11 11 1 8 920214 590 20 11 11 1 8 920214 580 20 11 11 1 8 920214 570 20 11 11 1 8 920214 560 20 11 11 1 8 920214 550 Figure 31 Cost vs Flow rate Margin (18 in) The 18 inch connection system also depicts this trend. One distinction between the two connection sizes is that the 18 inch system cost more than the 16 inch system (as expected since there is more material used in the pipe). Otherwise, the flow rate margin also goes as low as 550 gpm. Lastly, the only component in these component selection is the heat exchangers (the pump, valve, surge tank, and pipe remain constant). 400 500 600 700 800 900 1000 1100 860000 880000 900000 920000 940000 Flo w ra te M ar gin (G PM ) Cost (S) Cost vs Flow rate Margin (18 in connection) h11,pu11,v11,st1,pi8 h12,pu11,v11,st1,pi8 h13,pu11,v11,st1,pi8 h14,pu11,v11,st1,pi8 h15,pu11,v11,st1,pi8 h16,pu11,v11,st1,pi8 h17,pu11,v11,st1,pi8 h18,pu11,v11,st1,pi8 h19,pu11,v11,st1,pi8 h20,pu11,v11,st1,pi8 59 The power required by the centrifugal pump directly affects its flow rate capacity and amount of pressure it can overcome in the system. Since power is a limited resource, it is best to reduce its usage while also examining the affect it will have on the cost of the overall system. Table 5 Max Power analysis (18 inch) 18 in Connection Hx Pu V St Pi Obj mPow 11 11 11 1 8 874995 2000 11 11 11 1 8 874995 1990 11 11 11 1 8 874995 1980 11 11 11 1 8 874995 1970 11 11 11 1 8 874995 1960 11 11 11 1 8 874995 1950 11 11 11 1 8 874995 1940 11 11 11 1 8 874995 1930 11 11 11 1 8 874995 1920 11 11 11 1 8 874995 1910 11 11 11 1 8 874995 1900 11 11 11 1 8 874995 1890 11 11 11 1 8 874995 1880 11 11 11 1 8 874995 1870 13 12 11 1 8 886293 1860 60 Figure 32 Cost vs Max power (18 in) The results for the 18 inch connections result in a surprising discovery. One system configuration tends to dominate in terms of the lowest cost yet still meeting the power constraint. One other observation is the additional cost that would be added if the system had to be less than or equal to 1860 HP. Table 6 Max Power analysis (16 in) 16 inch connection Hx Pu V St Pi Obj mPow 1 1 1 1 3 794000 2000 1 1 1 1 3 794000 1990 1 1 1 1 3 794000 1980 1 2 1 1 3 797185 1970 1 2 1 1 3 797185 1960 1 2 1 1 3 797185 1950 1 2 1 1 3 797185 1940 1 3 1 1 3 799308 1930 1 4 1 1 3 800724 1920 1 5 1 1 3 801962 1910 3 7 1 1 3 813958 1900 5 8 1 1 3 824891 1890 7 9 1 1 3 835647 1880 1840 1860 1880 1900 1920 1940 1960 1980 2000 2020 870000 875000 880000 885000 890000 M ax P ow er (H P) Cost ($) Cost vs Max Power (18 in connection) h11,pu11,v11,st1,pi8 h13,pu12,v11,st1,pi8 61 Figure 33 Cost vs Max power (16 in) The results for the 16 inch connection also show the same trend that shows if the power is decrease the cost of the system will increase because the pump will be required to work at a higher efficiency (which costs more money). Another observation is that the main changes in system configuration involve the heat exchanger and pump, whereas the valve, surge tank, and pipe remain constant. 1860 1880 1900 1920 1940 1960 1980 2000 2020 790000 800000 810000 820000 830000 840000 M ax P ow er (H P) Cost ($) Cost vs Max Power (16 in connection) h1,pu1,v1,st1,pi3 h1,pu2,v1,st1,pi3 h1,pu3,v1,st1,pi3 h1,pu4,v1,st1,pi3 h1,pu5,v1,st1,pi3 h3,pu7,v1,st1,pi3 h5,pu8,v1,st1,pi3 h7,pu9,v1,st1,pi3 62 One of the main constraints applied to the system was the max amount of pressure drop that each component can have with respect to the discharge pressure the centrifugal pump can supply. Even though it would be ideal to have the pump working at its Best Efficiency Point (dependent on flow rate and pressure) throughout normal operation, there are always variations in system pressure and flow rate due to change in operation mode or system problems that require that the pump be sized higher than what it needs to be. This over sizing of the pump is defined as a ?margin?. The goal of the margin is to have it large enough to compensate for system variable, but not so much that the pump is operate at a very low efficiency (which reduces the pumps life span). The results on the pressure margin for the system are included in Figure 34 and Figure 35. Table 7 Pressure Margin Analysis (16 in) 16 inch Connection Hx Pu V St Pi Cost ($) Pressure Marg (ft) 1 6 1 1 3 803024 0 1 6 1 1 3 803024 -10 1 6 1 1 3 803024 -20 1 6 1 1 3 803024 -30 1 6 1 1 3 803024 -40 1 6 1 1 3 803024 -50 1 6 1 1 3 803024 -60 1 6 1 1 3 803024 -70 1 6 1 1 3 803024 -80 1 6 1 1 3 803024 -90 1 6 1 1 3 803024 -100 1 6 1 1 3 803024 -110 1 6 1 1 3 803024 -120 1 6 1 1 3 803024 -130 1 6 1 1 3 803024 -140 1 6 1 1 3 803024 -150 63 1 6 1 1 3 803024 -160 1 6 1 3 3 805114 -170 1 6 1 5 3 807204 -180 2 6 1 5 3 812228 -190 2 6 7 6 3 816365 -200 2 6 7 8 3 838455 -210 2 6 7 10 3 860545 -220 2 6 7 12 3 882635 -230 1 6 1 17 3 889744 -240 1 6 1 19 3 891834 -250 2 6 1 19 3 896858 -260 2 6 7 20 3 900995 -270 Figure 34 Cost vs Pressure Margin (16 in) The results show that a system configuration with a pressure margin of -160 ft is the optimal value in terms of cost for lower ranges of pressure margin, but as the pressure margin increases, so does the cost because the pump has to be sized to with stand higher pressures. All components were varied, except for the pipe. 900995, -270 -300 -250 -200 -150 -100 -50 0 78 00 00 80 00 00 82 00 00 84 00 00 86 00 00 88 00 00 90 00 00 92 00 00 Pr es su re M ar gin (f t) Cost ($) Cost vs Pressure Margin (16 inch connection) 16 inch Connection 64 Table 8 Pressure Margin Analysis (18 in) 18 inch Connection Hx Pu V St Pi Obj pMarg 11 11 11 1 8 874995 0 11 11 11 1 8 874995 -10 11 11 11 1 8 874995 -20 11 11 11 1 8 874995 -30 11 11 11 1 8 874995 -40 11 11 11 1 8 874995 -50 11 11 11 1 8 874995 -60 11 11 11 1 8 874995 -70 11 11 11 1 8 874995 -80 11 11 11 1 8 874995 -90 11 11 11 1 8 874995 -100 11 11 11 1 8 874995 -110 11 11 11 1 8 874995 -120 11 11 11 1 8 874995 -130 11 11 11 1 8 874995 -140 11 11 11 1 8 874995 -150 11 11 11 1 8 874995 -160 11 11 11 1 8 874995 -170 11 11 11 1 8 874995 -180 11 11 11 1 8 874995 -190 11 11 11 1 8 874995 -200 11 11 11 1 8 874995 -210 11 11 11 1 8 874995 -220 11 11 11 1 8 874995 -230 11 11 11 1 8 874995 -240 11 11 11 1 8 874995 -250 11 11 11 1 8 874995 -260 11 11 11 1 8 874995 -270 11 11 11 1 8 874995 -280 11 11 11 1 8 874995 -290 11 11 11 1 8 874995 -300 11 11 11 1 8 874995 -310 11 11 11 1 8 874995 -320 11 11 11 1 8 874995 -330 11 11 11 1 8 874995 -340 11 11 11 1 8 874995 -350 11 11 11 2 8 876040 -360 11 11 11 4 8 878130 -370 65 11 11 11 6 8 880220 -380 11 11 12 7 8 893917 -390 11 11 12 9 8 916007 -400 11 11 12 11 8 938097 -410 11 11 11 14 8 958580 -420 11 11 11 16 8 960670 -430 11 11 11 18 8 962760 -440 11 11 11 20 8 964850 -450 Figure 35 Cost vs Pressure Margin (18 in) For the 18 inch connection the pressure margin is much higher than the 16 inch configurations (almost by 200 ft), but also cost more. A pressure margin of -350 is the lowest possible pressure margin for the cost. The same trend still exists for the 18 inch as the 16 inch connection, which shows that as the pressure margin increases, so does the cost. 964850, -450 -500 -450 -400 -350 -300 -250 -200 -150 -100 -50 0 87 00 00 89 00 00 91 00 00 93 00 00 95 00 00 97 00 00 Pr es su re M ar gin (f t) Cost ($) Cost vs Pressure Margin (18 inch Connection) 18 inch Connection 66 Sensitivity Analysis with Pump Efficiencies For testing pump efficiencies effect on the system variables we examined different pressure margin trends for changing efficiencies. The results are shown for the 16 and 18 inch connection configurations. Figure 36 Pressure Margin Sensitivity Analysis (16 in) The results shown in Figure 36 show that the efficiency have an effect on the cost of the system, but not as strong a relation to pressure margin. Another observation is that the highest pressure margin is achieved by the least efficiency. Overall this shows the dependency between efficiency and cost. -350 -300 -250 -200 -150 -100 -50 0 750000 800000 850000 900000 950000 Pr es su re M ar gin (f t) Cost ($) Cost vs Pressure Margin w/ Varying Pump Efficiencies (16in connection) 0.6 0.62 0.64 0.66 0.68 67 Figure 37 Pressure Margin Sensitivity Analysis (18 in) For the 18 inch connection the results are different from those observed in the 16 inch connection. With the 18 inch configuration the pump efficiency does not have as big of an impact on cost or pressure margin until the max efficiency at 0.72. Each efficiency offer the same max pressure margin (-460 ft). -500 -450 -400 -350 -300 -250 -200 -150 -100 -50 0 850000 900000 950000 1000000 1050000 Pr es su re M ar gin (f t) Cost ($) Cost vs Pressure Margin w/ varing Pump Efficiences (18in connections) 0.6 0.62 0.64 0.66 0.68 0.7 0.72 68 Along with pressure margin, flow rate margin is something to analysis from the perspective of pump efficiencies. Pump efficiency and flow rate affect the pump required horsepower, so even though flow rate is not directly related to the pump efficiency, there will be some effect because of the power constraint. Below the sensitivity analysis for 16 inch and 18 inch configuration options are shown in Figure 38 and Figure 39. Figure 38 Flow rate Sensitivity Analysis (18 in) The results for the 18 inch configuration show that there is very little influence the efficiency has of the flow rate margin, and only at the highest efficiency (0.72) does the cost of the configuration rise, but still not deliver a flow rate margin as low as pumps of a lesser efficiency. This result is similar to the pressure margin sensitivity as well. 400 500 600 700 800 900 1000 1100 870000 880000 890000 900000 910000 920000 930000 Flo w ra te M ar gin (g pm ) Cost ($) Cost vs Flowrate Margin w/ varying Pump Efficiencies (18 in connections) 0.72 0.7 0.68 0.66 0.64 0.62 0.6 69 Figure 39 Flow rate Sensitivity Analysis (16 in) Unlike the results of the 18 inch configuration, the 16 inch system configurations are highly impacted by changes in pump efficiency. The trend is such that as the pump efficiency increases, so does the cost of the system. Also at higher efficiencies the flow rate margin cannot reach lower margins, whereas lower efficiency pumps can. CHL Trade Off and Traceability After reviewing the results from the flow margin, pressure margin, power, and efficiency curves, several options were found that satisfy minimizing the margins and increasing efficiency. 70 Table 9 System Configuration Choices (16 and 18 inch) 16 inch 18 inch hcvstp pressure marg flow marg power efficiency hcvstp pressure marg flow marg power Efficiency 1,1,1,1,3 -170 1000 - 0.6 11,11,11,1,8 -350 1000 1870 0.6 10,6,1,1,3 - 550 - 0.64 12,11,11,1,8 - 950 - 0.6 7,9,1,1,3 -150 1000 1880 0.68 15,13,11,1,8 -360 1000 - 0.72 1,3,1,1,3 -150 800 - 0.62 12,11,11,1,8 - 950 - 0.68 1,5,1,1,3 -150 900 - 0.64 20,11,11,1,8 - 550 - 0.7 3,7,1,1,3 -170 1000 1900 0.66 13,12,11,1,8 - - 1860 - Figure 40 Pressure Margin vs Pump Efficiency (16 inch) Analysis of the tradeoff shows that there is one pareto optimal point that satisfies minimizing pressure margin and maximizing efficiency. That point is -170, 0.6 -150, 0.68 -150, 0.62 -150, 0.64 -170, 0.66 0.58 0.6 0.62 0.64 0.66 0.68 0.7 -175 -170 -165 -160 -155 -150 -145 Pu m p Ef fic ie nc y Pressure Margin(ft) Pressure Margin vs Pump Efficiency (16 inch) 1,1,1,1,3 7,9,1,1,3 1,3,1,1,3 1,5,1,1,3 3,7,1,1,3 71 (7,9,1,1,3). This system configuration for 16 inch connection is considered a possible solution to the design requirements. Figure 41 Pressure Margin vs Flow Margin (16 inch) Analysis of the tradeoff for pressure and flow margin show that for minimizing both axes result in a pareto optimal point at system configuration (1,3,1,1,3). This system configuration for the 16 inch connection is a potential solution for the system. Also, it is interesting to see that the original solution proposed from the previous tradeoff graph is dominated in this tradeoff, so that shows the solution is not globally optimal. -150, 1000 -150, 800 -150, 900 -170, 1000 0 200 400 600 800 1000 1200 -175 -170 -165 -160 -155 -150 -145 Flo w M ar gin (G PM ) Pressure Margin (ft) Pressure Margin vs Flow Margin (16 inch) 7,9,1,1,3 1,3,1,1,3 1,5,1,1,3 3,7,1,1,3 72 Figure 42 Flow Margin vs Pump Efficiency (18 inch) For this tradeoff for flow margin and pump efficiency, that optimal point would minimize flow margin and maximize pump efficiency. From the tradeoff there are two non-dominated solutions (20,11,11,1,8) and (15,13,11,1,8). The first configuration is optimal because it provides the lowest margin, whereas the second configuration is optimal because it offers the best pump efficiency. 1000, 0.6950, 0.6 1000, 0.72 950, 0.68 550, 0.7 0.58 0.6 0.62 0.64 0.66 0.68 0.7 0.72 0.74 0 200 400 600 800 1000 1200 Pu m p Ef fic ie nc y Flow Margin (GPM) Flow Margin vs Pump Efficiency (18 inch) 11,11,11,1,8 12,11,11,1,8 15,13,11,1,8 12,11,11,1,8 20,11,11,1,8 73 Figure 43 Pressure Margin vs Pump Efficiency (18 inch) This last tradeoff graph for the 18 inch configuration relates pressure margin to pump efficiency. The selection from the last tradeoff (15,13,11,1,8) is again a non- dominating solution because it has the highest pump efficiency. The other point (11,11,11,1,8) is also non-dominating because it has the lowest pressure margin. The resulting potential configuration options for the 16 inch and 18 inch systems are: 16 Inch Options Cost ($) h7,pu9,v1,st1,pi3 835647 h1,pu3,v1,st1,pi3 799308 18 Inch Options Cost ($) h20,pu11,v11,st1,pi8 920214 h15,pu13,v11,st1,pi8 897591 h11,pu11,v11,st1,pi8 874995 -350, 0.6 -360, 0.72 0.58 0.6 0.62 0.64 0.66 0.68 0.7 0.72 0.74 -362 -360 -358 -356 -354 -352 -350 -348 Pu m p Ef fic ie nc y Pressure Margin (ft) Pressure Margin vs Pump Efficiency (18 inch) 11,11,11,1,8 15,13,11,1,8 74 The optimization tool also aides in making changes to specifications to satisfy changes in the high level requirements (tracing system requirement to component specification changes). Take for example; that the engineer makes a change to the amount of flow rate margin they want (reduce it from 550 gpm to 500 gpm). Previously there would have to be many recalculations of components to resize them in order to simulate the process and the resubmit new RFQ to all the involve suppliers in the procurement process. With the optimization tool, the high level constraint is traced to the parameter in a specific component that needs to be changed. In this case, the main change required that the connection size of a certain pump, valve, and pipe must be increased (it is minus because it subtracts from the right hand side of the constraint on the connection size) (see ). Vendor 8 for the pump, vendor 1 for the valve, and vendor 3 for the pipes was suggested to increase their connection size in order to find a solution that met the new flow margin requirement. After making the changes the new result cost for the system is $875,133.77. Figure 44 CPLEX Relaxation Suggestion 75 Negotiation aided by Optimization Another benefit to optimization is the information provided for negotiation between design engineer and supplier. The importance of using optimization for negotiation is because the optimization provides modifications to the system requirements and equipment parameters to meet negotiation criteria. Additionally, the optimization results show the positive and negatives of implementing the change. Negotiation is usually done with respect to cost, services, and transportation, but with optimization can be expanded to include engineering categories such as performance and reliability. This is because of the understanding of how equipment parameters are related to high level requirements. Therefore negotiation is another application of the collaborative requirement framework for verification. The method for implementing negotiation via optimization will include defining negotiation objectives, determining the key parameter and equipment for each objective, and evaluating the negotiation objectives with respect to the equipment and the system requirements. Defining negotiation objectives is critical to negotiation because it prevents purchases from conceding and accepting equipment and system designs that could be improved. In the collaborative requirement framework negotiation objectives will be implemented as constraints in the optimization problem. Afterward, optimization results should be used to determine the key equipment parameters that affect the negotiation objective the most. This will help focus on what suppliers and equipment need to be negotiated with to improve upon the negotiation objectives. Lastly the negotiation objectives are evaluated to determine their effect on one another and to determine what the next step in 76 negotiation should occur (if needed). An example of this negotiation method is shown below with the CHL system, evaluating several negotiation objectives. Using the CHL system, the negotiation method will show how cost, performance, and reliability can be improved. In particular, the centrifugal pump will be the main focus because of its main contribution to the performance, reliability, and cost of the system. For evaluation, the negotiated results are compared with the previous selections using optimization. In Table 10 the negotiation objectives are shown for the centrifugal pump. Table 10 CHL Negotiation Objectives Negotiation Objectives Criteria Parameter Baseline Desired Performance Efficiency 0.77 0.9 Reliability Specific Speed (rpm) 2100 1900 Cost Capacity Factor (gpm*psi) 60,000 ($15,000) 55,000 ($12,000) 77 The performance of the centrifugal pump (and the rest of the system) is highly related to the pump?s efficiency, best efficiency point (BEP) flowrate, and BEP differential pressure head. To represent pump performance the pump industry uses a pump characteristic curve similar to the one shown in Figure 45 [25]. These curves show the amount of discharge pressure head (y-axis) a pump can provide for a given volumetric flow rate (x-axis) and also show how the other components in the system increase in pressure head with the rise in volumetric flow rate (the red line). The pump curves and system curves play a role in selecting the best performing pump. Figure 45 Sample Pump Characteristic Curve Additionally there is a strong relationship between the efficiency, power, flowrate, and differential head (shown in equation 7.1). Therefore when negotiating with respect to efficiency ( there is an effect on the pump flowrate (Q), discharge head (H), and shaft power ( ( is the hydraulic power). 78  = ? ? ?.? 7.1  =  7.2 Another negotiation objective is equipment reliability. Equipment reliability can be defined for each component by efficiencies and material properties and from a system viewpoint. From a system viewpoint certain equipment have more priority than others equipment because of their functionality. In this case, equipment reliability also entails preventing critical equipment failure and improving the operation of critical equipment failures [26]. Equipment such as the centrifugal pump and heat exchanger are considered as critical equipment for the CHL system. To demonstrate reliability analysis for negotiation will be conducted on the centrifugal pump. The main reliability parameter for the centrifugal pump is the pump suction- specific speed. In the pump industry it is an empirically established stance that pump models with a specific speed less than 11,000 rpm has a more stable operation and are more reliable. So for pumps with a specific speed in the range of 8,000-11,000 operation should be safe. Otherwise pumps may experience impeller and casing erosion, shaft deflection and many other problems [27]. Therefore with respect to reliability, the lower the pump specific speed the better reliability of the pump. Equation 7.2 shows the relationship the specific speed (Ns) has to the pump speed (N), flow rate (Q), and discharge head (H).  =  ? / 7.3 79 The last negotiation objective is cost. For the centrifugal pump the main contributors to cost are the flowrate and discharge pressure. One parameter, the capacity factor [28] (the product of the flowrate and discharge pressure), is a good gauge of the cost of the pump. By negotiating the flowrate or discharge pressure down, the resulting cost of the pump will go down. Therefore the way that the pump cost will be negotiated is by reducing the capacity factor. After analysis of the different pump suppliers (with respect to the three negotiation objectives) three options arose for the 16 inch connection size and two potential options were determined for the 18 inch connection size. Table 11, Table 12, and Table 13 show the suppliers selected and their objective values. Also in Table 14 and Table 15 the objective values for each supplier is shown with respect to their connection size. Lastly, the results from the tables (for the 16 inch connections) are represented in Figure 46, Figure 47, and Figure 48. All this information will be used to guide negotiation. Specifically, the 16 inch connection options will be negotiated in this example. Table 11 Reliability (Specific speed) Objective Results Connection Size Centrifugal Pump Supplier Lowest Specific Speed 16 8 2040 18 11 2090 Table 12 Cost (Capacity Factor) Objective Results Connection Size Pump Supplier Lowest Pump Capacity 16 1 53200 18 11 68180 80 Table 13 Performance (Efficiency) Objective Results Connection Size Pump Supplier Maximum Efficiency 16 10 0.76 18 14 0.8 Table 14 Objective Values for 16 inch Connection 16 inch Pump Supplier Specific Speed Efficiency Capacity Factor Pump Cost System Cost 8 2040 0.74 63471.46 15,091 539037 10 2146.665 0.76 66587.93 15,828.50 539280 1 2590.02 0.67 53200 12,869.82 538215 Table 15 Objective Values for 18 inch Connection 18 inch Pump Supplier Specific Speed Efficiency Capacity Factor Pump Cost System Cost 11 2090 0.77 68180 16,220.70 505108 14 2237.027 0.8 72533.12 17,430.50 505398 Figure 46 Specific Speed vs Efficiency for 16 inch Connection 0.66 0.67 0.68 0.69 0.7 0.71 0.72 0.73 0.74 0.75 0.76 0.77 1900 2000 2100 2200 2300 2400 2500 2600 2700 Ef fic ie nc y Specific Speed Specific Speed vs Efficiency (16 in) 8 10 1 81 Figure 47 Capacity Factor vs Efficiency for 16 inch Connection Figure 48 Specific Speed vs Capacity Factor for 16 inch Connection When negotiating all the objective must be taken into consideration. This example shows that there is one supplier that satisfies the specific speed (reliability) criteria (supplier 8), one supplier that satisfies the capacity factor (cost) criteria (supplier 1), and no supplier that satisfies the efficiency criteria. Therefore focusing on efficiency, the design engineers want to know high the efficiency can be 0.66 0.67 0.68 0.69 0.7 0.71 0.72 0.73 0.74 0.75 0.76 0.77 50000 55000 60000 65000 70000 Ef fic ie nc y Capacity Factor Capacity Factor vs Efficiency (16 inch) 8 10 1 40000 45000 50000 55000 60000 65000 70000 1900 2100 2300 2500 2700 Ca pa cit y F ac to r Specific Speed Specific Speed vs Capacity Factor (16 inch) 8 10 1 82 negotiated without affecting the other negotiation criteria. For instance, if the efficiency of supplier 8 needed to be negotiated, the design engineers need to understand how much the efficiency is allowed to increase before it affects the cost and reliability of the pump. From the optimization results it is determined that the maximum efficiency the pump can be negotiated to is 0.81 before it effects the reliability (specific speed) criteria. Figure 49 and Figure 50 show the results. Figure 49 Cost vs Efficiency Negotiation Limit Figure 50 Specific Speed vs Efficiency Negotiation Limit 0.76 0.78 0.8 0.82 0.84 0.86 0.88 0.9 0.92 0 5000 10000 15000 20000 25000 Ef fic ie nc y Cost Cost vs Efficiency Negotiation Limit cost vs eff cost max 0.76 0.78 0.8 0.82 0.84 0.86 0.88 0.9 0.92 2100 2150 2200 2250 2300 2350 Ef fic ie nc y Specific Speed (rpm) Specific Speed vs Efficiency Negotiation Limit ss vs eff ss max 83 Chapter 8: Conclusion and Future Work Conclusion By the advent of SysML, MFM models, and CPLEX, this thesis shows there is a way of performing collaborative requirements engineering by using constraints that can be traced to high level requirements through MBSE. The CHL system served as a good baseline system to examine how CRE would work in component selection aspect of procurement. Through optimization best system could be configured base on the objective. In this case, there were tradeoffs that were identified that helped in selecting the right group of components to meet the system requirements. Overall the work will help in clarifying the related parameters and give designers more understanding on how changes in requirement will affect the configuration of components. Future Work Potential ways I can extend this research include applying this work to other areas of the system lifecycle instead of the procurement phase. From an engineering standpoint, further research could be applied to gather different mathematical models of the components and allow for more components to be connected. Additionally this research can look at how different simulation tools generate specifications from RFQ and the variation in the software tools supplier data (can be used to compare optimization results) To apply optimization techniques to not only the product selection, but process selection (how the component are connected and material used in construction), similar to the IPPD. 84 Appendices A: CPLEX Code //Data //System Data// range sv = 1..26; //Number of system variables float SystemVars[sv] = ...; //General Vendor Data// //range Ename= 1..5; //heatex, pump, valve, surgeTank, pipe range VendorNumb =1..20; //List of Vendors {string} Ename = ...; //Pump Data range peg = 1..7; float PumpData[peg][VendorNumb] = ...; //DesFlow,Pwer,Eff,DesDiffHead,DesPress,MaxDiffHead,MaxDiffPress,NPS Hr[8],connDia[9] //Pipe Data range pip = 1..9; float PipeData[pip][VendorNumb] = ...; //NomSiz,WallThick,Len,RoughCon,HLoss,TotHLoss,Wght,TWght,MaxFlow int PipeMat[VendorNumb] = ...;//Pipe Material //Heat Exchanger Data range hx = 1..34; float HxData[hx][VendorNumb] = ...; //Port diameter[8], Cooling Vol Flow[9], DiffHead[11], Efficiency[34] //Valve Data range vlv = 1..4; float ValveData[vlv][VendorNumb] = ...;//conDia[1],Cv[2],VolFlow[3],diffHead[4] //Surge Tank Data range st = 1..9; float SurgeTData[st][VendorNumb] = ...; //volume[1],fluidHght[2],wallThick[3],diameter[4],height[5],desHead[ 8] //Cost Data; float Cost[Ename][VendorNumb] = ...; //Variables dvar boolean x[Ename][VendorNumb]; //Objective minimize sum(e in Ename, v in VendorNumb) x[e][v]*Cost[e][v]; //Constraints subject to { OneVendor: forall(q in Ename) sum(z in VendorNumb) x[q][z] == 1; PipeConnection: 85 sum(y in VendorNumb) PipeData[1][y]*x["Pi"][y] == SystemVars[16];//Set requirement of pipe connection size HxConnection: sum(f in VendorNumb) HxData[8][f]*x["Hx"][f] == SystemVars[16]; ValveConnection: sum(t in VendorNumb) ValveData[1][t]*x["V"][t] == SystemVars[16]; PumpConnection: sum(w in VendorNumb) PumpData[7][w]*x["Pu"][w] == SystemVars[16]; PipeLength: sum(e in VendorNumb) PipeData[3][e]*x["Pi"][e] == SystemVars[21]; PipeMaterial: sum(e in VendorNumb) PipeMat[e]*x["Pi"][e] == SystemVars[22]; TankSupplyPumpHead: sum(h in VendorNumb)SurgeTData[8][h]*x["ST"][h]+SystemVars[25] >= sum(h in VendorNumb)PumpData[5][h]*x["Pu"][h]; forall(i in 23..23) HxReqEfficiency: sum(d in VendorNumb) HxData[34][d]*x["Hx"][d] >= SystemVars[i]; PumpPressLoss: sum(e in VendorNumb)HxData[11][e]*x["Hx"][e]+sum(e in VendorNumb)ValveData[4][e]*x["V"][e]+sum(e in VendorNumb)PipeData[6][e]*x["Pi"][e]-sum(e in VendorNumb)SurgeTData[8][e]*x["ST"][e] <= sum(e in VendorNumb)PumpData[4][e]*x["Pu"][e]; forall(i in 25..25) HxVolFlowrateTop: sum(c in VendorNumb)HxData[9][c]*x["Hx"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] >=-SystemVars[25]; forall(i in 25..25) HxVolFlowrateBot: sum(c in VendorNumb)HxData[9][c]*x["Hx"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] <=SystemVars[25]; forall(i in 25..25) ValveVolFlowrateTop: sum(c in VendorNumb)ValveData[3][c]*x["V"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] >=-SystemVars[25]; forall(i in 25..25) ValveVolFlowrateBot: sum(c in VendorNumb)ValveData[3][c]*x["V"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] <=SystemVars[25]; forall(i in 25..25) PipeVolFlowrateTop: sum(c in VendorNumb)PipeData[9][c]*x["Pi"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] >=-SystemVars[25]; forall(i in 25..25) PipeVolFlowrateBot: sum(c in VendorNumb)PipeData[9][c]*x["Pi"][c]-sum(c in VendorNumb)PumpData[1][c]*x["Pu"][c] <=SystemVars[25]; PumpEfficiency: sum(g in VendorNumb)PumpData[3][g]*x["Pu"][g]>=SystemVars[26]; 86 } main { thisOplModel.generate(); var chl = thisOplModel; var cMarg = chl.SystemVars[25]; //var best; var curr = Infinity; var ofile = new IloOplOutputFile("chl_cool_marg.txt"); while ( 1 ) { //best = curr; if ( cplex.solve() ) { curr = cplex.getObjValue(); writeln(); writeln("OBJECTIVE: ",curr); ofile.writeln(cMarg," ", curr); } else { writeln("No solution!"); break; } //if ( best==curr ) break; cMarg-=10; thisOplModel.HxVolFlowrateTop[25].LB = -cMarg; thisOplModel.HxVolFlowrateBot[25].UB = cMarg; thisOplModel.ValveVolFlowrateTop[25].LB = -cMarg; thisOplModel.ValveVolFlowrateBot[25].UB = cMarg; thisOplModel.PipeVolFlowrateTop[25].LB = -cMarg; thisOplModel.PipeVolFlowrateBot[25].UB = cMarg; } /* if (best != Infinity) { writeln("plan = ",produce.Plan); }*/ ofile.close(); 0; } 87 Appendices B: Component Engineering Data Pipe: Ve nd or # No min al Size (in) Wall thic kne ss (in) Le ngt h (ft) Haze n- Willi ams roug hnes s cons tant Hea d Loss (ft/1 00ft ) Tota l Hea d Loss (ft) Wei ght (lbs/ 100 ft) Tota l Wei ght (lbs) Ma x Flo wra te (gp m) Cost ($/ft ) Tota l Cost ($) Mat eria l Pres sure Dro p (psi) 1 16 0.375 20 0 120 3.09 039 618. 078 62.5 781 3 125. 156 3 750 0 35.0 437 5 700 87.5 1 268. 245 8 2 16 0.375 20 0 140 2.32 288 6 464. 577 1 62.5 781 3 125. 156 3 750 0 34.4 179 7 688 35.9 4 2 201. 626 5 3 16 0.375 20 0 130 2.66 461 1 532. 922 3 62.5 781 3 125. 156 3 750 0 30.6 632 8 613 26.5 6 3 231. 288 3 4 16 0.5 200 120 3.09 039 618. 078 82.7 7 165. 54 750 0 46.3 512 927 02.4 1 268. 245 8 5 16 0.5 200 140 2.32 288 6 464. 577 1 82.7 7 165. 54 750 0 45.5 235 910 47 2 201. 626 5 6 18 0.375 20 0 120 1.96 354 1 392. 708 1 70.5 881 3 141. 176 3 800 0 39.5 293 5 790 58.7 1 170. 435 3 7 18 0.375 20 0 140 1.47 589 1 295. 178 3 70.5 881 3 141. 176 3 800 0 38.8 234 7 776 46.9 4 2 128. 107 4 8 18 0.375 20 0 130 1.69 301 4 338. 602 7 70.5 881 3 141. 176 3 800 0 34.5 881 8 691 76.3 6 3 146. 953 6 88 9 18 0.562 20 0 120 1.96 354 1 392. 708 1 104. 665 7 209. 331 3 800 0 58.6 127 7 117 225. 5 1 170. 435 3 10 18 0.562 20 0 140 1.47 589 1 295. 178 3 104. 665 7 209. 331 3 800 0 57.5 661 2 115 132. 2 2 128. 107 4 11 16 0.656 20 0 120 3.89 658 1 779. 316 1 107. 501 3 215. 002 6 850 0 60.2 007 2 120 401. 4 1 338. 223 2 12 16 0.656 20 0 140 2.92 885 7 585. 771 5 107. 501 3 215. 002 6 850 0 59.1 257 1 118 251. 4 2 254. 224 8 13 16 0.656 20 0 130 3.35 972 9 671. 945 8 107. 501 3 215. 002 6 850 0 52.6 756 3 105 351. 3 3 291. 624 5 14 16 0.844 20 0 120 3.89 658 1 779. 316 1 136. 615 273. 229 9 850 0 76.5 043 8 153 008. 8 1 338. 223 2 15 16 0.844 20 0 140 2.92 885 7 585. 771 5 136. 615 273. 229 9 850 0 75.1 382 3 150 276. 5 2 254. 224 8 16 18 0.75 200 120 2.44 216 1 488. 432 3 138. 172 5 276. 345 900 0 77.3 766 154 753. 2 1 211. 979 6 17 18 0.75 200 140 1.83 564 6 367. 129 2 138. 172 5 276. 345 900 0 75.9 948 8 151 989. 8 2 159. 334 1 18 18 0.75 200 130 2.10 569 2 421. 138 5 138. 172 5 276. 345 900 0 67.7 045 3 135 409. 1 3 182. 774 1 19 18 0.938 20 0 120 2.44 216 1 488. 432 3 170. 924 4 341. 848 8 900 0 95.7 176 6 191 435. 3 1 211. 979 6 20 18 0.938 20 0 140 1.83 564 6 367. 129 2 170. 924 4 341. 848 8 900 0 94.0 084 1 188 016. 8 2 159. 334 1 89 Valve: Connection Diameter (in) Cv Max Allowable Flowrate (gpm) Differential Head (ft) Pressure Drop (psi) Cost ($) 1 16 2000 7500 32.40207 14.0625 2915.11 2 16 2020 7550 32.18855 13.96983 3599.69 3 16 2040 7600 31.9799 13.87928 4284.27 4 16 2060 7650 31.77596 13.79077 4968.85 5 16 2080 7700 31.57658 13.70423 5314.9 6 16 2100 7750 31.3816 13.61961 5660.95 7 16 2120 7800 31.19089 13.53685 6007 8 16 2140 7850 31.00431 13.45587 6353.05 9 16 2160 7900 30.82173 13.37663 6699.1 10 16 2180 7950 30.64302 13.29907 8910.1 11 18 2200 8000 30.46807 13.22314 11284.43 12 18 2220 8050 30.29675 13.14879 13935.89 13 18 2240 8100 30.12897 13.07597 16587.35 14 18 2260 8150 29.96461 13.00464 19238.81 15 18 2280 8200 29.80357 12.93475 21890.27 16 18 2300 8250 29.64575 12.86626 24541.73 17 18 2320 8300 29.49107 12.79912 27193.19 18 18 2340 8350 29.33942 12.73331 29844.65 19 18 2360 8400 29.19072 12.66877 32496.11 20 18 2380 8450 29.04489 12.60548 35147.57 90 Surge Tank/Compression Expansion Tank: Ven dor Volum eteric Capacit y (gal) Heigh t of fluid (in) Wall Thick ness (in) Nomi nal Diam eter (in) Nomi nal Heigh t (in) Corrosi on allowa nce (in) Desig n Press ure (psi) Des ign Hea d (ft) Critical Pressure (psi)(buc kling) Cost ($) 1 4385.711 360.7 274 0.196 609 60 370.7 27446 0.07 82.4 6 190 2.22344 6 100 300 2 4583.648 364.7 483 0.202 114 61 374.7 4828 0.07 84.6 3 195 2.29837 8 101 345 3 4788.252 368.8 377 0.207 731 62 378.8 37712 0.07 86.8 200 2.37628 9 102 390 4 4999.649 372.9 925 0.213 459 63 382.9 92508 0.07 88.9 7 205 2.45722 4 103 435 5 5217.964 377.2 096 0.219 299 64 387.2 09636 0.07 91.1 4 210 2.54122 4 104 480 6 5443.321 381.4 863 0.225 25 65 391.4 86251 0.07 93.3 1 215 2.62833 6 105 525 7 5675.846 385.8 197 0.231 314 66 395.8 19679 0.07 95.4 8 220 2.71860 6 116 570 8 5915.663 390.2 074 0.237 489 67 400.2 07408 0.07 97.6 5 225 2.81208 1 127 615 9 6162.899 394.6 471 0.243 776 68 404.6 47072 0.07 99.8 2 230 2.90881 2 138 660 10 6417.678 399.1 364 0.250 175 69 409.1 36443 0.07 101. 99 235 3.00884 8 149 705 11 6680.125 403.6 734 0.256 686 70 413.6 73419 0.07 104. 16 240 3.11223 9 160 750 12 6950.365 408.2 56 0.263 308 71 418.2 56019 0.07 106. 33 245 3.21903 9 171 795 13 7228.525 412.8 824 0.270 043 72 422.8 82369 0.07 108. 5 250 3.3293 182 840 14 7514.728 417.5 507 0.276 889 73 427.5 507 0.07 110. 67 255 3.44307 6 183 885 15 7809.101 422.2 593 0.283 847 74 432.2 59338 0.07 112. 84 260 3.56042 184 930 16 8111.769 427.0 067 0.290 917 75 437.0 06697 0.07 115. 01 265 3.68138 7 185 975 17 8422.856 431.7 913 0.298 099 76 441.7 91277 0.07 117. 18 270 3.80603 4 187 020 18 8742.489 436.6 117 0.305 393 77 446.6 11654 0.07 119. 35 275 3.93441 5 188 065 19 9070.793 441.4 665 0.312 799 78 451.4 66477 0.07 121. 52 280 4.06658 8 189 110 20 9407.894 446.3 545 0.320 317 79 456.3 54465 0.07 123. 69 285 4.20260 9 190 155 91 Plate Heat Exchanger: V e n d o r # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 P ro ce ss H e at Tran sfe rre d (B tu /h r) 683698423.4 683698423.4 683698423.4 683698423.4 683698423.4 683698423 683698423.4 683698423.4 683698423 683698423 683698423.4 683698423.4 6.84E+08 6.84E+08 683698423.4 683698423 6.84E+08 6.84E+08 6.84E+08 6.84E+08 C o o lan t H e at Tran sfe rre d (B tu /h r) 87017715.12 87742862.75 88468010.37 89193158 89918305.62 90643453.3 91368600.88 92093748.5 92818896.1 93544043.8 94269191.38 94994339.01 95719487 96444634 97169781.88 97894930 98620077 99345225 1E+08 1.01E+08 P ro ce ss In le t Te m p (F) 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 P ro ce ss O u tle t Te m p (F) 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 C o o lan t In le t Te m p e ratu re (F) 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 41 C o o lan t O u tle t Te m p (F) 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 70 Te m p e ratu re D iffe re n ce (F) 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 29 P o rt D iam e te r (in ) 16 16 16 16 16 16 16 16 16 18 16 18 18 18 18 18 18 18 18 18 C o o l V o l. Flo w rate (gp m ) 6000 6050 6100 6150 6200 6250 6300 6350 6400 6450 6500 6550 6600 6650 6700 6750 6800 6850 6900 6950 P ro ce ss V o l. Flo w rate (gp m ) 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 3000 To tal P re ssu re D ro p (fe e t) 160.8596956 148.8677789 150.3112778 151.7576003 153.2067336 154.658665 156.1133816 145.0287583 146.372778 147.719327 149.0683959 150.4199722 151.774 141.4869 142.7426023 144.00058 145.2608 146.5233 147.7881 138.2064 C o o l M ass Flo w rate (lb /h r) 3003614.481 3028644.601 3053674.722 3078704.843 3103734.963 3128765.08 3153795.205 3178825.325 3203855.45 3228885.57 3253915.687 3278945.808 3303976 3329006 3354036.17 3379066.3 3404096 3429127 3454157 3479187 P ro ce ss M ass Flo w rate (lb /h r) 1128210.122 1128210.122 1128210.122 1128210.122 1128210.122 1128210.12 1128210.122 1128210.122 1128210.12 1128210.12 1128210.122 1128210.122 1128210 1128210 1128210.122 1128210.1 1128210 1128210 1128210 1128210 Th e rm al C o n d u ctivity (B TU /h r ft F) 111.0098147 111.0098147 111.0098147 111.0098147 111.0098147 111.009815 111.0098147 111.0098147 111.009815 111.009815 111.0098147 111.0098147 111.0098 111.0098 111.0098147 111.00981 111.0098 111.0098 111.0098 111.0098 H e atin g A re a (ft2) 43065.36096 43424.23897 43783.11698 44141.99499 44500.87299 44859.751 45218.62901 45577.50702 45936.385 46295.263 46654.14104 47013.01905 47371.9 47730.78 48089.65307 48448.531 48807.41 49166.29 49525.17 49884.04 Effe ctive P late A re a (ft2) 4306.536096 4342.423897 4378.311698 4414.199499 4450.087299 4485.9751 4521.862901 4557.750702 4593.6385 4629.5263 4665.414104 4701.301905 4737.19 4773.078 4808.965307 4844.8531 4880.741 4916.629 4952.517 4988.404 Effe ctive P late Le n gth (ft) 75.62420359 75.89707047 76.16881212 76.43944234 76.70897465 76.9774223 77.24479832 77.51111539 77.776386 78.0406224 78.30383667 78.56604046 78.82725 79.08746 79.34670365 79.604979 79.8623 80.11868 80.37412 80.62864 Effe ctive P late W id th (ft) 56.94653156 57.21464438 57.48168543 57.74766748 58.01260308 58.2765045 58.53938387 58.80125294 59.0621233 59.3220064 59.5809133 59.83885502 60.09584 60.35189 60.60699545 60.861182 61.11445 61.36682 61.6183 61.86889 N u m b e r o f P late s 20 21 21 21 21 21 21 22 22 22 22 22 22 23 23 23 23 23 23 24 N u m b e r o f ch an n e ls p e r p ass 10 10 10 10 10 10 10 11 11 11 11 11 11 11 11 11 11 11 11 12 Flu id A re a p e r flu id (ft2) 5.32336177 5.629921007 5.656197846 5.68237048 5.708440143 5.73440805 5.760275373 6.075345454 6.10229858 6.1291497 6.155899963 6.182550501 6.209102 6.532488 6.560101188 6.5876143 6.615029 6.642345 6.669565 7.001084 C h an cro ss Se ctio n al A re a (ft2) 0.560353871 0.562992101 0.565619785 0.568237048 0.570844014 0.5734408 0.576027537 0.578604329 0.58117129 0.58372854 0.586276187 0.588814333 0.591343 0.593863 0.596372835 0.598874 0.601366 0.60385 0.606324 0.60879 C o o lin g Flu id M ass V e lo city (lb s/ft2/h r) 564232.6429 537955.0792 539881.172 541799.387 543709.8201 545612.565 547507.7146 523233.6745 525024.366 526808.077 528584.8872 530354.8766 532118.1 509607.7 511278.1151 512942.34 514600.4 516252.4 517898.4 496949.7 C o o lin g R e yn o ld s N u m b e r 12476.51507 11895.45613 11938.04659 11980.46285 12022.70703 12064.7812 12106.68744 11569.93114 11609.5276 11648.9696 11688.25908 11727.39772 11766.39 11268.63 11305.56551 11342.365 11379.03 11415.56 11451.95 10988.73 C o o lin g N u N u m b e r 260.4481022 252.4981904 253.0854513 253.6695819 254.2506238 254.828618 255.4036032 247.9850835 248.536404 249.084922 249.6306711 250.1736877 250.714 243.768 244.2871058 244.80367 245.3177 245.8293 246.3385 239.8151 C o o lin g H e at Tran sfe r C o e fficie n t (B tu /h r-ft2-F) 1614.621702 4299.712087 4309.712367 4319.659343 4329.553722 4339.3962 4349.187445 4222.859811 4232.24808 4241.58861 4250.882005 4260.128862 4269.33 4151.049 4159.888116 4168.6845 4177.438 4186.15 4194.821 4083.735 P ro ce ss Flu id M ass V e lo city (lb s/ft2/h r) 211935.6472 200395.3733 199464.4022 198545.682 197638.9511 196743.956 195860.4492 185703.0404 184882.812 184072.861 183272.9786 182482.961 181702.6 172707.6 171980.5975 171262.32 170552.6 169851.2 169158 161147.9 P ro ce ss R e yn o ld s N u m b e r 9522.587985 9004.06609 8962.236155 8920.95667 8880.215884 8840.00239 8800.305115 8343.917432 8307.06334 8270.67103 8234.731094 8199.234414 8164.172 7760.011 7727.347396 7695.0742 7663.184 7631.669 7600.523 7240.619 P ro ce ss N u N u m b e r 179.5341257 173.1176719 172.5944849 172.0773441 171.5661301 171.060727 170.5610218 164.7580956 164.284713 163.81654 163.3534793 162.8954343 162.4423 157.169 156.7386309 156.31282 155.8914 155.4744 155.0617 150.2486 P ro ce ss H e at Tran sfe r C o e fficie n t (B tu /h r-ft2-F) 108.1009909 104.2375192 103.9224981 103.6111176 103.3033057 102.998993 102.6981105 99.20405564 98.9190229 98.637127 98.35830882 98.08251099 97.80968 94.63449 94.37537989 94.118989 93.86527 93.61418 93.36566 90.46761 O ve rall H e at Tran sfe r C o e fficie n t(B tu /h r-ft2-F) 86.40428168 86.73327945 86.51910452 86.30715371 86.09738645 85.8897632 85.6842455 83.19058664 82.9936718 82.7987123 82.60567405 82.41452381 82.22523 79.92694 79.74528513 79.565359 79.38713 79.21058 79.03566 76.91063 O ve ral H e at Tran sfe rre d (M B tu /h r) 87.42678103 88.49100272 89.00201205 89.51171625 90.02012916 90.5272644 91.03313521 89.08526702 89.5741977 90.0619345 90.54848949 91.03387458 91.5181 89.63401 90.10269724 90.570293 91.0368 91.50224 91.96661 90.1424 C o st ($) 602915.0535 607939.3456 612963.6377 617987.9298 623012.2219 628036.514 633060.8061 638085.0982 643109.39 648133.682 653157.9746 658182.2667 663206.6 668230.9 673255.143 678279.44 683303.7 688328 693352.3 698376.6 Efficie n cy(H e at Tran s C o e f/O p tim al) 0.864042817 0.867332794 0.865191045 0.863071537 0.860973864 0.85889763 0.856842455 0.831905866 0.82993672 0.82798712 0.82605674 0.824145238 0.822252 0.799269 0.797452851 0.7956536 0.793871 0.792106 0.790357 0.769106 P re ssu re D ro p (p si) 69.81310787 64.60861603 65.23509456 65.86279853 66.49172236 67.1218605 67.75320763 62.94248109 63.5257856 64.1101881 64.6956838 65.28226795 65.86994 61.40532 61.95028938 62.496251 63.0432 63.59112 64.14003 59.98156 92 Centrifugal Pump: Vend or # DesignVolumet ric Flowrate (gpm) Power Req (HP) Rated Efficien cy Design Differenti al Head (ft) Design Differenti al Pressure (psi) Net Positive Suction Head Require d (ft) Nomina l Diamet er (in) 1 6500 1971.667 0.6 720 318.7296 200 16 2 6600 1930.891 0.61 706 312.5321 214 16 3 6700 1923.068 0.62 704 311.6467 216 16 4 6800 1915.333 0.63 702 310.7614 218 16 5 6900 1907.682 0.64 700 309.876 220 16 6 7000 1900.111 0.65 698 308.9906 222 16 7 7100 1892.616 0.66 696 308.1053 224 16 8 7200 1885.194 0.67 694 307.2199 226 16 9 7300 1877.842 0.68 692 306.3346 228 16 10 7400 1870.556 0.69 690 305.4492 230 16 11 7500 1863.333 0.7 688 304.5638 232 18 12 7600 1856.172 0.71 686 303.6785 234 18 13 7700 1849.069 0.72 684 302.7931 236 18 14 7800 1842.023 0.73 682 301.9078 238 18 15 7900 1835.03 0.74 680 301.0224 240 18 16 8000 1828.089 0.75 678 300.137 242 18 17 8100 1821.197 0.76 676 299.2517 244 18 18 8200 1814.354 0.77 674 298.3663 246 18 19 8300 1807.556 0.78 672 297.481 248 18 20 8400 1800.802 0.79 670 296.5956 250 18 93 CHL System: Values Design Cool Flowrate (gpm) 6500 Max Total Cool Flowrate (gpm) 8500 Min Total Cool Flowrate (gpm) 5000 Design Process Flowrate (gpm) 3000 Max Total Process Flowrate (gpm) 3500 Min Total Process Flowrate (gpm) 2700 Minimum Coolant Temp (F) 45 Maximum Coolant Temp (F) 75 Design Cool Supply Temp (F) 50 Design Cool Return Temp (F) 70 94 Design Process Supply Temp (F) 90 Design Process Return Temp (F) 65 Maximum Power (HP) 2000 Net Positive Suction Head Available (ft) 150 Total Water Volume (gallons) 16000 Min. Connection Diameter (in) 16 Max Differential Head (ft) 700 Max Differential Pressure (psi) 309.876 Design Differential Head (ft) 650 95 Design Differential Pressure (psi) 287.742 Req. Piping Length (ft) 200 Pipe Material Type 3 Hx Req Efficiency 0.74 Static Head (ft) 200 Cool Vol. Flow rate Margin (gpm) 1000 Pump Efficiency 0.6 Pressure Margin (ft) 0 96 Appendices C: SysML Diagrams 97 98 99 100 101 102 103 Appendices D: Tabular Requirements # Id Name Text 1 1 SystemPurpose The system shall transfer heat from three process fluids to a cooling fluid. 2 1.1 SystemHeatTransEquip The system shall require three equipment to transfer heat from the three process fluids. 3 1.1.1 SystemHeatExchanger The system shall include three heat exchangers to deliver heat from three process fluids and coolant fluid. 4 1.2 SystemFlowRate The system shall provide the necessary flowrate for the heat exchangers to cool the process fluid. 5 1.2.1 SystemFlowEquipment The system shall include an equipment that maintains the pressure and flowrate of the cooling fluid. 6 1.3 SystemCoolingFluid The system shall circulate cooling fluid. 7 1.3.1 SystemCoolingFluidType The system shall use Brine Refrigerant as a cooling fluid. 8 1.3.2 SystemCoolingFluidHeatRem The system shall remove heat feedback cooling fluid. 9 1.3.2.1 SystemRefrigerantSystem The system shall include a heat exchanger to reduce the temperature of feedback cooling fluid from 35 degrees F to 5 degrees F +- 10%. 10 1.3.3 SystemCoolingWaterVolume The system shall handle 2,000 m3 of cooling water (70,629.33 ft3). 11 1.4 SystemSafety The system shall be safe from temperature, pressure, and flow abnormalities. 104 12 1.4.1 SystemPressureProblems The system shall be able to withstand pressure deviations in the system. 13 1.4.2 SystemTemperatureProblems The system shall be able to handle fluctuations in the cooling fluid temperature. 14 1.4.3 SystemFlowrateProblems The system shall be able to handle flowrate fluctuations in the system. 15 1.5 SystemPower The system shall use offsite and onsite power. 16 1.5.1 SystemPowerType The system shall use Class 1E power supplies for onsite and offsite power. 17 1.5.2 SystemPowerUsage The system shall use a maximum of 10,000 Watts. 18 1.6 SystemCondensingVapor The system shall transfer heat from Condensing Vapor. 19 1.7.1 SystemCondVapHeatRemoval The system shall reduce the Condensing Vapor temperature from 200 deg F to 50 deg F+-1%. 20 1.7.2 SystemCondVapFlowrate The system shall handle Condensing Vapor at flowrates up to 150 gpm+-5%. 21 1.7 SystemCycloPentane The system shall transfer heat from Cyclo-Pentane. 22 1.8.1 SystemCycPenHeatRem The system shall reduce the temperature of Cyclo-Pentane from 300 deg F to 170 deg F+- 1%. 23 1.8.2 SystemCycPenFlowrate The system shall handle Cyclo-Pentane at flowrates up to 140 gpm+-5%. 105 24 1.8 SystemEthyleneGlycol The system shall transfer heat from 60% Ethylene Glycol. 25 1.9.1 SystemEthGlyHeatLoad The system shall reduce the temperature of 60% Ethylene Glycol from 270 deg F to 100 deg F+-1%. 26 1.9.2 SystemEthGlyFlowrate The system shall handle 60% Ethylene Glycol at flowrates up to 200 gpm+-5%. 27 1.9 SystemConnection The system shall be a closed loop system. 28 1.10 SystemOperation The system shall operate at normal conditions. # Id Name Text 1 2.0 PumpPurpose The centrifugal pump shall provide the necessary flow rate for the system. 2 2.0.1 CPMaintainFlow The centrifugal pump shall maintain constant flow rate to system. 3 2.1 PumpOperation The centrifugal pump shall handle operate under varying temperatures, pressures, and flow rates. 4 2.1.1 CPPressure The centrifugal pump shall have an input pressure no lower than 25 psi. 5 2.2 PumpSafety There shall be two centrifugal pumps. 6 2.2.1 CPArrangement The centrifugal pumps shall be connected in parallel. 106 # Id Name Text 1 3.0 Hx1Purpose 2 3.0.1 Hx1ProcessService Hx1 shall service Condensing Vapor process fluid. 3 3.0.1.1 Hx1 HeatLoad Hx1 shall provide sufficient heat load to reduce the temperature of Condensing Vapor from 200 deg F to 50 deg F+-1%. 4 3.0.1.2 Hx1 Flowrate Hx1 shall handle Condensing Vapor at a flowrate up to 150 gpm+-5%. 5 3.0.2 Hx1CoolantService Hx1 shall service Brine Refrigerant. 6 3.0.2.1 Hx1CoolantTemp Hx1 shall handle Brine Refrigerant temperatures of 5 degrees F +-10%. 7 3.1 Hx2Purpose 8 3.1.1 Hx2ProcessService Hx2 shall service Cyclo-Pentane process fluid. 9 3.1.1.1 Hx2 HeatLoad Hx2 shall provide sufficient heat load to reduce the temperature of Cyclo-Pentane from 300 deg F to 170 deg F+-1%. 10 3.1.1.2 Hx2 Flowrate Hx2 shall handle Cyclo-Pentane at a flowrate up to 140 gpm+-5%. 11 3.1.2 Hx2CoolantService Hx2 shall service Brine Refrigerant. 12 3.1.2.1 Hx2CoolantTemp Hx2 shall handle Brine Refrigerant temperatures of 5 degrees F +-10%. 107 13 3.2 Hx3Purpose 14 3.2.1 Hx3ProcessService Hx3 shall service 60% Ethylene Glycol process fluid. 15 3.2.1.1 Hx3 HeatLoad Hx3 shall provide sufficient heat load to reduce the temperature of 60% Ethylene Glycol from 270 deg F to 100 deg F+-1%. 16 3.2.1.2 Hx3 Flowrate Hx3 shall handle 60% Ethylene Glycol at flowrates up to 200 gpm+-5%. 17 3.2.2 Hx3CoolantService Hx3 shall service Brine Refrigerant. 18 3.2.2.1 Hx3CoolantTemp Hx3 shall handle Brine Refrigerant temperatures of 5 degrees F +-10%. # Id Name Text 1 5.0 ValvePurpose The valves shall control the flow rate of cooling fluid to each heat exchanger. 2 5.0.1 ValveFlowrate The valves shall be able to operate over a range of flow rates. 3 5.0.1.1 ValveMassFlowrate All valves shall be able to handle a maximum mass flowrates of 582,259 pounds/hour (264,108.24 kg/h)+- 5%. 4 5.0.2 ValveDifferentialPressure The valves shall have a differential pressure no greater than 30 psid (or a differential head no greater than 40 feet) +-5%. 108 # Id Name Text 1 4.0 SurgeTankPurpose The surge tanks shall hold and supply cooling fluid to the system. 2 4.0.1 SurgeTankNPSH The surge tank shall provide the npsh for the centrifugal pumps. 3 4.0.2 SurgeTankMaintainEquilibrium The surge tanks shall provide cooling fluid storage to compensate for temperature and pressure fluctuations in the system. 4 4.1 SurgeTankCost The max cost for the surge tank shall be a percentage of the maximum system cost. 109 Appendices E: CHL MFM Model 110 References [1] INCOSE, "What is System Engineeing," 14 June 2004. [Online]. Available: http://www.incose.org/practice/whatissystemseng.aspx. [2] Rockwell Automation, "What is Smart Manufacturing?," [Online]. Available: https://smart-process-manufacturing.ucla.edu/about/news/time-magazine -what-is-smart-manufactuing.pdf. [3] IBM, "Smart Manufacturing for chemicasl and petroleum," [Online]. Available: http://www.ibm.com/smarterplanet/us/en/oil_exploration/nextsteps/solution/ S338598H34497P21.html. [4] T. Hessman, "The Dawn of the SMART FACTORY," Industry Week, Febuary 2013. [5] C. Ebert, "Improving engineering efficiency with PLM/ALM," Software & Systems Modeling, vol. 12, no. 3, pp. 443-449, July 2013. [6] Siemens, "Industry Week," June 2013. [Online]. Available: http://www.industryweek- digital.com/industryweekmag/june_2013?pg=36#pg36. [7] A. Hewett, "Product Lifecycle Management (PLM): Critical," in Information Technology and Product Development, Springer, 2010, pp. 81-105. [8] J. F. Davis and Y. Yamashita, "Intelligent Systems in Process Engineering," in Perry's Chemical Engineering Handbook, McGraw-Hill, 1997, pp. 3-91. [9] P. A. Biello, "Critical Information," Mechanical Engineering, vol. 134, no. 6, pp. 32-35, June 2012. [10] J. A. Estefan, "Survey of Model-Based Systems Engineering (MBSE) Methodologies," INCOSE, Seattle, 2007. [11] J. Hirtz, R. B. Stone, D. A. McAdams, S. Szykman and K. L. Wood, "A Functional Basis for Engineering Design: Reconciling and Evolving Previous Efforts," NIST, Gaithersburg, 2002. [12] RDF Working Group, "Resource Description Framework (RDF)," W3C Semantic Web, 10 2 2004. [Online]. Available: http://www.w3.org/RDF/. [13] D. Nau, M. Ball and J. Baras, "Integrating AI Planning and Integer Programming for Use in Integrated Product and Proces Design," in AI in Engineering Design and Manufacturing, 2000. [14] M. Palmer, "Collaborative Requirements Engineering Project," 1 October 2011. [Online]. Available: http://www.nist.gov/el/msid/proceng/simca_cre.cfm. [15] R. Sinnot, Chemical Engineering Design, vol. 6, Oxford: Elsevier, 2005, p. 199. [16] Westinghouse, "AP1000 Design Control Document". [17] Mitsubishi Heavy Industries, "Design Control Document For The US-APWR," US- APWR, Tokyo, 2011. [18] OMG, "OMG Systems Modeling Language," 2012. [Online]. Available: http://www.omgsysml.org. 111 [19] OMG, "OMG Systems Modeling Language (OMG SysML) Tutorial," June 2010. [Online]. Available: http://retis.sssup.it/~marco/files/lesson21_SysML.pdf. [20] J. Baras and S.-A. Yang, Hands-On System Engineering Projects, 2012, pp. 221-261. [21] S. Friedenthal, A. Moore and R. Steiner, A Practical Guide to SysML: The Systems Modeling Language, Elsevier, 2012. [22] M. Lind, "An Introduction to Multilevel Flow Modeling," in Nuclear Safety and Simulation, 2011. [23] K. Urbaniec and C. McDermott, Optimal Design of Process Equipment, Warsaw: Ellis Horwood Limited, 1986. [24] IBM, "Modeling with OPL," IBM, [Online]. Available: http://www- 01.ibm.com/software/commerce/optimization/modeling/. [25] J. Evans, "Centrifugal Pump Efficiency-Preservation of Effiiciency," May 2012. [Online]. Available: http://www.pump-zone.com/topics/pumps/centrifugal- pumps/centrifugal-pump-efficiency-0512. [26] IAEA, "Guidance for optimizing nuclear power plant maintenance programmes," 2003. [27] P. Girdhar and O. Moniz, Practical Centrifugal Pumps: Design, Operation and Maintenance, Newnes, 2005. [28] M. S. Peters, K. D. Timmerhaus and R. E. West, Plant Design and Economics for Chemical Enigneers, 5th ed., New York: McGraw-Hill, 2003, p. 518. [29] INCOSE, "Survey of Model-Based Systems Engineering (MBSE) Methodologies," Seattle, 2008.