A -- Upgrade of the Intelligent System for Heat Exchanger Design (ISHED) to Enable Optimum Condenser Designs and Improve Robustness

Performance Work Statement (PWS) Upgrade of the Intelligent System for Heat Exchanger Design (ISHED) to Enable Optimum Condenser Designs and Improve Robustness A. Background Information In 2003, the HVAC&R Equipment Performance Group completed a simulation package containing first principles-based simulation models for finned-tube air-to-refrigerant evaporators and condensers, called EVAP-COND. Since its placement on the Building Fire Research Laboratory (BFRL) website (http://www2.bfrl.nist.gov/software/evap-cond/) in February 2003, over 2000 organizations (manufacturers, academia and consultants) from over 40 countries have downloaded this simulation package. EVAP-COND facilitates the development of refrigerant circuitry designs through a combination of sophisticated heat exchanger simulators and an effective visual user interface. The interface allows the design engineer to specify a refrigerant circuitry architecture in a graphical manner. Upon completion of the simulation run, the design engineer may use the interface to display a detailed set of performance parameters, which will help him/her to specify a better circuitry architecture for the next simulation run. A description of EVAP-COND is provided in the EVAP-COND?s Help menu. The recent developments in machine learning ? as incorporated in ISHED (Intelligent System for Heat Exchange Design, Domanski et al, 2004) ? allow an engineer to improve the heat exchanger design through automated optimization of refrigerant circuitry. In 2005, the HVAC&R Equipment Performance Group expanded the EVAP-COND interface to incorporate ISHED into an EVAP-COND software package, EVAP-COND/ISHED. At the same time, extensive testing of ISHED continued to cover the whole range of simulation capabilities of the EVAP-COND simulation models, including various refrigerants. At this point of development, ISHED?s operation is stable and robust during optimization runs of evaporator circuitries. However, our tests showed that ISHED has stability problems specific to condenser design and that it lacks the ability to readily explore multi-tube-merging circuitry configurations which are hardly ever used in evaporators but are common in condensers. The tests also demonstrated that some of the current design constraints enforced within ISHED are compatible with optimized evaporator designs but are inherently bad for condenser performance, if implemented. Consequently, an additional level of sophistication needs to be introduced into the ISHED module so that it can successfully optimize condenser circuitries while maintaining the current capability to optimize evaporator circuitries. B. Purpose and Objectives of the Procurement The purpose of this procurement is to upgrade ISHED in the following areas: Item (a) Implement a rule that generated condenser refrigerant circuitry architectures must locate all inlet tubes in the heat exchanger?s last dept row. Additionally, implement an optional rule, that all exit tubes in the condenser refrigerant circuitry must locate in the heat exchanger?s first depth row; ?optional rule? means that the user can enable or disable this rule via an input parameter COND_EXIT_FRONT read by ISHED from a file ISHED.PAR located in directory \ISHED. In all cases, the number of outlet tubes cannot exceed the number of inlet tubes, i.e., the generated condenser circuitries will have a merging or parallel flow configuration. Item (b) Experiment, define, and implement a new set of probabilities with which structure modifying operators are activated to encourage the design of architectures in which more than two tubes merge into a single tube. This new set of probabilities, implemented for condenser optimization, should differ from the set of probabilities used for the evaporators because multiple tubes merging into one tube is a much more favorable characteristic for condensers then evaporators due to the different volumetric flow reductions experienced. Item (c) Improve the robustness of ISHED for condenser optimization while maintaining the current ISHED?s capabilities to optimize the evaporator. C. Contractor Requirements Education and Experience - The contractor shall have a Bachelor?s degree or higher in computer science and at least 3 years experience with programming in C++ - The contractors shall have at least 3 years experience with Darwinian and symbolic learning-based methods used for optimization purposes. General requirements - The ISHED module should recognize the optimization case (evaporator or condenser) by reading the first line of the file ?EVAP.OPC? located in directory \ISHED\COILS\HX. This line will consist of the number ?3? if the optimization run is for an evaporator or ?12? if it is for a condenser. - The implemented code modifications should not affect ISHED?s compatibility with the current beta version of the EVAP-COND/ISHED C++ interface. - The implemented code modifications should not deteriorate ISHED?s current capability to optimize refrigerant circuitries for evaporators. Attachment A specifies the robustness criteria for evaporator optimization. Objective specific requirements Item (a) The requirement is as stated in Section B, Item (a). Item (b) The requirement is as stated in Section B, Item (b). Item (c) The Contractor is required to modify the existing code of ISHED to satisfy to the robustness acceptance criteria specified in Attachment B. The period of performance is 4.5 months, beginning on the date of contract award. The contractor shall present the deliverables to the COTR within 2 months from the date of contract award. The Contracting Officer?s Technical Representative (COTR) shall provide the contractor with his feedback within one month from the date the COTR received the deliverables. Then, the contactor will have one month to incorporate modifications to the deliverables based on the COTR feedback. The COTR will have 0.5 month for final verification. Completion of the project includes: - Implementation of all stipulated code modifications to effect the stipulated functional upgrades - Delivery of the executable element for the program (electronic format) - Delivery of the source code of the program (electronic version) - A written report documenting the changes made to the code and the resulting input/output requirements. The contractor is expected to consult with the COTR bi-weekly during the period of performance of this contract. For planning purposes, each meeting should last less than one (1) hour. Note: NIST will be the distribution source for ISHED; the contractor shall not distribute the program code received from the COTR or developed under this contract in any form. D. Government Responsibilities The Government will provide the contractor with the source code to be upgraded of the current version of ISHED, with the latest executable versions of the EVAP and COND heat exchanger simulators from the EVAP-COND package, and with the executable element of the current beta version of the EVAP-COND/ISHED interface. The COTR will be available to the contractor for consultation regarding design details of the EVAP-COND interface and its interaction with the simulation modules. E. Reporting Requirements and Deliverables See the ?Contractor Requirements? section above. F. Program Management and Control Requirements No special requirements for this contract. G. Inspection and Acceptance Criteria The COTR will evaluate the written report and the performance of the machine learning components of the ISHED package through testing of the complete ISHED system as to their conformance to the contract requirements and will approve or disapprove them within four weeks after delivery of the deliverables. I. Attachments Attachment A: Acceptance Criteria of the ISHED Package for Evaporator Optimization Attachment B: Acceptance Criteria of the ISHED Package for Condenser Optimization. Quality Assurance Surveillance Plan (QASP) I. Objective: This plan provides a basis for the COTR to evaluate the quality of the Contractor?s performance. The oversight provided for in the contract and in this plan will help to ensure that service levels reach and maintain the required levels throughout the contract term. Furthermore, this plan provides the COTR with a proactive way to avoid unacceptable or deficient performance, and provides verifiable input for the required annual past performance evaluations. II. Performance Standards: A. Quality Level: By monitoring the Contractor, the COTR will determine whether the performance levels set forth in the contract have been attained. Quality/performance standards for all tasks are specified herein. B. Management Responsiveness: The COTR will determine whether the Contractor has managed the task order effectively and efficiently, with successful and timely response to the quality standards set forth herein. The COTR will confirm whether the Contractor has satisfactorily met all reporting requirements. III. Evaluation Methods: The COTR will conduct performance evaluations based upon Section II above and the required performance levels set forth herein. Desired Outcomes Required Service Performance Standard Monitoring Methods that May be Used Result for Not Meeting the Performance Stds. The machine learning components of ISHED delivered in accordance with the stated period of performance Delivery date set forth in the PWS is met. 100% compliance is required for the acceptance criteria specified in Appendix A and B. Software tests will be conducted by COTR as described in Appendix A and B. For each percentage less than 100, a corresponding amount of payment shall be withheld until compliance is achieved. Attachment A Acceptance Criteria of the ISHED Package for Evaporator Optimization The COTR will verify that the robustness of the ISHED package has not been reduced for evaporator optimization by the additional work in this contract. The COTR will verify the requirements in this attachment by performing optimization runs for three test cases for which robustness was verified during prior work. 1. Evaporator with three depth rows having 43 tubes per depth row; population size = 20; 2. Evaporator with five depth rows having 17 tubes per depth row; population size = 20; 3. Evaporator with four depth rows having 32 tubes per depth row; population size = 20; For each heat exchanger, 5 optimization runs will be performed. The number of generations will be 500. All other ISHED.PAR parameters will be set to their default values, except where stated otherwise. The performance will be considered satisfactory if: - at least four out of five optimization runs for each heat exchanger completes evaluations of all 500 generations (i.e., 80 % completion rate is acceptable) - all optimization runs show increasing trend in capacities with the progression of the optmization run. The following are detailed specifications for the heat exchagers and operating conditions that will be used to evaluate the robustness of the machine learning components of the ISHED package. They are presented in the format compatable with the EVAP-COND package. Parameters for evaporator #1 (3 rows of 43 tubes = 129 total) Evaporator #1 Input DT1 file: Evaporator Test Case #1 9.220000,10.010000,25.400000,22.230000,454.000000,1099.510010,10.005000 2.004000,0.203200,0.221600,0.386001 2,1 1 ------------------------------------------------------------------- Evaporator #1 EVAP.OPC file: Use with Refrigerant R22 3 Air inlet temperature (C) ,26.666700 Air inlet pressure (kPa) ,101.324997 Air inlet relative humidity (-),0.500000 Refrig. outlet sat. temp. (C) ,7.000000 Refrig. superheat (C) ,5.000000 Refrig. inlet quality (-) ,0.200000 Refrig. mass flow rate (kg/h) ,350.000000 ------------------------------------------------------------------- Evaporator #1 Air flow profile, file DT3 1 0.1,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 1.67,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 80.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 8 Max_fed = 8 Parameters for evaporator #2 (5 rows of 17 tubes = 85 total) Evaporator #2 input DT1 file: Evaporator Test Case #2 9.220000,10.010000,25.400000,22.230000,454.000000,439.109985,10.005000 2.004000,0.203200,0.221600,0.386001 2,1 1 ------------------------------------------------------------------- Evaporator #2 EVAP.OPC file: Use with R22 3 Air inlet temperature (C) ,26.666700 Air inlet pressure (kPa) ,101.324997 Air inlet relative humidity (-),0.500000 Refrig. outlet sat. temp. (C) ,7.000000 Refrig. superheat (C) ,5.000000 Refrig. inlet quality (-) ,0.200000 Refrig. mass flow rate (kg/h) ,350.000000 ------------------------------------------------------------------- Evaporator #2 Air flow profile, file DT3 1 0.1,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 1.00,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 60.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 6 Max_fed = 8 Parameters for evaporator #3 (4 rows of 32 tubes = 128 total) Evaporator #3 input DT1 file Evaporator Test Case #4 9.200000,10.000000,25.400000,22.200001,500.000000,820.099976,10.000000 2.000000,0.200000,0.221569,0.386013 3,1 1 ------------------------------------------------------------------- Evaporator #3 EVAP.OPC FILE: Use with R600a 3 Air inlet temperature (C) ,26.670000 Air inlet pressure (kPa) ,101.324997 Air inlet relative humidity (-),0.500000 Refrig. outlet sat. temp. (C) ,7.000000 Refrig. superheat (C) ,5.000000 Refrig. inlet quality (-) ,0.200000 Refrig. mass flow rate (kg/h) ,90.000000 ------------------------------------------------------------------- Evaporator #3 Air flow profile. DT3 1 0.1,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 3.25,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 90.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 4 Max_inlets = 10 Maxsplitsallowed = 10 Attachment B Acceptance Criteria of the ISHED Package for Condenser Optimization The COTR will evaluate the robustness of the machine learning components of the ISHED package by executing optimization runs for a total of five heat exchagers of the following general specifications, more detail for each test case is located on the following pages of this attachment: 1. Condenser with three depth rows having 43 tubes per depth row; population size = 20; 2. Condenser with five depth rows having 26 tubes per depth row; population size = 20; 3. Condenser with four depth rows having 32 tubes per depth row; population size = 20; 4. Condenser with four depth rows having 32 tubes per depth row; population size = 20; 5. Condenser with two depth rows having 50 tubes per depth row; population size = 20; For each heat exchanger, 5 optimization runs will be performed. The number of generations will be 500. All other ISHED.PAR parameters will be set to their default values, except where stated otherwise. The performance will be considered satisfactory if: - at least four out of five optimization runs for each heat exchanger completes evaluations of all 500 generations (i.e., 80 % completion rate is acceptable) - all optimization runs show increasing trend in capacities with the progression of the optmization run. The following are detailed specifications for the heat exchagers and operating conditions that will be used to evaluate the robustness of the machine learning components of the ISHED package. They are presented in the format compatable with the EVAP-COND package. Parameters for condenser #1 (3 rows of 43 tubes = 129 total) Condenser #1 Input DT1 file Condenser Test Case #1 9.220000,10.010000,25.400000,22.230000,454.000000,1099.510010,10.005000 2.004000,0.203200,0.221600,0.386001 2,1 1 ------------------------------------------------------------------- Condenser #1 EVAP.OPC file: Use with Refrigerant R600a 12 Air inlet temperature (C) ,35.00000 Air inlet pressure (kPa) ,101.325 Air inlet relative humidity (-),0.500000 Refrig. inlet sat. temp. (C) ,45.000000 Refrig. superheat (C) ,9.7000000 Refrig. mass flow rate (kg/h) ,80.000000 1/0 = Y/N; iterate subcooling ,1 Target subcooling (C) ,5.000000 ------------------------------------------------------------------- Condenser #1 Input DT3 file 2 0.0,1098.9,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.03,5.30,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 80.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 10 Max_fed = 8 Parameters for condenser #2 (5 rows of 26 tubes = 130 total) Condenser #2 input DT1 file Condenser Test Case #2 9.220000,10.010000,25.400000,22.230000,454.000000,439.109985,10.005000 2.004000,0.203200,0.221600,0.386001 2,1 1 ------------------------------------------------------------------- Condenser #2 EVAP.OPC file: Use with R600a 12 Air inlet temperature (C) ,35.00000 Air inlet pressure (kPa) ,101.325 Air inlet relative humidity (-),0.500000 Refrig. inlet sat. temp. (C) ,45.000000 Refrig. superheat (C) ,9.7000000 Refrig. mass flow rate (kg/h) ,80.000000 1/0 = Y/N; iterate subcooling ,1 Target subcooling (C) ,5.000000 ------------------------------------------------------------------- Condenser #2 Air flow profile DT3 2 0.6,667.1,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.07,8.73,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 60.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 10 Max_fed = 8 Parameters for condenser #3 (4 rows of 32 tubes, = 128 total) Condenser #3 Input DT1 file Condenser Test Case #3 9.200000,10.000000,25.400000,22.200001,500.000000,769.299988,10.000000 2.000000,0.200000,0.221569,0.386013 3,1 1 ------------------------------------------------------------------- Condenser #3 EVAP.OPC FILE: Use with Refrigerant R32 12 Air inlet temperature (C) ,35.00000 Air inlet pressure (kPa) ,101.325 Air inlet relative humidity (-),0.500000 Refrig. inlet sat. temp. (C) ,45.000000 Refrig. superheat (C) ,46.9000000 Refrig. mass flow rate (kg/h) ,200.000000 1/0 = Y/N; iterate subcooling ,1 Target subcooling (C) ,5.000000 ------------------------------------------------------------------- Condenser #3 Input DT3 file 2 1.3,819.5,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.06,7.11,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 80.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 10 Max_fed = 8 Parameters for condenser #4 (4 rows of 32 tubes = 128 total) Condenser #4 input DT1 file Condenser Test Case #4 9.200000,10.000000,25.400000,22.200001,500.000000,820.099976,10.000000 2.000000,0.200000,0.221569,0.386013 3,1 1 ------------------------------------------------------------------- Condenser #4 EVAP.OPC file: Use with R600a 12 Air inlet temperature (C) ,35.00000 Air inlet pressure (kPa) ,101.325 Air inlet relative humidity (-),0.500000 Refrig. inlet sat. temp. (C) ,45.000000 Refrig. superheat (C) ,9.7000000 Refrig. mass flow rate (kg/h) ,80.000000 1/0 = Y/N; iterate subcooling ,1 Target subcooling (C) ,5.000000 ------------------------------------------------------------------- Condenser #4 Air flow profile. DT3 1 819.5,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 3.58,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 80.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 10 Max_fed = 8 Parameters for condenser #5 (2 rows of 50tubes = 100 total) Condenser #5 input DT1 file Condenser Test Case #5 9.200000,10.000000,25.400000,22.200001,500.000000,820.099976,10.000000 2.000000,0.200000,0.221569,0.386013 3,1 1 ------------------------------------------------------------------- Condenser #5 EVAP.OPC file: Use with R600a 12 Air inlet temperature (C) ,35.00000 Air inlet pressure (kPa) ,101.325 Air inlet relative humidity (-),0.500000 Refrig. inlet sat. temp. (C) ,45.000000 Refrig. superheat (C) ,9.7000000 Refrig. mass flow rate (kg/h) ,60.000000 1/0 = Y/N; iterate subcooling ,1 Target subcooling (C) ,5.000000 ------------------------------------------------------------------- Condenser #5 Air flow profile. DT3: 2 0.0,1276.7,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 0.03,3.59,0.0,0.0,0.0,0.0,0.0,0.0 0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0 Refrigerant heat transfer coefficient: 1.000000 Refrigerant pressure drop : 1.000000 Air-side heat transfer coefficient : 1.000000 Air vol. flow rate (m3/min) : 63.000000 Fan power (W) : 100.000000 ------------------------------------------------------------------- No seed files Min_inlets = 1 Max_inlets = 10 Max_fed = 8