Test scheduling optimization method and test scheduling optimization system based on mechanism model

Abstract

The present invention discloses a test scheduling optimization method and a test scheduling optimization system based on a mechanism model, which relates to the field of wind tunnel system tests, including: establishing a wind tunnel system test system mechanism model, defining its modular architecture according to the system composition, and describing the task execution and equipment response mechanism; using a discrete process planning algorithm or a scheduling algorithm to construct a scheduling algorithm model; calling the scheduling algorithm model and the wind tunnel test system mechanism model through a scheduling logic driver; associating and / or integrating with a software platform to construct a configured visual monitoring interface; integrating a scheduling logic driver through a graphical interface application software, and encapsulating mathematical models and mechanism models. The present invention provides a test scheduling optimization method and a test scheduling optimization system based on a mechanism model, which solves the problems of large calculation amount, long time consumption, and low efficiency in test plan formulation in the prior art by combining the mechanism model of wind tunnel test operation with the test scheduling optimization design.

Description

Technical Field

[0001] The present invention relates to the technical field of wind tunnel system test plan scheduling, and more specifically, to a test scheduling optimization method and a test scheduling optimization system based on a mechanism model. Background Art

[0002] The purpose of test plan scheduling is to obtain the optimal test execution sequence under different objectives (optimal efficiency or optimal cost, etc.) under the premise of meeting various constraints, generate a test plan for test analysis and decision-making, and be able to dynamically adjust and optimize the plan according to the actual progress. At this stage, some rules have been summarized for test plan scheduling, such as first determining the test sequence of the test object, and on this basis, adjusting the temperature first according to the consumption demand of liquid nitrogen, then adjusting the total pressure, then adjusting the Mach number, and finally adjusting the model attitude. It is worth noting that the determination of temperature and total pressure is not fixed, but can be dynamically adjusted as needed (factors such as sudden situations causing the test to be terminated, the speed of equipment warming up after the end of the test on the day, etc.). If the optimal goal is to minimize energy consumption, it is necessary to focus on factors such as electricity prices and liquid nitrogen consumption in different time periods.

[0003] At present, test plan scheduling mainly relies on experience accumulation. The test plan scheduling algorithms applicable in other industries are not applicable to wind tunnel scenarios. They can only be scheduled manually by experienced engineers. The calculation of economic and time factors such as power consumption and liquid nitrogen consumption in the test schedule also relies on manual calculation through formulas. The calculation is large and time-consuming, which seriously affects the efficiency of test plan formulation. Summary of the invention

[0004] An object of the present invention is to solve at least the above problems and / or disadvantages and to provide at least the advantages which will be described hereinafter.

[0005] In order to achieve these purposes and other advantages of the present invention, a method for optimizing test schedule based on a mechanism model is provided, comprising:

[0006] S1. Establish a wind tunnel system test system mechanism model, define its modular architecture according to the system composition, and describe the task execution and equipment response mechanism;

[0007] S2. Use discrete process planning algorithm or scheduling algorithm to build a scheduling algorithm model;

[0008] S3, calling the scheduling algorithm model and the wind tunnel test system mechanism model through the scheduling logic driver to realize automatic iterative optimization and generate the operation plan, and testing the operation effect through the verification plan;

[0009] Associate and / or integrate with the software platform to build a configurable visual monitoring interface;

[0010] S4, through the graphical interface application software, integrate the scheduling logic driver, encapsulate the mathematical model and mechanism model;

[0011] Among them, the operation of the scheduling algorithm script is based on the user-defined scheduling configuration, and multiple scheduling rules and multiple objectives are used to optimize the scheduling algorithm, so that the scheduling algorithm script can call the scheduling algorithm and equipment model in parallel, and generate the test operation plan in a multi-sample manner.

[0012] Preferably, in S1, the wind tunnel test system mechanism model includes: a task logic simulation sub-model of the equipment task execution process, a time and resource consumption sub-model required for various types of equipment to respond to different task processes, a sub-model of the impact of equipment response under different test conditions, and other equipment operation mechanism sub-models, as well as an overall test system model integrating the above sub-models.

[0013] Preferably, the task logic simulation sub-model is based on the joint relationship between the mechanical system, the electronic control system and the power system in the equipment, and maps the logical relationship between the equipment operations to the model level;

[0014] The task logic simulation sub-model includes: an electromechanical control simulation model, a multi-body dynamics simulation model, and a system-level simulation model;

[0015] The system-level simulation model is constructed by coupling at least two of a semi-flexible wall nozzle equipment-level model, a two-throat equipment-level model, a model transport isolation system equipment-level model, a full model vehicle equipment-level model, and a half-model vehicle equipment-level model.

[0016] Preferably, the coding format of the design variables in the scheduling algorithm model needs to be consistent with the task process sequence and equipment number information.

[0017] Preferably, the content of the scheduling algorithm model includes: setting constraint rules for business scenarios, setting schedulable elements, setting inputs for scheduling large processes, setting inputs and outputs for scheduling sub-processes, setting inputs for scheduling algorithms and their parameters, sub-process energy consumption calculations, and running sub-process sorting algorithms.

[0018] Preferably, the sub-process sorting algorithm performs sorting based on rules provided by the user, and only considers various resource consumptions of each test of the sub-sequence number in the rule sorting, and does not consider resource consumption of switching the sub-sequence number.

[0019] Preferably, the input of the scheduling process is: work number, work name, work time, nitrogen consumption, and power consumption;

[0020] The inputs of the scheduling subprocess are: test number, test total temperature, test Mach number, and test total pressure;

[0021] The output of the scheduling subprocess is the scheduled sequence.

[0022] Preferably, in S3, the content of the scheduling logic driver program includes:

[0023] S31, integrating the scheduling task process coding in the scheduling planning algorithm with the external input interface retained by the overall model of the test system, and parsing it into the test process logic parameters;

[0024] S32, integrating the system state requirement code in the scheduling algorithm with the external input interface retained by the overall model of the test system, and parsing it into system model parameters;

[0025] S33, integrating the objective function and constraint function in the scheduling algorithm with the external output interface retained by the overall model of the test system to realize the call of the simulation results.

[0026] A test scheduling optimization system is applied to a test scheduling optimization method based on a mechanism model. The architecture of the test scheduling optimization system includes:

[0027] An external input layer for accepting one or more optimization target parameters;

[0028] An interactive layer for intuitively displaying the input of external input parameters and the viewing of scheduling results;

[0029] The model layer is used to describe the system mechanism and can receive real-time data to simulate the system operation status;

[0030] The optimal scheduling solution output after the model layer iteration is completed is displayed in the output layer of the interface in the form of a color block diagram;

[0031] The model layer specifically includes: a scheduling algorithm model developed using the Python language, and a wind tunnel test system mechanism model developed using the Modelica language.

[0032] The present invention has at least the following beneficial effects:

[0033] The present invention solves the problems of heavy reliance on experience when manually carrying out test plan scheduling, large manual workload to obtain multiple scheduling schemes, cumbersome scheduling calculations, and easy errors in the process. By constructing a multi-constraint and multi-objective test scheduling optimization algorithm and a user interaction interface, the test task scheduling and resource allocation business model is converted into a mathematical model of constraint programming based on the optimization goals and constraints, and then the mathematical model is solved by a multi-objective evolutionary algorithm. It is realized that according to the scheduling rules and the working conditions and constraints input by the user, multiple scheduling schemes are automatically obtained and the time expenditure, electricity resource consumption, nitrogen resource consumption and other data that the user cares about are calculated, and the optimal scheduling scheme is recommended to the user.

[0034] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a structural schematic diagram of the test scheduling optimization system of the present invention;

[0036] Figure 2 A flow chart for the present invention and the scheduling logic driver program operation. DETAILED DESCRIPTION

[0037] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0038] A test scheduling optimization method, including a test scheduling optimization method based on a mechanism model, characterized in that it includes the following steps:

[0039] S1, establish the wind tunnel system test system mechanism model, describe the task execution and equipment response mechanism, with modular characteristics, including the task logic simulation sub-model of the equipment task execution process, the time and resource consumption sub-model required for various types of equipment to respond to different task processes, and the impact sub-model of the equipment response state under different test conditions. The whole of these models is called the overall test system model;

[0040] The process of establishing the mechanism model of the wind tunnel system test system includes:

[0041] S11, create an overall model architecture of the test system that can adapt to the requirements of different task processes, including the task logic simulation sub-model architecture, the time and resource consumption sub-model architecture, the equipment response impact sub-model architecture, and the interface model of the three;

[0042] S12, develop a task logic simulation sub-model template, define an external interaction interface, and control the task execution process or improve the task logic simulation sub-model through external instructions;

[0043] S13, develop sub-models for the time and resource consumption required for various types of equipment to respond to different task processes, which are used to define system parameters, can calculate the time, energy, and resource consumption required for the equipment based on parameters, working conditions, and operating parameters, and define external interactive interfaces, which can modify system model parameters and read system operating status and energy consumption information through external instructions. Calculate energy consumption based on various standard theoretical formulas such as the compressor shaft power formula and the liquid nitrogen consumption formula; when using the compressor shaft power formula, calculate the shaft power through the Mach number, total temperature, and total pressure, and then calculate the energy consumption based on the shaft power; when using the liquid nitrogen consumption formula, calculate the energy consumption through the Mach number, total temperature, and total pressure.

[0044] S14, develop a sub-model of the impact of equipment response under different test conditions, define test conditions such as test task curves, test environment requirements, load requirements, data collection points, and data processing methods as model operation parameters or initialization requirements, adjust model contents such as medium properties, equipment energy consumption, and resource consumption curves, and support comprehensive calculation of equipment operating costs and efficiency under different test conditions.

[0045] S15, integrate the above-mentioned models into the overall model of the test system and run them to confirm the correctness and completeness of the external interface. Through the integrated overall model of the test system, an external interface is provided, which can be integrated with the scheduling optimization algorithm to support the calculation of the objective function or constraint conditions of the iterative process.

[0046] S2, build a scheduling algorithm model, which can adopt various discrete process planning algorithms or scheduling algorithms including genetic algorithms. The coding form of the design variables of the scheduling algorithm needs to be consistent with the task process sequence, equipment number and other information; the specific content of the scheduling algorithm model includes:

[0047] 1. Set constraints for business scenarios;

[0048] 2. Set up schedulable elements;

[0049] 3. Set the input of the scheduling process. The input of the scheduling process includes work number, work name, work time, nitrogen consumption, and power consumption;

[0050] 4. Set the input and output of the scheduling sub-process. The input of the scheduling sub-process is the test sequence number, test total temperature, test Mach number, and test total pressure; the output of the scheduling sub-process is the sequence after scheduling;

[0051] 5. Set the input of the scheduling algorithm and its parameters;

[0052] 6. Calculation of energy consumption of sub-processes;

[0053] 7. Run the sub-process sorting algorithm. The sub-process sorting algorithm is based on rule sorting: the schedule is sorted according to the rules provided by the user; among them, the rule-based sorting only considers the various resource consumptions of each sub-sequence number test, and does not consider the resource consumption of sub-sequence number switching. The sub-process sorting algorithm is an optimization algorithm sorting, including the following contents:

[0054] 1) Problem realization: Use element-wise problem definition to define the optimization problem;

[0055] 2) Define the optimization algorithm: Select the GA genetic algorithm to instantiate an optimization problem solving algorithm object;

[0056] 3) Define termination conditions: set the termination conditions of the optimization algorithm;

[0057] 4) Optimization calculation: Use the defined optimization algorithm and termination conditions to optimize the problem.

[0058] S3, the scheduling algorithm is integrated with the equipment model. The task logic and equipment information required in the scheduling algorithm are consistent with the overall model of the test system. The scheduling algorithm can call the overall model of the test system, the task logic simulation sub-model, the time and resource consumption sub-model or the equipment response impact sub-model in parallel in the background. The overall model of the test system calculates the test process time, cost, resource consumption and other computing requirements, and returns them to the scheduling algorithm to support algorithm iteration and optimization;

[0059] Among them, in S3, the test plan scheduling interface includes: test basic information input area, scheduling function button area, historical query condition input area, scheduling list area, test schedule display area, historical schedule display area, scheduling algorithm area, and test details area.

[0060] The scheduling algorithm model and the equipment mechanism model are integrated through the scheduling logic driver, and the integration content includes:

[0061] Integrate the scheduling task process coding in the scheduling planning algorithm with the external input interface retained by the overall model of the test system, and parse it into test process logic parameters;

[0062] Integrate the system state requirement coding in the scheduling algorithm with the external input interface retained by the overall model of the test system and parse it into system model parameters;

[0063] Integrate the objective function and constraint function in the scheduling algorithm with the external output interface retained by the overall model of the test system to realize the call of simulation results;

[0064] S4, run the scheduling algorithm script: according to the user-defined scheduling configuration, multiple scheduling rules and multi-objective optimization scheduling algorithms are used for scheduling. The script can call the scheduling algorithm and equipment model in parallel, generate the test operation plan in a multi-sample manner, and quickly verify the operation effect of the plan through the overall model of the test system;

[0065] Among them, the running scheduling algorithm script can use multiple scheduling rules and multi-objective optimization scheduling algorithms for scheduling according to the user-defined scheduling configuration. The script can call the scheduling algorithm and equipment model in parallel, generate the test operation plan in a multi-sample manner, and quickly verify the operation effect of the plan through the overall model of the test system;

[0066] Running the scheduling algorithm script can deploy the custom script algorithm in the workstation or computing server, and if necessary, debug the communication interface between the workstation, server and the overall model server of the test system;

[0067] The running scheduling algorithm script can run the script in multi-threaded parallel form. If necessary, multiple test system overall model instance samples can be generated in advance to meet the efficiency requirements of parallel solution;

[0068] Running the scheduling algorithm script can perform iterative solutions based on the optimization algorithm and return the final result or a set of alternative solutions.

[0069] S5, deployment, operation and monitoring of scheduling algorithm scripts, combined with the test control and management system software platform, integrates the scheduling algorithm, test system mechanism model and software platform, builds a configurable visual monitoring interface, and realizes the simulation display and operation monitoring of the optimization results.

[0070] Among them, in S5, the visual monitoring interface includes the following contents: interaction between the scheduling algorithm script and the front-end user interface; front-end input of the large process; input of the energy consumption configuration table of the large process; front-end input of the sub-process; output after scheduling the large process; output after scheduling the sub-process.

[0071] A test scheduling optimization system includes a scheduling algorithm mathematical model, a wind tunnel test system mechanism model, a scheduling logic driver and a graphical interactive interface. The scheduling algorithm mathematical model generates a preliminary scheduling plan based on the scheduling conditions input by the user, the wind tunnel test system mechanism model simulates and solves the preliminary scheduling plan to obtain the scheduling output index, the scheduling logic driver is a simulation engine that drives the simulation operation of the above two types of models, and the graphical interactive interface visualizes the scheduling results.

[0072] The present invention solves the problems of heavy reliance on experience when manually carrying out test plan scheduling, large manual workload to obtain multiple scheduling schemes, cumbersome scheduling calculations, and easy errors in the process. A test scheduling algorithm model combining the overall model of the test system with the scheduling optimization algorithm is constructed, and the optimization objectives and constraints are obtained through the simulation results of the overall model of the test system, which solves the problem of inaccurate consumption of various energy and resources in the test task planning process. Then, it is iterated and optimized with optimization algorithms such as multi-objective evolution, and task working conditions and initialization conditions are generated according to scheduling rules and user input. Multiple scheduling schemes are automatically obtained and the time cost, electricity resource consumption, nitrogen resource consumption and other data that users are concerned about are calculated, and the optimal scheduling scheme is recommended to users, and visual numerical verification can be performed.

[0073] Example 1

[0074] The present invention designs a wind tunnel test scheduling optimization system based on a mechanism model. Its working principle is to construct a wind tunnel test system mechanism model and a scheduling algorithm mathematical model, take the optimization objectives and constraints under specific scenarios as input, use the mathematical model to generate multiple scheduling schemes, and call the mechanism model to solve the target parameters corresponding to each scheme. The mathematical model then performs iterative optimization based on these solution results until the optimal scheduling scheme is output.

[0075] The present invention provides a test scheduling optimization system, which includes a test system mechanism model constructed based on Modelica language and a scheduling algorithm mathematical model constructed based on Python language, and adopts C++ language and QT software development framework to customize the user interface, which can receive input parameters such as optimization objectives and constraints input by users and pass them to the model and algorithm to form an interactive test plan scheduling optimization system. It realizes the functions of enabling users to input parameters and configure scheduling algorithms through interface input, iterative mutual calling of scheduling algorithm mathematical models and mechanism models for scheduling processing, outputting the optimal scheduling scheme and visually presenting it on the software interface, and generating word reports and the like.

[0076] See also Figure 1 ,The system architecture layers include external input layer, interaction layer, model layer (including two types of models), and output layer.

[0077] 1. External input layer: specifically refers to the input parameters of the schedule, including one or more optimization target parameters (such as gas consumption, power consumption, time consumption), parameter priorities under multiple optimization target parameters, and scheduling and optimization constraints (such as the order of some processes is fixed and cannot be sorted)

[0078] 2. Interaction layer: The interface application developed using C++ language and QT software development framework is the software presentation of the test scheduling optimization system. Users can intuitively and visually interact to enter external input parameters and view scheduling results.

[0079] 3. Model layer: It includes two types of models. The first type is the mathematical model of scheduling algorithm developed in Python language, which uses a variety of discrete process planning algorithms to realize the generation of multiple scheduling schemes, and uses a variety of parameter optimization algorithms to realize the iterative optimization of scheduling schemes. The second type is the mechanism model of wind tunnel test system developed in Modelica language, which uses the real wind tunnel physical mechanism equation to deduce and solve the output index value (electricity consumption, time consumption, liquid nitrogen consumption, etc.) under each scheduling scheme. These two types of models are presented in the physical form of py script files and mo model files respectively.

[0080] 4. Output layer: After the model layer iteration is completed, the optimal scheduling solution output is displayed on the interface in the form of a color block diagram, and a report document in word format is generated.

[0081] like Figure 1 As shown in the figure, the specific steps to realize the multi-optimization and multi-objective wind tunnel test scheduling optimization system are as follows:

[0082] 1. Construct the mechanism model of wind tunnel system experiment system;

[0083] 1) Modeling language and technical route selection

[0084] The wind tunnel system is a complex system with multiple disciplines and professions coupled. Mechanical structures such as the two-throat and flexible-wall nozzle involve disciplines such as mechanics, control, and electrical engineering. The physical principles of its working process involve mathematical physics equations from multiple disciplines. A modeling language is needed to describe the interaction and coupling relationship between these multiple physical domains, the physical relationship between the motor control signal and the speed torque, the kinematic relationship between the driving element and the driven element of the transmission mechanism, and the control response relationship between the attitude and action of the electromechanical equipment in the control value field of the control system. The traditional technical means of independent modeling for joint simulation of each discipline are relatively backward, and the speed and solution accuracy of the joint simulation are difficult to meet the needs of the digital twin experiment of the wind tunnel system. This solution uses the multi-domain unified modeling specification language Modelica. The Modelica language uses mathematical equations to describe the physical laws of different disciplines. Based on the topological structure of the system, it builds and integrates models based on the inherent component connection mechanism of the language. It can model and simulate any problem that can be described by differential equations or algebraic equations.

[0085] 2) Task logic simulation sub-model of equipment operation

[0086] The wind tunnel system contains multiple devices. During the wind tunnel test, the devices operate in conjunction with each other. There is a joint relationship between the devices, such as the joint relationship between the mechanical system, the electronic control system and the power system in the equipment, which is mapped to the model level, that is, the joint simulation relationship between electromechanical control and multi-body dynamics. The task logic simulation sub-model includes the electromechanical control simulation model and the multi-body dynamics simulation model of several key devices, and multiple device-level models are coupled to build a system-level simulation model. The complete model structure includes a semi-flexible wall nozzle equipment-level model, a two-throat equipment-level model, a model transport isolation system equipment-level model, a full model vehicle equipment-level model, a half-model vehicle equipment-level model and a system-level simulation model.

[0087] 3) Equipment operation time and resource consumption simulation model

[0088] The consumption of resources such as time, electricity, and liquid nitrogen is the key output indicator of test scheduling. The lifting and lowering process of the model car during the operation of the equipment, and the process of the electric cylinder pushing the connecting rod to produce displacement to adjust the semi-flexible wall surface all consume time and electricity. A mathematical and physical equation model consistent with the mechanism of the real equipment is constructed, and based on the model, the theoretical calculated values ​​of the time and electricity consumed to complete the lifting of a model car of a given mass under the condition of a given model car lifting distance can be solved. Assisted by the corrected influencing factors obtained from the statistical analysis of a large amount of test data, the consumption of resources such as time, electricity, and liquid nitrogen caused by each action during the operation of the equipment can be calculated. These calculated values ​​will serve as an important basis for the result analysis and iterative optimization of the scheduling algorithm model, and these calculated values ​​have reliable theoretical support.

[0089] 4) Simulation model of the impact of test conditions on equipment operation

[0090] Wind tunnels are mainly used for blowing experiments to study the influence of aerodynamic forces on test models under test conditions. A set of test conditions refers to the combination of Mach number, total temperature, total pressure and other parameters required to achieve a specified Reynolds number. Based on the mechanism, the calculation relationship between equipment operating parameters such as compressor speed, flexible wall nozzle and second throat wall shape and test condition parameters is constructed, as well as the influence of aerodynamic experiments on the stable operation of these equipment.

[0091] 2. Construct a mathematical model for scheduling algorithms;

[0092] 1) Constraints and optimization goals of business scenarios

[0093] According to the energy consumption characteristics of wind tunnel operation, a reasonable and coordinated test plan can be formulated, which can not only reduce the wind tunnel occupancy time of the test, improve the test efficiency, but also reduce the energy consumption of wind tunnel operation. In the actual wind tunnel test scenario, scheduling generally has two constraint rules. The first constraint rule is to give priority to the lowest energy consumption, that is, to minimize energy consumption while satisfying the time and equipment constraints. The second constraint rule is that since heating and cooling have a relatively large impact on energy consumption, the wind tunnel does not allow rapid heating and cooling. Once the test starts, the test will generally not stop. The scheduling algorithm cannot arbitrarily arrange and combine all scheduling elements, and reasonable scheduling must be performed under the premise of satisfying the above constraints.

[0094] The optimization goal of scheduling is to minimize the occupancy time of the test wind tunnel, improve test efficiency, and make full use of natural conditions such as natural temperature rise at night under the condition of minimum energy consumption.

[0095] 2) Definition of Scheduling Terms

[0096] The object of test scheduling is the various process tasks included in the complete process of wind tunnel test, which are divided into main tasks and subtasks.

[0097] The main tasks include preparation, cleaning, cooling, testing, rewarming and other process tasks, which is the complete process of wind tunnel test. The order of these main tasks is generally not changeable, and the time consumption, power consumption and gas consumption of each main task usually have historical experience data.

[0098] The subtasks are different working condition test tasks in the "test" main task. Subtask scheduling is to schedule the sequence of these working condition test tasks. The difference between the working conditions is the difference in four parameters: temperature, total pressure, angle of attack, and sideslip angle. Therefore, the temperature sequence is prioritized during scheduling, and then the sequence of total pressure, angle of attack, and sideslip angle is prioritized at the same temperature. The power consumption, time consumption, and gas consumption of the subtask need to be derived by simulation of the mechanism model.

[0099] 3) Input of main task schedule

[0100] The input of the main task scheduling includes the process name, flow program number, process time, energy consumption, etc. These parameters are determined and input by the user on the front-end interface, and the results are passed to the scheduling algorithm, which accumulates the energy consumption. The energy consumption parameters such as the time consumption, power consumption and gas consumption of the main task do not need to be calculated by the mechanism model, but are entered by the user on the interface with reference to historical experience data.

[0101] 4) Input and output of subtask scheduling

[0102] The input of subtask scheduling is the test sequence number, test total temperature, test Mach number, test total pressure, etc. The output of subtask scheduling is the scheduled sequence.

[0103] 5) Subtask Scheduling Algorithm

[0104] There are two implementations of subtask scheduling algorithms: rule-based sorting and optimization algorithm sorting.

[0105] Sorting algorithm 1: rule-based sorting

[0106] The rule-based sorting algorithm sorts the schedule according to rules based on user experience, such as total pressure from high to low test (Mach number from low to high test). The rule-based scheduling only considers the energy consumption of each test of the sub-sequence number, and does not consider the energy consumption of switching sub-sequence numbers.

[0107] Sorting algorithm 2: Optimizing algorithm sorting

[0108] The optimization algorithm is implemented using pymoo in python. Using Pymoo to solve the optimization problem includes the following steps.

[0109] Problem definition: This method uses the element-wise problem definition method to define the optimization problem. It inherits the ElementwiseProblem class in Pymoo.

[0110] Define the optimization algorithm: Pymoo encapsulates many ready-made optimization problem-solving algorithms, all of which are based on object-oriented methods. This method chooses the GA genetic algorithm.

[0111] Define the termination condition: A common termination criterion is to limit the number of iterations of the function. This method sets the termination condition to 10 algorithm iterations. Users can also customize the termination condition in the front-end interface, and the corresponding parameter is smaerStopCondition.

[0112] Optimization calculation: Optimize the problem using the defined algorithm and termination conditions.

[0113] 6) Energy consumption calculation of subtask scheduling

[0114] The calculated values ​​of energy consumption indicators such as time consumption, electricity consumption and gas consumption are simulated and solved using a mechanism model. The mechanism formulas involved in the calculation in the model include the compressor shaft power formula and the liquid nitrogen consumption formula. The shaft power is calculated through the Mach number, total temperature and total pressure based on the compressor shaft power calculation formula, and the energy consumption (electricity) is calculated based on the shaft power. The energy consumption (nitrogen) is calculated through the Mach number, total temperature and total pressure based on the liquid nitrogen consumption calculation formula.

[0115] The complete test energy consumption includes not only the energy consumption corresponding to each test condition, but also the energy consumption during the test condition switching process. Only in this way can the intelligent optimization algorithm calculate the energy consumption of different test sequences, thereby comparing the energy consumption of different test sequences and obtaining the optimal sequence with the minimum energy consumption.

[0116] 3. Develop a scheduling logic driver that can call mathematical models and mechanism models and automatically iterate and optimize;

[0117] The scheduling logic driver is developed based on Python language. Its main function is to select appropriate subtask sorting algorithms based on the constraints defined in the input parameters and business characteristics, and call the scheduling algorithm mathematical model to generate a set of scheduling schemes that meet the constraints, and call the mechanism model to deduce and calculate the energy consumption calculation value of each scheduling scheme, and feed these energy consumption calculation values ​​back to the scheduling algorithm to select or modify the scheduling scheme. This process also includes some parameter sensitivity analysis, parameter automatic adjustment, genetic algorithm optimization and other calculation processes. Until the optimal scheduling scheme is formed. The scheduling logic driver mainly relies on computer programs to implement this iterative process.

[0118] When applied to specific projects, it also supports automatic acquisition of input parameters through rule template import and test outline and test task book analysis, as well as output of the optimal scheduling plan to the experimental scheduling plan document and other input and output interface packaging processes. The scheduling logic driver operation flow chart is shown in Figure 2 .

[0119] 4. Develop graphical interface application software, integrate scheduling logic drivers, encapsulate mathematical models and mechanism models, and form software programs that can interact with humans and machines.

[0120] The experimental scheduling optimization system has a configurable graphical user interface, which includes input controls and a monitoring interface. It receives the optimization goals and constraints of user income through a series of input controls, selects the scheduling optimization algorithm, and enters some configuration parameters required by the algorithm. Then, the scheduling is started with one click through the interface controls. At this time, the scheduling logic driver will be called to start the scheduling iterative optimization process of the mathematical model and mechanism model. The final scheduling deduction results will also be intuitively presented in the form of a color block diagram.

[0121] The scheduling algorithm script is developed in Python. Its deployment process mainly involves the installation and deployment of the Python interpreter. The operation and monitoring of the scheduling algorithm can be viewed directly in the graphical user interface. The output logs and iteration information during the algorithm operation will be printed in real time in the interface monitoring window.

[0122] In terms of technical approach, the transmission of input information generates a json file through the interface module, the scheduling algorithm script file reads the json file, the output result of the scheduling algorithm processing is written into the json file, and the interface module reads the json file for visual display.

[0123] The above solution is only an illustration of a preferred embodiment, but is not limited thereto. When implementing the present invention, appropriate replacement and / or modification can be performed according to user needs.

[0124] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and the embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily realized. Therefore, without departing from the general concept defined by the claims and equivalent scope, the present invention is not limited to the specific details and the illustrations shown and described here.

Claims

1. A test scheduling optimization method based on a mechanism model, characterized in that: include: S1. Establish a wind tunnel system test system mechanism model, define its modular architecture according to the system composition, and describe the task execution and equipment response mechanism; S2. Use discrete process planning algorithm or scheduling algorithm to build a scheduling algorithm model; S3, calling the scheduling algorithm model and the wind tunnel test system mechanism model through the scheduling logic driver to realize automatic iterative optimization and generate the operation plan, and testing the operation effect through the verification plan; Associate and / or integrate with the software platform to build a configurable visual monitoring interface; S4, through the graphical interface application software, integrate the scheduling logic driver, encapsulate the mathematical model and mechanism model; Among them, the operation of the scheduling algorithm script is based on the user-defined scheduling configuration, and multiple scheduling rules and multiple objectives are used to optimize the scheduling algorithm, so that the scheduling algorithm script can call the scheduling algorithm and equipment model in parallel to generate the test operation plan in a multi-sample manner; In S1, the wind tunnel test system mechanism model includes: a task logic simulation sub-model of the equipment task execution process, a time and resource consumption sub-model required for various types of equipment to respond to different task processes, a sub-model of the impact of equipment response under different test conditions, and an overall test system model integrating the above sub-models; The task logic simulation sub-model is based on the joint relationship between the mechanical system, the electronic control system and the power system in the equipment, and maps the logical relationship between the equipment operations to the model level; The task logic simulation sub-model includes: an electromechanical control simulation model, a multi-body dynamics simulation model, and a system-level simulation model; The system-level simulation model is constructed by coupling at least two of a semi-flexible wall nozzle equipment-level model, a two-throat equipment-level model, a model transport isolation system equipment-level model, a full model vehicle equipment-level model, and a half-model vehicle equipment-level model; The contents of the scheduling algorithm model include: setting the constraint rules of the business scenario, setting the schedulable elements, setting the input of the scheduling process, setting the input and output of the scheduling sub-process, setting the input of the scheduling algorithm and its parameters, sub-process energy consumption calculation, running the sub-process sorting algorithm, and the sub-process sorting algorithm is an optimization algorithm sorting, including the following contents: Use element-wise problem definition to define the optimization problem; Select GA genetic algorithm to instantiate an optimization problem solving algorithm object; Set 10 algorithm iterations as the termination condition of the optimization algorithm; Optimize the problem using the defined optimization algorithm and termination conditions; The sub-process sorting algorithm is sorted based on the rules provided by the user, and only the resource consumption of each sub-sequence number test is considered in the rule sorting, and the resource consumption of sub-sequence number switching is not considered; The input of the scheduling process is: work number, work name, work time, nitrogen consumption, and power consumption; The inputs of the scheduling subprocess are: test number, test total temperature, test Mach number, and test total pressure; The output of the scheduling subprocess is the scheduled sequence; In S3, the content of the scheduling logic driver includes: S31, integrating the scheduling task process coding in the scheduling planning algorithm with the external input interface retained by the overall model of the test system, and parsing it into the test process logic parameters; S32, integrating the system state requirement code in the scheduling algorithm with the external input interface retained by the overall model of the test system, and parsing it into system model parameters; S33, integrating the objective function and constraint function in the scheduling algorithm with the external output interface retained by the overall model of the test system to realize the call of the simulation results.

2. The test scheduling optimization method based on the mechanism model as claimed in claim 1, characterized in that: The coding format of the design variables in the scheduling algorithm model needs to be consistent with the task process sequence and equipment number information.

3. A test scheduling optimization system, applied to the test scheduling optimization method based on the mechanism model as described in any one of claims 1-2, characterized in that: The architecture of the test scheduling optimization system includes: An external input layer for accepting one or more optimization target parameters; An interactive layer for intuitively displaying the input of external input parameters and the viewing of scheduling results; The model layer is used to describe the system mechanism and can receive real-time data to simulate the system operation status; The optimal scheduling solution output after the model layer iteration is completed is displayed in the output layer of the interface in the form of a color block diagram; The model layer includes: a scheduling algorithm model developed using Python language, and a wind tunnel test system mechanism model developed using Modelica language.

Images