Data-driven environment-friendly gas arc extinguishing system multi-objective optimization method and related equipment

Abstract

The invention discloses a data-driven environment-friendly gas arc extinguishing system multi-objective optimization method and related equipment, and the method comprises the steps: firstly inputting a collected key design variable of an environment-friendly gas arc extinguishing system into a parameterized circuit breaker cold state simulation model for simulation, and outputting an objective function value; constructing and training a data driving model by using the key design variables and the corresponding objective function values to predict and evaluate a design space; and finally, the data driving model is combined with an optimization objective function used for representing the breaking capacity of the circuit breaker, a multi-objective genetic algorithm is adopted to output an optimal design variable, and optimization of the environment-friendly gas arc extinguishing system is achieved. According to the method, based on multi-objective collaborative optimization and a data driving model, the problems of reliability and efficiency in traditional design are solved, the structure of the circuit breaker is modified in a small range so that the breaking performance of the circuit breaker can be remarkably improved, and the reliability and efficiency of the design process of the circuit breaker are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of high-voltage circuit breakers, and in particular relates to a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system and related equipment. Background Art

[0002] High-voltage gas circuit breakers are important equipment for power system control and protection. To ensure the safe and reliable operation of the power grid, the circuit breaker must have the ability to switch capacitive and inductive loads and effectively interrupt the rated short-circuit current, especially in the event of short-circuit faults and circuit breaker terminal faults. During the current interruption, as the moving arc and static arc contacts separate, a high-temperature arc is generated between the electrodes. Gas circuit breakers usually use the strong blowing effect of high-pressure gas in the upstream chamber of the nozzle to achieve arc cooling and extinguishing. Due to the complex structure of the gas circuit breaker, the gas flow in the gas circuit breaker exhibits instability due to the influence of the variable flow path formed by the contacts, nozzles and other structural components. Different nozzle and gas chamber structures and sizes lead to different gas flow fields after contact separation. At the same time, the change of the nozzle structure, under the influence of the transient recovery voltage, will change the electric field distribution between the contacts, thereby affecting the dielectric recovery strength. In addition, when the large current is interrupted, the dynamic distribution of temperature and gas density caused by the hot gas must also be considered. Therefore, gas properties such as temperature, pressure, density, and dielectric strength jointly affect the interruption performance. These characteristics may conflict, and improving only one factor will not meet the current circuit breaker design requirements.

[0003] On the one hand, the existing circuit breaker design method relies solely on experience to make a large number of parameter adjustments and manual testing of key components such as nozzles, which has the defects of poor reliability and relatively low efficiency; on the other hand, the previous optimization goals focused on traditional SF6 gas circuit breakers, and the optimization method will not be applicable after the gas medium is replaced. As a strong greenhouse gas with a global warming potential (GWP) of 23,500 times that of CO2, the use of SF6 is subject to increasingly stringent restrictions. In recent years, C4F7N has become the most promising SF6 alternative gas, with a GWP value of only 2210 and a dielectric strength of about twice that of SF6. Traditional methods such as increasing the filling pressure and reducing the capacity are also insufficient to fully tap the potential of C4F7N in circuit breaker interruption applications.

[0004] It can be seen that the existing circuit breaker design method only relies on experience to perform a large number of parameter adjustments and manual testing of key components such as nozzles, which has the defects of poor reliability and relatively low efficiency. Summary of the invention

[0005] The present invention provides a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system and related equipment to solve the technical problems that the existing circuit breaker design method only relies on experience to perform a large number of parameter adjustments and manual testing of key components such as nozzles, which has poor reliability and relatively low efficiency.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system, comprising:

[0008] S1: Collect key design variables of the environmentally friendly gas arc extinguishing system, input the key design variables into a pre-built parameterized circuit breaker cold state simulation model, and output the objective function value;

[0009] S2: Build and train data-driven models based on key design variables and corresponding objective function values ​​to predict and evaluate the design space;

[0010] S3: Combining the trained data-driven model and the pre-built optimization objective function, a multi-objective genetic algorithm is used to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0011] Furthermore, in S1, the Morris method is used to perform sensitivity analysis on the key design variables collected, and the key design variables are divided into insensitive variables and highly sensitive variables;

[0012] Keep the insensitive variables in the initial data, and use the Latin hypercube sampling method to randomly generate multiple sample points for the highly sensitive variables;

[0013] All sample points generated by highly sensitive variables are input into the pre-built parameterized circuit breaker cold-state simulation model, and the objective function value is output.

[0014] Furthermore, the Morris method is used to perform sensitivity analysis on the key design variables collected, wherein the average value μ and standard deviation σ of the corresponding trajectory basic effect EE are used as global sensitivity indexes to identify the design variables that have a significant impact on the breaking performance; the formulas for the average value μ and standard deviation σ are as follows:

[0015]

[0016] Where X=(x1,x2,...,x k ) is the k-dimensional independent input, Y(X) is the output response, Δ is a value in {1 / (p-1),…,1-1 / (p-1)}, p is the number of layers; i is the i-th key design variable; r is the number of trajectories.

[0017] Further, in S1, the steps of constructing the parameterized circuit breaker cold state simulation model are as follows:

[0018] Draw a two-dimensional model of the circuit breaker according to the initial structure of the circuit breaker and perform mesh generation, set key design variables as variable parameters to build a parametric circuit breaker geometric model, and realize automatic geometric structure modeling and autonomous mesh update of the parametric circuit breaker geometric model;

[0019] The complete physical parameter data of environmentally friendly gases are encoded in arrays and combined with a linear interpolation algorithm to build a real gas model; the standard k-ε turbulence model is used to solve the exchange relationship between the momentum and energy of the gas medium under the turbulence inside the circuit breaker;

[0020] The parameterized circuit breaker geometric model, real gas model and turbulence model are coupled to construct a parameterized circuit breaker cold-state simulation model.

[0021] Furthermore, in S2, a data set with key design variables as input and objective function values ​​as output is established, and a data-driven model is constructed using the support vector regression method, wherein the correlation coefficient R is used. 2 As an indicator of the accuracy of data-driven models:

[0022]

[0023] Where N is the sample size, is the predicted value, y i is the actual value, is the average of the actual values.

[0024] Furthermore, in S3, the optimized design variables and the optimized objective function values ​​are output based on the trained data-driven model, and the trade-offs between multiple optimization objectives are achieved through the Pareto boundary in combination with the pre-built optimization objective function, and the optimal design variables are output;

[0025] The optimization objective of the optimization objective function includes minimizing the ratio of electric field intensity to density and maximizing the average pressure in the compressed air chamber, which is specifically expressed as:

[0026]

[0027] Where Ω is the design space, g(X) represents the volume of the optimized air chamber V Opt , no greater than the reference design V Base ; X is the key design variable;

[0028] Minimize Obj.func.f1(X)=(E / ρ) max ;

[0029]

[0030] E / ρ represents the ratio of electric field strength to gas density; Indicates the average pressure in the compression chamber.

[0031] Furthermore, it also includes:

[0032] S4: Generate an optimized circuit breaker simulation model based on the optimal design variables, and perform real CFD simulation verification on the optimized circuit breaker simulation model.

[0033] A data-driven environmental gas arc extinguishing system multi-objective optimization system, used to implement the steps of the data-driven environmental gas arc extinguishing system multi-objective optimization method, including:

[0034] A simulation module, used for collecting key design variables of the environmentally friendly gas arc extinguishing system, inputting the key design variables into a pre-built parameterized circuit breaker cold state simulation model, and outputting an objective function value;

[0035] A data-driven model building module, used to build and train a data-driven model based on key design variables and corresponding objective function values ​​to predict and evaluate the design space;

[0036] The optimal design variable output module is used to combine the trained data-driven model and the pre-built optimization objective function, and adopt a multi-objective genetic algorithm to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0037] A device comprising:

[0038] Memory for storing computer programs;

[0039] A processor is used to implement the steps of the above-mentioned data-driven environmentally friendly gas arc extinguishing system multi-objective optimization method when executing the computer program.

[0040] A computer-readable storage medium stores a computer program, which, when executed by a processor, is used to implement the steps of the above-mentioned data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The present invention also provides a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system. The method first inputs the key design variables of the environmentally friendly gas arc extinguishing system collected into a parameterized circuit breaker cold simulation model for simulation, thereby outputting the objective function value; the key design variables and the corresponding objective function values ​​are used to construct and train a data-driven model to predict and evaluate the design space; the data-driven model is finally combined with an optimization objective function for characterizing the breaking capacity of the circuit breaker, and a multi-objective genetic algorithm is used to output the optimal design variables, thereby realizing the optimization of the environmentally friendly gas arc extinguishing system; the method is based on multi-objective collaborative optimization and data-driven models, solves the reliability and efficiency problems in traditional designs, slightly modifies the circuit breaker structure to achieve significant improvements in the circuit breaker breaking performance, and improves the reliability and efficiency of the circuit breaker design process. By modifying the objective function and design variables, the method can be further applied to the optimization design process of other electrical equipment, and has wide applicability.

[0043] Preferably, in the present invention, the Morris method is introduced to perform sensitivity analysis on key design variables, the design variables are divided into insensitive variables and highly sensitive variables, and only the highly sensitive variables are sampled and simulated; this method reduces unnecessary calculations, improves simulation efficiency, and ensures the accuracy of the optimization results.

[0044] Further preferably, in the present invention, by calculating the mean value and standard deviation of the basic effect as the global sensitivity index, the design variables that have a significant impact on the breaking performance can be more accurately identified, which provides an important basis for subsequent optimization work and helps to further improve the optimization effect.

[0045] Preferably, in the present invention, the steps of constructing a parameterized circuit breaker cold-state simulation model include parameterization of a geometric model, construction of a real gas model, and solution of a turbulence model, which can more accurately simulate the actual working state of the circuit breaker and provide a reliable basis for subsequent simulation and optimization.

[0046] Preferably, in the present invention, a support vector regression method is used to construct a data-driven model, and the accuracy of the model is evaluated by the correlation coefficient, which can efficiently process a large amount of data and build a high-precision prediction model, providing strong support for subsequent optimization work.

[0047] Preferably, in the present invention, a trade-off between multiple optimization objectives is achieved through the Pareto boundary, that is, the optimal design variables are iteratively calculated and output through a multi-objective genetic algorithm, which can comprehensively consider multiple optimization objectives and find the optimal compromise solution, thereby meeting the overall performance requirements of the system.

[0048] Preferably, in the present invention, at the end of the optimization process, an optimized circuit breaker simulation model is generated based on the optimal design variables, and a real CFD simulation verification is performed, which ensures the reliability and practicality of the optimization results and provides strong support for actual production. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 A schematic diagram of a parameterized circuit breaker cold state simulation model provided by an embodiment of the present invention;

[0050] Figure 2 Max (E / ρ) and Time distribution diagram; where (a) is SF6; (b) is C4F7N mixed gas.

[0051] Figure 3 The Morris sensitivity index μ, μ* and σ provided in the embodiment of the present invention are relative to the output Max (E / ρ) and Among them, (a) is SF6; (b) is C4F7N mixed gas; Among them, the left first and second in (a) are the Max (E / ρ) corresponding to SF6; the right first and second are the Max (E / ρ) corresponding to SF6 (b) The first and second left images are the Max(E / ρ) corresponding to the C4F7N mixed gas; the first and second right images are the Max(E / ρ) corresponding to the C4F7N mixed gas

[0052] Figure 4 The objective function prediction accuracy graph of SF6 provided in the embodiment of the present invention, wherein (a) and (b) use the random forest prediction accuracy graph; (c) and (d) use the SVR data-driven model prediction accuracy graph;

[0053] Figure 5 A convergence process diagram of the Pareto front generated by the optimization process of SF6 and C4F7N gases provided in an embodiment of the present invention; wherein (a) is SF6; (b) is a C4F7N mixed gas;

[0054] Figure 6 A comparison diagram of simulation results of the optimized design and basic design of SF6 and C4F7N gases provided in an embodiment of the present invention; wherein (a) is SF6; (b) is C4F7N mixed gas;

[0055] Figure 7 A flow chart of a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system provided by an embodiment of the present invention;

[0056] Figure 8 A flow chart of a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system provided by the present invention;

[0057] Fig. 9 A structural schematic diagram of a data-driven environmentally friendly gas arc extinguishing system multi-objective optimization system provided by the present invention. DETAILED DESCRIPTION

[0058] The present invention provides a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system, such as Figure 8 As shown, the following steps are included:

[0059] S1: Collect the key design variables of the environmentally friendly gas arc extinguishing system, input the key design variables into the pre-built parameterized circuit breaker cold simulation model, and output the objective function value; specifically, use the Morris method to perform sensitivity analysis on the collected key design variables, and divide the key design variables into insensitive variables and highly sensitive variables;

[0060] Keep the insensitive variables in the initial data, and use the Latin hypercube sampling method to randomly generate multiple sample points for the highly sensitive variables;

[0061] All sample points generated by highly sensitive variables are input into the pre-built parameterized circuit breaker cold-state simulation model, and the objective function value is output.

[0062] Here, the Morris method is used to perform sensitivity analysis on the key design variables collected. The average value μ and standard deviation σ of the corresponding trajectory basic effect EE are used as global sensitivity indexes to identify the design variables that have a significant impact on the breaking performance. The formulas for the average value μ and standard deviation σ are as follows:

[0063]

[0064] Where X=(x1,x2,...,x k ) is the k-dimensional independent input, Y(X) is the output response, Δ is a value in {1 / (p-1),…,1-1 / (p-1)}, p is the number of layers; i is the i-th key design variable; r is the number of trajectories.

[0065] The construction steps of the parameterized circuit breaker cold state simulation model are as follows:

[0066] Draw a two-dimensional model of the circuit breaker according to the initial structure of the circuit breaker and perform mesh generation, set key design variables as variable parameters to build a parametric circuit breaker geometric model, and realize automatic geometric structure modeling and autonomous mesh update of the parametric circuit breaker geometric model;

[0067] The complete physical parameter data of environmentally friendly gases are encoded in arrays and combined with a linear interpolation algorithm to build a real gas model; the standard k-ε turbulence model is used to solve the exchange relationship between the momentum and energy of the gas medium under the turbulence inside the circuit breaker;

[0068] The parameterized circuit breaker geometric model, real gas model and turbulence model are coupled to construct a parameterized circuit breaker cold-state simulation model.

[0069] S2: Build and train data-driven models based on key design variables and corresponding objective function values ​​to predict and evaluate the design space.

[0070] In S2, a data set with key design variables as input and objective function values ​​as output is established, and a data-driven model is constructed using the support vector regression method. 2 As an indicator of the accuracy of data-driven models:

[0071]

[0072] Where N is the sample size, is the predicted value, y i is the actual value, is the average of the actual values.

[0073] S3: Combining the trained data-driven model and the pre-built optimization objective function, a multi-objective genetic algorithm is used to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; among which, the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0074] In S3, the optimized design variables and the optimized objective function values ​​are output based on the trained data-driven model. Combined with the pre-built optimization objective function, the trade-offs between multiple optimization objectives are achieved through the Pareto boundary, and the optimal design variables are output;

[0075] Among them, the optimization objectives of the optimization objective function include minimizing the ratio of electric field intensity to density and maximizing the average pressure in the compressed air chamber, which can be specifically expressed as:

[0076]

[0077] Where Ω is the design space, g(X) represents the volume of the optimized air chamber V Opt , no greater than the reference design V Base ; X is the key design variable;

[0078] Minimize Obj.func.f1(X)=(E / ρ) max ;

[0079]

[0080] E / ρ represents the ratio of electric field strength to gas density; Indicates the average pressure in the compression chamber.

[0081] S4: Generate an optimized circuit breaker simulation model based on the optimal design variables, and perform real CFD simulation verification on the optimized circuit breaker simulation model.

[0082] like Fig. 9 As shown, the present invention also provides a data-driven multi-objective optimization system for an environmentally friendly gas arc extinguishing system, including: a simulation module, used to collect key design variables of the environmentally friendly gas arc extinguishing system, input the key design variables into a pre-built parameterized circuit breaker cold-state simulation model, and output the objective function value; a data-driven model construction module, used to construct and train a data-driven model based on the key design variables and the corresponding objective function values ​​to predict and evaluate the design space; an optimal design variable output module, used to combine the trained data-driven model and the pre-built optimization objective function, adopt a multi-objective genetic algorithm, and output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein, the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0083] The present invention also provides a device, comprising: a memory for storing a computer program; a processor for implementing the steps of the data-driven environmentally friendly gas arc extinguishing system multi-objective optimization method when executing the computer program.

[0084] When the processor executes the computer program, the steps of implementing the above-mentioned data-driven multi-objective optimization of the environmentally friendly gas arc extinguishing system are implemented, for example: collecting key design variables of the environmentally friendly gas arc extinguishing system, inputting the key design variables into a pre-built parameterized circuit breaker cold simulation model, and outputting the objective function value; constructing and training a data-driven model based on the key design variables and the corresponding objective function values ​​to predict and evaluate the design space; combining the trained data-driven model and the pre-built optimization objective function, using a multi-objective genetic algorithm, outputting the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0085] Alternatively, when the processor executes the computer program, it implements the functions of each module in the above-mentioned system, for example: a simulation module, which is used to collect key design variables of the environmentally friendly gas arc extinguishing system, input the key design variables into a pre-built parameterized circuit breaker cold-state simulation model, and output the objective function value; a data-driven model construction module, which is used to construct and train a data-driven model based on key design variables and corresponding objective function values ​​to predict and evaluate the design space; an optimal design variable output module, which is used to combine the trained data-driven model and the pre-built optimization objective function, and use a multi-objective genetic algorithm to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein, the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0086] Exemplarily, the computer program can be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete preset functions, and the instruction segments are used to describe the execution process of the computer program in the data-driven environmental gas arc extinguishing system multi-objective optimization device. For example, the computer program can be divided into a simulation module, a data-driven model construction module, and an optimal design variable output module; the specific functions of each module are as follows: a simulation module, which is used to collect key design variables of the environmental gas arc extinguishing system, input the key design variables into a pre-built parameterized circuit breaker cold simulation model, and output the objective function value; a data-driven model construction module, which is used to build and train a data-driven model based on key design variables and corresponding objective function values ​​to predict and evaluate the design space; an optimal design variable output module, which is used to combine the trained data-driven model and the pre-built optimization objective function, and use a multi-objective genetic algorithm to output the optimal design variables to achieve the optimization of the environmental gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

[0087] The data-driven environmental gas arc extinguishing system multi-objective optimization device can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The data-driven environmental gas arc extinguishing system multi-objective optimization device can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above is an example of a data-driven environmental gas arc extinguishing system multi-objective optimization device, and does not constitute a limitation on the data-driven environmental gas arc extinguishing system multi-objective optimization device, and can include more components than the above, or a combination of certain components, or different components. For example, the data-driven environmental gas arc extinguishing system multi-objective optimization device can also include input and output devices, network access devices, buses, etc.

[0088] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The processor is the control center of the multi-objective optimization of the data-driven environmental gas arc extinguishing system, and uses various interfaces and lines to connect various parts of the multi-objective optimization device of the entire data-driven environmental gas arc extinguishing system.

[0089] The memory can be used to store the computer program and / or module, and the processor implements various functions of the data-driven environmentally friendly gas arc extinguishing system multi-objective optimization device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0090] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0091] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the data-driven multi-objective optimization method of an environmentally friendly gas arc extinguishing system are implemented.

[0092] The module / unit integrated in the data-driven environmentally friendly gas arc extinguishing system multi-objective optimization system can be stored in a computer-readable storage medium if it is implemented in the form of a software functional unit and sold or used as an independent product.

[0093] Based on such understanding, the present invention implements all or part of the processes in the above-mentioned data-driven environmental gas arc extinguishing system multi-objective optimization method, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned data-driven environmental gas arc extinguishing system multi-objective optimization method can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or preset intermediate form, etc.

[0094] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0095] It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0096] The present invention will be further described below in conjunction with the embodiments and drawings:

[0097] Example 1

[0098] As described in the background technology, there are reliability and efficiency problems in traditional optimization methods; and traditional methods such as increasing the filling pressure and reducing the capacity are not enough to fully tap the potential of C4F7N in circuit breaker interruption applications; therefore, it is necessary to propose an efficient and reliable multi-objective optimization design method for new environmentally friendly gas circuit breakers to support the equipment development and engineering application of high-performance high-voltage environmentally friendly gas switches.

[0099] In order to solve the above problems, the present invention provides a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system. The method first determines the optimization objective functions of two circuit breakers, and uses the Morris method to quickly screen the design variables to facilitate the dimensionality reduction of the design space; secondly, a parameterized GCB geometric model is established, and a real CFD simulation is performed after reasonable sampling of the design variables; subsequently, an accurate data-driven model is constructed to predict and evaluate the design space; finally, the multi-objective optimal GCB geometry is iteratively searched based on the response value of the data-driven model and the multi-objective genetic algorithm; the CFD evaluation of the optimization results confirms the effectiveness of the design framework.

[0100] like Figure 7As shown, this embodiment provides a data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system, specifically including:

[0101] Step 1: Propose multiple optimization objective functions that can characterize the breaking performance of the circuit breaker: that is, propose two optimization objective functions. One is to reduce the ratio of electric field strength to gas density, which can increase the breakdown voltage, that is, improve the ability of the circuit breaker to break capacitive current; the other is to increase the average air pressure in the compressed air chamber, which can form a strong air blowing effect at the current zero zone to accelerate the recovery of gas medium strength and improve the ability of the circuit breaker to break large short-circuit current. Specifically:

[0102] (1) When interrupting a small capacitive short-circuit current, the arc effect has little effect on the gas flow. The circuit breaker needs to have good electrical breaking capacity and withstand a recovery voltage of up to twice the rated voltage on both sides without dielectric breakdown. The interaction between the gas flow density and the electrical stress in the nozzle area determines the dielectric breakdown of the gas circuit breaker when interrupting capacitive current. According to the empirical formula, the critical breakdown voltage U BD It is usually a function of the ratio of the electric field strength E to the gas density p (E / ρ):

[0103]

[0104] Where a and b are experimental coefficients.

[0105] Therefore, one of the optimization goals is to improve the circuit breaker's ability to break capacitive current, that is, to increase the breakdown voltage and reduce E / ρ. The optimization objective function expression is:

[0106] Minimize Obj.func.f1(X)=(E / ρ) max

[0107] (2) When interrupting large-capacity short-circuit currents of up to tens of kA, the circuit breaker needs to have good thermal breaking capability and be able to withstand the rapidly rising transient recovery voltage between the contacts without arc reignition. Therefore, the thermal breaking performance requires a strong gas cooling effect to reduce the dielectric conductivity and minimize the post-arc current. The gas cooling effect depends on the pressure upstream of the arc area or the pressure ratio between the upstream and downstream areas. The higher the pressure in the gas chamber, the faster the gas flow rate for convection cooling and the faster the dielectric heat dissipation, thereby improving the interruption capacity of the circuit breaker. The second optimization objective function is determined as:

[0108]

[0109] Step 2: Determine the key design variables of the circuit breaker and their value ranges, and screen and sort the design variables based on sensitivity analysis;

[0110] (1) According to the actual structure of the circuit breaker and previous experience in optimization design, the key design variables of the circuit breaker are determined, including the upstream and downstream lengths, radius and pressure cylinder volume of the nozzle. At the same time, the reference value and value range of the variables are determined by considering the typical circuit breaker structure foundation and the actual production allowable range.

[0111] (2) Through sensitivity analysis, the uncertainty in the model output can be qualitatively attributed to the model input, thereby identifying the design variables with less impact on the output and determining the interaction effects of the model input variables. The Morris method based on the random OAT (one variable at a time) experiment is used to implement global sensitivity analysis with the mean μ and standard deviation σ of the corresponding trajectory elementary effects (EE) as the global sensitivity index, and the design variables are ranked by sensitivity.

[0112]

[0113] Where X=(x1,x2,...,x k ) is the k-dimensional independent input of the model, Y(X) is the output response, Δ is a value in {1 / (p-1),…,1-1 / (p-1)}, and p is the number of layers.

[0114] The input parameters are randomly sampled at p levels of the k-dimensional unit hypercube. The operation is repeated for each input at a random starting point, called a trajectory, and the corresponding EE can then be calculated. For the r constructed trajectories, the sensitivity index can be derived:

[0115]

[0116]

[0117] In the above formula, the arithmetic mean and absolute mean of EE of the i-th variable are expressed as μ i and The larger the value, the stronger the sensitivity of the variable. i is the standard deviation of the EE, which is used to measure the nonlinearity and / or interaction effects of the ith variable. i If , it is large, it is considered to have a nonlinear effect or to interact with at least one other variable; the Morris method involves fewer model executions and can provide qualitative sensitivity results.

[0118] Step 3: Based on the initial structure of the circuit breaker, a parameterized circuit breaker cold state simulation model is constructed considering key design variables;

[0119] (1) According to the initial structure of the circuit breaker, the ANSYS modeling tool is used to draw the two-dimensional model of the circuit breaker and perform mesh division. The key design variables are set as variable parameters, and a parametric circuit breaker geometric model is constructed to achieve automatic modeling of the model's geometric structure and autonomous mesh update. The model structure can be dynamically modified and the regional mesh can be automatically updated according to the value range of the design variables.

[0120] (2) Build a real gas model, encode the complete physical parameter data of environmental gas into an array and combine it with a linear interpolation algorithm to import the physical parameter data of gas at different temperatures and pressures (including gas density ρ, enthalpy h, speed of sound c, viscosity coefficient μ, thermal conductivity λ, electrical conductivity σ and constant pressure specific heat Cp, etc.) into the control equations for calculation and solution. The standard k-ε turbulence model is used to solve the exchange relationship between the momentum and energy of the gas medium under the turbulence inside the circuit breaker.

[0121] (3) The above-mentioned parameterized circuit breaker geometric model is coupled with mathematical models such as real gas model and turbulence model to construct the corresponding circuit breaker magnetohydrodynamic cold simulation model.

[0122] Step 4: Effectively sample the design variables and perform real CFD simulation to obtain the corresponding relationship between the objective function and the design variables;

[0123] (1) Based on the sensitivity ranking of the design variables in step 2, the insensitive variables are kept at the base value, while the highly sensitive variables are subjected to Latin hypercube experimental design to generate 100 uniform design variable sample sets. At the same time, a model sample set corresponding to the design variable samples is generated based on the parameterized model.

[0124] (2) For all sampling points, computational fluid dynamics software is used to perform real CFD simulation, that is, input into the parameterized circuit breaker cold simulation model for simulation; the dynamic distribution characteristics of the pressure field, airflow field, electromagnetic field, etc. inside the arc extinguishing chamber of the high-voltage circuit breaker are obtained, and the corresponding relationship between the objective function and the design variables is extracted.

[0125] Step 5: Use the design variable values ​​and objective function values ​​at the sampling points as training data to build a data-driven model to predict and evaluate the design space;

[0126] Based on the real simulation results obtained in step 4, a data set with design variable values ​​as input and objective function values ​​as output is established, and the data set is randomly divided into training data and test data in a ratio of 80:20. In order to solve the challenge of fitting small sample data, the support vector regression (SVR) method is used to build a data-driven model, map the input space to a high-dimensional feature space, and realize low-dimensional nonlinear regression prediction by determining the optimal hyperplane.

[0127] The design variables were used as input, normalized using the min-max method, and the two objective functions were used as output. The optimal hyperparameters of the data-driven model were determined based on the Bayesian optimization principle. In addition, the k-fold (k=5) cross validation (CV) method was used to ensure that the model had good generalization ability and avoid overfitting. The correlation coefficient (R 2 ) as an indicator of the accuracy of the data-driven model. 2 The closer it is to 1, the higher the prediction accuracy of the model.

[0128]

[0129] Where N is the sample size, is the predicted value, y i is the actual value, y i is the average of the actual values.

[0130] Step 6: Determine the optimal design variables of the circuit breaker by combining the response value of the data-driven model with the multi-objective genetic algorithm;

[0131] The data-driven model established in step 5 outputs a large amount of corresponding data of the optimized design variables and the optimized objective function values, and the trade-offs between multiple optimization objectives are achieved through the Pareto boundary to select the optimal design variable solution. The optimization goal of the present invention is to minimize the ratio of electric field intensity to density and maximize the average pressure in the compressed air chamber, which is expressed as:

[0132]

[0133] Where Ω is the design space, g(X) represents the volume of the optimized air chamber V Opt Not greater than the reference design V Base .

[0134] Step 7: Perform a real CFD simulation on the circuit breaker structure model based on the optimal design variables to verify the reliability of the results.

[0135] The optimized circuit breaker simulation model is generated based on the optimal design variable values ​​obtained by the multi-objective optimization model, and a real CFD simulation is carried out to observe whether the values ​​of the optimization objective function obtained by the simulation correspond to the values ​​calculated by the data-driven model, so as to verify the reliability of the results.

[0136] Example 2

[0137] In this embodiment, the data-driven environmental gas arc extinguishing system multi-objective optimization method provided in Embodiment 1 is implemented and applied in combination with a specific example, as follows:

[0138] Based on the structure of a compressed air circuit breaker, ANSYS Workbench software is used to establish Figure 1The parameterized geometric model shown in the figure is meshed. Combining design experience and actual engineering requirements, the design variables that affect the airflow and electric field distribution include the upstream section of the nozzle (radius X4, length X2, inclination X3), the downstream section (radius X5, length X6 and X7, inclination X1) and the volume of the compressor cylinder (length X8, height X9). The value ranges of each variable are listed in Table 1.

[0139] Table 1 shows the range of key design variables.

[0140]

[0141] Firstly, based on the basic structure of the circuit breaker, a real gas model of SF6 and C4F7N mixed gas is introduced, and a real CFD cold simulation is carried out to understand the physical processes related to the optimization objectives. Figure 2 The maximum E / ρ value and the average pressure change of the compressed gas chamber of the SF6 and C4F7N mixed gas after the moving and static contacts are separated. The C4F7N mixed gas shows obvious pressure oscillations, which is mainly due to the high sound velocity of the main component CO2, which is conducive to rapid flow. This causes the pressure to increase rapidly to a higher level and then decay faster than SF6.

[0142] Then, the Morris method based on random OAT (one variable at a time) experiment is used to perform sensitivity analysis on the circuit breaker design variables under different gas conditions. The mean values ​​μ, μ* and standard deviations σ of the corresponding trajectory elementary effects (EE) of different design variables of the two gases are as follows: Figure 3 As shown. Figure 3 The analysis shows that under the SF6 gas range, the design variables X7 and X8 are both located in the lower left corner of the target 1 scatter plot, which means that they have an effect on (E / ρ) max The influence of X8 on the overall output of the model is relatively small. However, in objective 2, X8 shows stronger sensitivity, while the less sensitive variables are X4 and X7. Considering these two objectives, we determined that the influence of X7 on the overall output of the model is negligible. Similarly, the same analysis was performed on the C4F7N mixed gas, where X6, X7 and X8 were considered to be relatively insensitive design variables for the two objectives. Therefore, the design variables with less influence were kept as initial data, and the highly sensitive design variables were randomly generated with 100 sample points through the Latin hypercube sampling method. As shown in Table 2.

[0143] Table 2 is a table of design variables for different gas high sensitivity

[0144]

[0145] Realistic CFD simulations were performed at all sampling points to generate data sets and build data-driven models. Figure 4 This is the prediction accuracy diagram of the SVR data-driven model in SF6. The diagonal line indicates the highest prediction accuracy. It can be seen that the prediction accuracy of SVR is very high. The R 2 They are 0.9329 and 0.9997 respectively.

[0146] Based on the constructed SVR data-driven model, the multi-objective optimization of the circuit breaker is carried out. The convergence history of the Pareto frontier generated by the two gases is shown in Figure 5 As shown, the points of different colors represent the state of convergence of the optimization at different generations. The NSGA-Ⅱ algorithm performs 1000 iterations for each gas type, and 50 individuals are evaluated in each iteration. Therefore, the performance of up to 50,000 different circuit breaker geometries was evaluated, and convergence was observed at the 500th iteration. The green star on the Pareto frontier represents the optimal design selected based on the actual engineering applicability. For comparison, the objective function performance of the initial baseline design is represented by a red star. By visually comparing the geometric distance between the two stars, it can be seen that the optimization effect is obvious, and the two conflicting objectives can be considered comprehensively. In order to further clarify the effectiveness of the data-driven model and the optimization procedure, a real CFD evaluation was performed on the selected optimized design. The actual values ​​obtained by CFD simulation and the predicted values ​​of the SVR model are shown in Table 3. The maximum error of the predicted value is 0.573% and the minimum error is 0.097%, which once again confirms the accuracy of the SVR data-driven model.

[0147] Table 3 is the optimization design results table

[0148]

[0149] like Figure 5 As shown, Figure 5 The simulation results of the two objective functions under the optimized design structure and the original design structure. After optimization, compared with the baseline design, the (E / ρ) of SF6 max The average pressure in the cavity increased by about 14.1%. Similarly, for the C4F7N mixed gas, (E / ρ) max It decreased by about 3.4%, and the average air pressure in the compression chamber increased by about 4.2%.

[0150] The above embodiment is only one of the implementation methods that can realize the technical solution of the present invention. The scope of protection claimed by the present invention is not limited only to this embodiment, but also includes changes, replacements and other implementation methods that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention.

Claims

1. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system, characterized in that: include: S1: Collect key design variables of the environmentally friendly gas arc extinguishing system, input the key design variables into a pre-built parameterized circuit breaker cold state simulation model, and output the objective function value; S2: Build and train data-driven models based on key design variables and corresponding objective function values ​​to predict and evaluate the design space; S3: Combining the trained data-driven model and the pre-built optimization objective function, a multi-objective genetic algorithm is used to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

2. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 1, characterized in that: In S1, the Morris method is used to perform sensitivity analysis on the key design variables collected, and the key design variables are divided into insensitive variables and highly sensitive variables; Keep the insensitive variables in the initial data, and use the Latin hypercube sampling method to randomly generate multiple sample points for the highly sensitive variables; All sample points generated by highly sensitive variables are input into the pre-built parameterized circuit breaker cold-state simulation model, and the objective function value is output.

3. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 2, characterized in that: The Morris method is used to perform sensitivity analysis on the key design variables collected, wherein the average value μ and standard deviation σ of the corresponding trajectory basic effect EE are used as global sensitivity indexes to identify the design variables that have a significant impact on the breaking performance; the formulas for the average value μ and standard deviation σ are as follows: Where X=(x1,x2,...,x k ) is the k-dimensional independent input, Y(X) is the output response, Δ is a value in {1 / (p-1),…,1-1 / (p-1)}, p is the number of layers; i is the i-th key design variable; r is the number of trajectories.

4. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 1, characterized in that: In S1, the steps of constructing the parameterized circuit breaker cold state simulation model are as follows: Draw a two-dimensional model of the circuit breaker according to the initial structure of the circuit breaker and perform mesh generation, set key design variables as variable parameters to build a parametric circuit breaker geometric model, and realize automatic geometric structure modeling and autonomous mesh update of the parametric circuit breaker geometric model; The complete physical parameter data of environmentally friendly gases are encoded in arrays and combined with a linear interpolation algorithm to build a real gas model; the standard k-ε turbulence model is used to solve the exchange relationship between the momentum and energy of the gas medium under the turbulence inside the circuit breaker; The parameterized circuit breaker geometric model, real gas model and turbulence model are coupled to construct a parameterized circuit breaker cold-state simulation model.

5. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 1, characterized in that: In S2, a data set with key design variables as input and objective function values ​​as output is established, and a data-driven model is constructed using the support vector regression method. 2 As an indicator of the accuracy of data-driven models: Where N is the sample size, is the predicted value, y i is the actual value, is the average of the actual values.

6. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 1, characterized in that: In S3, the optimized design variables and the optimized objective function values ​​are output based on the trained data-driven model. Combined with the pre-built optimization objective function, the trade-offs between multiple optimization objectives are achieved through the Pareto boundary, and the optimal design variables are output; The optimization objective of the optimization objective function includes minimizing the ratio of electric field intensity to density and maximizing the average pressure in the compressed air chamber, which is specifically expressed as: Where Ω is the design space, g(X) represents the volume of the optimized air chamber V Opt , no greater than the reference design V Base ; X is the key design variable; Minimize Obj.func.f1(X)=(E / ρ) max ; E / ρ represents the ratio of electric field intensity to gas density; Indicates the average pressure in the compression chamber.

7. A data-driven multi-objective optimization method for an environmentally friendly gas arc extinguishing system according to claim 1, characterized in that: Also includes: S4: Generate an optimized circuit breaker simulation model based on the optimal design variables, and perform real CFD simulation verification on the optimized circuit breaker simulation model.

8. A data-driven environmental gas arc extinguishing system multi-objective optimization system, used to implement the steps of the data-driven environmental gas arc extinguishing system multi-objective optimization method according to any one of claims 1 to 7, characterized in that: include: A simulation module, used for collecting key design variables of the environmentally friendly gas arc extinguishing system, inputting the key design variables into a pre-built parameterized circuit breaker cold state simulation model, and outputting an objective function value; A data-driven model building module, used to build and train a data-driven model based on key design variables and corresponding objective function values ​​to predict and evaluate the design space; The optimal design variable output module is used to combine the trained data-driven model and the pre-built optimization objective function, and adopt a multi-objective genetic algorithm to output the optimal design variables to achieve the optimization of the environmentally friendly gas arc extinguishing system; wherein the optimization objective function is used to characterize the breaking capacity of the circuit breaker.

9. A device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the data-driven environmentally friendly gas arc extinguishing system multi-objective optimization method described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it is used to implement the steps of the data-driven environmentally friendly gas arc extinguishing system multi-objective optimization method described in any one of claims 1-7.

Images