Switch cabinet non-intrusive temperature monitoring method based on adaptive genetic algorithm and parameter optimization system

Abstract

The invention discloses a switch cabinet non-intrusive temperature monitoring method based on an adaptive genetic algorithm and a parameter optimization system. The method comprises the following steps: establishing a switch cabinet heat transfer model and carrying out a model verification experiment; designing confirmation indexes considering the model and experimental data; model parameters are optimized through an adaptive genetic algorithm, and crossover and mutation probabilities are dynamically adjusted to improve precision; and finally completing model authentication. The passive wireless sensor is adopted to collect the temperature of the cabinet surface of the switch cabinet, the heat conduction equation and the global optimization algorithm are combined, real-time online monitoring and high-precision prediction of the temperature of equipment and a contact in the switch cabinet are achieved, the defect that a temperature measuring sensor needs to be installed in a power failure mode is overcome, and follow-up maintenance is more convenient.

Description

Technical Field

[0001] The present invention belongs to the technical field of switchgear temperature monitoring, and particularly relates to a non-invasive temperature monitoring method and parameter optimization system for switchgear based on an adaptive genetic algorithm. Background Art

[0002] There are a large number of electrical connection points in each link of the "generation, transformation, transmission, distribution, and use" of the power system. Due to reasons such as contact resistance and mechanical vibration, local overheating is likely to occur. According to statistics, about 40% of electrical accidents are caused by overheating of connection points. If such faults are not detected in time, it may lead to arc discharge, equipment damage, and even fire, seriously threatening the safe operation of the power grid. Although existing technologies such as optical fiber temperature measurement or infrared imaging have been partially applied to temperature monitoring, they still have significant defects in terms of real-time performance, accuracy, and adaptability. Optical fiber temperature measurement requires power outage for installation, while infrared imaging has high requirements for surface cleanliness.

[0003] Currently, the temperature monitoring of power equipment mainly relies on manual inspections or fixed sensors. Manual inspections are inefficient and difficult to cover unattended scenarios; traditional sensors such as contact thermocouples require complex wiring and are vulnerable to electromagnetic interference. Although wireless sensors do not require wiring, they require power outage of the switchgear, and it is difficult to meet the installation conditions. In the existing market, although the infrared temperature measurement system occupies the mainstream of the market, with a proportion of over 70%, its static model is difficult to adapt to dynamic thermal load changes, resulting in insufficient long-term monitoring stability. In addition, the infrared temperature measurement has a poor effect on monitoring the internal temperature of the switchgear and is difficult to reflect the temperature of the internal contacts of the switchgear. Moreover, existing non-invasive temperature measurement technologies mostly rely on single-parameter calibration and lack comprehensive modeling of complex heat transfer paths such as radiation and convection, further restricting the improvement of accuracy.

[0004] In the prior art, for example, Patent CN202110968187.4 proposes a non-invasive temperature measurement technology based on the fusion of infrared and visible light images, which extracts device state information through dual-frequency image analysis and calculates the lumped heat source temperature. However, this method relies on complex image processing algorithms, is sensitive to device surface fouling and lighting conditions, and does not solve the problem of adaptive adjustment of model parameters under dynamic working conditions. In addition, Li Pengcheng pointed out in "Defect Diagnosis and Analysis of Substation Equipment Based on Infrared Temperature Measurement Technology" that although traditional infrared temperature measurement technology can achieve non-contact monitoring, it has limitations such as the temperature measurement accuracy being interfered by the ambient temperature and being unable to penetrate the cabinet to monitor internal heat sources.

[0005] Therefore, it is necessary to propose a non-invasive temperature monitoring method and parameter optimization system for switchgear based on an adaptive genetic algorithm to solve the above problems. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a non-invasive temperature monitoring method and parameter optimization system for switchgear based on an adaptive genetic algorithm, which can real-time correct model parameters under dynamic working conditions and take into account the multi-modal characteristics of the heat transfer path to achieve accurate prediction of the internal temperature of the switchgear and fault warning, thereby solving the problems of complex wiring of optical fiber temperature measurement for existing switchgear and inconvenient installation requiring power outage.

[0007] To achieve the above technical effects, the technical solution adopted by the present invention is as follows: A non-invasive temperature monitoring method for switchgear based on an adaptive genetic algorithm, comprising the following steps: S1, establish a non-invasive temperature measurement model based on the heat transfer principle, and carry out model confirmation and certification tests; S2, analyze the corresponding relationship between the calculation results of the non-invasive temperature measurement model and the certification test results; S3, based on the existing test data, construct a temperature measurement model for model parameter optimization and accuracy improvement; S4, based on the proposed model confirmation index, carry out model confirmation on the parameter-optimized model again.

[0008] Preferably, in step S1, the certification test is carried out under multiple test settings. The response measurement of the certification verification experiment is carried out at three positions to collect the temperature responses at different positions on the surface. And the on-site test switchgear is operated under different loads, and the contact temperatures of the busbar chamber, circuit breaker chamber, cable chamber and instrument panel chamber inside the switchgear are monitored and recorded in real time through the optical fiber sensor or contact sensor inside the switchgear; for different working conditions of the switchgear, the temperature change data is regularly collected and recorded at a data collection period of 1 h.

[0009] Preferably, in step S2, analyzing the corresponding relationship between the calculation results of the non-invasive temperature measurement model and the certification test results specifically includes: Design a model confirmation index, and the model confirmation index is designed based on all corresponding relationships and the accuracy of the calculation model and test data; Based on the non-invasive temperature measurement model and test data, estimate the index values under different model confirmation scenarios, and judge whether the compliance degree and accuracy of the model for the test meet the requirements. If not, further correct the model parameters through the adaptive genetic algorithm to improve the accuracy of the model for temperature monitoring; the error index selects the mean absolute error MAE to be less than 2 °C and the RMSE to be less than 3 °C in the switchgear temperature prediction scenario.

[0010] Preferably, in step S3, constructing a temperature measurement model for model parameter optimization and accuracy improvement based on the existing test data includes: With the minimization of the confirmation index as the optimization goal, and the thermal conductivity, geometric parameters, and environmental factors as the design variables, the model parameter identification and update are realized through the adaptive genetic algorithm, and the crossover probability Pc and the mutation probability Pm are dynamically adjusted according to the fitness.

[0011] Preferably, in step S4, based on the proposed model confirmation index, the model confirmation for the parameter-optimized model is carried out again, including: After the compliance is met, model certification is carried out, including accuracy verification; when the error index MAE is less than 2 °C, the model coefficients and reliability are confirmed; the stability and consistency of the model under different conditions, different times, or different input parameters are evaluated.

[0012] Preferably, establishing a non-invasive temperature measurement model based on the heat transfer principle specifically includes: According to the first law of thermodynamics, the heat balance equation is obtained: imported heat + internal heat source heat generation = increase in internal energy, and its differential form is: ; Furthermore, the control equation for the heat conduction of the switch cabinet is obtained as: ; The radiation heat formula is: ; Then, according to the heat conduction principle, the heat flux density is expressed as ; Ignoring the influence of thermal radiation and solving, the steady-state non-invasive temperature measurement model is obtained: ; The non-steady-state non-invasive temperature measurement model of the switch cabinet is obtained as follows: ; In the formula, is the thermal conductivity, is the unit normal vector of the interface outward, is the heat generation rate per unit volume of the internal heat source, is the density, is the specific heat capacity, is the temperature of the object, is the temperature, is the heat flux density, is the heat source, is the radiation heat, is the emissivity of the object surface, is the radiation surface area, is the Stefan-Boltzmann constant, 、 are the temperatures of the two radiators, is the cabinet surface temperature, is the temperature of the heat source inside the cabinet, is the ambient temperature, is the thermal conductivity, is the convective heat transfer, is the relevant distance, is the specific coefficient.

[0013] Preferably, optimizing the model parameters through an adaptive genetic algorithm includes: Defining the mean square error between the experimental data and the model measurement data as the objective function, and using the reciprocal of the objective function as the fitness; Initializing the parameters of the algorithm, setting the population size of the adaptive genetic algorithm to , using floating-point encoding for the encoding method, the thermal conductivity coefficient is , the maximum number of iterations is , initializing the size of the population , setting the search space to , the initial population is randomly distributed within, the crossover probability , the mutation probability ; Calculating the fitness value of each individual in the population , dynamically adjusting the coefficients in the heat transfer model according to the individual characteristics of the current generation, applying the adjusted model to the test calculation, and then calculating the root mean square error based on the experimental calculation results , using the reciprocal of the root mean square error as the fitness value.

[0014] Preferably, optimizing the model parameters through an adaptive genetic algorithm further includes: Selection operation, using the roulette wheel method according to the fitness value, selecting individuals from the current population to form the next generation population, and dynamically allocating the weight of each individual to the overall optimization direction in combination with the fitness value and the model adjustment strategy; Crossover operation, performing crossover operation on the selected individuals according to the crossover probability , automatically analyzing the fitness trend of the attached individuals, and adjusting the crossover point to improve the optimization effect; Mutation operation, performing mutation operation on the selected individuals according to the mutation probability , when the adjustment range of the model coefficients is large or the fitness convergence speed is slow, automatically increasing the mutation amplitude, and conversely, when the model tends to be stable, decreasing the mutation amplitude.

[0015] Preferably, characterized in that, based on the adjustment results of the model coefficients and the fitness change trend, dynamically updating the parameters of the genetic algorithm, the adjustment strategy is: , ; Among them, and is an adjustment function, which is adaptively generated based on the model coefficients and fitness changes; Recalculate the fitness values of the population individuals, gradually update to obtain the optimal solution, and judge the algorithm to see if the maximum number of iterations is reached ; if so, output the optimal solution and end the iteration; otherwise, transfer to the next step to continue the next coefficient adaptive dynamic optimization; Output the global optimal solution, which is the optimal thermal conductivity coefficient of the non-invasive temperature measurement model.

[0016] Preferably, a non-invasive temperature measurement monitoring parameter optimization system for switchgear based on an adaptive genetic algorithm is characterized in that it is used to execute the non-invasive temperature measurement monitoring method for switchgear based on an adaptive genetic algorithm described in claims 1 to 9; the system includes: A model establishment module, which is used to establish a non-invasive temperature measurement model based on the heat transfer principle and carry out model confirmation and certification tests; A model analysis and correction module, which is used to analyze the correspondence between the calculation results and test results of the non-invasive temperature measurement model, design model confirmation indicators, judge whether the model meets the requirements, and if not, correct the model parameters through an adaptive genetic algorithm; An algorithm optimization module, which is used to based on the existing test data, with minimizing the confirmation index as the optimization goal, and with the thermal conductivity coefficient, geometric parameters, and environmental factors as design variables, realize model parameter identification and update through an adaptive genetic algorithm; A model certification module, which is used to carry out model confirmation again on the model with optimized parameters based on the model confirmation index, conduct accuracy verification and reliability evaluation, and complete model certification after the compliance is met.

[0017] The beneficial effects of the present invention are as follows: 1. The present invention establishes a non-invasive temperature measurement model for switchgear based on heat transfer, which solves the problems that when the existing traditional discoloration temperature measurement patches or infrared thermometers are used for temperature monitoring, their effects are greatly affected by the surface shape and fouling degree of the measured object, with low accuracy, low automation, and the need for a large amount of manpower.

[0018] 2. The present invention uses an adaptive genetic algorithm to optimize the parameters of the temperature measurement model for switchgear, effectively improving the accuracy and stability of the model. Through experimental verification, the root mean square error (RMSE) of the temperature measurement model of the present invention is ±1.2 °C, with a 65% improvement in accuracy compared to the traditional infrared temperature measurement method (±3.5 °C). Moreover, it requires no manual intervention and is applicable to unattended substations. The genetic algorithm automatically adjusts the model parameters by simulating the process of natural selection, improving the accuracy of the temperature measurement model, reducing the computational complexity, or enhancing the operating efficiency, avoiding the local optimum problem in traditional optimization methods, and thus being able to more accurately predict the temperature distribution of the switchgear, thereby improving the safety and reliability of equipment operation. Description of the Drawings

[0019] Figure 1 is the flowchart for model verification of the present invention; Figure 2 is the experimental structure diagram for the temperature measurement of switchgear in the embodiment of the present invention; Figure 3 is the flowchart of the adaptive genetic algorithm in the embodiment of the present invention; Figure 4 is the curve graph of the fitness and parameter changes during the iterative process of the adaptive genetic algorithm in the embodiment of the present invention. Detailed Embodiments

[0020] Embodiment 1: As Figure 1 shown, the non-intrusive temperature monitoring method for switchgear based on the adaptive genetic algorithm includes the following steps: S1. Establish a non-intrusive temperature measurement model based on the heat transfer principle, and conduct model confirmation and certification tests; S2. Analyze the correspondence between the calculation results of the non-intrusive temperature measurement model and the results of the certification tests; S3. Based on the existing test data, construct a temperature measurement model for model parameter optimization and accuracy improvement; S4. Based on the proposed model confirmation indicators, conduct model confirmation again on the model with optimized parameters.

[0021] Preferably, in step S1, the certification tests are carried out under multiple test settings. The response measurements of the certification verification experiments are conducted at three positions to collect the temperature responses at different positions on the surface. Moreover, the on-site test switchgear is operated under different loads, and the contact temperatures in the busbar chamber, circuit breaker chamber, cable chamber, and instrument panel chamber inside the switchgear are monitored and recorded in real time through the optical fiber sensors or contact sensors inside the switchgear; for different operating conditions of the switchgear, the temperature change data is collected and recorded regularly with a data acquisition period of 1 h.

[0022] Preferably, in step S2, analyzing the correspondence between the calculation results of the non-intrusive temperature measurement model and the results of the certification tests specifically includes: Design model confirmation metrics are designed based on all correspondences and the accuracy of the computational model and test data; Estimate the metric values under different model confirmation scenarios based on the non-invasive temperature measurement model and test data, and determine whether the compliance of the model with the test and its precision ability meet the requirements. If not, further correct the model parameters through the adaptive genetic algorithm to improve the accuracy of the model for temperature monitoring; the error metrics selected are that the mean absolute error MAE is less than 2°C and the RMSE is less than 3°C in the switchgear temperature prediction scenario.

[0023] Preferably, in step S3, based on the existing test data, the temperature measurement model for model parameter optimization and accuracy improvement includes: With minimizing the confirmation metric as the optimization objective, and the thermal conductivity, geometric parameters, and environmental factors as the design variables, realize model parameter identification and update through the adaptive genetic algorithm, and the crossover probability Pc and mutation probability Pm are dynamically adjusted with the fitness.

[0024] Preferably, in step S4, based on the proposed model confirmation metrics, conduct model confirmation on the model with optimized parameters again, including: Conduct model certification after the compliance is met, including accuracy verification; confirm the model coefficients and reliability when the error metric MAE is less than 2°C; evaluate and examine the stability and consistency of the model under different conditions, at different times, or with different input parameters.

[0025] Preferably, establishing a non-invasive temperature measurement model based on the heat transfer principle specifically includes: Obtain the heat balance equation according to the first law of thermodynamics: imported heat + internal heat source heat generation = increase in internal energy, and its differential form is: ; Furthermore, obtain the control equation for heat conduction of the switchgear as: ; The radiation heat formula is: ; Then, according to the heat conduction principle, obtain the heat flux density expressed as ; Neglect the influence of heat radiation and solve to obtain the steady-state non-invasive temperature measurement model: ; Obtain the non-steady-state non-invasive temperature measurement model of the switchgear as follows: ; In the formula, is the thermal conductivity, is the unit normal vector of the interface outward, is the heat generation rate per unit volume of the internal heat source, is the density, is the specific heat capacity, is the temperature of the object, is the temperature, is the heat flux density, is the heat source, is the radiant heat, is the emissivity of the object surface, is the radiation surface area, is the Stefan - Boltzmann constant, 、 are the temperatures of two radiators, is the temperature of the cabinet surface, is the temperature of the heat - generating point inside the cabinet, is the ambient temperature, is the thermal conductivity, is the convective heat transfer, is the relevant distance, is the specific coefficient.

[0026] Preferably, optimizing the model parameters by the adaptive genetic algorithm includes: Defining the mean square error between the experimental data and the model measurement data as the objective function, and using the reciprocal of the objective function as the fitness; Initializing the parameters of the algorithm, setting the population size of the adaptive genetic algorithm to , adopting the floating - point encoding method, the thermal conductivity coefficient is , the maximum number of iterations is , initializing the group size , setting the search space to , randomly distributing the initial population within, the crossover probability , the mutation probability ; Calculating the fitness value of each individual in the population , dynamically adjusting the coefficients in the heat transfer model according to the individual characteristics of the current generation, applying the adjusted model to the test calculation, and then calculating the root - mean - square error based on the experimental calculation results, and using the reciprocal of the root - mean - square error as the fitness value.

[0027] Preferably, optimizing the model parameters by the adaptive genetic algorithm further includes: Selection operation, using the roulette wheel method according to the fitness value to select individuals from the current population to form the next - generation population, and dynamically allocating the weight of each individual to the overall optimization direction in combination with the fitness value and the model adjustment strategy; Crossover operation, according to the crossover probability Perform crossover operation on the selected individuals, automatically analyze the fitness trend of the attached individuals, and adjust the crossover point to improve the optimization effect; Mutation operation, according to the mutation probability Perform mutation operation on the selected individuals. When the adjustment range of the model coefficients is large or the fitness convergence speed is slow, automatically increase the mutation amplitude. Conversely, when the model tends to be stable, reduce the mutation amplitude.

[0028] Preferably, it is characterized in that, based on the adjustment result of the model coefficients and the fitness change trend, dynamically update the parameters of the genetic algorithm, and the adjustment strategy is: , ; wherein, and are adjustment functions, which are adaptively generated based on the model coefficients and the fitness change; Calculate the fitness values of the population individuals again, gradually update to obtain the optimal solution, judge the algorithm to see if the maximum number of iterations is reached ; if so, output the optimal solution and end the iteration; otherwise, go to the next step to continue the next coefficient adaptive dynamic optimization; Output the global optimal solution, and the global optimal solution is the optimal thermal conductivity coefficient of the non-invasive temperature measurement model.

[0029] Preferably, an optimization system for non-invasive temperature measurement monitoring parameters of switchgear based on an adaptive genetic algorithm is characterized in that it is used to execute the non-invasive temperature measurement monitoring method of switchgear based on an adaptive genetic algorithm described in claims 1 to 9; the system includes: A model establishment module, which is used to establish a non-invasive temperature measurement model based on the heat transfer principle and carry out model confirmation and certification tests; A model analysis and correction module, which is used to analyze the correspondence between the calculation results and the test results of the non-invasive temperature measurement model, design model confirmation indicators, judge whether the model meets the requirements, and if not, correct the model parameters through an adaptive genetic algorithm; An algorithm optimization module, which is used to, based on the existing test data, take minimizing the confirmation index as the optimization goal, take the thermal conductivity coefficient, geometric parameters, and environmental factors as design variables, and realize model parameter identification and update through an adaptive genetic algorithm; A model certification module, which is used to carry out model confirmation again on the model with optimized parameters based on the model confirmation index, conduct accuracy verification and reliability assessment, and complete model certification after the compliance is met.

[0030] Embodiment 2: The non-invasive temperature measurement model of switchgear and the method for optimizing the parameters of the adaptive genetic algorithm of the present invention are specifically implemented according to the following steps: Step 1: Establish a non-invasive temperature measurement model for the switchgear based on the principles of heat transfer and determine the model boundary conditions; Step 1 is specifically implemented according to the following steps: Step 1.1: The heat transfer in the switchgear is a complex process, consisting of three basic modes: conduction, convection, and radiation. According to the first law of thermodynamics, the sum of the net heat flux introduced into the object and the heat generation rate of the internal heat source in the object should be equal to the increase in the internal energy of the object, thus obtaining the heat balance equation: introduced heat + internal heat source heat generation = increase in internal energy. The differential form is expressed as: ; where, is the thermal conductivity, and n is the unit normal vector of the interface; is the heat generation rate per unit volume of the internal heat element; is the density of the object, c is the specific heat capacity, t is the temperature of the object, is the time; and are the surface area and volume of the object, respectively.

[0031] Step 1.2: According to the heat balance principle, the control equation for heat conduction in the switchgear can be expressed as: ; ; In the formula: ρ is the density; Cp is the constant pressure heat capacity; v is the flow velocity; T is the temperature; t is the time; q is the heat flux density; Q is the heat source; λ is the thermal conductivity; Φ is the radiation heat; is the emissivity of the object surface, and A is the radiation surface area; is the Stefan-Boltzmann constant; T1 and T2 are the temperatures of the two radiators.

[0032] For the establishment of the non-invasive temperature measurement model, according to the heat conduction principle, the heat flux density can be expressed as: ; In the formula, Ts is the cabinet surface temperature, Th is the temperature of the internal heat source in the cabinet (the internal heat source in the cabinet is evenly distributed or a concentrated point source), T0 is the ambient temperature, k is the thermal conductivity, and h is the convective heat transfer.

[0033] Ignoring the influence of thermal radiation and solving, we get: ; From this, the steady-state non-invasive temperature measurement model of the switchgear can be written as ; The non-steady-state non-invasive temperature measurement model of the switchgear ; Step 2 is specifically implemented according to the following steps: Step 2: Obtain a series of experimental data on material characterization, verification, and certification related to the calculation model. The schematic diagram of the temperature measurement positions of the switchgear is as Figure 2 shown. The verification experiment is carried out at four electrical contact points. For different working conditions, the measuring device measures the temperature rise under each working condition. The certification experiment measures the temperature rise data of the device at three different positions.

[0034] As Figure 3 and Figure 4 shown, Step 3 is specifically implemented according to the following steps: Step 3: Optimize the model parameters through the adaptive genetic algorithm. Define the mean square error between the experimental data and the model measurement data as the objective function. Use the reciprocal of the objective function as the fitness. The higher the fitness, the smaller the error.

[0035] Step 3.1: Initialize the parameters of the algorithm. Set the population size of the adaptive genetic algorithm to N, use floating-point encoding for the encoding method, the thermal conductivity coefficient to k, and the maximum number of iterations to tmax; Step 3.2: Initialize the population size N = 100, set the search space to [10, 400], and the initial population is randomly distributed within [10, 400], with the crossover probability Pc and the mutation probability Pm.

[0036] Step 3.3: Calculate the fitness value f of each individual in the population. According to the individual characteristics of the current generation, dynamically adjust the coefficients in the heat transfer model, such as the initial temperature, thermal boundary conditions, thermal conductivity range, etc. Apply the adjusted model to the test calculation, and then calculate the root mean square error RMSE according to the experimental calculation results: ; Use the reciprocal of the root mean square error as the fitness value to evaluate the quality of individuals.

[0037] Step 3.4: Selection operation. Use the roulette wheel method according to the fitness value to select N individuals from the current population to form the next generation population; combine the fitness value and the model adjustment strategy to dynamically allocate the weight of each individual to the overall optimization direction.

[0038] Step 3.5: Crossover operation. Perform the crossover operation on the selected individuals according to the crossover probability Pc, automatically analyze the fitness trend of the attached individuals, and adjust the crossover points to improve the optimization effect.

[0039] Step 3.6: Mutation operation. Perform the crossover operation on the selected individuals according to the crossover probability Pm. When the adjustment range of the model coefficients is large or the fitness convergence speed is slow, automatically increase the mutation amplitude; conversely, when the model tends to be stable, reduce the mutation amplitude to prevent the algorithm from falling into a local optimum.

[0040] Step 3.7: Dynamically update the parameters of the genetic algorithm based on the adjustment results of the model coefficients and the trend of fitness change.

[0041] The adjustment strategy is as follows: ; ; Among them, and are adjustment functions, which are adaptively generated based on the model coefficients and fitness change.

[0042] Step 3.8: Calculate the fitness values of the population individuals again, and gradually update to obtain the optimal solution; Step 3.9: Judge the algorithm to see if it has reached the maximum number of iterations tmax. If so, output the optimal solution and end. Otherwise, go to Step 4 to continue the next coefficient adaptive dynamic optimization.

[0043] Step 3.10: Output the global optimal solution Xbest, that is, the optimal thermal conductivity coefficient of the non-invasive temperature measurement model.

Claims

1. A non-invasive temperature monitoring method for switchgear based on an adaptive genetic algorithm, characterized in that It includes the following steps: S1. Establish a non-invasive temperature measurement model based on the principle of heat transfer, and conduct model confirmation and certification tests; S2. Analyze the corresponding relationship between the calculation results of the non-invasive temperature measurement model and the results of the certification tests; S3. Based on the existing test data, construct a temperature measurement model for model parameter optimization and accuracy improvement; S4. Based on the proposed model confirmation indicators, conduct model confirmation on the model with optimized parameters again.

2. The non-invasive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 1, wherein In step S1, the certification tests are carried out under multiple test settings. The response measurements of the certification verification experiments are carried out at three positions to collect the temperature responses at different positions on the surface. And the on-site test switchgear is operated under different loads, and the contact temperatures in the busbar chamber, circuit breaker chamber, cable chamber and instrument panel chamber inside the switchgear are monitored and recorded in real time through the optical fiber sensors or contact sensors inside the switchgear; for different working conditions of the switchgear, the temperature change data is collected and recorded regularly at fixed data acquisition periods.

3. The non-intrusive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 1, characterized in that, In step S2, analyzing the corresponding relationship between the calculation results of the non-invasive temperature measurement model and the results of the certification tests specifically includes: Design model confirmation indicators, which are designed based on all corresponding relationships and the accuracy of the calculation model and test data; Based on the non-invasive temperature measurement model and test data, estimate the indicator values in different model confirmation scenarios, and judge whether the compliance degree of the model with the test and the accurate ability meet the requirements. If not, further correct the model parameters through the adaptive genetic algorithm to improve the accuracy of the model for temperature monitoring.

4. The non-invasive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 1, characterized in that, In step S3, constructing a temperature measurement model for model parameter optimization and accuracy improvement based on the existing test data includes: Taking the minimization of the confirmation indicator as the optimization goal, taking the thermal conductivity, geometric parameters, and environmental factors as design variables, and realizing model parameter identification and update through the adaptive genetic algorithm. The crossover probability Pc and the mutation probability Pm are dynamically adjusted according to the fitness.

5. The non-invasive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 1, characterized in that In step S4, conducting model confirmation on the model with optimized parameters again based on the proposed model confirmation indicators includes: Carry out model certification after the compliance is met, including accuracy verification; when the error indicator MAE is less than the set threshold, confirm the model coefficients and reliability; evaluate and examine the stability and consistency of the model under different conditions, different times or different input parameters.

6. The non-intrusive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 2, characterized in that, Establishing a non-invasive temperature measurement model based on the principle of heat transfer specifically includes: Obtain the heat balance equation according to the first law of thermodynamics: imported heat + internal heat source heat generation = increase in internal energy, and its differential form is: ； Furthermore, obtain the control equation for heat conduction of the switchgear as: ； The radiation heat formula is: ； Then, according to the heat conduction principle, obtain the heat flux density expressed as ； Neglect the influence of thermal radiation and solve to obtain the steady-state non-invasive temperature measurement model: ； Obtain the non-steady-state non-invasive temperature measurement model of the switchgear as follows: ； In the formula, is the thermal conductivity, is the unit normal vector of the interface, is the heat generation rate per unit volume of the internal heat source, is the density, is the specific heat capacity, is the temperature of the object, is the temperature, is the heat flux density, is the heat source, is the radiant heat, is the emissivity of the object surface, is the radiant surface area, is the Stefan-Boltzmann constant, 、 are the temperatures of two radiators, is the temperature of the cabinet surface, is the temperature of the heat generation point inside the cabinet, is the ambient temperature, is the thermal conductivity, is the convective heat transfer, is the relevant distance, is the specific coefficient.

7. The non-intrusive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 4, characterized in that, Optimizing the model parameters through the adaptive genetic algorithm includes: Define the mean square error between the experimental data and the model measurement data as the objective function, and use the reciprocal of the objective function as the fitness; Initialize the parameters of the algorithm, and set the population size of the adaptive genetic algorithm to , adopt floating-point encoding for the encoding method, and the thermal conductivity coefficient is , the maximum number of iterations is , initialize the size of the population, set the search space, and the initial population is randomly distributed within, and the crossover probability , the mutation probability ; Calculate the fitness value of each individual in the population , dynamically adjust the coefficients in the heat transfer model according to the individual characteristics of the current generation, apply the adjusted model to the test calculation, and then calculate the root mean square error based on the experimental calculation results , and use the reciprocal of the root mean square error as the fitness value.

8. The non-intrusive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 7, wherein Optimizing the model parameters through the adaptive genetic algorithm also includes: Selection operation: Using the roulette wheel method based on the fitness value, select individuals from the current population to form the next generation population, and dynamically allocate the weight of each individual to the overall optimization direction by combining the fitness value and the model adjustment strategy; Crossover operation, according to the crossover probability Perform crossover operations on the selected individuals, automatically analyze the fitness trends of the attached individuals, and adjust the crossover points to improve the optimization effect; Mutation operation, according to the mutation probability Perform mutation operation on the selected individuals. When the adjustment range of the model coefficients is large or the fitness convergence speed is slow, automatically increase the mutation amplitude; conversely, when the model tends to be stable, decrease the mutation amplitude.

9. The non-invasive temperature monitoring method for switchgear based on the adaptive genetic algorithm according to claim 8, characterized in that, Based on the adjustment results of the model coefficients and the change trend of the fitness, dynamically update the parameters of the genetic algorithm, and the adjustment strategy is: ， ； Among them, and are adjustment functions, which are adaptively generated based on model coefficients and fitness changes; Recalculate the fitness values of the population individuals, gradually update to obtain the optimal solution, and judge the algorithm to see if the maximum number of iterations has been reached ; if so, output the optimal solution and end the iteration; otherwise, go to the next step to continue the next coefficient self-adaptive dynamic optimization; Output the global optimal solution, and the global optimal solution is the optimal thermal conductivity of the non-invasive temperature measurement model.

10. A non-intrusive temperature measurement and monitoring parameter optimization system for switchgear based on an adaptive genetic algorithm, characterized in that, For implementing the non-intrusive temperature measurement and monitoring method of switchgear based on the adaptive genetic algorithm described in claims 1 to 9; the system includes: A model establishment module, which is used to establish a non-intrusive temperature measurement model based on the heat transfer principle, and conduct model confirmation and certification tests; A model analysis and correction module, which is used to analyze the correspondence between the calculation results and test results of the non-intrusive temperature measurement model, design model confirmation indicators, judge whether the model meets the requirements, and if not, correct the model parameters through the adaptive genetic algorithm; An algorithm optimization module, which is used to based on the existing test data, with minimizing the confirmation indicator as the optimization goal, and with the thermal conductivity, geometric parameters, and environmental factors as the design variables, realize model parameter identification and update through the adaptive genetic algorithm; A model certification module, which is used to conduct model confirmation again on the model with optimized parameters based on the model confirmation indicators, perform accuracy verification and reliability assessment, and complete model certification after the compliance is met.

Images