Method for monitoring and regulating oil temperature of oil-immersed transformer

Abstract

The invention discloses an oil temperature monitoring and regulating method for an oil-immersed transformer, which belongs to the technical field of control optimization, and comprises the following steps: constructing an oil temperature PID control system of the oil-immersed transformer, introducing an improved candle flame optimization algorithm, and setting oil temperature PID control parameters by using the improved candle flame optimization algorithm. And setting three optimal control parameters obtained through optimization by using an improved candle flame optimization algorithm as parameters of an oil temperature PID controller in the PID control system of the oil-immersed transformer, and optimizing the oil temperature control effect of the oil-immersed transformer. According to the method, the oil temperature PID control system of the oil-immersed transformer is constructed, and the improved candle flame optimization algorithm is introduced, so that the oil temperature of the oil-immersed transformer can be recovered to be stable more quickly when working condition disturbance such as load sudden change or environment temperature change occurs, the insulation aging risk caused by long-time high-temperature operation can be reduced, and the service life of the oil-immersed transformer is prolonged. And the operation safety and reliability of the oil-immersed transformer are enhanced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of control optimization, and particularly relates to an oil-immersed transformer oil temperature monitoring and regulation method. BACKGROUND

[0002] As a core equipment in the power transmission and distribution system, the oil-immersed transformer undertakes the key task of voltage transformation and power transmission, and its safe, stable and long-acting operation state is the cornerstone of guaranteeing the reliability of the entire power grid. During operation, the windings and the core inside the transformer continuously generate a large amount of heat due to losses, and these heat is mainly transferred to the cooling system for dissipation through the circulation of insulating oil. Therefore, the temperature of the insulating oil directly affects the aging rate of the solid insulating material inside the transformer. An excessively high oil temperature can sharply accelerate the aging of the insulation, shorten the service life of the transformer, and even cause serious faults and lead to major safety accidents. Therefore, more accurate and reliable control of the top layer oil temperature of the oil-immersed transformer has always been the core requirement of the oil temperature control system.

[0003] At present, the oil temperature control of the oil-immersed transformer generally adopts a closed-loop regulation scheme based on a PID controller, which realizes the dynamic stability of the oil temperature by adjusting the operating power of the cooling system. However, since the oil-immersed transformer is a complex thermal object with large inertia, pure lag and nonlinearity, the temperature response is relatively slow and is easily disturbed by various internal and external factors such as environmental temperature and load fluctuations. Therefore, when the parameters of the traditional PID controller are not well tuned, it is difficult to achieve ideal control effect, resulting in slow oil temperature response, long regulation time, and even large temperature rise overshoot under disturbance.

[0004] Traditional PID parameter tuning methods such as engineering tuning method or manual trial and error method rely heavily on the professional experience of engineers, which not only consumes time and effort, but also cannot guarantee that the obtained parameters are globally optimal solutions. Especially when facing complex nonlinear systems such as transformers, the control performance is often unsatisfactory. Although some studies have attempted to apply standard meta-heuristic optimization algorithms such as particle swarm algorithm and genetic algorithm to PID parameter optimization, these standard algorithms also have inherent defects such as easy to fall into local optimum and slow convergence speed, and the optimization performance and stability still need to be improved. SUMMARY

[0005] In order to overcome the technical problems described in the background art, the present application provides an oil-immersed transformer oil temperature monitoring and regulation method. By constructing an oil-immersed transformer oil temperature PID control system and introducing an improved candle flame optimization algorithm, the oil-immersed transformer oil temperature can be quickly restored to stability when facing load mutation or environmental temperature change and other working condition disturbances, the risk of insulation aging caused by long-time high-temperature operation can be reduced, and the operation safety and reliability of the oil-immersed transformer are enhanced.

[0006] The technical scheme of the present application is as follows: an oil-immersed transformer oil temperature monitoring and regulation method, comprising the following steps: S1, constructing an oil-immersed transformer oil temperature PID control system, including an oil temperature error calculation module, an oil temperature PID controller module, an improved candle flame optimization algorithm module, an oil temperature adjustment module, and an oil temperature monitoring module; S2, introducing an improved candle flame optimization algorithm, including the following specific improvements: S21, introducing a nonlinear adaptive parameter adjustment strategy, dynamically adjusting the heat propagation coefficient through the current iteration number t and the random displacement factor , which can use a wide range of global exploration at the beginning of iteration, and automatically switch to fine local exploitation of the optimal solution region at the later stage of iteration; S22, introducing an elite-guided heat propagation strategy, selecting the top p% individuals with the best fitness from the current population to form an elite pool , and the ordinary flame individuals will randomly select an elite target from the elite pool for learning and moving, thereby enhancing population diversity and expanding the effective search area; S23, introducing a random displacement strategy based on flight mechanism, the generated step vector L has both high-frequency short steps and low-frequency long-distance jumping characteristics, so that the flame individuals can not only deeply exploit the advantage area, but also jump to a new area for exploration with a certain probability, improving the global search performance; S3, using the improved candle flame optimization algorithm to set the oil temperature PID control parameters in the oil-immersed transformer oil temperature PID control system, and obtaining the best control parameters through optimization 、 、 ; S4, setting the three best control parameters obtained by using the improved candle flame optimization algorithm as the parameters of the oil temperature PID controller in the oil-immersed transformer PID control system, and optimizing the oil-immersed transformer oil temperature control effect.

[0007] Further, in the oil-immersed transformer temperature PID control system constructed in step S1, the top layer oil temperature is collected by the temperature monitoring module and transmitted to the temperature error calculation module, the oil temperature error calculation module receives the set target oil temperature and calculates the error between the target oil temperature and the top layer oil temperature, and then outputs the real-time error to the oil temperature PID controller module, which calculates the control amount according to the real-time error and outputs it to the oil temperature adjustment module to realize oil temperature adjustment.

[0008] Furthermore, in step S1, the thermodynamic characteristics of the oil-immersed transformer are modeled, and a mathematical model of the response of the top oil temperature of the oil-immersed transformer to the power regulation of the cooling system is constructed: , in, This is the process gain, representing the degree Celsius change in oil temperature caused by a 1% change in cooling power. Since increasing cooling power leads to a temperature decrease, it is a negative value. Pure time delay represents the time interval between a change in the input and the start of a response in the output. is the process time constant, representing the magnitude of the system response inertia, and s is the independent variable of the Laplace transform.

[0009] Furthermore, in step S21, the nonlinear adaptive parameter harmonicization strategy includes the following steps: S2101. At the beginning of each iteration, in order to dynamically balance the exploration and mining phases of the algorithm, calculate the nonlinear incremental heat propagation coefficient used to modulate the heat propagation process. , , in, The heat propagation coefficient for the current iteration, and These are the preset minimum and maximum values ​​of the coefficient, respectively, where t is the current iteration number. This represents the maximum number of iterations. S2102. Calculate the nonlinear decreasing random displacement factor used to dominate the random displacement intensity of Lévy flight. This guides the algorithm to smoothly transition its search focus from global exploration to local mining. , in, These are the preset maximum and minimum values ​​of the factor, respectively, where t is the current iteration number. This represents the maximum number of iterations.

[0010] Furthermore, in step S22, the elite-guided heat propagation strategy includes the following steps: S2201. In each iteration, the current flame population is sorted by fitness, and the top p% of flame individuals with the best fitness are selected to form an elite pool. , where p is the preset elite ratio; S2202, Execute elite-guided heat propagation: For each Flame individual i in the population, from the elite pool Randomly select an elite Flame as the guiding target. And perform a learning move toward that target. , in, For flame individuals Calculated potential new locations, For flame individuals Current location For adaptive thermal propagation coefficient, For flame individuals The normalized temperature is calculated as follows: ,in For individuals fitness value, and These are the best and worst fitness values ​​for the current population, respectively. A randomly selected elite fire-guided target.

[0011] Further, in step S23, based on The stochastic displacement strategy of the flight mechanism includes the following steps: S2301, Generate a conforming A stably distributed random step size vector L, , Where L is the generated multidimensional step size vector, and λ is the stability index that controls the distribution shape and takes a value between (1,2]. S2302. For each individual flame, its current position and the adaptive factor it has passed through are compared. Scaling Step vector Add them together to perform an efficient perturbation operation. , in, For flame individuals Calculated potential new locations, For flame individuals Current location Let L be the adaptive random displacement factor. Flight step length vector.

[0012] Furthermore, step S3, which uses an improved candle flame optimization algorithm to tune the oil temperature PID control parameters in the oil-immersed transformer oil temperature PID control system, includes the following steps: S31. Set the control parameters of the improved candle flame optimization algorithm, and randomly generate an initial population containing multiple flame individuals within the preset PID parameter search space. S32. Before reaching the preset maximum number of iterations, repeatedly update the position of each flame individual in the population, and update the global optimal solution based on the updated position; S33. After the iterative optimization step is terminated, the final global optimal solution is output as the optimal combination of oil temperature PID control parameters.

[0013] Furthermore, step S32 also includes a step of performing a reset operation on the stagnant flame individual, specifically: S321. Set a stagnation counter for each flame individual. If the fitness of an individual fails to improve after completing the two independent update attempts, its counter is incremented by one; otherwise, it is reset to zero. S322. After all individuals in the population have completed their update attempts, check all stall counters. If the count value of an individual reaches the preset stall threshold k, then the position of that individual will be randomly reinitialized within the PID parameter search space.

[0014] The beneficial effects of this invention due to the adoption of the above-mentioned technology are as follows: By constructing a PID control system for oil-immersed transformer oil temperature and introducing an improved candle flame optimization algorithm, this invention effectively overcomes the defect of the candle flame optimization algorithm being prone to getting trapped in local optima. Compared with the candle flame optimization algorithm, it has a faster convergence speed and better optimization accuracy, and can find a better parameter combination for the PID controller, thereby improving the control performance of the entire oil-immersed transformer oil temperature control system, shortening the system adjustment time, and enabling the oil temperature of the oil-immersed transformer to recover and stabilize more quickly when facing operating disturbances such as sudden load changes or changes in ambient temperature. This effectively reduces the risk of insulation aging caused by long-term high-temperature operation and enhances the operational safety and reliability of the oil-immersed transformer. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the present invention.

[0016] Figure 2 This is a curve comparing the optimal fitness values ​​of the candle flame optimization algorithm and the improved candle flame optimization algorithm.

[0017] Figure 3 The optimization parameter values ​​of the candle flame optimization algorithm and the improved candle flame optimization algorithm. Process diagram.

[0018] Figure 4 The optimization parameter values ​​of the candle flame optimization algorithm and the improved candle flame optimization algorithm. Process diagram.

[0019] Figure 5The optimization parameter values ​​of the candle flame optimization algorithm and the improved candle flame optimization algorithm. Process diagram.

[0020] Figure 6 This is a graph showing the evolution of adaptive parameters of the improved candle flame optimization algorithm of this invention.

[0021] Figure 7 This is a comparison chart of the optimized PID control effects of the candle flame optimization algorithm and the improved candle flame optimization algorithm. Detailed Implementation

[0022] Example 1: As Figures 1-7 As shown, the present invention provides a method for monitoring and controlling the oil temperature of an oil-immersed transformer, comprising the following steps: S1. Construct an oil-immersed transformer oil temperature PID control system, including an oil temperature error calculation module, an oil temperature PID controller module, an improved candle flame optimization algorithm module, an oil temperature regulation module, and an oil temperature monitoring module; S2. An improved candle flame optimization algorithm is introduced, with specific improvements including: S21. Introduce a nonlinear adaptive parameter harmonicization strategy to dynamically adjust the heat propagation coefficient based on the current iteration number t. With random displacement factor It can use a large-scale global exploration in the early stage of iteration, and automatically switch to fine-grained local mining of the optimal solution region in the later stage of iteration. S22. Introduce an elite-led heat propagation strategy to select the top p% of individuals with the best fitness from the current population to form an elite pool. A regular fire-type individual will randomly select an elite target from this elite pool. They learn and move, thereby enhancing population diversity and expanding the effective search area; S23, Introducing based on The random displacement strategy of the flight mechanism generates a step vector L that combines high-frequency short step size with low-frequency long-distance jump characteristics, enabling the flame individual to both deeply mine advantageous areas and jump to new areas with a certain probability to explore, thus improving global search performance. S3. The improved candle flame optimization algorithm is used to tune the PID control parameters of the oil temperature in the oil-immersed transformer oil temperature PID control system, and the optimal control parameters are obtained through optimization. , , ; S4. The three optimal control parameters obtained by using the improved candle flame optimization algorithm are set as the parameters of the oil temperature PID controller in the oil-immersed transformer PID control system to optimize the oil temperature control effect of the oil-immersed transformer.

[0023] In the oil-immersed transformer temperature PID control system constructed in step S1, the top oil temperature is collected by the temperature monitoring module and transmitted to the temperature error calculation module. The oil temperature error calculation module receives the set target oil temperature, calculates the error between the target oil temperature and the top oil temperature, and outputs the real-time error to the oil temperature PID controller module. The oil temperature PID controller module calculates the control quantity based on the real-time error and outputs it to the oil temperature regulation module to achieve oil temperature regulation. The oil temperature regulation module refers to the cooling system of the oil-immersed transformer, including the oil pump and the radiator outside the oil tank. The oil temperature is regulated by adjusting the power of the cooling system. When the fan speed increases or the oil pump starts, the amount of air or oil flowing through the radiator increases, and the heat carried away per unit time increases, thereby reducing the top oil temperature of the oil-immersed transformer and thus achieving closed-loop control.

[0024] In step S1, the thermodynamic characteristics of the oil-immersed transformer are modeled, and a mathematical model of the response of the top oil temperature of the oil-immersed transformer to the power regulation of the cooling system is constructed: , in, This is the process gain, representing the degree Celsius change in oil temperature caused by a 1% change in cooling power. Since increasing cooling power leads to a temperature decrease, it is a negative value. Pure time delay represents the time interval between a change in the input and the start of a response in the output. is the process time constant, representing the magnitude of the system response inertia, and s is the independent variable of the Laplace transform.

[0025] The constructed fitness function J is: , in, Here, ITAE is the weighting coefficient, and ITAE is the integral of time multiplied by the absolute error. This is an overshoot penalty item. To adjust the time penalty item, This is a steady-state error penalty term.

[0026] In step S21, the nonlinear adaptive parameter harmonicization strategy includes the following steps: S2101. At the beginning of each iteration, in order to explore and develop the dynamic equilibrium algorithm, the nonlinear incremental heat propagation coefficient used to modulate the heat propagation process is calculated. , , in, The heat propagation coefficient for the current iteration, and These are the preset minimum and maximum values ​​of the coefficient, respectively, where t is the current iteration number. This represents the maximum number of iterations. S2102. Calculate the nonlinear decreasing random displacement factor used to dominate the random displacement intensity of Lévy flight. This guides the algorithm to smoothly transition its search focus from global exploration to local mining. , in, These are the preset maximum and minimum values ​​of the factor, respectively, where t is the current iteration number. This represents the maximum number of iterations.

[0027] In step S22, the elite-guided heat propagation strategy includes the following steps: S2201. In each iteration, the current flame population is sorted by fitness, and the top p% of flame individuals with the best fitness are selected to form an elite pool. , where p is the preset elite ratio; S2202, Execute elite-guided heat propagation: For each Flame individual i in the population, from the elite pool Randomly select an elite Flame as the guiding target. And perform a learning move toward that target. , in, For flame individuals Calculated potential new locations, For flame individuals Current location For adaptive thermal propagation coefficient, For flame individuals The normalized temperature is calculated as follows: ,in For individuals fitness value, and These are the best and worst fitness values ​​for the current population, respectively. A randomly selected elite fire-guided target.

[0028] In step S23, based on The stochastic displacement strategy of the flight mechanism includes the following steps: S2301, Generate a conforming A stably distributed random step size vector L, , Where L is the generated multidimensional step size vector, and λ is the stability index that controls the distribution shape and takes a value between (1,2]. S2302. For each individual flame, its current position and the adaptive factor it has passed through are compared. Scaling Step vector Add them together to perform an efficient perturbation operation. , in, For flame individuals Calculated potential new locations, For flame individuals Current location Let L be the adaptive random displacement factor. Flight step length vector.

[0029] Step S3, which uses an improved candle flame optimization algorithm to tune the PID control parameters of the oil temperature in the oil-immersed transformer oil temperature PID control system, includes the following steps: S31, Initialize the improved candle flame optimization algorithm, including the following sub-steps: S3101. Set the internal control parameters of the algorithm, that is, the proportion of elite individuals used to select elite individuals. Stability index used to generate Lévy flight step length Stagnation reset threshold used to determine whether an individual has entered a stagnation state. Boundary values ​​of parameters used to control adaptive strategies , , and ; S3102. Initialize the population and historical data container in the computer memory, that is, create a position matrix to store the positions of multiple flame individuals, create a fitness vector to store the fitness values ​​of the flame individuals and initialize it to infinity, create a counter vector to record the number of times each individual stagnates and initialize it to zero, and create a data structure to record the algorithm convergence process and parameter evolution history. S3103. Generate the initial population, which means for each individual flame to be generated, search the preset PID parameter space. Within, its initial position vector is generated through a uniform random distribution. It then calls the fitness function to calculate its initial fitness value; S3104. Initialize the globally optimal record, that is, set the globally optimal fitness value. Initialize to infinity and set the global optimal solution vector Initialize to a zero vector; S32, execute the iterative optimization main loop, until the preset maximum number of iterations is reached. Before proceeding, repeat the following sub-steps: S3201. Update the nonlinear adaptive strategy parameters, that is, based on the current iteration number. With maximum number of iterations The current heat transfer coefficient is calculated using the following formula. With random displacement factor : , , S3202. Sort the current population and update the global optimal solution, that is, sort all flame individuals in ascending order based on their fitness values. If the optimal fitness value in the current population is better than the recorded global optimal fitness value... Then Update the fitness value to the current optimal value and update its corresponding position vector to the new global optimum. ; S3203, Building an Elite Pool That is, based on the preset elite ratio Based on the ranking results, select the top-fittest individuals from the current population. Individuals constitute the elite pool ; S3204. Traverse and update every Flame individual in the population. For each individual... Two independent update attempts are executed sequentially: S32041. Execute the first update attempt, namely the elite-guided heat propagation strategy, and randomly select an elite target from the elite pool. And calculate a potential new location. , , in For individuals The normalized temperature; for After boundary treatment, assess its fitness; if its fitness is better than that of the individual... The current fitness of the individual will then be considered. Location updated to ; S32042, Continue executing the second update attempt, which is based on The random displacement strategy of flight generates a Flight step vector And calculate another potential new location. , , right After boundary treatment, assess its fitness; if its fitness is better than that of the individual... The current fitness will then affect the individual Location updated to ; S32043, Update the individual's stagnation counter, that is, if the individual If the fitness value fails to improve after the two update attempts in this round, its corresponding stagnation counter is incremented by 1; otherwise, it is reset to zero. S3205. Perform a reset operation on stagnant individuals. That is, after all individuals in the population have completed their update attempts, iterate through and check the stagnant counters of all individuals. If an individual's counter value reaches the preset stagnant reset threshold... Then the location of that individual in the search space The system performs random reinitialization and resets its corresponding stall counter. S3206. Record the optimization data of the current iteration, and record the global optimal fitness value under the current iteration round. With the global optimal solution Store it in the corresponding historical data container; S33, When the iteration loop terminates after reaching the maximum number of iterations, the final globally optimal solution will be recorded. The output is the parameter combination that optimizes the performance of the PID control system for oil temperature in oil-immersed transformers.

[0030] By converting the above technical solutions into project code that runs in the Matlab environment, a mathematical model for the PID control system of the top oil temperature of an oil-immersed transformer is established. The time constant T = 450 seconds, and the pure time delay is... =90 seconds, and initialize the parameters of the improved candle flame optimization algorithm (AELCFO) and the standard candle flame optimization algorithm (CFO), including the population size N=50, the dimension of the variable to be optimized d=3 (corresponding to Kp, Ki, Kd), the maximum number of iterations max_iter=100, and the upper limit of the search space ub=[-5,-0.01,-100] and the lower limit lb=[-40,-0.5,-1500] set according to the characteristics of the controlled object of the oil-immersed transformer.

[0031] After running the project code, a comparison graph of the convergence curves of the two algorithms is obtained, as shown below. Figure 2As shown, the improved candle flame optimization algorithm has basically converged to the optimal fitness value of 9788 by about the 30th iteration. Compared with the standard candle flame optimization algorithm, which finally converges to 1.963e+04, its optimization accuracy has been improved by 50.1%, and the convergence speed is faster. This proves that the improved strategy can effectively avoid the algorithm from getting trapped in local optima and find high-performance parameters more efficiently for the PID control system of oil temperature of oil-immersed transformer.

[0032] like Figure 5 As shown, the adaptive parameter evolution curve of the improved candle flame optimization algorithm demonstrates the algorithm's inherent dynamic balance mechanism of exploration and utilization. By tuning the parameters of the PID controller in the temperature PID control system of an oil-immersed transformer, two sets of optimal parameters are obtained: the optimal PID control parameters obtained by the candle flame optimization algorithm are as follows: =-13.181, =-0.033, =-455.112; The optimal PID control parameters obtained by the improved candle flame optimization algorithm are as follows: =-13.088, =-0.027, =-510.559, and by inputting these two sets of optimal control parameters into the closed-loop simulation model of the PID control system for oil temperature of an oil-immersed transformer, the corresponding controller step response effect is obtained, such as Figure 6 As shown, the test uses a step signal with a target value of -1 to simulate the setpoint tracking process of a negative gain system. The PID controller tuned by the improved candle flame optimization algorithm has a closed-loop system settling time. The response time was 378.48 seconds, far superior to the 660.07 seconds achieved by the optimized candle flame algorithm, representing a 42.66% improvement in response speed. Although the overshoot of the improved candle flame optimization algorithm was... =5.76%, slightly higher than the 5.11% of the candle flame optimization algorithm, but its advantage in response speed allows the transformer oil temperature to reach a new steady state more quickly after the operating conditions change. This results in better control effects on ensuring transformer safety and extending insulation life, fully verifying the effectiveness and superiority of the improved candle flame optimization algorithm.

Claims

1. A method for monitoring and controlling the oil temperature of an oil-immersed transformer, characterized in that: Includes the following steps: S1. Construct an oil-immersed transformer oil temperature PID control system, including an oil temperature error calculation module, an oil temperature PID controller module, an improved candle flame optimization algorithm module, an oil temperature regulation module, and an oil temperature monitoring module; S2. An improved candle flame optimization algorithm is introduced, with specific improvements including: S21. Introduce a nonlinear adaptive parameter harmonicization strategy to dynamically adjust the heat propagation coefficient based on the current iteration number t. With random displacement factor It can use a large-scale global exploration in the early stage of iteration, and automatically switch to fine-grained local mining of the optimal solution region in the later stage of iteration. S22. Introduce an elite-led heat propagation strategy to select the top p% of individuals with the best fitness from the current population to form an elite pool. A regular fire-type individual will randomly select an elite target from this elite pool. They learn and move, thereby enhancing population diversity and expanding the effective search area; S23, Introducing based on The random displacement strategy of the flight mechanism generates a step vector L that combines high-frequency short step size with low-frequency long-distance jump characteristics, enabling the flame individual to both deeply mine advantageous areas and jump to new areas with a certain probability to explore, thus improving global search performance. S3. The improved candle flame optimization algorithm is used to tune the PID control parameters of the oil temperature in the oil-immersed transformer oil temperature PID control system, and the optimal control parameters are obtained through optimization. , , ; S4. The three optimal control parameters obtained by using the improved candle flame optimization algorithm are set as the parameters of the oil temperature PID controller in the oil-immersed transformer PID control system to optimize the oil temperature control effect of the oil-immersed transformer.

2. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: In the oil-immersed transformer temperature PID control system constructed in step S1, the top oil temperature is collected by the temperature monitoring module and transmitted to the temperature error calculation module. The oil temperature error calculation module receives the set target oil temperature, calculates the error between the target oil temperature and the top oil temperature, and outputs the real-time error to the oil temperature PID controller module. The control quantity is calculated based on the real-time error and output to the oil temperature regulation module to achieve oil temperature regulation.

3. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: In step S1, the thermodynamic characteristics of the oil-immersed transformer are modeled, and a mathematical model of the response of the top oil temperature of the oil-immersed transformer to the power regulation of the cooling system is constructed: ， in, This is the process gain, representing the degree Celsius change in oil temperature caused by a 1% change in cooling power. Since increasing cooling power leads to a temperature decrease, it is a negative value. Pure time delay represents the time interval between a change in the input and the start of a response in the output. is the process time constant, representing the magnitude of the system response inertia, and s is the independent variable of the Laplace transform.

4. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: In step S21, the nonlinear adaptive parameter harmonicization strategy includes the following steps: S2101. At the beginning of each iteration, in order to dynamically balance the exploration and mining phases of the algorithm, calculate the nonlinear incremental heat propagation coefficient used to modulate the heat propagation process. , ， in, The heat propagation coefficient for the current iteration, and These are the preset minimum and maximum values ​​of the coefficient, respectively, where t is the current iteration number. This represents the maximum number of iterations. S2102. Calculate the nonlinear decreasing random displacement factor used to dominate the random displacement intensity of Lévy flight. This guides the algorithm to smoothly transition its search focus from global exploration to local mining. ， in, These are the preset maximum and minimum values ​​of the factor, respectively, where t is the current iteration number. This represents the maximum number of iterations.

5. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: In step S22, the elite-guided heat propagation strategy includes the following steps: S2201. In each iteration, the current flame population is sorted by fitness, and the top p% of flame individuals with the best fitness are selected to form an elite pool. , where p is the preset elite ratio; S2202, Execute elite-guided heat propagation: For each Flame individual i in the population, from the elite pool Randomly select an elite Flame as the guiding target. And perform a learning move toward that target. ， in, For flame individuals Calculated potential new locations, For flame individuals Current location For adaptive thermal propagation coefficient, For flame individuals The normalized temperature is calculated as follows: ,in For individuals fitness value, and These are the best and worst fitness values ​​for the current population, respectively. A randomly selected elite fire-guided target.

6. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: In step S23, based on The stochastic displacement strategy of the flight mechanism includes the following steps: S2301, Generate a conforming A stably distributed random step size vector L, ， Where L is the generated multidimensional step size vector, and λ is the stability index that controls the distribution shape and takes a value between (1,2]. S2302. For each individual flame, its current position and the adaptive factor it has passed through are compared. Scaling Step vector Add them together to perform an efficient perturbation operation. ， in, For flame individuals Calculated potential new locations, For flame individuals Current location Let L be the adaptive random displacement factor. Flight step length vector.

7. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 1, characterized in that: Step S3, which uses an improved candle flame optimization algorithm to tune the PID control parameters of the oil temperature in the oil-immersed transformer oil temperature PID control system, includes the following steps: S31. Set the control parameters of the improved candle flame optimization algorithm, and randomly generate an initial population containing multiple flame individuals within the preset PID parameter search space. S32. Before reaching the preset maximum number of iterations, repeatedly update the position of each flame individual in the population, and update the global optimal solution based on the updated position; S33. After the iterative optimization step is terminated, the final global optimal solution is output as the optimal combination of oil temperature PID control parameters.

8. The method for monitoring and controlling the oil temperature of an oil-immersed transformer according to claim 7, characterized in that: Step S32 also includes a step of resetting the stagnant flame individual, specifically: S321. Set a stagnation counter for each flame individual. If the fitness of an individual fails to improve after completing the two independent update attempts, its counter is incremented by one; otherwise, it is reset to zero. S322. After all individuals in the population have completed their update attempts, check all stall counters. If the count value of an individual reaches the preset stall threshold k, then the position of that individual will be randomly reinitialized within the PID parameter search space.

Images