Thermal protection PID control method based on improved hypersonic flight vehicle
Through the improved thermal protection PID control method of hypersonic aircraft, the improved Lotus effect algorithm is used to optimize the PID controller parameters, and the thermal protection problem of hypersonic aircraft in extreme high temperature environments is solved, achieving more efficient and reliable thermal protection control.
Patent Information
- Application Number
- CN202411691947.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-25
- Publication Date
- 2025-03-14
AI Technical Summary
Hypersonic aircraft faces extreme high-temperature environments during high-speed flights. The existing thermal protection measures are difficult to meet the needs of the entire flight, and the high-temperature materials are costly.
The improved thermal protection PID control method of hypersonic aircraft is adopted, and the PID control model is constructed, the Lotus effect algorithm is improved, the parameters of the PID controller are optimized and adjusted, and the effectiveness of the thermal protection parameters is verified through MATLAB and Simulink.
The adaptive control of the thermal protection system of the hypersonic aircraft is realized, which significantly improves the system's adjustment accuracy and response speed, and enhances the stability and reliability in complex environments.
Smart Images
Figure CN119620589A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of hypersonic aircraft thermal protection, and in particular to a PID control method based on improved hypersonic aircraft thermal protection. Background Art
[0002] In the modern aerospace industry, as the complexity and performance requirements of flight missions continue to increase, hypersonic aircraft face extreme thermal environment challenges when performing high-speed and long-distance flight missions. These challenges mainly come from the high temperatures generated by the friction between the aircraft and the high-speed air during flight, which can damage the structural materials of the aircraft.
[0003] For hypersonic aircraft, existing thermal protection measures, such as the use of passive thermal protection materials and structures, are often unable to meet the full flight requirements of hypersonic aircraft. Conventional high-temperature resistant materials are difficult to withstand extreme high temperatures, and those materials that can withstand extreme high temperatures are often accompanied by significant cost increases. Therefore, it is now necessary to develop an improved hypersonic aircraft thermal protection system to ensure that the temperature of the aircraft surface can always be carefully managed. Summary of the invention
[0004] The embodiment of the present application provides a PID control method based on an improved hypersonic aircraft thermal protection system, which can dynamically balance global search and local development, and more efficiently find the global optimal solution, which not only speeds up the parameter optimization speed, but also improves the adjustment accuracy and response speed of the system, effectively enhances its stability and reliability in complex environments, and provides a more sensitive and reliable control strategy for the hypersonic aircraft thermal protection system, which can significantly improve the robustness and adaptability of the system.
[0005] The embodiment of the present application provides a PID control method for thermal protection of an improved hypersonic aircraft, comprising the following steps:
[0006] S10, constructing a PID control model for a hypersonic vehicle thermal protection system;
[0007] S20, improved lotus effect algorithm;
[0008] S30, using the improved lotus effect algorithm to optimize and adjust the parameters of the PID controller;
[0009] S40, simulating the PID control model to determine the optimal solution, that is, the optimal three parameters Kp, Ki, and Kd, applying the tuned thermal protection parameters to the PID control model to verify whether the temperature of the thermal protection system of the hypersonic aircraft can be maintained within a predetermined temperature target range, thereby achieving adaptive control of the thermal protection system of the hypersonic aircraft;
[0010] Wherein, step S20 is specifically as follows:
[0011] S21, in the population initialization stage, a dual-track excitation kernel initialization population is introduced based on the guiding field array idea. Specifically, two sets of parameter trajectories are introduced, each set of trajectories is optimized for different exploration goals, where the first set of trajectories focuses on exploring in a wide parameter space to identify potentially favorable areas, and the second set of trajectories focuses on in-depth exploration of the identified areas to fine-tune the best performing potential solutions. The formula is as follows:
[0012] Positions i,j =rand1×(ub j -lb j )+lb j +DE(F,npop,x all ) (1)
[0013]
[0014] In formula (1), Positions i，j is the position of the i-th individual in the j-th dimension, where i is the individual index and j is the index of the problem dimension, all possible values of which are 1, 2, and 3. rand1 is a random number generated in the interval [0, 1], which is used to randomly determine the initial position of the individual in each dimension. ub j and lb j They represent the upper and lower bounds of the j-th dimension parameter, respectively. In formula (2), DE is the dual-rule excitation kernel, F is the scaling factor used to adjust the influence of the trajectory, npop is the size of the population, and represents the total number of individuals in the solution set. x_all is the set of all candidate solutions. randi(npop) Indicates the position of an individual randomly selected from the population for the double-rule excitation operation. rand is a random number generated in the interval [0, 1], and p is the probability threshold. When rand is greater than p, the double-rule excitation is performed, otherwise no operation is performed.
[0015] S22, uses elastic adaptive activation method to dynamically balance global search and local development, the formula is:
[0016]
[0017] Where Levy(x) is the random step value, λ is the elastic adjustment factor, which dynamically adjusts the step size according to the current individual fitness, fit(x) is the fitness value of the current individual, r1 and r2 are two independent random variables, r1 is used to generate the direction of the step, r2 is used to generate the size of the step, σ is the step scale parameter, which is used to control the scale of the step in the Levy distribution, β is the shape parameter, which determines the tail characteristics of the Levy distribution, and is used to control the probability of jumping and affect the balance of global and local searches. is the gamma function, which is used to calculate the constants related to the distribution, fit(g * ) is the fitness value of the global optimal solution;
[0018] S23. By using the collaborative control fusion adjustment strategy, in the iterative update phase, the agent dynamically adjusts the behavior mode according to the neighboring agents and environmental feedback to achieve search and adaptation in a complex environment. The formula is:
[0019]
[0020] Where t is the current iteration number, represents the position change of individual i in the t+1th iteration, represents the position change of individual i in the tth iteration, a is the alignment coefficient, which is used to adjust the degree of alignment of individual i with the average speed of its neighboring individuals, C is the aggregation coefficient, which is used to adjust the degree of movement of individual i to the geometric center of the neighboring group, f is the food attraction coefficient, which is used to adjust the degree of attraction of individual i to the known food source position, e is the enemy avoidance coefficient, which is used to adjust the degree of avoidance of individual i from potential threats or enemy positions, S is the separation coefficient, which is used to adjust the spatial distance of individual i to avoid being too close to neighboring individuals and maintain an appropriate spatial distance, β is the scaling factor, which is used to control the intensity of the coefficient acfe, and the tanh function is used to perform nonlinear transformation on each behavioral factor coefficient to ensure that the output value is in the range of [-1, 1], W is the inertia weight, which controls the intensity of the position influence at the previous moment, is the influence of the separation, alignment, aggregation, food attraction, and enemy avoidance behaviors of individual i in the tth iteration. The specific calculation formulas are shown in Equations (7), (8), (9), (10), and (11);
[0021]
[0022]
[0023] Where t is the current iteration number, is the position of the ith individual at the tth iteration, is the position of the jth individual at the tth iteration, N is the total number of individuals in the population, is the best position found so far, This is the worst location found so far.
[0024] In a possible implementation, in step S10, the PID control model includes a data acquisition module, a signal processing module, an improved lotus effect optimization algorithm module, an intelligent PID control module, an actuator execution module, and a feedback module;
[0025] The data acquisition module is used to collect the temperature of the aircraft surface, and uses a pressure sensor and a velocity sensor to monitor the pressure and airflow velocity on the aircraft surface;
[0026] The signal processing module is used to filter and calibrate the collected environmental data;
[0027] The improved lotus effect optimization algorithm module is used to obtain the three parameters Kp, Ki, and Kd, and pass the three parameters to the intelligent PID control module;
[0028] The intelligent PID control module is used to dynamically adjust the thermal protection measures using the PID algorithm according to the processed environmental data;
[0029] The actuator execution module is used to control the hypersonic aircraft thermal protection system according to the output of the intelligent PID control module to adjust the thermal protection state of the aircraft in real time;
[0030] The feedback module is used to monitor the effect of the control execution result to optimize the thermal protection measures.
[0031] In a possible implementation, in the data acquisition module, the data acquisition module uses a B-type thermocouple to collect the surface temperature of the aircraft.
[0032] In a possible implementation, step S30 is specifically as follows:
[0033] S31, parameter initialization: Based on the idea of guiding field array, the dual-track excitation kernel is introduced to initialize the lotus population position, the upper bound ub and lower bound lb of the algorithm search, the solution space dimension dim, the population size N, and the maximum number of population iterations L. The guiding field array formula is shown in equations (1) and (2);
[0034] S32, select a fitness function for evaluating the adaptability of the solution in a dynamic environment, and the fitness function formula is:
[0035] J = ∫0 ∞ e -λt sin(te(t))dt (12);
[0036] where \(t\) is the time variable, \(e(t)\) is the real-time error between the actual temperature and the target temperature of the hypersonic vehicle thermal protection system, \(\lambda\) is the attenuation factor, which is used to adjust the attenuation rate of the error over time and improve the sensitivity of the system to early errors, \(e\) -λt is the exponential weight factor, which decays over time and is used to quickly correct the latest deviation, and \(J\) is the fitness function value;
[0037] S33. In the ecological exploration and refinement stage, the elastic activation adaptation method is used to dynamically adjust the search step size. When the calculated fitness value is small, that is, when the quality of the solution is poor, a larger step size is adopted for exploration to avoid falling into local optima; when the calculated fitness value is large, that is, when the quality of the solution is good, the step size is reduced to make the search more concentrated in a small range around the current solution, which helps for more detailed exploration to find a more accurate optimal solution. The elastic adaptation activation method formulas are shown in Eqs. (3), (4), and (5);
[0038] S34. In the development stage, the selection factor \(p\) is calculated and compared with the selection threshold \(p1\). If \(p < p1\), the cooperative control fusion adjustment strategy is adopted, and the behavior pattern is dynamically adjusted according to the neighboring agents and environmental feedback, and the position of the search agent in space is updated. The update formula of the cooperative control fusion adjustment strategy is shown in Eq. (6), and the update formulas are shown in Eqs. (7), (8), (9), (10), and (11);
[0039] Otherwise, execute:
[0040]
[0041]
[0042] where \(p1\) is the experimental setting value, taking 0.5, \(t\) is the current iteration number, is the position of the \(i\)-th individual at the \((t + 1)\)-th iteration, is the position of the \(i\)-th individual at the current \(t\)-th iteration, \(g\) * is the best position found in all iterations up to the current iteration number, \(R\) is the growth region, which shrinks as the algorithm iterates, and \(L\) is the final iteration number;
[0043] S35. In the development and strengthening stage, the movement of droplets on lotus leaves is simulated as a local search. After the water droplets are formed on the leaves, they will be placed in the local optimal position closest to them. This local optimum is called a pit. The water droplets will move to the pits on the leaves. When they are filled, the water will overflow until the next pit is found and filled and overflowed. Each pit has its depth, and the depth determines its capacity. The capacity evaluates the quality of the solution, so the deepest pit is considered to be the most valuable pit, which is the location of the local optimal solution. The location of the deepest pit where the water droplets are located is the local optimal solution in the iteration process. At this stage, local search is performed and the capacity of each pit is calculated. The calculation formula for the pit capacity is as follows:
[0044]
[0045] in, is the pit capacity of the ith solution at the tth iteration, indicating the quality of the solution, is the fitness value of the i-th solution in the t-th iteration, f Max is the maximum fitness value of all solutions in the current iteration process, f Min is the minimum fitness value of all solutions in the current iteration process, and const is a constant used to adjust the influence of pit capacity;
[0046] S36. In the development and enhancement stage, the movement of droplets on the surface of lotus leaves is used to simulate local search. After calculating the capacity of each potential position, that is, the pit, the droplet moves to the pit. If any pit is filled with water and overflows, the excess water will flow to other pits. The dynamics of the droplet includes the update of speed and position. The formulas for updating the speed and position of the droplet moving to the pit and the movement speed and position of the droplet when the pit overflows are as follows:
[0047]
[0048] In the formula, is the moving speed of the ith droplet at the t+1th iteration, q is the attenuation coefficient, and the control speed gradually decreases over time, which helps stabilize the algorithm and avoid overshoot. is the moving speed of the ith droplet at the tth iteration, is the position of the i-th droplet after the t+1th iteration, is the position of the ith droplet after the tth iteration; Rand2 is a random number generated in the interval [0, 1], which is used to enhance the local search capability of the algorithm. The location of the deepest pit searched up to the tth iteration is the local optimal solution;
[0049] S37, evaluating the fitness value of the new solution, if the fitness value of the new solution is higher than the current optimal solution, setting it as the new optimal solution and updating it;
[0050] S38. Check whether the maximum number of iterations has been reached. If so, the algorithm stops iterating and decomposes the current optimal solution into three parameters, Kp, Ki, and Kd. These parameters are then used in the PID controller to optimize the PID parameters. If the maximum number of iterations has not been reached, return to step S30 to continue exploring for a better solution.
[0051] In a possible implementation, in step S40, MATLAB and Simulink are used to simulate the PID control model of the hypersonic vehicle thermal protection system.
[0052] In a possible implementation, in step S40, the thermal protection PID controller obtains the deviation e(t) between the desired temperature setpoint and the actual temperature through the deviation calculation module, calculates the Kp, Ki, and Kd parameters calculated by the improved lotus effect algorithm, outputs the control quantity u(t) to the execution unit module, and then sends a command to the controlled object module to control the controlled object module to output a control command to maintain the actual temperature to reach the desired temperature in a short time and stabilize near the desired temperature. In the simulation process, the system transfer function is used as the controlled object to verify the significant advantages of the proposed improved algorithm in optimizing the PID controller parameters. The system transfer function formula is:
[0053]
[0054] Where G(s) is the transfer function and S is the function variable.
[0055] Beneficial effects: Compared with the prior art, the improved hypersonic vehicle thermal protection PID control method provided by the present application is based on the original lotus effect algorithm, uses the guiding field array idea, introduces a dual-track excitation core initialization population, and effectively enhances the algorithm's exploration breadth and solution diversity;
[0056] Through the elastic adaptive activation method and the collaborative control fusion adjustment strategy, the algorithm can dynamically balance between global search and local development, and can more efficiently find the global optimal solution. It not only speeds up the parameter optimization, but also improves the adjustment accuracy and response speed of the system, effectively enhances its stability and reliability in complex environments, and provides a more sensitive and reliable control strategy for the thermal protection system of hypersonic aircraft, significantly improving the robustness and adaptability of the system.
[0057] In contrast, although traditional PID control methods and standard lotus effect algorithms provide certain basic performance, they have problems such as slow convergence, easy to fall into local optimality and insufficient optimization accuracy, which is particularly obvious when facing rapidly changing environmental conditions (such as hypersonic flight). The improved hypersonic aircraft thermal protection PID control method provided in this application effectively overcomes these shortcomings through dynamic adjustment and optimization strategies.
[0058] These and other objects, features and advantages of the present invention will be fully reflected in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] Figure 1 A schematic diagram of the PID control process of the present application based on the improved hypersonic aircraft thermal protection PID control method is shown.
[0060] Figure 2 A schematic diagram of the PID controller model in this application is shown.
[0061] Figure 3 The system block diagram of the improved hypersonic aircraft thermal protection PID control method of the present application is shown.
[0062] Figure 4 This is a comparison chart of the optimization of the proportional gain Kp of PID using the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm.
[0063] Figure 5 This is a comparison chart of the optimization of the integral gain Ki of PID using the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm.
[0064] Figure 6 This is a comparison chart of the optimization of the differential gain Kd of PID using the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm.
[0065] Figure 7 A comparison chart of the changes in fitness values during the optimization process of the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm.
[0066] Figure 8 Comparison chart of control effects of PID optimization using improved lotus effect optimization algorithm and basic lotus effect optimization algorithm. DETAILED DESCRIPTION
[0067] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.
[0068] Those skilled in the art should understand that, in the disclosure of the specification, the orientation or position relationship indicated by the terms "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc. are based on the orientation or position relationship shown in the drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the above terms should not be understood as limiting the present invention.
[0069] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term "one" should not be understood as a limitation on the quantity.
[0070] refer to Figures 1 to 8 The embodiment of the present application provides a PID control method based on thermal protection of an improved hypersonic aircraft, which uses an improved lotus effect optimization algorithm to quickly find the optimal solution and efficiently determine the PID control parameters suitable for the thermal protection system of the hypersonic aircraft, thereby significantly improving the stability, response speed, reliability and control accuracy of the aircraft thermal protection system.
[0071] Specifically, a PID control method for thermal protection of an improved hypersonic vehicle comprises the following steps:
[0072] S10, construct a PID control model for the thermal protection system of a hypersonic vehicle. The model is as follows Figure 3 As shown;
[0073] S20, an improved lotus effect algorithm, includes the use of the guiding field array idea, the introduction of a dual-track excitation core to initialize the population, the use of elastic adaptive activation method, dynamic balance of global search and local development, and the use of a collaborative control fusion adjustment strategy to dynamically adjust the population behavior mode, thereby effectively improving the ability to find the global optimal solution;
[0074] S30, using the improved lotus effect algorithm to optimize and adjust the parameters of the PID controller;
[0075] S40, simulating the PID control model using MATLAB and Simulink, determining the optimal solution, i.e., the optimal three parameters Kp, Ki, and Kd, by continuous iteration and calculation of fitness values, applying the tuned thermal protection parameters to the PID control model, verifying whether the temperature of the thermal protection system of the hypersonic aircraft can be maintained within a predetermined temperature target range, and realizing adaptive control of the thermal protection system of the hypersonic aircraft;
[0076] The algorithm involved in step S20 includes three parts, specifically:
[0077] S21, in the population initialization stage, based on the idea of guiding field array, a dual-track excitation kernel is introduced to initialize the population. More specifically, two sets of parameter trajectories are introduced, each set of trajectories is optimized for different exploration goals, where the first set of trajectories focuses on exploring in a wide parameter space to identify potentially favorable areas, and the second set of trajectories focuses on in-depth mining of identified areas to fine-tune the best performing potential solutions. The formula is as follows:
[0078] Positions i，j =rand1×(ub j -lb j )+lb j +DE(F,npop,x all ) (1)
[0079]
[0080] In formula (1), Positions i，j is the position of the ith individual in the jth dimension. This is an individual position matrix in a population, where i is the individual index and j is the index of the problem dimension. Since the goal is to adjust the three parameters of the PID controller, the problem dimension in this application scenario is 3, and all possible values of j are 1, 2, and 3. rand1 is a random number generated in the interval [0, 1], which is used to randomly determine the initial position of the individual in each dimension. ub j and lb j They represent the upper and lower bounds of the j-th dimension parameter, respectively, and are used to limit the range of the parameter. In formula (2), DE is the dual-rule excitation kernel, F is the scaling factor, which is used to adjust the influence of the trajectory, npop is the size of the population, and represents the total number of individuals in the solution set. x_all is the set of all candidate solutions, and x_all randi(npop) Indicates the position of an individual randomly selected from the population for the double-rule excitation operation. rand is a random number generated in the interval [0, 1], and p is the probability threshold. When rand is greater than p, the double-rule excitation is performed, otherwise no operation is performed.
[0081] S22, uses elastic adaptive activation method to dynamically balance global search and local development, the formula is:
[0082]
[0083] Where Levy(x) is the random step value, λ is the elastic adjustment factor, which dynamically adjusts the step size according to the current individual fitness, fit(x) is the fitness value of the current individual, r1 and r2 are two independent random variables, r1 is used to generate the direction of the step, r2 is used to generate the size of the step, σ is the step scale parameter, which is used to control the scale of the step in the Levy distribution, β is the shape parameter, which determines the tail characteristics of the Levy distribution, and is used to control the probability of jumping and affect the balance of global and local searches. is the gamma function, which is used to calculate the constants related to the distribution, fit(g * ) is the fitness value of the global optimal solution;
[0084] S23. By using the collaborative control fusion adjustment strategy, in the iterative update phase, the agent dynamically adjusts the behavior mode according to the neighboring agents and environmental feedback to achieve search and adaptation in a complex environment. The formula is:
[0085]
[0086] Where t is the current iteration number, represents the position change of individual i in the t+1th iteration, represents the position change of individual i in the tth iteration, a is the alignment coefficient, which is used to adjust the degree of alignment of individual i with the average speed of its neighboring individuals, C is the aggregation coefficient, which is used to adjust the degree of movement of individual i to the geometric center of the neighboring group, f is the food attraction coefficient, which is used to adjust the degree of attraction of individual i to the known food source position, e is the enemy avoidance coefficient, which is used to adjust the degree of avoidance of individual i from potential threats or enemy positions, S is the separation coefficient, which is used to adjust the spatial distance of individual i to avoid being too close to neighboring individuals and maintain an appropriate spatial distance, β is the scaling factor, which is used to control the intensity of the coefficient acfe, and the tanh function is used to perform nonlinear transformation on each behavioral factor coefficient to ensure that the output value is in the range of [-1, 1], W is the inertia weight, which controls the intensity of the position influence at the previous moment, is the effect of separation, alignment, aggregation, food attraction, and enemy avoidance behaviors of individual i in the tth iteration, The specific calculation formulas of are shown in formula (7), formula (8), formula (9), formula (10) and formula (11);
[0087]
[0088]
[0089] Where t is the current iteration number, is the position of the ith individual at the tth iteration, is the position of the jth individual at the tth iteration, N is the total number of individuals in the population, is the best position found so far, This is the worst location found so far.
[0090] In one embodiment, in step S10, the PID control model includes a data acquisition module, a signal processing module, an improved lotus effect optimization algorithm module, an intelligent PID control module, an actuator execution module, and a feedback module;
[0091] The data acquisition module is used to collect the surface temperature of the aircraft, and uses a pressure sensor and a velocity sensor to monitor the pressure on the surface of the aircraft and the airflow velocity, wherein the data acquisition module can use a B-type thermocouple to collect the surface temperature of the aircraft;
[0092] The signal processing module is used to filter and calibrate the collected environmental data;
[0093] The improved lotus effect optimization algorithm module is used to obtain the three parameters Kp, Ki, and Kd, and transfer the three parameters to the intelligent PID control module;
[0094] The intelligent PID control module is used to dynamically adjust the thermal protection measures using the PID algorithm according to the processed environmental data;
[0095] The actuator execution module is used to control the hypersonic aircraft thermal protection system according to the output of the intelligent PID control module to adjust the thermal protection state of the aircraft in real time;
[0096] The feedback module is used to monitor the effect of the control execution result to optimize the thermal protection measures.
[0097] In one embodiment, step S30 is specifically as follows:
[0098] S31, parameter initialization: Based on the idea of guiding field array, the dual-track excitation kernel is introduced to initialize the lotus population position, the upper bound ub and lower bound lb of the algorithm search, the solution space dimension dim, the population size N, and the maximum number of population iterations L. The guiding field array formula is shown in equations (1) and (2);
[0099] S32, select a fitness function for evaluating the adaptability of the solution in a dynamic environment, and the fitness function formula is:
[0100] J = ∫0 ∞ e-λt ∫sin(te(t))dt (12);
[0101] In the formula, t is the time variable, e(t) is the real-time error between the actual temperature and the target temperature of the thermal protection system of the hypersonic vehicle, λ is the attenuation factor, which is used to adjust the attenuation rate of the error over time and improve the sensitivity of the system to early errors, e -λt is the exponential weight factor, which decays over time and is used to quickly correct the latest deviation, and J is the fitness function value;
[0102] S33. In the ecological exploration and refinement stage, the elastic activation adaptation method is used to dynamically adjust the search step size. When the calculated fitness value is small, that is, when the quality of the solution is poor, a larger step size is used for exploration to avoid falling into local optima; when the calculated fitness value is large, that is, when the quality of the solution is good, the step size is reduced to make the search more concentrated in a small range around the current solution, which helps for more detailed exploration to find a more accurate optimal solution. The formula of the elastic adaptation activation method is shown in Equations (3), (4) and (5);
[0103] S34. In the development stage, to prevent falling into local optima, the selection factor p is calculated and compared with the selection threshold p1. If p < p1, the cooperative control fusion adjustment strategy is adopted, and the behavior pattern is dynamically adjusted according to the neighboring agents and environmental feedback, and the position of the search agent in space is updated. The update formula of the cooperative control fusion adjustment strategy is shown in Equation (6), and the update formulas are shown in Equations (7), (8), (9), (10) and (11); otherwise, execute:
[0104]
[0105] where p1 is the experimental setting value, taking 0.5, t is the current iteration number, is the position of the i-th individual at the (t + 1)-th iteration, is the position of the i-th individual at the current t-th iteration, g * is the best position found in all iteration processes up to the current iteration number, R is the growth region, which shrinks as the algorithm iterates, and L is the final iteration number;
[0106] S35. In the development and strengthening stage, the movement of droplets on lotus leaves is simulated as a local search. After the water droplets are formed on the leaves, they will be placed in the local optimal position closest to them. This local optimum is called a pit. The water droplets will move to the pits on the leaves. When they are filled, the water will overflow until the next pit is found and filled and overflowed. Each pit has its depth, and the depth determines its capacity. The capacity evaluates the quality of the solution, so the deepest pit is considered to be the most valuable pit, which is the location of the local optimal solution. The location of the deepest pit where the water droplets are located is the local optimal solution in the iteration process. At this stage, local search is performed and the capacity of each pit is calculated. The calculation formula for the pit capacity is as follows:
[0107]
[0108] in, is the pit capacity of the ith solution at the tth iteration, indicating the quality of the solution, f i t is the fitness value of the i-th solution in the t-th iteration, f Max is the maximum fitness value of all solutions in the current iteration process, f Min is the minimum fitness value of all solutions in the current iteration process, and const is a constant used to adjust the influence of pit capacity;
[0109] S36. In the development and enhancement stage, the movement of droplets on the surface of lotus leaves is used to simulate local search. After calculating the capacity of each potential position, that is, the pit, the droplet moves to the pit. If any pit is filled with water and overflows, the excess water will flow to other pits. The dynamics of the droplet includes the update of speed and position. The formulas for updating the speed and position of the droplet moving to the pit and the movement speed and position of the droplet when the pit overflows are as follows:
[0110]
[0111]
[0112] Where V i t+1 is the moving speed of the ith droplet at the t+1th iteration, q is the attenuation coefficient, and the control speed gradually decreases over time, which helps to stabilize the algorithm and avoid overshoot. V i t is the moving speed of the ith droplet at the tth iteration, is the position of the i-th droplet after the t+1th iteration, is the position of the ith droplet after the tth iteration; Rand2 is a random number generated in the interval [0, 1], which is used to enhance the local search capability of the algorithm. The location of the deepest pit searched up to the tth iteration is the local optimal solution;
[0113] S37, evaluating the fitness value of the new solution, if the fitness value of the new solution is higher than the current optimal solution, setting it as the new optimal solution and updating it;
[0114] S38. Check whether the maximum number of iterations has been reached. If so, the algorithm stops iterating and decomposes the current optimal solution into three parameters, Kp, Ki, and Kd. These parameters are then used in the PID controller to optimize the PID parameters. If the maximum number of iterations has not been reached, return to step S30 to continue exploring for a better solution.
[0115] In one embodiment, in step S40, the thermal protection PID controller obtains the deviation e(t) between the desired temperature setpoint and the actual temperature through the deviation calculation module, calculates the Kp, Ki, and Kd parameters calculated by the improved lotus effect algorithm, outputs the control quantity u(t) to the execution unit module, and then sends a command to the controlled object module to control the controlled object module to output a control command to maintain the actual temperature to reach the desired temperature in a short time and stabilize near the desired temperature. In the simulation process, the system transfer function is used as the controlled object to verify the significant advantages of the proposed improved algorithm in optimizing the PID controller parameters. The system transfer function formula is:
[0116]
[0117] Where G(s) is the transfer function and S is the function variable.
[0118] By analyzing Figure 4 , Figure 5 and Figure 6 From the performance of the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm in the process of optimizing PID control parameters shown in the figure, it is not difficult to see that the improved lotus effect optimization algorithm can approach the most ideal Kp, Ki, and Kd values in fewer iterations and achieve the optimal control effect. This shows that compared with the basic version of the lotus effect optimization algorithm, the improved version of the algorithm is faster in optimization and convergence speed, and at the same time improves the stability of the system. This makes the improved lotus effect optimization algorithm more effective in rapidly changing application environments, and can quickly adapt to and optimize the parameters of the PID controller to achieve better control accuracy and responsiveness.
[0119] The fitness value is usually used to evaluate the quality of the solution in the population. Generally speaking, the smaller the fitness value, the higher the quality of the solution. Therefore, by comparing Figure 7From the fitness data of the improved lotus effect optimization algorithm and the basic lotus effect optimization algorithm shown, it can be seen that the fitness value of the improved lotus effect optimization algorithm is smaller. In addition, the fitness value of the improved lotus effect optimization algorithm reaches a stable state faster, which indicates that the improved lotus effect optimization algorithm can effectively converge to the optimal solution within a smaller number of iterations, thereby providing faster and more stable optimization performance. Therefore, compared with the basic version, the improved lotus effect optimization algorithm exhibits superior performance.
[0120] from Figure 8 The results show that the overshoot of the improved lotus effect optimization algorithm in the PID control system is significantly lower than that of the basic lotus effect optimization algorithm. In addition, the improved algorithm provides a faster tracking response and shortens the iteration time required to reach a stable state. This shows that in a complex dynamic environment, the improved lotus effect optimization algorithm can enable the system to reach a steady state faster, demonstrate better response speed and control performance, and have stronger robustness. These characteristics enable the algorithm to perform well in the application of hypersonic vehicle thermal protection systems, effectively improving the overall efficiency and stability of the system.
[0121] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The advantages of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.
Claims
1. A PID control method for thermal protection of an improved hypersonic vehicle, characterized in that: The following steps are involved: S10, constructing a PID control model for a hypersonic vehicle thermal protection system; S20, improved lotus effect algorithm; S30, using the improved lotus effect algorithm to optimize and adjust the parameters of the PID controller; S40, simulating the PID control model to determine the optimal solution, that is, the optimal three parameters Kp, Ki, and Kd, applying the tuned thermal protection parameters to the PID control model to verify whether the temperature of the thermal protection system of the hypersonic aircraft can be maintained within a predetermined temperature target range, thereby achieving adaptive control of the thermal protection system of the hypersonic aircraft; Wherein, step S20 is specifically as follows: S21, in the population initialization stage, a dual-track excitation kernel initialization population is introduced based on the guiding field array idea. Specifically, two sets of parameter trajectories are introduced, each set of trajectories is optimized for different exploration goals, where the first set of trajectories focuses on exploring in a wide parameter space to identify potentially favorable areas, and the second set of trajectories focuses on in-depth exploration of the identified areas to fine-tune the best performing potential solutions. The formula is as follows: Positions i，j =rand1×(ub j -lb j )+lb j +DE(F,npop,x all ) (1) In formula (1), Positions i，j is the position of the i-th individual in the j-th dimension, where i is the individual index and j is the index of the problem dimension, all possible values of which are 1, 2, and 3. rand1 is a random number generated in the interval [0,1], which is used to randomly determine the initial position of the individual in each dimension. ub j and lb j They represent the upper and lower bounds of the j-th dimension parameter, respectively. In formula (2), DE is the dual-rule excitation kernel, F is the scaling factor used to adjust the influence of the trajectory, npop is the size of the population, and represents the total number of individuals in the solution set. x_all is the set of all candidate solutions. randi(npop) Indicates the position of an individual randomly selected from the population for the double-rule excitation operation. rand is a random number generated in the interval [0,1], and p is the probability threshold. When rand is greater than p, the double-rule excitation is performed, otherwise no operation is performed. S22, using elastic adaptive activation method to dynamically balance global search and local development, the formula is: Where Levy(x) is the random step value, λ is the elastic adjustment factor, which dynamically adjusts the step size according to the current individual fitness, fit(x) is the fitness value of the current individual, r1 and r2 are two independent random variables, r1 is used to generate the direction of the step, r2 is used to generate the size of the step, σ is the step scale parameter, which is used to control the scale of the step in the Levy distribution, β is the shape parameter, which determines the tail characteristics of the Levy distribution, and is used to control the probability of jumping and affect the balance of global and local searches. is the gamma function, which is used to calculate the constants related to the distribution, fit(g * ) is the fitness value of the global optimal solution; S23. By using the collaborative control fusion adjustment strategy, in the iterative update phase, the agent dynamically adjusts the behavior mode according to the neighboring agents and environmental feedback to achieve search and adaptation in a complex environment. The formula is: Where t is the current iteration number, represents the position change of individual i in the t+1th iteration, represents the position change of individual i in the tth iteration, a is the alignment coefficient, which is used to adjust the degree of alignment of individual i with the average speed of its neighboring individuals, c is the aggregation coefficient, which is used to adjust the degree of movement of individual i to the geometric center of the neighboring group, f is the food attraction coefficient, which is used to adjust the degree of attraction of individual i to the known food source position, e is the enemy avoidance coefficient, which is used to adjust the degree of avoidance of individual i from potential threats or enemy positions, s is the separation coefficient, which is used to adjust the spatial distance of individual i to avoid being too close to neighboring individuals and maintain an appropriate spatial distance, β is the scaling factor, which is used to control the intensity of the coefficient acfe, and the tanh function is used to perform nonlinear transformation on each behavioral factor coefficient to ensure that the output value is in the range of [-1,1], w is the inertia weight, which controls the intensity of the position influence at the previous moment, is the influence of the separation, alignment, aggregation, food attraction, and enemy avoidance behaviors of individual i in the tth iteration. The specific calculation formulas are shown in Equations (7), (8), (9), (10), and (11); Where t is the current iteration number, is the position of the ith individual at the tth iteration, is the position of the jth individual at the tth iteration, N is the total number of individuals in the population, is the best position found so far, This is the worst location found so far.
2. The improved hypersonic vehicle thermal protection PID control method according to claim 1, characterized in that: In step S10, the PID control model includes a data acquisition module, a signal processing module, an improved lotus effect optimization algorithm module, an intelligent PID control module, an actuator execution module, and a feedback module; The data acquisition module is used to collect the temperature of the aircraft surface, and uses a pressure sensor and a velocity sensor to monitor the pressure and airflow velocity on the aircraft surface; The signal processing module is used to filter and calibrate the collected environmental data; The improved lotus effect optimization algorithm module is used to obtain the three parameters Kp, Ki, and Kd, and transfer the three parameters to the intelligent PID control module; The intelligent PID control module is used to dynamically adjust the thermal protection measures using the PID algorithm according to the processed environmental data; The actuator execution module is used to control the hypersonic aircraft thermal protection system according to the output of the intelligent PID control module to adjust the thermal protection state of the aircraft in real time; The feedback module is used to monitor the effect of the control execution result to optimize the thermal protection measures.
3. The improved hypersonic vehicle thermal protection PID control method according to claim 2, characterized in that: In the data acquisition module, the data acquisition module uses a B-type thermocouple to collect the surface temperature of the aircraft.
4. The improved hypersonic vehicle thermal protection PID control method according to claim 1, characterized in that: Step S30 is specifically as follows: S31, parameter initialization: Based on the idea of guiding field array, the dual-track excitation kernel is introduced to initialize the lotus population position, the upper bound ub and lower bound lb of the algorithm search, the solution space dimension dim, the population size N, and the maximum number of population iterations L. The guiding field array formula is shown in equations (1) and (2); S32, select a fitness function for evaluating the adaptability of the solution in a dynamic environment, the fitness function formula is: J=∫0 ∞ it -λt sin(te(t))dt (12); Where t is the time variable, e(t) is the real-time error between the actual temperature and the target temperature of the thermal protection system of the hypersonic vehicle, λ is the attenuation factor, which is used to adjust the attenuation rate of the error over time and improve the system's sensitivity to early errors, and e -λt is an exponential weight factor, which decays over time and is used to quickly correct the latest deviations, and J is the fitness function value; S33. In the ecological exploration and refinement stage, the elastic activation adaptation method is used to dynamically adjust the search step size. When the calculated fitness value is small, that is, the quality of the solution is poor, a larger step size is used for exploration to avoid falling into the local optimum; when the calculated fitness value is large, that is, the quality of the solution is good, the step size is reduced to make the search more concentrated on a small range around the current solution, which is conducive to more detailed exploration to find a more accurate optimal solution. The formula of the elastic adaptive activation method is shown in formula (3), formula (4) and formula (5); S34. In the development stage, calculate the selection factor p, compare the selection factor with the selection threshold p1. If p < p1, then adopt the collaborative control fusion adjustment strategy, dynamically adjust the behavior pattern according to the neighboring agents and environmental feedback, update the position of the search agent in space, and update The collaborative control fusion adjustment strategy is shown in Equation (6), and the update formulas are shown in Equations (7), (8), (9), (10) and (11); Otherwise execute: Among them, p1 is the experimental setting value, which is 0.5, and t is the current number of iterations. is the position of the ith individual at the t+1th iteration, is the position of the i-th individual at the current t-th iteration, g * is the best position found in all iterations up to the current iteration number, R is the growth area, which shrinks as the algorithm iterates, and L is the final iteration number; S35. In the development and strengthening stage, the movement of droplets on lotus leaves is simulated as a local search. After the water droplets are formed on the leaves, they will be placed in the local optimal position closest to them. This local optimum is called a pit. The water droplets will move to the pits on the leaves. When they are filled, the water will overflow until the next pit is found and filled and overflowed. Each pit has its depth, and the depth determines its capacity. The capacity evaluates the quality of the solution, so the deepest pit is considered to be the most valuable pit, which is the location of the local optimal solution. The location of the deepest pit where the water droplets are located is the local optimal solution in the iteration process. At this stage, local search is performed and the capacity of each pit is calculated. The calculation formula for the pit capacity is as follows: in, is the pit capacity of the ith solution at the tth iteration, indicating the quality of the solution, f i t is the fitness value of the i-th solution in the t-th iteration, f Max is the maximum fitness value of all solutions in the current iteration process, f Min is the minimum fitness value of all solutions in the current iteration process, and const is a constant used to adjust the influence of pit capacity; S36. In the development and enhancement stage, the movement of droplets on the surface of lotus leaves is used to simulate local search. After calculating the capacity of each potential position, that is, the pit, the droplet moves to the pit. If any pit is filled with water and overflows, the excess water will flow to other pits. The dynamics of the droplet includes the update of speed and position. The formulas for updating the speed and position of the droplet moving to the pit and the movement speed and position of the droplet when the pit overflows are as follows: Where V i t+1 is the moving speed of the ith droplet at the t+1th iteration, q is the attenuation coefficient, and the control speed gradually decreases over time, which helps to stabilize the algorithm and avoid overshoot. V i t is the moving speed of the ith droplet at the tth iteration, is the position of the i-th droplet after the t+1th iteration, is the position of the ith droplet after the tth iteration; Rand2 is a random number generated in the interval [0,1], which is used to enhance the local search capability of the algorithm. The location of the deepest pit searched up to the tth iteration is the local optimal solution; S37, evaluating the fitness value of the new solution, if the fitness value of the new solution is higher than the current optimal solution, setting it as the new optimal solution and updating it; S38. Check whether the maximum number of iterations has been reached. If so, the algorithm stops iterating and decomposes the current optimal solution into three parameters, Kp, Ki, and Kd. These parameters are then used in the PID controller to optimize the PID parameters. If the maximum number of iterations has not been reached, return to step S30 to continue exploring for a better solution.
5. The improved hypersonic vehicle thermal protection PID control method according to claim 1, characterized in that: In step S40, MATLAB and Simulink are used to simulate the PID control model of the hypersonic vehicle thermal protection system.
6. The improved hypersonic vehicle thermal protection PID control method according to claim 5, characterized in that: In step S40, the thermal protection PID controller obtains the deviation e(t) between the desired temperature setpoint and the actual temperature through the deviation calculation module, and calculates the Kp, Ki, and Kd parameters calculated by the improved lotus effect algorithm, outputs the control quantity u(t) to the execution unit module, and then sends a command to the controlled object module to control the controlled object module to output a control command to maintain the actual temperature to the desired temperature in a short time and stabilize it near the desired temperature. In the simulation process, the system transfer function is used as the controlled object to verify the significant advantages of the proposed improved algorithm in optimizing the PID controller parameters. The system transfer function formula is: Where G(s) is the transfer function and S is the function variable.