A control optimization method for air-cooled compression condensing machine

Abstract

The invention discloses a control optimization method for an air-cooled compression condensing machine, and belongs to the technical field of PID control optimization. The steps include: first, designing an AC asynchronous motor controller, and using Simulink to establish a speed loop PI controller simulation model; second, establishing an improved capuchin search algorithm model through Matlab, and the improvements include introducing a Latin hypercube operator, fusing a particle swarm algorithm, and an adaptive factor improvement algorithm, and iteratively optimizing the best Kp and Ki parameters; finally, outputting the optimized best Kp and Ki parameters to the speed loop PI controller simulation model; adjusting the speed loop PI control parameters through the improved algorithm, improving the dynamic response performance, reducing overshoot and steady-state errors, and ensuring that the compression condensing machine maintains a stable refrigeration effect under different working conditions.

Description

Technical Field

[0001] The invention belongs to the technical field of PID control optimization, and in particular relates to a control optimization method for an air-cooled compression condenser. Background Art

[0002] An air-cooled compression condensing machine is a refrigeration equipment mainly used for cooling and dehumidification applications. It achieves the cooling effect by compressing cold air and taking away the heat from the cold air using an air-cooled heat exchanger. The main components of an air-cooled compression condensing machine are a compressor, an air-cooled condenser, a fan, an expansion valve, an evaporator, a refrigerant pipeline, a liquid storage tank and a drying filter. The compressor is the core component of the air-cooled compression condensing machine and is responsible for compressing the refrigerant gas. Since AC asynchronous motors have high starting torque and operating stability and are suitable for continuous long-term operation, manufacturers usually choose AC asynchronous motors as the drive motors of compressors.

[0003] In industrial control, AC asynchronous motor controllers are mainly controlled by FOC (rotor flux oriented vector control); in FOC, the steady-state error of the controller is eliminated through speed PI control, ensuring the precise control of the speed of the AC asynchronous motor; the main advantages of PI control are simple structure, good dynamic performance and high control efficiency, and it is suitable for various industrial control processes; in the traditional PI control parameter setting process, the proportional coefficient (Kp) and integral coefficient (Ki) of PI control are mainly adjusted based on the experience of engineers, which is not only time-consuming and labor-intensive, but also due to the limitation of human brain computing power, the accuracy of the adjusted control parameters is not high; the AC asynchronous motor controller is a multivariable, highly coupled, time-varying nonlinear control system, and traditional PI control is difficult to achieve satisfactory results when dealing with such a complex control system.

[0004] Applying intelligent optimization algorithms to the tuning of the parameters of the PI controller of AC asynchronous motors is an ingenious method. The Capuchin Search Algorithm (CapSA) is a new intelligent optimization algorithm proposed in 2021. The algorithm is inspired by the social behavior and survival strategies of capuchins. Capuchins are highly social primates that exhibit a series of complex and efficient behavioral patterns in the process of finding food and avoiding enemies. The algorithm searches and optimizes by simulating the jumping, climbing, following, and clustering behaviors of capuchins. It has the advantages of strong global search capabilities, easy implementation, and applicability to a variety of optimization problems; but it also has the disadvantages of being easily trapped in local optimal solutions and having slow convergence speed. Summary of the invention

[0005] The purpose of the present invention is to propose a control optimization method for an air-cooled compression condensing machine, which automatically adjusts the speed loop PI control parameters of an AC asynchronous motor controller by improving the capuchin search algorithm, thereby improving the dynamic response performance of the AC asynchronous motor controller, enabling the controller to quickly respond to speed change instructions, reduce overshoot and steady-state errors, and thereby ensure that the compressor can quickly adjust the speed and power under different working conditions, thereby maintaining a stable refrigeration effect.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A control optimization method for an air-cooled compression condensing machine, comprising the following specific steps:

[0008] Step 1: Design the AC asynchronous motor controller of the compressor condensing machine and use Simulink software to establish a speed loop PI controller simulation model.

[0009] Step 2: Improve the capuchin search algorithm, specifically: (1) Introduce the Latin hypercube operator to improve the algorithm initialization position formula; (2) Integrate the particle swarm algorithm to improve the speed update formula of the capuchin search algorithm; (3) Introduce an adaptive factor to improve the capuchin life factor.

[0010] Step 3: Convert the AC asynchronous motor speed loop PI control parameter tuning problem into a mathematical model parameter problem to be solved, establish the mathematical model of the improved capuchin search algorithm through Matlab software, and iteratively optimize the optimal Kp and Ki parameters.

[0011] Step 4: Output the optimal Kp and Ki parameters to the speed loop PI controller simulation model to verify the superiority of the improved capuchin search algorithm.

[0012] Furthermore, in the step 1, the AC asynchronous motor controller of the compression condensing machine includes: an AC asynchronous motor module, a three-phase inverter module, a voltage and current acquisition module, a filter module, a CLARK transformation module, a PARK transformation module, a position and speed observer module, a subtractor module, an improved capuchin search algorithm module, a speed loop PI controller module, a current loop PI controller module, an inverse PARK transformation module, and a SVPWM control module.

[0013] Furthermore, in step 2, the capuchin search algorithm is improved, specifically:

[0014] Step 1: Introduce the Latin hypercube operator to improve the algorithm initialization position formula and enhance the diversity of the algorithm population distribution. The improved formula is:

[0015]

[0016] Where, i=1,…,nPop, j=1,…,Dim, Ub is the upper bound of the population search space, Lb is the lower bound of the population search space, nPop is the number of algorithm populations, γ is a random number with a value between (0, nPop], and r1 is a random number with a value between (0, 1];

[0017] Step 2: Integrate the particle swarm algorithm to improve the speed update formula of the capuchin search algorithm to help the algorithm jump out of the local optimal solution. The improved formula is:

[0018] v i，j (t+1)=ρv i，j (t)+τa 1 (x best，j (t)-x i，j (t))r 2 +τa 2 (F j -x i，j (t))r 3 ; (2)

[0019]

[0020] In formula (2), v i，j (t+1) is the updated velocity, v i，j (t) is the speed of the current iteration, x bast，i (t) is the position of the best individual in the current iteration, F j is the position of the food in the jth dimension, x i，j (t) is the position of the jth dimension of the i-th individual in the population, ρ is the inertia weight, τ is the life factor, a 1 a 2 The value is 1.0, t is the current iteration number, r 2 、r 3 is a random number with a value between (0, 1];

[0021] In formula (3), fit i is the fitness of the i-th individual in the population, fit best is the population's optimal position fitness, fit worse is the worst position fitness of the population;

[0022] Step 3: Introduce an adaptive factor to improve the lifespan factor of the capuchin monkey, balance the global search and local search capabilities of the algorithm, and the improved lifespan factor formula is:

[0023]

[0024] Where τ is the life factor, β 0is the scaling factor, with a value of 2, α controls the decline rate of the lifespan factor, with a value of 10, β 1 controls the minimum value of the lifespan factor, with a value of 0.4, t is the current iteration number, and T is the final iteration number.

[0025] Furthermore, in the third step, a mathematical model of the improved capuchin monkey search algorithm is established, and the optimal Kp and Ki parameters are iteratively optimized. The specific steps are as follows:

[0026] S1. Select ITAE as the performance index of the PI controller. ITAE evaluates the control effect of the system by integrating the time-weighted absolute error, can comprehensively consider the response time and steady-state error of the system, and provides a balanced evaluation criterion for the dynamic performance and steady-state performance of the system. The formula of ITAE is:

[0027]

[0028] In the formula, J is the value of the PI performance index controller, e(τ) is the error value between the target speed and the actual speed, and τ is the running time of the AC asynchronous motor controller of the compression condenser;

[0029] S2. Encode the Kp and Ki parameters of the speed loop PI controller as the dimension values of the population solutions of the capuchin monkey search algorithm;

[0030] S3. Initialize the algorithm parameters, including the total number of iterations T, the population size nPop, the upper and lower bounds of the search space Ub and Lb, and the dimension value Dim of the algorithm solution, etc. Select ITAE as the objective function of the algorithm;

[0031] S4. Initialize the initial position of the improved capuchin monkey search algorithm. The formula for initializing the position of the improved algorithm is the same as above;

[0032] S5. Calculate the fitness of all individual positions in the population through the objective function. According to the principle that the smaller the individual fitness, the better the performance of the solution, record the individual position with the smallest fitness as the best individual in the population, and the individual with the largest fitness as the worst individual in the population;

[0033] S6. Divide the capuchin monkey population into two groups, half as the leader group and the other half as the follower group. When i < nPop, the i-th individual belongs to the leader group, and the meaning of i is the same as above. The update method of the leader is:

[0034] Q1. The probability size of ε is used to select the position update method of the capuchin monkey. ε is a random number with a value in (0, 1];

[0035] Q2. When 0.1 ≤ ε ≤ 0.2, simulate the behavior of the capuchin monkeys in the leader group jumping on the tree for position update. The formula for position update at this stage is:

[0036]

[0037] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, which is the optimal solution P of the algorithm. bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, θ is the jumping angle of the capuchin monkey, which is a random number between (0, 1.5], g is the gravitational acceleration with a value of 9.81, and i, j, and t have the same meanings as above;

[0038] Q3, when 0.2≤ε≤0.3, the capuchin monkeys in the simulated leader group jump and walk on the ground to update their positions. The formula for position update at this stage is:

[0039]

[0040] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, P ef The probability that a capuchin monkey bounces when jumping on the ground is 9, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, θ is the jumping angle of the capuchin monkey, which is a random number between (0, 1.5], g is the gravitational acceleration, which is 9.81, and the meanings of i, j, and t are the same as above;

[0041] Q4, when 0.3≤ε≤0.5, the capuchin monkeys in the leader group are simulated to update their positions by walking normally on the ground. The formula for position update at this stage is:

[0042] x i，j (t+1)=x i，j (t)+v i，j (t);

[0043] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x i，j (t) is the j-th dimension value of the position of the i-th individual in the current iteration, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, and i, j, and t have the same meanings as above;

[0044] Q5. When 0.5≤ε≤0.75, the capuchin monkeys in the leader group are simulated to update their positions by swinging on the tree while foraging. The formula for position update at this stage is:

[0045] x i，j (t+1)=F j (t)+τP bf ×sin(2θ);

[0046] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, τ is the lifespan factor of the capuchin monkey, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the speed of the capuchin monkey, θ is the jumping angle of the capuchin monkey, is a random number between (0, 1.5], g is the gravitational acceleration with a value of 9.81, and i, j, and t have the same meanings as above;

[0047] Q6. When ε≥0.75, the capuchin monkeys in the simulated leader group update their positions when climbing trees and their tops while foraging. The formula for position update at this stage is:

[0048] x i，j (t+1)=F j (t)+τP bf (v i，j (t)-v i，j-1 (t));

[0049] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, τ is the lifespan factor of the capuchin monkey, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the capuchin monkey’s velocity in the current iteration, v i，j-1 (t) is the previous dimension value of the jth dimension of the velocity of the capuchin monkey in the current iteration, g is the gravitational acceleration in the current iteration and takes the value of 9.81, and i, j, and t have the same meanings as above;

[0050] Q7. When ε < 0.1, the capuchin monkeys simulating the leader group update their positions in random migration behavior. The formula for position update at this stage is:

[0051] x i，j (t+1)=τ×[Lb j +ε×(Ub j-Lb j )];

[0052] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, τ is the lifespan factor of the capuchin monkey, Ub j is the jth dimension of the algorithm search space upper bound, Lb j is the jth dimension of the lower bound of the algorithm search space, and ε, i, j, and t have the same meanings as above;

[0053] S7. When i ≥ nPop, the i-th individual belongs to the follower group, i has the same meaning as above, and the update method of the followers is:

[0054] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x l，j (t) is the lth individual in the leader group, x i-1，j (t) is the j-th dimension value of the position of the i-1th individual in the current iteration;

[0055] S8. Determine the food location by greedy selection and record the fitness of the food location. The formula for greedy selection is:

[0056]

[0057] In the formula, F j (t+1) is the updated value of the jth dimension of the food position, F j (t) is the j-th dimension value of the food position in the current iteration, x best，j (t) is the j-th dimension value of the best individual position in the current iteration, fit F is the fitness of the food position, fit beest is the fitness of the best individual position in the current iteration;

[0058] S9. Determine whether t is greater than the final number of iterations T. If so, the iteration is terminated, the optimal solution of the algorithm is output, and the values ​​of each dimension of the optimal solution of the algorithm are decoded and passed to the speed loop PI controller. If not, the algorithm returns to S5 to continue iterating.

[0059] Furthermore, in step 4, the optimal Kp and Ki parameters are output to the speed loop PI controller simulation model to verify the superiority of the improved capuchin search algorithm. The specific steps are:

[0060] Ss1, the voltage and current acquisition module collects three-phase voltage and current signals from the AC asynchronous motor module, and filters the voltage and current signals through the filter module;

[0061] Ss2, CLARK conversion module for three-phase stator voltage signal u a 、u b 、u c and the three-phase stator current signal i a 、i b 、i c Perform CLARK transformation and output the two-phase stator voltage signal u in the stationary coordinate system sα 、u sβ and two-phase stator current signal i sα 、i sβ ;

[0062] Ss3, will u sα 、u sβ 、i sα 、i sβ The input position and speed observer module obtains the rotor flux ψ by decoupling the mathematical model of the AC asynchronous motor. rα and ψ rβ , calculate the rotor flux to get the estimated rotor angle And then Derivative to get the estimated motor speed Ss4, PARK conversion module for two-phase stator current signal i sα 、i sβ Perform PARK transformation to obtain the two-phase stator current i in the rotating coordinate system sd 、i sq ;

[0063] Ss5, given AC asynchronous motor controller target speed ω ref and d-axis target current I sd , calculate the target speed ω through the error calculation module ref The actual speed The error value is input into the speed loop PI controller module;

[0064] Ss6. Improve the capuchin search algorithm module to optimize and adjust the two parameters of the speed loop PI controller, and transfer the iterative optimal Kp and Ki parameters to the speed loop PI controller module of the compressor condensing machine AC asynchronous motor controller;

[0065] Ss7, output q-axis target current I sq , calculate I through the error calculation module sq with i sq The error value and I sd with i sd The error value is input into the two current loop PI controller modules, and the two-phase stator voltage v is obtained through PI control. sd 、v sq ;

[0066] Ss8, the inverse PARK transformation module treats the incoming v sd 、v sq Perform inverse PARK transformation to obtain v sα and v sβ And transmitted to the SVPWM control module;

[0067] Ss9, SVPWM control module generates three pairs of complementary PWM signals and outputs them to the three-phase inverter module;

[0068] S10, three-phase inverter module outputs three-phase voltage v a 、v b 、v c To the AC asynchronous motor module, double closed-loop control of the compressor condensing machine AC asynchronous motor controller is realized.

[0069] By adopting the above technical solution, the present invention achieves the following beneficial effects:

[0070] The present invention proposes a control optimization method for an air-cooled compression condensing machine. For the first time, the speed loop PI controller parameters are adjusted by improving the capuchin search algorithm, the position initialization formula of the algorithm is improved by introducing the Latin hypercube operator, the speed update formula of the capuchin search algorithm is improved by fusing the particle swarm algorithm, and the life factor of the capuchin is improved by introducing an adaptive factor. The method avoids the density of population solutions in the initial stage, avoids the capuchin search algorithm from falling into a local optimal solution, and dynamically balances the global and local search capabilities of the algorithm, thereby improving the dynamic response performance of the AC asynchronous motor controller, enabling the controller to quickly respond to speed change instructions, reduce overshoot and steady-state errors, and further ensure that the compressor can quickly adjust the speed and power under different working conditions, thereby maintaining a stable refrigeration effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Flowchart for tuning the speed loop PI controller parameters for improving the capuchin search algorithm.

[0072] Figure 2 Model diagram of AC asynchronous motor controller for compressor condensing machine optimized by improved capuchin search algorithm.

[0073] Figure 3 Parameter curve of speed loop PI controller for improving capuchin search algorithm.

[0074] Figure 4 Graph of tuning the speed loop PI controller for the improved Capuchin search algorithm.

[0075] Figure 5 This is a comparison chart of the best fitness between the ordinary capuchin search algorithm and the improved capuchin search algorithm. DETAILED DESCRIPTION

[0076] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the present invention. The following embodiments are only a part of the embodiments of the present invention and are used to more clearly illustrate the technical solutions of the present invention. It can be understood that the specific embodiments are only used to explain the present invention, but the implementation methods of the present invention are not limited thereto.

[0077] The present invention proposes a control optimization method for an air-cooled compression condenser, such as Figure 1 As shown, it is divided into the following four steps:

[0078] Step 1. Design the compressor condensing machine AC asynchronous motor controller model, and use Simulink software to establish the speed loop PI controller simulation model. The compressor condensing machine AC asynchronous motor controller model includes: AC asynchronous motor module, three-phase inverter module, voltage and current acquisition module, filter module, CLARK transformation module, PARK transformation module, position and speed observer module, subtractor module, improved capuchin search algorithm module, speed loop PI controller module, current loop PI controller module, inverse PARK transformation module, SVPWM control module.

[0079] Step 2: Improve the capuchin search algorithm, specifically:

[0080] Step 1: Introduce the Latin hypercube operator to improve the algorithm initialization position formula. The improved formula is:

[0081]

[0082] Where, i=1,…,nPop, j=1,…,Dim, Ub is the upper bound of the population search space, Lb is the lower bound of the population search space, nPop is the number of algorithm populations, γ is a random number with a value between (0, nPop], and r1 is a random number with a value between (0, 1];

[0083] Step 2, integrate the particle swarm algorithm to improve the speed update formula of the capuchin search algorithm. The improved formula is:

[0084] v i，j (t+1)=ρv i，j (t)+τa 1 (x best，j (t)-x i，j (t))r 2 +τa 2 (F j -x i，j (t))r 3 ; (2)

[0085]

[0086] In formula (2), v i，j (t+1) is the updated velocity, v i，j (t) is the speed of the current iteration, x best，i (t) is the position of the best individual in the current iteration, F j is the position of the food in the jth dimension, x i，j (t) is the position of the jth dimension of the i-th individual in the population, ρ is the inertia weight, τ is the life factor, a 1 a 2 The value is 1.0, t is the current iteration number, r 2 、r 3 is a random number with a value between (0, 1];

[0087] In formula (3), fit i is the fitness of the i-th individual in the population, fit best is the population's optimal position fitness, fit worse is the worst position fitness of the population;

[0088] Step 3: Introduce adaptive factors to improve the lifespan factor of capuchin monkeys. The improved lifespan factor formula is:

[0089]

[0090] Where τ is the life factor, β 0 is the scaling factor, the value is 2, α controls the rate of decrease of the life factor, the value is 10, β 1 Controls the minimum value of the life factor, which is 0.4. t is the current iteration number, and T is the final iteration number.

[0091] Step 3: Convert the AC asynchronous motor speed loop PI control parameter tuning problem into a mathematical model parameter problem to be solved. Use Matlab software to establish the mathematical model of the improved capuchin search algorithm, and iteratively optimize the optimal Kp and Ki parameters. The specific steps are as follows:

[0092] S1. Set the model running time to 2s, the sampling time to 0.001, and select ITAE as the performance indicator of the PI controller. The formula of ITAE is:

[0093]

[0094] In the formula, J is the value of the PI performance index controller, e(τ) is the error value between the target speed and the actual speed, and τ is the running time of the AC asynchronous motor controller of the compressor condenser;

[0095] S2, encoding the Kp and Ki parameters of the speed loop PI controller into the dimension values ​​of the population solution of the capuchin search algorithm;

[0096] S3. Initialize the algorithm parameters. Set the total number of iterations \(T = 20\), the population size \(nPop = 30\), the upper bound of the search space \(Ub=[20, 200]\), the lower bound of the search space \(Lb = [0.001, 0.001]\), the dimension value \(Dim = 2\), and select ITAE as the objective function of the algorithm;

[0097] S4. Initialize the initial positions of the improved capuchin monkey search algorithm. The formula for initializing the positions of the improved algorithm is the same as above;

[0098] S5. Calculate the fitness of all individual positions in the population through the objective function. Record the position of the individual with the minimum fitness as the best individual in the population, and the individual with the maximum fitness as the worst individual in the population;

[0099] S6. Divide the capuchin monkey population into two groups. Half are the leader group and the other half are the follower group. When \(i < nPop\), the \(i\)-th individual belongs to the leader group. The meaning of \(i\) is the same as above. The update method for the leader is:

[0100] Q1. Select the probability size of \(\epsilon\) to choose the position update method of the capuchin monkey. \(\epsilon\) is a random number with a value in the range \((0, 1]\);

[0101] Q2. When \(0.1\leqslant\epsilon\leqslant0.2\), simulate the behavior of the capuchin monkeys in the leader group jumping on the tree to update the positions. The formula for position update in this stage is:

[0102]

[0103] In the formula, \(x i，j (t + 1)\) is the \(j\)-th dimension value of the updated position of the \(i\)-th individual in the population, \(F j (t)\) is the \(j\)-th dimension value of the position of the food in the current iteration, that is, the optimal solution \(P\) of the algorithm bf is the probability of maintaining balance during the capuchin monkey's jump, with a value of \(0.8\), \(v i，j (t)\) is the \(j\)-th dimension value of the velocity of the capuchin monkey in the current iteration, \(\theta\) is the jumping angle of the capuchin monkey, which is a random number with a value in the range \((0, 1.5]\), \(g\) is the acceleration due to gravity, with a value of \(9.81\). The meanings of \(i\), \(j\), and \(t\) are the same as above;

[0104] Q3. When \(0.2\leqslant\epsilon\leqslant0.3\), simulate the behavior of the capuchin monkeys in the leader group jumping and walking on the ground to update the positions. The formula for position update in this stage is:

[0105]

[0106] In the formula, \(x i，j (t + 1)\) is the \(j\)-th dimension value of the updated position of the \(i\)-th individual in the population, \(Fj (t) is the j-th dimension value of the food position in the current iteration, P ef The probability that a capuchin monkey bounces when jumping on the ground is 9, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, θ is the jumping angle of the capuchin monkey, which is a random number between (0, 1.5], g is the gravitational acceleration with a value of 9.81, and i, j, and t have the same meanings as above;

[0107] Q4, when 0.3≤ε≤0.5, the capuchin monkeys in the leader group are simulated to update their positions by walking normally on the ground. The formula for position update at this stage is:

[0108] x i，j (t+1)=x i，j (t)+v i，j (t);

[0109] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x i，j (t) is the j-th dimension value of the position of the i-th individual in the current iteration, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, and i, j, and t have the same meanings as above;

[0110] Q5. When 0.5≤ε≤0.75, the capuchin monkeys in the leader group are simulated to update their positions by swinging on the tree while foraging. The formula for position update at this stage is:

[0111] x i，j (t+1)=F j (t)+τP bf ×sin(2θ);

[0112] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, τ is the lifespan factor of the capuchin monkey, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the speed of the capuchin monkey, θ is the jumping angle of the capuchin monkey, is a random number between (0, 1.5], g is the gravitational acceleration with a value of 9.81, and i, j, and t have the same meanings as above;

[0113] Q6. When ε≥0.75, the capuchin monkeys in the simulated leader group update their positions when climbing trees and their tops while foraging. The formula for position update at this stage is:

[0114] x i，j (t+1)=F j (t)+τP bf (v i，j (t)-v i，j-1 (t));

[0115] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, τ is the lifespan factor of the capuchin monkey, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the capuchin monkey’s velocity in the current iteration, v i，j-1 (t) is the previous dimension value of the jth dimension of the velocity of the capuchin monkey in the current iteration, g is the gravitational acceleration in the current iteration and takes the value of 9.81, and i, j, and t have the same meanings as above;

[0116] Q7. When ε < 0.1, the capuchin monkeys simulating the leader group update their positions in random migration behavior. The formula for position update at this stage is:

[0117] x i，j (t+1)=τ×[Lb j +ε×(Ub j -Lb j )];

[0118] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, τ is the lifespan factor of the capuchin monkey, Ub j is the jth dimension of the algorithm search space upper bound, Lb j is the jth dimension of the lower bound of the algorithm search space, and ε, i, j, and t have the same meanings as above;

[0119] S7. When i ≥ nPop, the i-th individual belongs to the follower group, i has the same meaning as above, and the update method of the followers is:

[0120] In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x l，j (t) is the lth individual in the leader group, x i-1，j (t) is the j-th dimension value of the position of the i-1th individual in the current iteration;

[0121] S8. Determine the food location by greedy selection and record the fitness of the food location. The formula for greedy selection is:

[0122]

[0123] In the formula, F j (t+1) is the updated value of the jth dimension of the food position, F j (t) is the j-th dimension value of the food position in the current iteration, x best，j (t) is the j-th dimension value of the best individual position in the current iteration, fit F is the fitness of the food position, fit beest is the fitness of the best individual position in the current iteration;

[0124] S9, determine whether t is greater than the final iteration number T, if so, the iteration is terminated, the optimal solution of the algorithm is output, and the values ​​of each dimension of the optimal solution of the algorithm are decoded and transmitted to the speed loop PI controller, if not, the algorithm returns to S5 to continue iteration;

[0125] S10, such as Figure 3 As shown, the algorithm inputs the best Kp and Ki parameters found through iterations into the model. The best Kp parameter found through iterations is 18.875, and the Ki parameter is 118.763.

[0126] Step 4: Output the optimal Kp and Ki parameters to the speed loop PI controller simulation model, such as Figure 2 As shown, the specific steps are:

[0127] Ss1, the voltage and current acquisition module collects three-phase voltage and current signals from the AC asynchronous motor module, and filters the voltage and current signals through the filter module;

[0128] Ss2, CLARK conversion module for three-phase stator voltage signal u a 、u b 、u c and the three-phase stator current signal i a 、i b 、i c Perform CLARK transformation and output the two-phase stator voltage signal u in the stationary coordinate system sα 、u sβ and two-phase stator current signal i sα 、i sβ ;

[0129] Ss3, will u sα 、u sβ 、i sα 、i sβ The input position and speed observer module obtains the rotor flux ψ by decoupling the mathematical model of the AC asynchronous motor. rα and ψ rβ, calculate the rotor flux to get the estimated rotor angle And then Derivative to get the estimated motor speed

[0130] Ss4, PARK conversion module for two-phase stator current signal i sα 、i sβ Perform PARK transformation to obtain the two-phase stator current i in the rotating coordinate system sd 、i sq ;

[0131] Ss5, given AC asynchronous motor controller target speed ω ref and d-axis target current I sd , calculate the target speed ω through the error calculation module ref The actual speed The error value is input into the speed loop PI controller module;

[0132] Ss6. Improve the capuchin search algorithm module to optimize and adjust the two parameters of the speed loop PI controller, and transfer the iterative optimal Kp and Ki parameters to the speed loop PI controller module of the compressor condensing machine AC asynchronous motor controller;

[0133] Ss7, output q-axis target current I sq , calculate I through the error calculation module sq with i sq The error value and I sd with i sd The error value is input into the two current loop PI controller modules, and the two-phase stator voltage v is obtained through PI control. sd 、v sq ;

[0134] Ss8, the inverse PARK transformation module treats the incoming v sd 、v sq Perform inverse PARK transformation to obtain v sα and v sβ And transmitted to the SVPWM control module;

[0135] Ss9, SVPWM control module generates three pairs of complementary PWM signals and outputs them to the three-phase inverter module;

[0136] S10, three-phase inverter module outputs three-phase voltage v a 、v b 、v c To the AC asynchronous motor module, double closed-loop control of the compressor condensing machine AC asynchronous motor controller is realized.

[0137] like Figure 4As shown in the figure, the fitness curves of the ordinary capuchin search algorithm and the improved capuchin search algorithm when adjusting the parameters of the speed loop PI controller are compared and analyzed. In the early stage of iteration, the convergence speed of the improved algorithm is significantly faster than that of the ordinary algorithm, while the ordinary capuchin search algorithm only jumps out of the local optimum and continues to converge at the 10th generation. Relatively speaking, the improved algorithm can still perform a large-scale search. In the later stage of iteration, the improved capuchin search algorithm continues to maintain efficient global search and local accuracy capabilities, and finally reaches the best fitness of 1.485, which is better than the ordinary algorithm's 1.492. This shows that the improved capuchin search algorithm has obvious performance advantages and optimization effects in adjusting the parameters of the speed loop PI controller, and significantly improves the control accuracy and dynamic response capability of the system.

[0138] like Figure 5 As shown in the figure, the response speed, overshoot and stabilization time of the ordinary capuchin search algorithm and the improved capuchin search algorithm when adjusting the speed loop PI controller parameters are compared and analyzed. It can be seen that the speed loop PI controller adjusted by the improved capuchin search algorithm has no obvious difference in response speed compared with the speed loop PI controller adjusted by the ordinary capuchin search algorithm, but the stabilization time is significantly shortened, and the overshoot of the improved algorithm is significantly smaller than that of the ordinary algorithm. This shows that the improved capuchin search algorithm is effective for adjusting the parameters of the speed loop PI controller of the AC asynchronous motor of the compressor condenser.

[0139] In summary, the present invention provides a control optimization method for an air-cooled compression condensing machine. The method adjusts the speed loop PI control parameters of the AC asynchronous motor controller by improving the capuchin search algorithm, which significantly improves the dynamic response performance of the controller, so that the controller can quickly respond to speed change instructions and reduce overshoot and steady-state errors, thereby ensuring that the AC asynchronous motor can quickly adjust the speed and power under different working conditions and maintain a stable refrigeration effect. Through this optimization method, the overall performance of the compression condensing machine is effectively improved.

Claims

1. A control optimization method for an air-cooled compression condenser, characterized in that: The specific steps are as follows: Step 1: Design an AC asynchronous motor controller for the compression condenser, and use Simulink software to establish a simulation model of the speed loop PI controller; Step 2: Improve the capuchin monkey search algorithm, specifically as follows: Step 1: Introduce the Latin hypercube operator to improve the formula for initializing the position of the algorithm. The improved formula is: In the formula, i = 1,..., nPop, j = 1,..., Dim, Ub is the upper bound of the population search space, Lb is the lower bound of the population search space, nPop is the number of algorithm populations, Dim is the dimension value of the algorithm solution, γ is a random number with a value in (0, nPop], and r1 is a random number with a value in (0, 1]; Step 2: Incorporate the particle swarm algorithm to improve the velocity update formula of the capuchin monkey search algorithm. The improved formula is: v i，j (t+1)=ρv i，j (t)+τa1(x best，j (t)-x i，j (t))r2+τa2(F j -x i，j (t))r3; (2) In formula (2), v i，j (t+1) is the updated velocity, v i，j (t) is the speed of the current iteration, x best,i (t) is the position of the best individual in the current iteration, F j is the position of the food in the jth dimension, x i，j (t) is the position of the jth dimension of the i-th individual in the population, ρ is the inertia weight, τ is the life factor, a1a2 is 1.0, t is the current iteration number, r2 and r3 are random numbers with values ​​in (0, 1]; In formula (3), fit i is the fitness of the i-th individual in the population, fit best is the population's optimal position fitness, fit worse is the worst position fitness of the population; Step 3: Introduce an adaptive factor to improve the lifespan factor of the capuchin monkey. The improved formula is: In the formula, τ is the lifespan factor, β0 is the scaling factor with a value of 2, α controls the descent speed of the lifespan factor with a value of 10, β1 controls the minimum value of the lifespan factor with a value of 0.4, t is the current iteration number, and T is the final iteration number; Step 3: Convert the problem of tuning the PI control parameters of the AC asynchronous motor speed loop into a problem of solving the parameters of the mathematical model. Use Matlab software to establish a mathematical model of the improved capuchin monkey search algorithm, and iteratively optimize to find the optimal Kp and Ki parameters; Step 4: Output the optimized optimal Kp and Ki parameters to the simulation model of the speed loop PI controller to verify the superiority of the improved capuchin monkey search algorithm.

2. A control optimization method for an air-cooled compression condenser according to claim 1, characterized in that: In the above Step 3, when establishing the mathematical model of the improved capuchin monkey search algorithm and iteratively optimizing to find the optimal Kp and Ki parameters, the specific steps are as follows: S1: Select ITAE as the performance index of the PI controller, and encode the Kp and Ki parameters of the speed loop PI controller as the dimension values of the population solution of the capuchin monkey search algorithm; S2: Initialize the algorithm parameters, including the total number of iterations T, the number of populations nPop, the upper and lower bounds of the search space Ub and Lb, and the dimension value Dim of the algorithm solution. Select ITAE as the objective function of the algorithm; S3: Initialize the initial position of the improved capuchin monkey search algorithm; S4: Calculate the fitness of all individual positions in the population through the objective function, record the position of the individual with the minimum fitness as the best individual in the population, and the individual with the maximum fitness as the worst individual in the population; S5: Divide the capuchin monkey population into two groups: leaders and followers. When i < nPop, the i-th individual belongs to the leader group; S6: When i ≥ nPop, the i-th individual belongs to the follower group. The update method of the follower is: In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x l，j (t) is the lth individual in the leader group, x i-1，j (t) is the j-th dimension value of the position of the i-1th individual in the current iteration; S7: Determine the food position through greedy selection and record the fitness of the food position; S8: Judge whether t is greater than T. If so, the iteration terminates, and output the optimal solution of the algorithm to the speed loop PI controller. If not, return to S4 to continue the iteration.

3. A control optimization method for an air-cooled compression condenser according to claim 1, characterized in that: In the above Step 4, when outputting the optimized optimal Kp and Ki parameters to the simulation model of the speed loop PI controller, the specific steps are as follows: Ss1, the voltage and current acquisition module collects the three-phase voltage and current from the AC asynchronous motor module, and filters the voltage and current through the filter module; Ss2, three-phase stator voltage signal u a 、u b 、u c and the three-phase stator current signal i a 、i b 、i c Perform CLARK transformation and output two-phase stator voltage u sα 、u sβ and the two-phase stator current i sα 、i sβ ; Ss3, will u sα 、u sβ 、i sα , β sβ Input position and speed observer module, decouple the AC asynchronous motor mathematical model to obtain rotor flux ψ rα and ψ rβ , calculate the rotor flux to get the estimated rotor angle Then the estimated motor speed is obtained Ss4, for two-phase stator current i sα 、i sβ Perform PARK transformation to obtain the two-phase stator current i sd 、i sq ; Ss5, given target speed ω ref and the target current I sd , calculate the target speed ω ref The actual speed The error value is input into the speed loop PI controller module; Ss6. Improve the capuchin search algorithm module to adjust the parameters of the speed loop PI controller, and transfer the iterative optimal Kp and Ki parameters to the speed loop PI controller module of the compressor condensing machine AC asynchronous motor controller; Ss7. Calculate the output target current I sq , calculate I through the error calculation module sq with i sq The error value and I sd with i sd The error value is input into the two current loop PI controller modules, and the two-phase stator voltage v is obtained through PI control. sd 、v sq ; Ss8, for the incoming v sd 、v sq Perform inverse PARK transformation to obtain v sα and v sβ And transmitted to the SVPWM control module; Ss9, the SVPWM control module generates three pairs of complementary PWM signals and outputs them to the three-phase inverter module; Ss10, three-phase inverter module outputs three-phase voltage v a 、v b 、v c To the AC asynchronous motor module, double closed-loop control of the compressor condensing machine AC asynchronous motor controller is realized.

4. A control optimization method for an air-cooled compression condenser according to claim 2, characterized in that: In step 3, S1, the formula of ITAE is: Where J is the value of the PI performance index controller, e(τ) is the error between the target speed and the actual speed, and τ is the running time of the AC asynchronous motor controller of the compressor condensing machine.

5. A control optimization method for an air-cooled compression condenser according to claim 2, characterized in that: In step 3, S6, the leader is updated in the following manner: Q1, ε are random numbers with values ​​between (0, 1], when 0.1≤ε≤0.2, the position update formula is: In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, P bf The value is 0.7, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, θ is a random number between (0, 1.5], and g is 9.81; Q2, when 0.2≤ε≤0.5, the capuchin monkeys in the simulated leader group move on the ground and update their positions; Q3. When 0.5≤ε≤0.75, the formula for position update is: x i，j (t+1)=F j (t)+τP bf ×sin(2θ); In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, τ is the life span factor of the capuchin monkey, and g is the gravitational acceleration with a value of 9.81; Q4. When ε≥0.75, the position update formula is: x i，j (t+1)=F j (t)+τP bf (v i，j (t)-v i，j-1 (t)); In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, v i，j (t) is the j-th dimension value of the capuchin monkey’s velocity in the current iteration, v i，j-1 (t) is the previous dimension value of the jth dimension of the velocity of the capuchin monkey in the current iteration, and g is the gravitational acceleration with a value of 9.81; Q5. When ε<0.1, the position update formula is: x i，j (t+1)=τ×[Lb j +ε×(Ub j -Lb j )]; In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, τ is the lifespan factor of the capuchin monkey, Ub j is the jth dimension of the algorithm search space upper bound, Lb i The j-th dimension of the algorithm search space lower bound.

6. A control optimization method for an air-cooled compression condenser according to claim 2, characterized in that: In step 3, the method of updating the behavior position of the capuchin monkeys simulating the leader group moving on the ground in S6 is: Q1. When 0.2≤ε≤0.3, the position update formula is: In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, F j (t) is the j-th dimension value of the food position in the current iteration, P ef The probability that a capuchin monkey bounces when jumping on the ground is 9, P bf is the probability of the capuchin monkey maintaining balance during jumping, with a value of 0.8, v i，j (t) is the j-th dimension value of the speed of the capuchin monkey in the current iteration, θ is the jumping angle of the capuchin monkey, which is a random number between (0, 1.5], and g is the gravitational acceleration with a value of 9.81; Q2, when 0.3≤ε≤0.5, the position update formula is: x i，j (t+1)=x i，j (t)+v i，j (t); In the formula, x i，j (t+1) is the j-th dimension value of the updated position of the i-th individual in the population, x i，j (t) is the j-th dimension value of the position of the i-th individual in the current iteration, v i，j (t) is the j-th dimension value of the capuchin monkey’s velocity in the current iteration.

7. A control optimization method for an air-cooled compression condenser according to claim 2, characterized in that: In step 3, S8, the formula for greedy selection is: In the formula, F j (t+1) is the updated value of the jth dimension of the food position, F j (t) is the j-th dimension value of the food position in the current iteration, x best，j (t) is the j-th dimension value of the best individual position in the current iteration, fit F is the fitness of the food position, fit beest is the fitness of the best individual position in the current iteration.

Images