Photovoltaic array fault identification method based on improved war strategy optimization support vector machine

Abstract

The invention relates to the technical field of photovoltaic array fault recognition, and particularly discloses a photovoltaic array fault recognition method based on an improved war strategy optimization support vector machine. The method comprises the following steps: normalizing extracted five-parameter data of the photovoltaic array; optimizing penalty factors and kernel function parameters of a support vector machine algorithm by utilizing an improved war strategy algorithm; and establishing a photovoltaic array fault identification model based on the improved war strategy optimization support vector machine, training the photovoltaic array fault identification model, and identifying the photovoltaic array fault by using the five parameters of the photovoltaic array. According to the method, the parameters of the support vector machine can be effectively optimized, so that the accuracy of photovoltaic array fault recognition is improved, and the accuracy is as high as 98.77%.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic array fault identification, and specifically relates to photovoltaic array fault identification using an artificial intelligence algorithm. Background Art

[0002] Amidst growing environmental concerns and energy crises, solar energy has gained widespread adoption as a sustainable and environmentally friendly energy source. As a core component of photovoltaic systems, photovoltaic arrays (PV) often operate for extended periods in harsh outdoor environments, making them susceptible to faults such as aging, shadowing, hot spots, short circuits, and open circuits. These faults can reduce efficiency, shorten service life, and even cause fires. Therefore, developing accurate and rapid methods for detecting PV array faults is crucial.

[0003] In recent years, artificial intelligence algorithms have gained popularity in the field of photovoltaic array fault detection. Their principle is to extract features by learning and analyzing photovoltaic array information to achieve high-precision detection. Commonly used artificial intelligence algorithms include using neural networks to detect photovoltaic array faults [Cao Shenglei, Wang Zun, Yang Jianwei, et al. A photovoltaic array fault diagnosis method based on BP neural network [J]. Automation Expo, 2022, 39(11): 67-71.]. All algorithms pursue high-precision identification of photovoltaic array faults. Summary of the Invention

[0004] To accurately identify photovoltaic array faults, this paper proposes a photovoltaic array fault identification method based on an improved war strategy optimized support vector machine for array fault diagnosis. Key features include using an improved war strategy algorithm to extract five array parameters for fault identification; optimizing the support vector machine performance; and using a chaotic map to initialize the war strategy algorithm population and controlling the update rate of soldiers through a nonlinear function.

[0005] The present invention is implemented by the following technical solution, a photovoltaic array fault identification method based on an improved war strategy optimized support vector machine, comprising the following steps:

[0006] S1. Using the improved war strategy algorithm to extract the five-parameter data of the photovoltaic array, and normalizing the extracted five-parameter data set of the photovoltaic array;

[0007] S2. Optimize the penalty factor and kernel function parameters of the support vector machine algorithm using the improved war strategy algorithm;

[0008] S3. Establish and train a photovoltaic array fault identification model based on the improved war strategy optimization support vector machine, and use the five parameters of the photovoltaic array to identify photovoltaic array faults.

[0009] Furthermore, the five-parameter data set of the photovoltaic array in S1 includes photocurrent, reverse saturation current, ideality factor, series resistance, and parallel resistance.

[0010] Furthermore, the normalization method in S1 adopts minimum-maximum normalization to normalize the data set to the range of [0, 1].

[0011] Furthermore, the S2 includes the following steps:

[0012] Step a: Set the basic parameters of the support vector machine algorithm, including population size, dimension, maximum number of iterations, penalty factor, and the range of kernel function parameters.

[0013] Step b: Initialize the penalty factor and kernel function parameters of the support vector machine algorithm; use chaos mapping to initialize the war strategy algorithm population, calculate the fitness value of the population individuals, and determine the current individual local optimal value and global optimal value.

[0014] Step c: Determine whether to defend or attack, and use Levy flight to update the soldier's position. The soldier's update speed is controlled by a nonlinear function.

[0015] Step d: Determine whether the war strategy algorithm meets the termination condition. If the war strategy algorithm reaches the maximum number of iterations or the accuracy requirement, then end and go to step e, otherwise go to step b to calculate the fitness value of the individual.

[0016] Step e: The obtained optimal solution is used as the penalty factor and kernel function parameter of the support vector machine.

[0017] Furthermore, the improvements of the Tent-Logistic-Cosine chaotic mapping are as follows:

[0018] The improved Tent formula is as follows:

[0019]

[0020] where r∈[0,1];

[0021] The improved Logistic formula is as follows:

[0022] x i+1 =3.78x i (1-x i )

[0023] The improved Cosine formula is as follows:

[0024] x i+1 =cos(π(F(a,x i )+G(b,x i )+β))

[0025] where F(a, xi) and G(b, xi) are two seed maps, a and b are their control parameters, and the variable β is a shift constant.

[0026] Furthermore, the S3 includes the following steps:

[0027] Step A: After obtaining the optimal penalty factor and kernel function parameters of the support vector machine algorithm, input the normalized photovoltaic array five-parameter sample for training;

[0028] Step B: Input the normalized test samples into the trained photovoltaic array fault recognition model to obtain the recognition results.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] This paper proposes a photovoltaic array fault identification method based on an improved war strategy optimized support vector machine. The method's key features include using an improved war strategy algorithm to extract five photovoltaic array parameters for fault identification; optimizing the support vector machine's performance; and using a chaotic map to initialize the war strategy algorithm population. The improved war strategy algorithm also controls the update speed of soldiers through a nonlinear function, thereby improving the accuracy of photovoltaic array fault identification. Experimental results demonstrate an accuracy of up to 98.77%, demonstrating its high practical engineering application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of the photovoltaic array fault identification method based on the improved war strategy optimization support vector machine. DETAILED DESCRIPTION

[0032] The present invention will be further described in detail below with reference to the examples.

[0033] Reference Figure 1 ,A photovoltaic array fault identification method based on improved war strategy optimization support vector machine,the steps are:

[0034] S1. Using the improved war strategy algorithm to extract the five-parameter data of the photovoltaic array, and normalizing the extracted five-parameter data set of the photovoltaic array;

[0035] S2. Optimize the penalty factor and kernel function parameters of the support vector machine algorithm using the improved war strategy algorithm;

[0036] S3. Establish and train a photovoltaic array fault identification model based on an improved war strategy optimized support vector machine to identify photovoltaic array faults.

[0037] Furthermore, the five-parameter data set of the photovoltaic array in S1 includes photocurrent, reverse saturation current, ideality factor, series resistance, and parallel resistance, and is obtained by extracting the photovoltaic array volt-ampere characteristic curve using an improved war strategy algorithm.

[0038] Furthermore, the normalization method in S1 adopts minimum-maximum normalization to normalize the data set to the range of [0, 1].

[0039] Furthermore, the S2 is implemented according to the following steps:

[0040] Step a: Set the basic parameters of the support vector machine algorithm, including population size, dimension, maximum number of iterations, penalty factor, and the range of kernel function parameters.

[0041] Step b: Initialize the penalty factor and kernel function parameters of the support vector machine algorithm; use the Tent-Logistic-Cosine chaotic map to initialize the war strategy algorithm population, calculate the fitness value of the population individuals, and use the adaptive weight to determine the current individual local optimal value and global optimal value.

[0042] Step c: Determine whether to defend or attack, and use Levy flight to update the position of the soldiers respectively. max ) 2 Controls the update speed of soldiers. It is the current iteration number, it max is the maximum number of iterations.

[0043] Step d: Determine whether the war strategy algorithm meets the termination condition. If the war strategy algorithm reaches the maximum number of iterations or the accuracy requirement, then end and go to step e, otherwise go to step b to calculate the fitness value of the individual.

[0044] Step e: The obtained optimal solution is used as the penalty factor and kernel function parameter of the support vector machine.

[0045] Furthermore, the improvements of the Tent-Logistic-Cosine chaotic mapping are as follows:

[0046] The improved Tent formula is as follows:

[0047]

[0048] where r∈[0,1];

[0049] The improved Logistic formula is as follows:

[0050] x i+1 =3.78x i (1-x i )

[0051] The improved Cosine formula is as follows:

[0052] x i+1 =cos(π(F(a,x i )+G(b,x i )+β))

[0053] Where F(a, x i ) and G(b, x i ) are two seed maps, a and b are their control parameters, and the variable β is a shift constant.

[0054] Furthermore, the S3 is implemented according to the following steps:

[0055] Step A: After obtaining the optimal penalty factor and kernel function parameters of the support vector machine algorithm, input the normalized photovoltaic array five-parameter sample for training;

[0056] Step B: Input the normalized test samples into the trained photovoltaic array fault recognition model to obtain the recognition results.

[0057] The most significant feature of this method is that the improved war strategy algorithm can effectively optimize the penalty factor and kernel function parameters of the support vector machine algorithm, thereby improving the accuracy of photovoltaic array fault identification. Experimental results show that the accuracy is as high as 98.77%, as shown in Table 1. This method has high practical engineering application value.

[0058] Table 1 Algorithm recognition results

[0059]

[0060] The above description merely represents the preferred embodiments of the present invention, and while the description is relatively specific and detailed, it should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make various modifications, improvements, and substitutions without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A photovoltaic array fault identification method based on an improved war strategy optimized support vector machine, characterized by: The following steps are involved: S1. Using the improved war strategy algorithm to extract the five parameter data of the photovoltaic array, and normalizing the extracted five parameter data of the photovoltaic array; S2. Optimize the penalty factor and kernel function parameters of the support vector machine algorithm using the improved war strategy algorithm; S3. Establish and train a photovoltaic array fault identification model based on the improved war strategy optimization support vector machine, and use the five parameters of the photovoltaic array to identify photovoltaic array faults.

2. The photovoltaic array fault identification method based on improved war strategy optimization support vector machine according to claim 1 is characterized by: The five parameter data of the photovoltaic array in S1 include photocurrent, reverse saturation current, ideality factor, series resistance, and parallel resistance, and are obtained by extracting the photovoltaic array volt-ampere characteristic curve using an improved war strategy algorithm.

3. The photovoltaic array fault identification method based on improved war strategy optimization support vector machine according to claim 1 is characterized by: The normalization method in S1 adopts minimum-maximum normalization to normalize the data set to the range of [0, 1].

4. The photovoltaic array fault identification method based on improved war strategy optimization support vector machine according to claim 1 is characterized by: The S2 comprises the following steps: Step a: Set the basic parameters of the support vector machine algorithm, including population size, dimension, maximum number of iterations, penalty factor, and the range of kernel function parameters; Step b: Initialize the penalty factor and kernel function parameters of the support vector machine algorithm; use the Tent-Logistic-Cosine chaotic map to initialize the war strategy algorithm population, calculate the fitness value of the population individuals, and use the adaptive weight to determine the current individual local optimal value and global optimal value; Step c: Determine whether to defend or attack, and use Levy flight to update the position of the soldiers respectively; and use the nonlinear function R = 1-(it / it max ) 2 Controls the update speed of soldiers; it is the current iteration number, it max is the maximum number of iterations; Step d: Determine whether the war strategy algorithm meets the termination condition; if the war strategy algorithm reaches the maximum number of iterations or the accuracy requirement, then end and go to step e, otherwise go to step b to calculate the fitness value of the individual; Step e: The obtained optimal solution is used as the penalty factor and kernel function parameter of the support vector machine.

5. The photovoltaic array fault identification method based on improved war strategy optimization support vector machine according to claim 4 is characterized in that: The improvements of the Tent-Logistic-Cosine chaotic mapping are as follows: Tent improved formula is as follows: where r∈[0,1]; The improved Logistic formula is as follows: x i+1 =3.78x i (1-x i ) The improved Cosine formula is as follows: x i+1 =cos(π(F(a,x i )+G(b,x i )+b)) Where F(a, x i ) and G(b, x i ) are two seed maps, a and b are their control parameters, and the variable β is a shift constant.

6. The photovoltaic array fault identification method based on improved war strategy optimization support vector machine according to claim 1 is characterized by: The S3 includes the following steps: Step A: After obtaining the optimal penalty factor and kernel function parameters of the support vector machine algorithm, input the normalized photovoltaic array five-parameter sample for training; Step B: Input the normalized test samples into the trained photovoltaic array fault recognition model to obtain the recognition results.

Images