Solar global radiation model identification method based on improved chaotic SCSO algorithm

Abstract

The application discloses a kind of based on the identification method of solar daily total radiation model of improved chaotic SCSO algorithm, belong to energy engineering system identification technical field.Solve the technical problem that solar daily total radiation model parameter identification precision is not high.Its technical scheme is as follows: including the following steps: step 1) establish the Box-Jenkins model of solar daily total radiation;Step 2) the identification process of improved chaotic SCSO algorithm is constructed.The beneficial effects of the application are as follows: the improved chaotic SCSO algorithm proposed in the application is a heuristic swarm intelligence optimization algorithm, which is better than the traditional algorithm for the identification effect of solar daily total radiation model, and also has higher convergence accuracy than the unimproved SCSO algorithm, which can well identify the proposed Box-Jenkins model.

Description

Technical Field

[0001] The present invention relates to the technical field of energy engineering system identification, and in particular to a method for identifying a solar daily total radiation model based on an improved chaotic SCSO algorithm. Background Art

[0002] Solar radiation is a crucial component of Earth's ecology and a source of energy. Therefore, it has gradually become a research hotspot in the field of energy engineering due to its crucial impact on solar power plants and ecosystems. In addition to collecting specific data for actual calculations, several classic simulation methods are currently being used both domestically and internationally. These methods establish different models for research and analysis, including the Hottel model and the Berliand model. Parameter identification of these models has also become a hot topic and holds high research value. Therefore, it is necessary to establish corresponding system models for total daily solar radiation and identify the parameters of these models. To this end, many researchers have proposed various identification methods, such as least squares, stochastic gradient descent, and particle swarm optimization.

[0003] As a classic parameter optimization algorithm, the least squares method often suffers from certain data saturation in actual engineering applications; although stochastic gradient descent is simple and easy to implement, its identification accuracy is not high and its convergence speed is not fast enough; although the particle swarm algorithm has good convergence effect, its parameters have certain coupling and parameter sensitivity are also relatively large, and local optimality is prone to occur in complex situations.

[0004] How to solve the above technical problems is the subject faced by the present invention. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for identifying the solar daily total radiation model based on an improved chaotic SCSO algorithm. The improved chaotic SCSO algorithm proposed in the present invention is a swarm intelligence optimization algorithm. Compared with traditional algorithms, its convergence accuracy and convergence speed are very high. At the same time, the computational complexity and parameter selection difficulty are less than most intelligent algorithms. It can be well applied to the parameter identification of the solar daily total radiation model.

[0006] The present invention is achieved by the following measures: a method for identifying a solar daily total radiation model based on an improved chaotic SCSO algorithm, comprising the following steps:

[0007] Step 1) establishing a Box-Jenkins model of the solar daily total radiation model;

[0008] Step 2) Construct the identification process of the improved chaotic SCSO algorithm.

[0009] As a further optimization scheme of the solar daily total radiation model identification method based on the improved chaotic SCSO algorithm provided by the present invention, the specific modeling steps of step 1) are as follows:

[0010] Step 1-1) Construct the Box-Jenkins model of the solar daily total radiation model: In formula (1), y(t) is the output of the system, x(t) is the non-interference output of the system, and w(t) is the colored noise output:

[0011] y(t)=x(t)+w(t), (1)

[0012] Where u(t) and v(t) are the input and error of the system respectively, and x(t) and w(t) are expressed as:

[0013]

[0014]

[0015] A(z -1 ), B(z -1 ), C(z -1 ) and D(z -1 ) is about the backshift operator z -1 Polynomials of :

[0016]

[0017]

[0018]

[0019]

[0020] Step 1-2) According to equations (2) and (3), the relationship between the output y(t) and the input u(t), the colored noise w(t), the interference-free output x(t) and the error v(t) can be obtained, where

[0021]

[0022]

[0023] Then the Box-Jenkins system of the solar daily total radiation model can be written as:

[0024]

[0025] The parameter estimation vector and system input and output data vector are

[0026]

[0027] As a further optimization scheme of the solar daily total radiation model identification method based on the improved chaotic SCSO algorithm provided by the present invention, the steps of constructing the identification process of the improved chaotic SCSO algorithm in step 2) are as follows:

[0028] Step 2-1) Initialize the population, for d m The identification problem of n parameters in a dimension generates a m ×n initial candidate population, and perform chaotic mapping on the initial candidate population;

[0029] Step 2-2) using the sunshine percentage as input data of the solar daily total radiation amount model, and using the solar daily total radiation amount as output data of the solar daily total radiation amount model;

[0030] Step 2-3) Define the fitness function

[0031]

[0032] Where: l is the data length, t is the time, is the estimated value of the system output, and y(t) is the actual output value of the system;

[0033] Step 2-4) According to formula (6), the parameter value r of the improved chaotic SCSO algorithm is linearly decreased from 2 to 0. G :

[0034]

[0035] Where: s M is the parameter value that simulates the auditory characteristics of the sand cat, which is set to 2 here;

[0036] Step 2-5) According to formula (7) and the obtained decrement parameter value r G Find the main parameter R that controls the transition between exploration and development of SandCat:

[0037] R=2×r G ×rand(0,1)-r G , (7)

[0038] Step 2-6) Calculate the search sensitivity range r of the sand dune cat exploration and development stage according to formula (8):

[0039] r=r G ×rand(0,1), (8)

[0040] Step 2-7) The best position, which is the position with the smallest fitness value, is recorded as θ b , and record the current position as θ c ;

[0041] Step 2-8) Use formula (9) to calculate the improved global reduction factor λ and determine the value of parameter R. If R≤1, enter the development stage and use formula (10) to calculate the next position, that is, the parameter estimation vector If R>1, it enters the exploration stage and uses formula (11) to calculate the next position, that is, the parameter estimation vector

[0042]

[0043]

[0044] Where: θ rd is a random position, α is a random angle between 0° and 360°;

[0045]

[0046] Where: c p is a random number smaller than the population size, θ bc is the assumed best next position;

[0047] Step 2-9) determines whether the maximum number of recursions has been reached. If not, the program jumps to step 2-4). If so, it proceeds to step 2-10);

[0048] Step 2-10) Output the identified parameter estimation vector Complete identification.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] (1) The present invention establishes a Box-Jenkins model of the total solar radiation model, takes the percentage sunshine rate as input data, and uses the improved chaotic SCSO algorithm to identify the model parameters; Figure 4 It can be seen that the algorithm can identify the model parameters very well.

[0051] (2) Compared with the SCSO algorithm, the improved chaotic SCSO algorithm in the present invention adds a global linear decreasing factor to the SCSO algorithm and adopts a chaotic mapping conversion function to map the initial value of the sand cat to between (0 and 1). This increases the search range of the sand cat, improves the convergence accuracy, and accelerates the convergence speed. Therefore, the error estimation is small, and the Box-Jenkins model of the solar daily total radiation can be well identified. At the same time, it also shows that this identification method has good applicability to the solar daily total radiation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0053] Figure 1 This is the overall flow chart of the solar daily total radiation model identification method based on the improved chaotic SCSO algorithm provided by the present invention.

[0054] Figure 2 Schematic diagram of the solar radiation process of the solar daily total radiation model identification method based on the improved chaotic SCSO algorithm provided by the present invention.

[0055] Figure 3 Schematic diagram of the general model of the Box-Jenkins system of the solar daily total radiation model identification method of the improved chaotic SCSO algorithm provided by the present invention.

[0056] Figure 4 Schematic diagram of the error between the identification parameters of the present invention and the true value. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. Of course, the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0058] Example 1

[0059] See also Figures 1 to 4 This embodiment provides a technical solution for identifying a solar daily total radiation model based on an improved chaotic SCSO algorithm. The specific steps are as follows:

[0060] Step 1) establishing a Box-Jenkins model of the solar daily total radiation model;

[0061] Step 2) Construct the identification process of the improved chaotic SCSO algorithm.

[0062] Specifically, the specific modeling steps of step 1) are as follows:

[0063] Step 1-1) Construct the Box-Jenkins model of the solar daily total radiation model: In formula (1), y(t) is the output of the system, x(t) is the non-interference output of the system, and w(t) is the colored noise output:

[0064] y(t)=x(t)+w(t), (1)

[0065] Where u(t) and v(t) are the input and error of the system respectively, and x(t) and w(t) are expressed as:

[0066]

[0067]

[0068] A(z -1 ), B(z -1 ), C(z -1 ) and D(z -1 ) is about the backshift operator z -1 Polynomials of :

[0069]

[0070]

[0071]

[0072]

[0073] Step 1-2) According to equations (2) and (3), the relationship between the output y(t) and the input u(t), the colored noise w(t), the interference-free output x(t) and the error v(t) can be obtained, where

[0074]

[0075]

[0076] Then the Box-Jenkins system of the solar daily total radiation model can be written as:

[0077]

[0078] The parameter estimation vector and system input and output data vector are

[0079]

[0080] Specifically, the steps of constructing the identification process of the improved chaotic SCSO algorithm in step 2) are as follows:

[0081] Step 2-1) Initialize the population, for d m The identification problem of n parameters in a dimension generates a m ×n initial candidate population, and perform chaotic mapping on the initial candidate population;

[0082] Step 2-2) using the sunshine percentage as input data of the solar daily total radiation amount model, and using the solar daily total radiation amount as output data of the solar daily total radiation amount model;

[0083] Step 2-3) Define the fitness function

[0084]

[0085] Where: l is the data length, t is the time, is the estimated value of the system output, and y(t) is the actual output value of the system;

[0086] Step 2-4) According to formula (6), the parameter value r of the improved chaotic SCSO algorithm is linearly decreased from 2 to 0. G :

[0087]

[0088] Where: s M is the parameter value that simulates the auditory characteristics of the sand cat, which is set to 2 here;

[0089] Step 2-5) According to formula (7) and the obtained decrement parameter value r G Find the main parameter R that controls the transition between exploration and development of SandCat:

[0090] R=2×r G ×rand(0,1)-r G , (7)

[0091] Step 2-6) Calculate the search sensitivity range r of the sand dune cat exploration and development stage according to formula (8):

[0092] r=r G ×rand(0,1), (8)

[0093] Step 2-7) The best position, which is the position with the smallest fitness value, is recorded as θ b , and record the current position as θ c ;

[0094] Step 2-8) Use formula (9) to calculate the improved global reduction factor λ and determine the value of parameter R. If |R|≤1, enter the development stage and use formula (10) to calculate the next position, that is, the parameter estimation vector If |R|>1, it enters the exploration phase and uses formula (11) to calculate the next position, that is, the parameter estimation vector

[0095]

[0096]

[0097] Where: θ rd is a random position, α is a random angle between 0° and 360°;

[0098]

[0099] Where: c p is a random number smaller than the population size, θ bc is the assumed best next position;

[0100] Step 2-9) determines whether the maximum number of recursions has been reached. If not, the program jumps to step 2-4). If so, it proceeds to step 2-10);

[0101] Step 2-10) Output the identified parameter estimation vector Complete identification.

[0102] The present embodiment is further described in conjunction with the accompanying drawings and embodiments. This embodiment is only used to illustrate this embodiment, but the scope of this embodiment is not limited to this embodiment. The schematic diagram of the total solar radiation used in this embodiment is as follows: Figure 2 shown.

[0103] Based on the general Box-Jenkins model mentioned above, this embodiment can be modeled as follows:

[0104] y(t)=x(t)+w(t),

[0105]

[0106]

[0107] Comparing the above model with step 1), we can get

[0108] a1=-1.3, a2=1.1, b1=0.6, b2=-1.3, c1=-1.24, c2=0.98, d1=1.0, d2=-1.1,

[0109] Determine a fitness function for the above model In order to be used in the improved chaotic SCSO algorithm, the fitness function is defined as follows:

[0110]

[0111] Where, is the estimated value of the output, and y(t) is the actual value of the output.

[0112] In order to conveniently substitute the parameters to be identified into the improved chaotic SCSO algorithm, the parameters to be identified are combined into a parameter vector θ, and the parameters to be identified are as follows:

[0113] θ=[a1,a2,b1,b2,c1,c2,d1,d2] T ∈Rn

[0114] Initialize the population according to step 2-1) and perform chaotic mapping at the same time;

[0115] Collect input and output data according to step 2-2);

[0116] Define the fitness function according to steps 2-3)

[0117] According to steps 2-4), the linear decreasing weight r of the improved chaotic SCSO algorithm is calculated. G ;

[0118] Calculate the conversion parameter R according to steps 2-5);

[0119] Calculate the sand cat's search sensitivity range r according to steps 2-6);

[0120] According to step 2-7) record the position θ with the minimum fitness b and the current position θ c ;

[0121] According to steps 2-8) determine the size of R and calculate the parameter estimation vector

[0122] Complete the loop according to steps 2-9) and 2-10) and output the parameter estimation vector obtained by identification Complete identification.

[0123] Among them, when setting the linear decreasing parameter value r of the sand cat G Initial values, population size, and parameter values ​​s for simulating the auditory characteristics of sand cats M Several issues need to be considered when setting the initial value: For the linearly decreasing parameter value r G The initial value is usually 2 and the parameter value decreases linearly to 0. The parameter value s is selected to simulate the auditory characteristics of the sand cat. M It is usually set to 2. When the population size is too large, the accuracy will be improved because its search ability is enhanced, but its search time will be longer. Conversely, when the population size is small, the search time will be shortened, but its search accuracy cannot be guaranteed. Therefore, for models with different structures and different input and output data, some reasonable changes can be made when actually selecting these parameter values.

[0124] The parameter identification results of the solar daily total radiation model identification method based on the improved chaotic SCSO algorithm of this embodiment are as follows: Figure 4As shown in the figure, this method has high identification accuracy. Furthermore, as it is derived from a heuristic intelligent optimization algorithm, its search capability is superior to traditional algorithms. Therefore, it converges with high accuracy and consumes very little search time. This also demonstrates the good applicability of this identification method to the global solar radiation model.

[0125] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for identifying the solar daily total radiation model based on an improved chaotic SCSO algorithm, characterized in that: The following steps are involved: Step 1) establishing a Box-Jenkins model of the solar daily total radiation model; Step 2) constructing the identification process of the improved chaotic SCSO algorithm; The steps of step 2) constructing the identification process of the improved chaotic SCSO algorithm are as follows: Step 2-1) Initialize the population, for d m The identification problem of n parameters in a dimension generates a m ×n initial candidate population, and perform chaotic mapping on the initial candidate population; Step 2-2) using the sunshine percentage as input data of the solar daily total radiation amount model, and using the solar daily total radiation amount as output data of the solar daily total radiation amount model; Step 2-3) Define the fitness function Where: l is the data length, t is the time, is the estimated value of the system output, and y(t) is the actual output value of the system; Step 2-4) According to formula (6), the parameter value r of the improved chaotic SCSO algorithm is linearly decreased from 2 to 0. G : Where: s M is the parameter value that simulates the auditory characteristics of the sand cat; Step 2-5) According to formula (7) and the obtained decrement parameter value r G Find the main parameter R that controls the transition between exploration and development of SandCat: R=2×r G ×rand(0,1)-r G , (3) Step 2-6) Calculate the search sensitivity range r of the sand dune cat exploration and development stage according to formula (8): r=r G ×rand(0,1), (4) Step 2-7) The best position, which is the position with the smallest fitness value, is recorded as θ b , and record the current position as θ c ; Step 2-8) Use formula (9) to calculate the improved global reduction factor λ and determine the value of parameter R. If |R|≤1, enter the development stage and use formula (10) to calculate the next position, that is, the parameter estimation vector If |R|>1, it enters the exploration phase and uses formula (11) to calculate the next position, that is, the parameter estimation vector Where: θ rd is a random position, α is a random angle between 0° and 360°; Where: c p is a random number smaller than the population size, θ bc is the assumed best next position; Step 2-9) determines whether the maximum number of recursions has been reached. If not, the program jumps to step 2-4). If so, it proceeds to step 2-10); Step 2-10) Output the identified parameter estimation vector Complete identification.

2. The method for identifying the solar daily total radiation model based on the improved chaotic SCSO algorithm according to claim 1 is characterized in that: The modeling steps of step 1) are as follows: Step 1-1) Construct the Box-Jenkins model of the solar daily total radiation model: In formula (1), y(t) is the output of the system, x(t) is the non-interference output of the system, and w(t) is the colored noise output: y(t)=x(t)+w(t), (8) Where u(t) and v(t) are the input and error of the system respectively, and x(t) and w(t) are expressed as: A(z -1 ), B(z -1 ), C(z -1 ) and D(z -1 ) is about the backshift operator z -1 Polynomials of : Step 1-2) According to equations (2) and (3), the relationship between the output y(t) and the input u(t), the colored noise w(t), the interference-free output x(t) and the error v(t) is obtained, where The Box-Jenkins system of the solar daily total radiation model is: The parameter estimation vector and system input and output data vector are:

Images