Method and system for predicting electric energy substitution potential of marine fishery base

Abstract

The invention relates to the field of electric energy substitution potential prediction, in particular to an electric energy substitution potential prediction method and system for a marine fishery base. Analyzing the multi-dimensional data to obtain key influence factors influencing the electric energy substitution quantity under each dimension; obtaining the electric energy substitution amount of the marine fishery base every year based on the historical data; training a machine learning model by taking the key influence factors as input and the electric energy substitution quantity as output to obtain a trained machine learning model; and inputting the key influence factors influencing the electric energy substitution quantity under each dimension in the to-be-measured data of the marine fishery base into the trained machine learning model to obtain the corresponding electric energy substitution quantity. According to the invention, the actual fluctuation of the marine fishery base can be captured timely and sensitively, and the accuracy of short-term electric energy substitution prediction of marine fishery can be improved.

Description

Technical Field

[0001] The invention relates to the field of electric energy substitution potential prediction, and in particular to a method and system for predicting the electric energy substitution potential of an offshore fishery base. Background Art

[0002] At present, the global energy transition is accelerating, and the research on electric energy substitution is of great significance. Marine fisheries are an important part of human utilization and development of marine resources, and they bear the heavy responsibility of meeting human demand for aquatic products. However, the high energy consumption and pollution problems of traditional fishing methods have seriously affected the atmospheric environment and marine ecosystems. Small-scale fisheries are gradually replaced by industrialized fisheries, and electrical equipment is more widely used in fisheries. Therefore, it is urgent to study the potential prediction system or device of electric energy substitution for traditional energy in marine fisheries to promote the sustainable development of fisheries.

[0003] Compared with other fishing bases, offshore fishing bases are special in that they can make full use of renewable energy, such as tidal energy, wind energy and photovoltaic energy, to reduce dependence on fossil fuels. Therefore, electric energy substitution is an important way to improve the energy efficiency of fishing vessels and promote sustainable development. However, the existing prediction methods for electric energy substitution of traditional energy are all based on predictions based on single historical data conditions, and usually use statistical analysis and machine learning methods for resource assessment, resulting in the prediction methods being not sensitive enough to short-term changes and difficult to fully capture the actual fluctuations of offshore fishing bases. Therefore, the predictions are inaccurate and unreasonable. Summary of the invention

[0004] In order to solve the problem that the existing technology cannot timely and sensitively capture the actual fluctuations of offshore fishery bases, the first aspect of the present invention proposes a method for predicting the electric energy substitution potential of offshore fishery bases, comprising:

[0005] S1: classifying the historical data of the offshore fishery base to obtain multi-dimensional data, and analyzing the multi-dimensional data to obtain key influencing factors affecting the amount of electric energy substitution in each dimension;

[0006] S2: Obtaining the annual electric energy replacement amount of the offshore fishery base based on the historical data;

[0007] S3: taking the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output to train the machine learning model, so as to obtain a trained machine learning model;

[0008] S4: Use the method in step S1 to obtain the key influencing factors that affect the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

[0009] Optionally, the dimensions include technical dimension, economic dimension and policy dimension.

[0010] Optionally, the key influencing factors under the technical dimension include the annual GDP total and / or the growth rate of employment in the fishery aquaculture industry of offshore fishery bases.

[0011] Optionally, the total GDP for the year is calculated as:

[0012] GDP sub (t+1)=K1×GDP sum (t)×h

[0013] Among them, GDP sub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture;

[0014] The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is:

[0015]

[0016] Among them, G sub is the employment population growth rate of fishery aquaculture in offshore fishery bases, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, is the average population in year t+1.

[0017] Optionally, the key influencing factors under the economic dimension include the amount of terminal fossil energy consumption replaced by electricity.

[0018] Optionally, the calculation formula for the amount of terminal fossil energy consumption replaced by electric energy is:

[0019]

[0020] Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e is the average thermal efficiency of electricity.

[0021] Optionally, the key influencing factors under the policy dimension include the proportion of fixed investment in electricity.

[0022] Optionally, the calculation formula for the proportion of fixed investment in electricity is:

[0023]

[0024] Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

[0025] Optionally, the calculation formula for the annual electric energy replacement amount of the offshore fishery base is:

[0026]

[0027] Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

[0028] Optionally, the machine learning model is an improved support vector machine regression model, and the improved support vector machine regression model is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model using a particle swarm algorithm.

[0029] Optionally, the construction process of the improved support vector machine regression model includes:

[0030] The key influencing factors in each dimension obtained from the historical data of offshore fishery bases are used as independent variables, and the corresponding electric energy substitution amount is used as the dependent variable to construct a regression function of the support vector machine regression model including the weight vector and the offset degree.

[0031] Transforming the regression function to obtain an optimization objective function regarding a weight vector and a degree of deviation and corresponding geometric constraints;

[0032] A data set {x i ,y i}, based on the data set {x i ,y i The optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample;

[0033] Mapping the regression function determined by the parameters using a kernel function to obtain a primary regression function;

[0034] A particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

[0035] Optionally, the step of using a particle swarm algorithm to optimize the penalty factor, tolerance, and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model includes:

[0036] Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle and assigning an initial value to the particle update speed;

[0037] The speed and position of the particle are updated using the preset speed and position update formula until the position of a particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained;

[0038] The fitness function is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error between .

[0039] Optionally, the speed and position update formula is:

[0040]

[0041] in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,d] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

[0042] Optionally, the kernel function is a radial basis kernel function.

[0043] The second aspect of the present application provides a system for predicting the potential of electric energy substitution for an offshore fishery base, comprising:

[0044] The first acquisition module is used to classify the historical data of the offshore fishery base to obtain multi-dimensional data, and analyze the multi-dimensional data to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension;

[0045] The second acquisition module is used to obtain the annual electric energy replacement amount of the offshore fishery base based on the historical data;

[0046] Training module: used to train the machine learning model with the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output, so as to obtain a trained machine learning model;

[0047] Prediction module: used to use the first acquisition module to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

[0048] Optionally, the dimensions in the first acquisition module include technical dimensions, economic dimensions and policy dimensions.

[0049] Optionally, the key influencing factors under the technical dimension in the first acquisition module include the annual GDP total and / or the growth rate of the employment population in the fishery farming industry of the offshore fishery base.

[0050] Optionally, the calculation formula for the total GDP of the year in the first acquisition module is:

[0051] GDP sub (t+1)=K1×GDP sum (t)×h

[0052] Among them, GDP sub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture;

[0053] The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is:

[0054]

[0055] Among them, G subis the employment population growth rate of fishery aquaculture in offshore fishery bases, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, is the average population in year t+1.

[0056] Optionally, the key influencing factors under the economic dimension in the first acquisition module include the consumption of terminal fossil energy replaced by electricity.

[0057] Optionally, the calculation formula for the amount of terminal fossil energy consumption replaced by electric energy in the first acquisition module is:

[0058]

[0059] Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e is the average thermal efficiency of electricity.

[0060] Optionally, the key influencing factors under the policy dimension in the first acquisition module include the proportion of fixed investment in electricity.

[0061] Preferably, the calculation formula of the proportion of fixed investment in electricity in the first acquisition module is:

[0062]

[0063] Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

[0064] Optionally, the calculation formula for the annual electric energy replacement amount of the offshore fishery base in the second acquisition module is:

[0065]

[0066] Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

[0067] Optionally, the machine learning model in the training module is an improved support vector machine regression model, and the improved support vector machine regression model is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model using a particle swarm algorithm.

[0068] Optionally, the construction process of the improved support vector machine regression model in the training module includes:

[0069] The key influencing factors in each dimension obtained from the historical data of offshore fishery bases are used as independent variables, and the corresponding electric energy substitution amount is used as the dependent variable to construct a regression function of the support vector machine regression model including the weight vector and the offset degree.

[0070] Transforming the regression function to obtain an optimization objective function regarding a weight vector and a degree of deviation and corresponding geometric constraints;

[0071] A data set {x i ,y i}, based on the data set {x i ,y i The optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample;

[0072] Mapping the regression function determined by the parameters using a kernel function to obtain a primary regression function;

[0073] A particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

[0074] Optionally, the training module uses a particle swarm algorithm to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model, and the steps include:

[0075] Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle and assigning an initial value to the particle update speed;

[0076] The speed and position of the particle are updated using the preset speed and position update formula until the position of a particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained;

[0077] The fitness function is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error between .

[0078] Optionally, the speed and position update formula in the training module is:

[0079]

[0080] in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,D] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

[0081] Optionally, the kernel function in the training module is a radial basis kernel function.

[0082] The third aspect of the present application provides a computer device, including: a data acquisition module, a prediction algorithm module and a data storage module, wherein:

[0083] Data collection module, used to collect historical data and data to be tested at the offshore fishery base;

[0084] A prediction algorithm module is used to execute the above-mentioned electric energy substitution potential prediction method of the offshore fishery base according to the historical data and the data to be tested collected by the data collection module to obtain the prediction result;

[0085] The data storage module is used to store the collected historical data, the data to be tested and the prediction results.

[0086] Optionally, the computer device further includes a user interface, a power detection module and a communication interface, wherein:

[0087] A user interface for visually displaying the historical data, the data to be tested, and the prediction results;

[0088] An electric energy detection module, used for real-time monitoring of the electric power usage of offshore fisheries to obtain the historical data and the data to be tested;

[0089] A communication interface is used to send control instructions to the power equipment of each offshore fishery base based on the prediction results.

[0090] The fourth aspect of the present application provides a computer-readable storage medium, characterized in that an execution program is stored thereon, and when the execution program is executed, the above-mentioned method for predicting the electric energy substitution potential of the offshore fishery base is implemented. Compared with the prior art, the beneficial effects of the present invention are:

[0091] The present invention provides a method and system for predicting the electric energy substitution potential of an offshore fishery base, comprising classifying historical data of the offshore fishery base to obtain multi-dimensional data, analyzing the multi-dimensional data to obtain key influencing factors affecting the amount of electric energy substitution in each dimension; obtaining the annual electric energy substitution amount of the offshore fishery base based on the historical data of the offshore fishery base; training a machine learning model with the key influencing factors as input and the amount of electric energy substitution as output to obtain a trained machine learning model; inputting the key influencing factors affecting the amount of electric energy substitution in each dimension in the data to be tested of the offshore fishery base into the trained machine learning model to obtain the corresponding amount of electric energy substitution; analyzing the electric energy substitution potential from multiple dimensions, thereby improving the comprehensiveness of the data, and thus being able to capture the actual fluctuations of the offshore fishery base in a timely and sensitive manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] Figure 1 A flow chart of a method for predicting electric energy substitution for an offshore fishery base proposed by the present invention;

[0093] Figure 2 A schematic diagram of the construction process of the improved support vector machine regression model proposed in the present invention;

[0094] Figure 3 The present invention proposes Figure 2 A detailed schematic diagram of step S5 in FIG.

[0095] Figure 4 A schematic diagram of the application flow of the electric energy substitution prediction method for an offshore fishery base proposed by the present invention;

[0096] Figure 5 A schematic diagram of a device for implementing the electric energy substitution prediction method for an offshore fishery base proposed by the present invention;

[0097] Figure 6It is a structural schematic diagram of the electric energy substitution potential prediction system of the offshore fishery base of the present invention;

[0098] Figure 7 This is a schematic diagram of the structure of the electronic device proposed by the present invention. DETAILED DESCRIPTION

[0099] The present invention proposes a method and system for predicting the electric energy substitution potential of an offshore fishery base. The present application comprehensively considers the multi-dimensional influencing factors that affect the electric energy substitution potential, can indirectly reflect the complex changes in the resource endowment of renewable energy and the marine environment, can predict the short-term changes of the offshore fishery base, fully capture the actual fluctuations of the offshore fishery base, and improve the accuracy and timeliness of the electric energy substitution potential prediction.

[0100] Embodiment 1:

[0101] A method for predicting electric energy substitution for offshore fishery bases, such as Figure 1 As shown, the process includes the following steps S1 to S4.

[0102] S1: Classify the historical data of the offshore fishery base to obtain multi-dimensional data, and analyze the multi-dimensional data to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension.

[0103] The complex changes in the marine environment will directly or potentially affect all aspects of marine fisheries. Therefore, it is necessary to obtain a variety of historical data of marine fishery bases, including the gross output value of fishery farms, the population of fishery farms, the total energy consumption of fishery farms, and the electricity consumption of fishery farms, calculate the proportion of new power investment, etc., exclude outliers and incomplete data, ensure data quality, and ensure the authority and accuracy of data sources. These historical data are classified in terms of technical dimensions, economic dimensions, and policy dimensions, and then the multi-dimensional data are analyzed to obtain the key influencing factors affecting the amount of electricity substitution in each dimension. For example, the key influencing factors of the technical dimension include the annual GDP total and / or the employment population growth rate of the fishery aquaculture industry in the marine fishery base, which to a certain extent reflects the technical level of the industry. The higher the technical level, the greater the gross output value and can drive the growth of the relevant employment population, while improving a certain level of electrification. The key influencing factors under the economic dimension include the amount of terminal fossil energy consumption replaced by electricity, and the key influencing factors under the policy dimension include the proportion of fixed investment in electricity and fixed investment in energy.

[0104] In a further preferred solution, the annual GDP calculation formula is:

[0105] GDP sub (t+1)=K1×GDP sum (t)×h

[0106] Among them, GDPsub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture;

[0107] The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is:

[0108]

[0109] Among them, G sub is the employment population growth rate of fishery aquaculture in offshore fishery bases, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, is the average population in year t+1;

[0110] In terms of economic dimension, the equivalent calorific value method is used to calculate the amount of terminal fossil energy consumption replaced by electricity to reflect the impact of the economic dimension. The calculation formula is:

[0111]

[0112] Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, such as coal, oil, gas, etc., H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e The average thermal efficiency of electricity is 22 J / kg for coal calorific value and 0.4 for average thermal efficiency; the calorific value of crude oil is 42 MJ / kg and 0.5 for average thermal efficiency; the calorific value of electricity is 3.6 MJ / kWh and 0.9 for average thermal efficiency. It can be seen that the resource endowment of the offshore fishery base, such as the type of resources it possesses, will directly affect the amount of terminal fossil energy consumption replaced by electricity.

[0113] In terms of policy influencing factors, including the annual proportion of fixed investment in electricity, the main consideration is the policies related to electricity investment and construction. In recent years, it has been proposed that new energy installed capacity will become the absolute main body of new additions, and power investment will accelerate the green and low-carbon transformation. Therefore, the embodiment of this application selects the proportion of fixed investment in electricity as a quantitative indicator of the influencing factors of the electricity substitution policy, and the calculation formula is:

[0114]

[0115] Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

[0116] Therefore, the multiple dimensional data of this application can directly or indirectly reflect the complex changing impacts of the marine environment. The GDP and employment situation of marine fisheries reflect the complex changing impacts of the marine environment in the technical dimension, the resource endowment of marine fishery bases reflects the complex changing impacts of the marine environment in the economic dimension, and power investment reflects the complex changing impacts of the marine environment in the policy dimension, so the consideration dimensions are more comprehensive.

[0117] S2: Obtain the annual electricity replacement amount of the offshore fishery base based on the historical data.

[0118] In a further preferred solution, the calculation formula for obtaining the annual electric energy replacement amount of the offshore fishery base based on historical data is:

[0119]

[0120] Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

[0121] S3: Taking the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output, the machine learning model is trained to obtain a trained machine learning model.

[0122] In a further preferred scheme, the machine learning model in the training module is an improved support vector machine regression model, and the improved support vector machine regression model is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model using a particle swarm algorithm.

[0123] In a further preferred embodiment, Figure 2 As shown, the construction process of the improved support vector machine regression model includes the following steps S31 to S35:

[0124] S31: Taking the key influencing factors in each dimension obtained based on the historical data of the offshore fishery base as the independent variables and the corresponding electric energy substitution as the dependent variable, a regression function of the support vector machine regression model including the weight vector and the degree of offset is constructed.

[0125] The regression function in this embodiment is:

[0126]

[0127] Among them, f(x) is the predicted value of the electric energy replacement amount returned by the regression function, ω is the weight vector, is a one-dimensional high-dimensional nonlinear mapping function, b is the degree of offset, ω and b are the parameters to be determined, and x is the key influencing factor of the input.

[0128] S32: Transform the regression function to obtain an optimization objective function regarding the weight vector and the degree of offset and the corresponding geometric constraints.

[0129] For the support vector regression model, the optimization goal is to minimize the complexity of the model while ensuring that the regression function is as close to the training data points as possible within a given tolerance range. Therefore, the support vector regression problem can be transformed into:

[0130]

[0131] in, is the regularization term, n is the total number of samples, f(x i ) is the predicted value of the electric energy replacement of the i-th sample, y i is the true value of the i-th sample, Lε(f(x i ),y i ) is the loss function,

[0132] After the relaxation factor is introduced, the corresponding optimization objective function is:

[0133]

[0134] C is the penalty factor, ξ i is the distance from the true value to the upper boundary of the predicted value, is the true value y i The distance to the lower boundary of the predicted value, i is the i-th sample, and n is the total number of samples.

[0135] The constraint condition in this embodiment requires that the predicted value and the true value are within the tolerance range ε, and the excess is determined by the slack variable ξ i and Control, the geometric constraints are:

[0136]

[0137] is the nonlinear mapping function corresponding to the i-th sample.

[0138] S33: Constructing a data set {x based on the key influencing factors corresponding to the historical data and the corresponding electric energy substitution amount i ,y i}, based on the data set {x i ,y iThe optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample.

[0139] The dataset {x i ,y i Electricity substitution potential in i The calculation formula is:

[0140]

[0141] Among them, D sub,i (t+1) is the amount of electricity replacement in the ith sample in the t+1th year, E i (t+1) is the actual electricity consumption of the ith sample in the t+1th year, D sum,i (t+1) is the total energy consumption of the i-th sample in the t+1th year, E i (t) is the actual electricity consumption of the ith sample in the tth year, D sum，i (t) is the total energy consumption of the ith sample in the tth year.

[0142] Furthermore, the dataset {x i ,y i The training set is used to perform Lagrangian solution to obtain the primary regression function, and the test set is used to perform testing to optimize the penalty factor, tolerance, and kernel function parameters in the primary regression function.

[0143] In order to minimize the optimization objective function, according to the geometric constraints, based on the training set, Lagrange multiplication is used and Karush-Kuhn-Tucker (KKT) conditions are applied to obtain the dual problem of the original constraint problem. The regression function with parameters determined by solving the dual problem based on the data set is:

[0144]

[0145] in, and α i is the Lagrange coefficient.

[0146] S34: using a kernel function to map the regression function determined by the parameters to obtain a primary regression function;

[0147] The regression function determined by the parameters is mapped using a kernel function. The kernel function can map the nonlinear relationship in the one-dimensional space to the multidimensional space and transform it into a linear relationship. The kernel function in the present application is preferably a radial basis kernel function:

[0148]

[0149] Among them, σ is the kernel function parameter, K(x,x i ) is the radial basis kernel function.

[0150] The corresponding primary regression function is:

[0151]

[0152] S35: Using a particle swarm algorithm to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

[0153] The fitness function of the particle swarm algorithm is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error (MSE) between them.

[0154]

[0155] In a further preferred solution, a particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model, such as Figure 3 As shown, the steps include S351 and S352:

[0156] S351: Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle, and assigning an initial value to the particle update speed;

[0157] Based on the test set, the size of the particle swarm population is set, and the fitness function in the particle swarm optimization process is determined. The penalty factor, tolerance and kernel function parameters are used as the particle position vector, so the dimension of the particle position is 3. First, the penalty factor, tolerance and kernel function parameters are assigned initial values ​​to initialize the particle position, and the initial value of the particle update speed is assigned.

[0158] S352: The speed and position of the particle are updated using a preset speed and position update formula until the position of a certain particle makes the fitness function less than the required value or reaches a predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained;

[0159] In a further preferred solution, the speed and position update formula in step S352 is:

[0160]

[0161] in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,D] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

[0162] S4: Use the method in step S1 to obtain the key influencing factors that affect the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

[0163] In the application process, based on the test data of the current offshore fishery base, the method in step S1 is used to obtain the key influencing factors, and the key influencing factors are input into the trained machine learning model to obtain the corresponding electric energy substitution amount.

[0164] This embodiment optimizes the key parameters in the support vector machine regression model by using a particle swarm algorithm, thereby improving the accuracy and generalization ability of the support vector machine regression model, thereby improving the accuracy of short-term electricity substitution prediction for marine fisheries.

[0165] like Figure 4As shown in the figure, in the optimization process, the influencing factors and the calculated electric energy substitution amount and other data are first input, the particle swarm optimization parameters are set and optimized and initialized, and then the particle speed and position are updated. During the update process, for each particle, the corresponding fitness value is compared with the fitness value corresponding to the optimal position of the previous iteration of its individual. If the point is better, the current position replaces the optimal position of the previous iteration, and the particle position is continuously updated until the position of a certain particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor C, tolerance ε and kernel function parameter σ are obtained. The parameter-optimized support vector regression (SVR) model is used to predict the electric energy substitution potential. After optimizing the penalty factor C, tolerance ε and kernel function parameter σ by this method, the prediction performance and accuracy of the support vector regression model are significantly improved, the local optimal problem is avoided, and the fitting ability and generalization ability of the support vector regression model are improved, thereby improving the accuracy and stability of electric energy substitution prediction.

[0166] Then, according to the electricity consumption characteristics and electricity substitution potential of the offshore fishery base, the power management module can be used to finely control various types of aquaculture equipment. According to the characteristics and substitution potential of the power equipment in the fishery, combined with the actual power consumption data of various types of equipment, an optimized alternative power solution is produced, and control strategy instructions are sent to each power equipment through the communication interface. According to the control instructions issued by the system, the power management module starts and stops various types of aquaculture equipment or adjusts the power. Each power management module is responsible for controlling a single type of aquaculture equipment to ensure the efficient implementation of the power substitution solution.

[0167] The method in the present application can be implemented using a prediction device, such as Figure 5 As shown, the prediction device may include a data determination unit, a load prediction unit and a result output unit, wherein the data determination unit is used to perform real-time data analysis and historical data collection, the load prediction unit is used to perform statistical analysis on the collected historical data to train the machine learning model, and finally the result output unit visualizes, stores and outputs the calculation.

[0168] Embodiment 2:

[0169] The present invention based on the same inventive concept also provides a system for predicting the potential of electric energy substitution for an offshore fishery base, such as Figure 6 As shown, including:

[0170] The first acquisition module is used to classify the historical data of the offshore fishery base to obtain multi-dimensional data, and analyze the multi-dimensional data to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension;

[0171] The second acquisition module is used to obtain the annual electric energy replacement amount of the offshore fishery base based on the historical data;

[0172] Training module: used to train the machine learning model with the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output, so as to obtain a trained machine learning model;

[0173] Prediction module: used to use the first acquisition module to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

[0174] In a further preferred solution, the dimensions in the first acquisition module include technical dimensions, economic dimensions and policy dimensions.

[0175] In a further preferred embodiment, the key influencing factors under the technical dimension in the first acquisition module include the annual GDP total and / or the growth rate of the employment population in the fishery farming industry of the offshore fishery base.

[0176] Preferably, the calculation formula for the total GDP of the year in the first acquisition module is:

[0177] GDP sub (t+1)=K1×GDP sum (t)×h

[0178] Among them, GDP sub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture;

[0179] The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is:

[0180]

[0181] Among them, G sub is the employment population growth rate of fishery aquaculture in offshore fishery bases, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, is the average population in year t+1.

[0182] In a further preferred solution, the key influencing factors under the economic dimension in the first acquisition module include the amount of terminal fossil energy consumption replaced by electric energy.

[0183] In a further preferred solution, the calculation formula for the amount of terminal fossil energy consumption replaced by electric energy in the first acquisition module is:

[0184]

[0185] Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e is the average thermal efficiency of electricity.

[0186] In a further preferred solution, the key influencing factors under the policy dimension in the first acquisition module include the proportion of fixed investment in electricity.

[0187] In a further preferred solution, the calculation formula for the proportion of fixed investment in electricity in the first acquisition module is:

[0188]

[0189] Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

[0190] In a further preferred solution, the calculation formula for the annual electric energy replacement amount of the offshore fishery base in the second acquisition module is:

[0191]

[0192] Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

[0193] In a further preferred scheme, the machine learning model in the training module is an improved support vector machine regression model, and the improved support vector machine regression model is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model using a particle swarm algorithm.

[0194] In a further preferred embodiment, the construction process of the improved support vector machine regression model in the training module includes:

[0195] The key influencing factors in each dimension obtained from the historical data of offshore fishery bases are used as independent variables, and the corresponding electric energy substitution amount is used as the dependent variable to construct a regression function of the support vector machine regression model including the weight vector and the offset degree.

[0196] Transforming the regression function to obtain an optimization objective function regarding a weight vector and a degree of deviation and corresponding geometric constraints;

[0197] A data set {x i ,y i}, based on the data set {x i ,y i The optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample;

[0198] Mapping the regression function determined by the parameters using a kernel function to obtain a primary regression function;

[0199] A particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

[0200] In a further preferred solution, the training module uses a particle swarm algorithm to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model, and the steps include:

[0201] Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle and assigning an initial value to the particle update speed;

[0202] The speed and position of the particle are updated using the preset speed and position update formula until the position of a particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained;

[0203] The fitness function is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error between .

[0204] In a further preferred embodiment, the speed and position update formula in the training module is:

[0205]

[0206] in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,D] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

[0207] In a further preferred solution, the kernel function in the training module is a radial basis kernel function.

[0208] Example 3

[0209] A computer device such as Figure 7 As shown, it includes: a data acquisition module, a prediction algorithm module and a data storage module, wherein:

[0210] Data collection module, used to collect historical data and data to be tested at the offshore fishery base;

[0211] A prediction algorithm module is used to execute the above-mentioned electric energy substitution potential prediction method of the offshore fishery base according to the historical data and the data to be tested collected by the data collection module to obtain the prediction result;

[0212] The data storage module is used to store the collected historical data, the data to be tested and the prediction results. It provides data support for subsequent analysis and evaluation, and can also be used to train prediction algorithms to further improve equipment performance.

[0213] In a further preferred embodiment, the computer device further includes a user interface, a power detection module and a communication interface, wherein:

[0214] A user interface for visually displaying the historical data and forecast results;

[0215] An electric energy detection module, used for real-time monitoring of the electric power usage of offshore fisheries to obtain the historical data and the data to be tested;

[0216] A communication interface is used to send control instructions to the power equipment of each offshore fishery base based on the prediction results.

[0217] Example 4

[0218] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and, of course, extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a method for predicting the potential for electric energy substitution of an offshore fishery base in the above embodiment.

[0219] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0220] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0221] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0222] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0223] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for predicting the electric energy substitution potential of an offshore fishery base, characterized in that: include: S1: classifying the historical data of the offshore fishery base to obtain multi-dimensional data, and analyzing the multi-dimensional data to obtain key influencing factors affecting the amount of electric energy substitution in each dimension; S2: Obtaining the annual electric energy replacement amount of the offshore fishery base based on the historical data; S3: taking the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output to train the machine learning model, so as to obtain a trained machine learning model; S4: Use the method in step S1 to obtain the key influencing factors that affect the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

2. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 1, characterized in that: The dimensions include technical dimension, economic dimension and policy dimension.

3. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 2, characterized in that: The key influencing factors under the technical dimension include the annual GDP total and / or the growth rate of employment in the fishery aquaculture industry in offshore fishery bases.

4. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 3, characterized in that: The GDP calculation formula for the year in question is: GDP sub (t+1)=K1×GDP sum (t)×h Among them, GDP sub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture; The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is: Among them, G sub is the growth rate of employed population in fishery breeding industry of offshore fishery base, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, and R is the average population in year t+1.

5. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 2, characterized in that: The key influencing factors under the economic dimension include the amount of terminal fossil energy consumption replaced by electricity.

6. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 5, characterized in that: The calculation formula for the amount of terminal fossil energy consumption replaced by electric energy is: Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e is the average thermal efficiency of electricity.

7. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 2, characterized in that: The key influencing factors under the policy dimension include the proportion of fixed investment in electricity.

8. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 7, characterized in that: The calculation formula for the proportion of electricity fixed investment is: Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

9. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 1, characterized in that: The calculation formula for the annual electric energy replacement of the offshore fishery base is: Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

10. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 1, characterized in that: The machine learning model is an improved support vector machine regression model, which is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model using a particle swarm algorithm.

11. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 10, characterized in that: The construction process of the improved support vector machine regression model includes: The key influencing factors in each dimension obtained from the historical data of offshore fishery bases are used as independent variables, and the corresponding electric energy substitution amount is used as the dependent variable to construct a regression function of the support vector machine regression model including the weight vector and the offset degree. Transforming the regression function to obtain an optimization objective function regarding a weight vector and a degree of deviation and corresponding geometric constraints; A data set {x i ,y i }, based on the data set {x i ,y i The optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample; Mapping the regression function determined by the parameters using a kernel function to obtain a primary regression function; A particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

12. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 11, characterized in that: The step of optimizing the penalty factor, tolerance and kernel function parameters in the primary regression function by using the particle swarm algorithm to obtain an improved support vector machine regression model comprises: Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle and assigning an initial value to the particle update speed; The speed and position of the particle are updated using the preset speed and position update formula until the position of a particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained; The fitness function is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error between .

13. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 12, characterized in that: The speed and position update formula is: in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,d] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

14. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 11, characterized in that: The kernel function is a radial basis kernel function.

15. A system for predicting the potential of electric energy substitution for an offshore fishery base, characterized in that: include: The first acquisition module is used to classify the historical data of the offshore fishery base to obtain multi-dimensional data, and analyze the multi-dimensional data to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension; The second acquisition module is used to obtain the annual electric energy replacement amount of the offshore fishery base based on the historical data; Training module: used to train the machine learning model with the key influencing factors corresponding to the historical data as input and the electric energy substitution amount corresponding to the historical data as output, so as to obtain a trained machine learning model; Prediction module: used to use the first acquisition module to obtain the key influencing factors affecting the amount of electric energy substitution in each dimension of the test data of the offshore fishery base, and input the key influencing factors corresponding to the test data into the trained machine learning model to obtain the corresponding electric energy substitution.

16. The electric energy substitution potential prediction system for an offshore fishery base according to claim 15, characterized in that: The dimensions in the first acquisition module include technical dimension, economic dimension and policy dimension.

17. The electric energy substitution potential prediction system for an offshore fishery base according to claim 16, characterized in that: The key influencing factors under the technical dimension in the first acquisition module include the annual GDP total and / or the growth rate of the employment population in the fishery farming industry of the offshore fishery base.

18. The method for predicting the electric energy substitution potential of an offshore fishery base according to claim 17, characterized in that: The calculation formula for the total GDP of the year in the first acquisition module is: GDP sub (t+1)=K1×GDP sum (t)×h Among them, GDP sub (t+1) is the total GDP of the offshore fishery base in the year t+1, K1 is the elasticity coefficient of GDP development of offshore fishery aquaculture, GDP sum (t) is the total GDP of the offshore fishery base in year t, and h is the production contribution rate of offshore fishery aquaculture; The formula for calculating the growth rate of employment in fishery aquaculture in the offshore fishery base is: Among them, G sub is the employment population growth rate of fishery aquaculture in offshore fishery bases, K2 is the elasticity coefficient of population growth, R1 is the population at the end of year t+1, R is the population at the beginning of year t+1, is the average population in year t+1.

19. The electric energy substitution potential prediction system for an offshore fishery base according to claim 16, characterized in that: The key influencing factors under the economic dimension in the first acquisition module include the amount of terminal fossil energy consumption replaced by electric energy.

20. The electric energy substitution potential prediction system for an offshore fishery base according to claim 19, characterized in that: The calculation formula of the amount of terminal fossil energy consumption replaced by electric energy in the first acquisition module is: Among them, T1 is the terminal fossil energy consumption replaced by electricity, q a is the terminal fossil energy consumption of the a-th energy source, a represents different energy types, H is the energy calorific value, θ is the average thermal efficiency of energy utilization, H e is the electric heating value, θ e is the average thermal efficiency of electricity.

21. The electric energy substitution potential prediction system for an offshore fishery base according to claim 16, characterized in that: The key influencing factors under the policy dimension in the first acquisition module include the proportion of fixed investment in electricity.

22. The electric energy substitution potential prediction system for an offshore fishery base according to claim 21, characterized in that: The calculation formula for the proportion of fixed investment in electricity in the first acquisition module is: Among them, G r It represents the proportion of electricity fixed investment, K is the elasticity coefficient, F1 is electricity fixed investment, and F2 is non-electric energy fixed investment.

23. The electric energy substitution potential prediction system for an offshore fishery base according to claim 15, characterized in that: The calculation formula for the annual electric energy replacement amount of the offshore fishery base in the second acquisition module is: Among them, D sub (t+1) is the amount of electricity replacement in the t+1th year, E(t+1) is the actual electricity consumption in the t+1th year, and D sum (t+1) is the total energy consumption in year t+1, E(t) is the actual electricity consumption in year t, and D sum (t) is the total energy consumption in year t.

24. The electric energy substitution potential prediction system for an offshore fishery base according to claim 15, characterized in that: The machine learning model in the training module is an improved support vector machine regression model, and the improved support vector machine regression model is obtained by optimizing the penalty factor, tolerance and kernel function parameters of the support vector machine regression model through a particle swarm algorithm.

25. The electric energy substitution potential prediction system for an offshore fishery base according to claim 24, characterized in that: The construction process of the improved support vector machine regression model in the training module includes: The key influencing factors in each dimension obtained from the historical data of offshore fishery bases are used as independent variables, and the corresponding electric energy substitution amount is used as the dependent variable to construct a regression function of the support vector machine regression model including the weight vector and the offset degree. Transforming the regression function to obtain an optimization objective function regarding a weight vector and a degree of deviation and corresponding geometric constraints; A data set {x i ,y i }, based on the data set {x i ,y i The optimization objective function and the corresponding geometric constraints are solved by Lagrange multiplication to obtain the weight vector and the offset degree in the regression function, and the weight vector and the offset degree are substituted into the regression function to obtain the regression function with parameters determined, where x i is the key influencing factor of the i-th sample, y i is the electric energy replacement corresponding to the i-th sample; Mapping the regression function determined by the parameters using a kernel function to obtain a primary regression function; A particle swarm algorithm is used to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model.

26. The electric energy substitution potential prediction system for an offshore fishery base according to claim 25, characterized in that: The training module uses a particle swarm algorithm to optimize the penalty factor, tolerance and kernel function parameters in the primary regression function to obtain an improved support vector machine regression model, and the steps include: Taking the penalty factor, tolerance and kernel function parameters as the position vector of the particle, assigning initial values ​​to the penalty factor, tolerance and kernel function parameters in the primary regression function to initialize the position of the particle and assigning an initial value to the particle update speed; The speed and position of the particle are updated using the preset speed and position update formula until the position of a particle makes the fitness function less than the required value or reaches the predetermined number of iterations, and the optimal values ​​of the penalty factor, tolerance and kernel function parameters are obtained; The fitness function is the predicted value of the primary regression function and the data set {x i ,y i The true value y in i The mean square error between .

27. The electric energy substitution potential prediction system for an offshore fishery base according to claim 26, characterized in that: The speed and position update formula in the training module is: in, is the velocity of the k+1th iteration of the dth particle, is the speed of the kth iteration of the dth particle, d∈[1,D], D is the size of the population, k is the number of iterations, s is the dimension of the particle, c1 is the individual learning factor, c2 is the population learning factor, r1,r2∈[1,D] are random numbers, p is the inertia weight, P bestsd is the individual optimal position value of the d-th particle in the first k iterations, G bestsd is the global optimal position value of all particles in the first k iterations; is the k+1th iteration position of the dth particle, is the position of the d-th particle at the k-th iteration.

28. The electric energy substitution potential prediction system for an offshore fishery base according to claim 25, characterized in that: The kernel function in the training module is a radial basis kernel function.

29. A computer device, characterized in that: include: Data acquisition module, prediction algorithm module and data storage module, including: Data collection module, used to collect historical data and data to be tested at the offshore fishery base; A prediction algorithm module, used to execute the electric energy substitution potential prediction method of the offshore fishery base according to any one of claims 1 to 14 according to the historical data and the data to be tested collected by the data collection module to obtain a prediction result; The data storage module is used to store the collected historical data, the data to be tested and the prediction results.

30. The computer device according to claim 19, characterized in that It also includes a user interface, an electric energy detection module and a communication interface, wherein: A user interface for visually displaying the historical data, the data to be tested, and the prediction results; An electric energy detection module, used for real-time monitoring of the electric power usage of offshore fisheries to obtain the historical data and the data to be tested; A communication interface is used to send control instructions to the power equipment of each offshore fishery base based on the prediction results.

31. A computer-readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, the method for predicting the electric energy substitution potential of an offshore fishery base as described in any one of claims 1 to 14 is implemented.

Images