Energy scheduling system based on wind and light output prediction

Abstract

The invention discloses a wind and light output prediction-based energy scheduling system, which comprises a data processing module, a control module and a power supply module, and is characterized in that the data processing module predicts wind power and photovoltaic power generation according to collected wind and light output data through a CNN-BiLSTM prediction model, and the wind and light output is predicted according to the predicted wind and light output and a scheduling strategy of the lowest system operation cost; the global search capability is improved by introducing a spiral search strategy, so that the situation that spider bees fall into local optimum when selecting a search space is avoided, an optimal system operation scheme is given by adopting an improved spider bee algorithm based on wind and light output prediction data and an energy management strategy, and safe operation of the system is ensured. The generating capacity of the new energy power station is regulated and controlled from the perspective of wind and light output, effective absorption of new energy and effective supply of electric energy required by users are achieved, and the reliability of the system can be guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of new energy dispatching, and in particular to an energy dispatching system based on wind and solar power output prediction. Background Art

[0002] With the continuous growth of energy demand and the improvement of environmental protection awareness, the development and utilization of new energy sources have received more and more attention. In particular, wind power and solar energy, as clean and renewable energy sources, their development and utilization are of great significance to reducing greenhouse gas emissions and alleviating energy crises. However, due to the intermittent and unstable nature of wind and photovoltaic power generation, how to effectively predict and dispatch these energy sources to ensure their coordinated operation with the power grid has become an urgent problem to be solved.

[0003] Traditional energy dispatch systems often rely on experience or simple prediction models. These methods often have difficulty making accurate predictions and dispatch decisions when faced with complex and changeable wind and solar power output. In addition, existing dispatch algorithms also have problems such as low efficiency and easy to fall into local optimality when dealing with large-scale, multi-objective, and multi-constrained optimization problems. How to improve the energy forecasting and dispatching system and enhance the absorption capacity and utilization efficiency of new energy is a direction worthy of research. Summary of the invention

[0004] Purpose of the invention: In view of the problems raised in the background technology, the present invention provides an energy dispatching system based on wind and solar output prediction. The system uses CNN and BiLSTM to build a prediction model, accurately captures the time series characteristics and spatial distribution characteristics of wind and solar output, and achieves high-precision prediction of wind and solar output. Combined with the improved spider bee algorithm, the energy dispatching strategy is optimized to achieve the lowest system operating cost while ensuring the safe and stable operation of the system. This system not only improves the absorption capacity of new energy, but also promotes the green and efficient use of energy.

[0005] Technical solution:

[0006] The present invention discloses an energy dispatching system based on wind and solar power output prediction, the system comprising a data processing module, a control module and a power supply module;

[0007] The data processing module has a built-in CNN-BiLSTM prediction model, which is used to collect historical wind and solar power output data, and predict future wind and solar power output data through the trained CNN-BiLSTM prediction model;

[0008] The control module has built-in improved spider bee algorithm and energy scheduling strategy, which is used to receive the future wind and solar power output data predicted by the data processing module, and provide the optimal system operation plan by introducing the improved spider bee algorithm with spiral search strategy and energy scheduling strategy;

[0009] The power supply module includes a solar array panel, a wind turbine, a battery and a power grid, and is used to receive and implement the optimal system operation plan obtained by the control module.

[0010] Furthermore, the CNN-BiLSTM prediction model establishment process is as follows:

[0011] CNN includes input layer, convolution layer, pooling layer and fully connected layer. The convolution layer is used to extract the temporal and spatial features of wind and solar power output data. Its output is I It can be expressed as:

[0012]

[0013] Among them, W I and b I They represent the weight and bias of the I-th convolutional layer respectively; * represents the convolution operation; f is the activation function, usually a rectified linear unit is used;

[0014] BiLSTM consists of forward LSTM and backward LSTM, and the calculation formula is as follows:

[0015]

[0016] in, is the wind and solar power output data output by the forward LSTM layer; x t The current wind and solar power output data; The wind and solar power output data output by the previous node in the forward LSTM layer; The wind and solar power output data output by the backward LSTM layer; The wind and solar power output data output by the previous node in the backward LSTM layer; y t The final output data of wind and solar power; is the forward weight matrix; is the backward weight matrix; b y is the offset of the linear relationship.

[0017] Furthermore, the energy scheduling strategy is expressed as:

[0018] f t =F mc +F b +F q

[0019]

[0020]

[0021] Among them, F t is the system operation cost; F mcF is the start-up and shutdown maintenance cost of the unit equipment; b F is the system energy purchase cost; q is the system energy abandonment cost; j is the operation and maintenance cost of unit j; P j (t) is the power output of unit j at time t; e t The price of electricity purchased for the system; P ebuy (t) is the amount of electricity purchased by the system at time t; pv and pw are the penalty coefficients for abandoning solar power and wind power, respectively; PV(t) and PV′(t) are the predicted photovoltaic output and actual photovoltaic output at time t, respectively; PW(t) and PW′(t) are the predicted wind turbine output and actual wind turbine output at time t, respectively; The upper and lower limits for purchasing power for the system.

[0022] Furthermore, the amount of electricity purchased by the system at time t, P ebuy (t), actual photovoltaic output PV′(t), actual wind turbine output PW′(t) are optimization variables, and the spider bee algorithm is used for optimization, including the following steps:

[0023] 1) Initialize the population and parameters. The population matrix of the spider bee algorithm is as follows:

[0024]

[0025] Where X is the spider wasp population; X i is the i-th spider bee; x ij is the value of the jth question variable posed by the i-th spider bee; N is the total number of members of the spider bee population; m is the number of decision variables;

[0026] 2) Bring the randomly generated particle positions into the objective function for evaluation to generate a set of corresponding system operation costs;

[0027] 3) Select the search space and locate the area that is most likely to contain the lowest system operation cost. The location update formula is as follows:

[0028]

[0029] In the formula, is the latest position of the spider bee, which means that this position is the latest objective function value; is the position of the i-th spider bee; are the positions of the ath and bth spider bees, respectively. a and b are two indicators randomly selected from the population to determine the search direction; rn is a random number that conforms to the normal distribution; r 1 、r 2 、r 3 、r 4is a random number uniformly distributed between [0,1]; is the position of the c-th spider bee, c is a randomly selected index in the population; l is a random number uniformly distributed between [-2,1]; H is the upper limit of the objective function; L is the lower limit of the objective function;

[0030] 4) Search and track prey. At the beginning of optimization, all spider bees perform a global search within the domain of the optimization problem to find the area that may contain the lowest system operation cost. The position update formula is as follows:

[0031]

[0032] Where t is the current iteration number; t max is the maximum number of iterations; r 5 、r 6 is a random number uniformly distributed between [0,1]; vc is a random number normally distributed between [-k,k],

[0033] 5) Capture prey and build a nest. The spider bee chases the spider and pulls the paralyzed spider to the area containing the lowest system operation cost, and considers it the best place to build a nest. The formula for simulating nesting behavior is as follows:

[0034]

[0035] Where, X * is the best solution for the current system operation; γ is a number generated during the flight of the spider bee; U is a binary number, which is distributed as follows:

[0036] 6) Determine whether the termination condition is met. If the termination condition is not met, return to step 3); if the constraint condition is met, go to step 7);

[0037] 7) Output the optimal objective function value, that is, the minimum system operating cost.

[0038] Furthermore, the global search capability is improved by introducing a spiral search strategy to prevent the spider bee from falling into a local optimum when selecting a search space. The improved formula of step 3) is as follows:

[0039]

[0040] In the formula, is the latest position of the improved spider bee, and β is the defined search step size.

[0041] Beneficial effects:

[0042] 1. The present invention establishes a CNN-BiLSTM prediction model to predict future wind and solar power output data. By fully learning the sequence data of the past and future moments through bidirectional LSTM, the accuracy of the prediction data is further enhanced, and the energy utilization efficiency and scheduling capabilities are enhanced.

[0043] 2. The present invention utilizes the improved spider bee algorithm to improve the global search capability by introducing a spiral search strategy, thereby jumping out of the local optimum and providing a global optimal scheduling solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a structural schematic diagram of the present invention;

[0045] Figure 2 It is the flow chart of the spider bee algorithm of the present invention. DETAILED DESCRIPTION

[0046] The present application is further described in detail below in conjunction with the accompanying drawings. It is necessary to point out here that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technical personnel in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0047] like Figure 1 As shown, this embodiment discloses an energy dispatching system based on wind and solar power output prediction, the system includes a data processing module, a control module and a power supply module;

[0048] The data processing module has a built-in CNN-BiLSTM prediction model, which is used to collect historical wind and solar power output data and predict future wind and solar power output data through the trained CNN-BiLSTM prediction model;

[0049] The process of building the CNN-BiLSTM prediction model includes the following steps:

[0050] 1) Convolutional Neural Networks (CNN) mainly includes input layer, convolution layer, pooling layer and fully connected layer. The convolution layer is used to extract the temporal and spatial characteristics of wind and solar power output data. Its output is I It can be expressed as:

[0051]

[0052] Among them, W I and b I They represent the weight and bias of the I-th convolutional layer respectively; * represents the convolution operation; f is the activation function, usually a rectified linear unit is used;

[0053] 2) Bidirectional Long Short-Term Memory (BiLSTM) is mainly composed of forward LSTM and backward LSTM. Because of the addition of reverse LSTM, BiLSTM can fully learn the sequence data of the past and future moments. The calculation formula is as follows:

[0054]

[0055] In the formula, is the wind and solar power output data output by the forward LSTM layer; x t The current wind and solar power output data; The wind and solar power output data output by the previous node in the forward LSTM layer; The wind and solar power output data output by the backward LSTM layer; The wind and solar power output data output by the previous node in the backward LSTM layer; y t The final output data of wind and solar power; is the forward weight matrix; is the backward weight matrix; b y is the offset of the linear relationship.

[0056] Through the CNN-BiLSTM model, accurate prediction of wind and solar power output within a certain period of time can be obtained.

[0057] The control module has built-in improved spider bee algorithm and energy dispatch strategy, which is used to receive the future wind and solar output data predicted by the data processing module, and give the optimal system operation plan by introducing the improved spider bee algorithm with spiral search strategy and energy dispatch strategy;

[0058] The optimization process is as follows:

[0059] ① Initialize the population and parameters. The population matrix of the spider bee algorithm is as follows:

[0060]

[0061] Where X is the spider wasp population; X i is the i-th spider bee; x ij is the value of the j-th question variable posed by the i-th spider bee; N is the total number of members of the spider bee population; m is the number of decision variables.

[0062] ② The randomly generated particle positions are brought into the objective function for evaluation to generate a set of corresponding system operating costs.

[0063] ③ Select the search space and locate the area that is most likely to contain the lowest system operation cost. The position update formula is as follows:

[0064]

[0065] In the formula, is the latest position of the spider bee, which means that this position is the latest objective function value; is the position of the i-th spider bee; are the positions of the ath and bth spider bees, respectively. a and b are two indicators randomly selected from the population to determine the search direction; rn is a random number that conforms to the normal distribution; r 1 、r 2 、r 3 、r 4 is a random number uniformly distributed between [0,1]; is the position of the c-th spider bee, c is a randomly selected index in the population; l is a random number uniformly distributed between [-2,1]; H is the upper limit of the objective function; L is the lower limit of the objective function.

[0066] ④ Search and track prey. At the beginning of optimization, all spider bees perform a global search within the domain of the optimization problem to find the area that may contain the lowest system operation cost. The position update formula is as follows:

[0067]

[0068] Where t is the current iteration number; t max is the maximum number of iterations; r 5 、r 6 is a random number uniformly distributed between [0,1]; vc is a random number normally distributed between [-k,k],

[0069] ⑤ Capture prey and build a nest. The spider bee chases the spider and pulls the paralyzed spider to the area containing the lowest system operation cost, and considers it the best place to build a nest. The formula for simulating nesting behavior is as follows:

[0070]

[0071] Where, X * is the best solution for the current system operation; γ is a number generated during the flight of the spider bee; U is a binary number, which is distributed as follows:

[0072] ⑥ Determine whether the termination condition is met. If the termination condition is not met, return to step 3); if the constraint condition is met, go to step 7).

[0073] ⑦ Output the optimal objective function value, that is, the minimum system operating cost.

[0074] In this embodiment, in order to prevent the spider bee from falling into the local optimum when selecting the search space and obtain the global optimal wind and solar output scheduling plan, a spiral search strategy is introduced in step ③ to improve the global search capability and improve the spider bee algorithm. The improved formula is as follows:

[0075]

[0076] In the formula, is the latest position of the improved spider bee, and β is the defined search step size.

[0077] The power supply module includes a solar array panel, a wind turbine, a battery and a power grid, and is used to receive and implement the optimal system operation plan obtained by the control module.

[0078] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. An energy dispatching system based on wind and solar power output prediction, characterized in that: The system includes a data processing module, a control module and a power supply module; The data processing module has a built-in CNN-BiLSTM prediction model, which is used to collect historical wind and solar power output data, and predict future wind and solar power output data through the trained CNN-BiLSTM prediction model; The control module has built-in improved spider bee algorithm and energy scheduling strategy, which is used to receive the future wind and solar power output data predicted by the data processing module, and provide the optimal system operation plan by introducing the improved spider bee algorithm with spiral search strategy and energy scheduling strategy; The power supply module includes a solar array panel, a wind turbine, a battery and a power grid, and is used to receive and implement the optimal system operation plan obtained by the control module.

2. The energy dispatching system based on wind and solar power output prediction according to claim 1 is characterized in that: The CNN-BiLSTM prediction model establishment process is as follows: CNN includes input layer, convolution layer, pooling layer and fully connected layer. The convolution layer is used to extract the temporal and spatial features of wind and solar power output data. Its output is I It can be expressed as: O I =fW I * I+b I Among them, W I and b I They represent the weight and bias of the I-th convolutional layer respectively; * represents the convolution operation; f is the activation function, usually a rectified linear unit is used; BiLSTM consists of forward LSTM and backward LSTM, and the calculation formula is as follows: in, is the wind and solar power output data output by the forward LSTM layer; x t The current wind and solar power output data; The wind and solar power output data output by the previous node in the forward LSTM layer; It is the wind and solar power output data output by the backward LSTM layer; The wind and solar power output data output by the previous node in the backward LSTM layer; y t The final output data of wind and solar power; is the forward weight matrix; is the backward weight matrix; b y is the offset of the linear relationship.

3. The energy dispatching system based on wind and solar power output prediction according to claim 1 is characterized in that: The energy scheduling strategy expression is: minF t =F mc +F b +F q Among them, F t is the system operation cost; F mc F is the start-up and shutdown maintenance cost of the unit equipment; b F is the system energy purchase cost; q is the system energy abandonment cost; j is the operation and maintenance cost of unit j; P j (t) is the power output of unit j at time t; e t The price of electricity purchased for the system; P ebuy (t) is the amount of electricity purchased by the system at time t; pv and pw are the penalty coefficients for abandoning solar power and wind power, respectively; PV(t) and PV′(t) are the predicted photovoltaic output and actual photovoltaic output at time t, respectively; PW(t) and PW′(t) are the predicted wind turbine output and actual wind turbine output at time t, respectively; The upper and lower limits for purchasing power for the system.

4. The energy dispatching system based on wind and solar power output prediction according to claim 3 is characterized in that: The amount of electricity purchased by the system at time t is P ebuy (t), actual photovoltaic output PV′(t), actual wind turbine output PW′(t) are optimization variables, and the spider bee algorithm is used for optimization, including the following steps: 1) Initialize the population and parameters. The population matrix of the spider bee algorithm is as follows: Where X is the spider wasp population; X i is the i-th spider bee; x ij is the value of the jth question variable posed by the i-th spider bee; N is the total number of members of the spider bee population; m is the number of decision variables; 2) Bring the randomly generated particle positions into the objective function for evaluation to generate a set of corresponding system operation costs; 3) Select the search space and locate the area that is most likely to contain the lowest system operation cost. The location update formula is as follows: In the formula, is the latest position of the spider bee, which means that this position is the latest objective function value; is the position of the i-th spider bee; are the positions of the ath and bth spider bees, respectively. a and b are two indicators randomly selected from the population to determine the search direction; rn is a random number that conforms to the normal distribution; r1, r2, r3, and r4 are random numbers uniformly distributed between [0,1]; is the position of the c-th spider bee, c is a randomly selected index in the population; l is a random number uniformly distributed between [-2,1]; H is the upper limit of the objective function; L is the lower limit of the objective function; 4) Search and track prey. At the beginning of optimization, all spider bees perform a global search within the domain of the optimization problem to find the area that may contain the lowest system operation cost. The position update formula is as follows: Where t is the current iteration number; t max is the maximum number of iterations; r5 and r6 are random numbers uniformly distributed between [0,1]; vc is a random number normally distributed between [-k,k], 5) Capture prey and build a nest. The spider bee chases the spider and pulls the paralyzed spider to the area containing the lowest system operation cost, and considers it the best place to build a nest. The formula for simulating nesting behavior is as follows: Where, X * is the best solution for the current system operation; γ is a number generated during the flight of the spider bee; U is a binary number, which is distributed as follows: 6) Determine whether the termination condition is met. If not, return to step 3); if the constraint condition is met, go to step 7); 7) Output the optimal objective function value, that is, the minimum system operating cost.

5. The energy dispatching system based on wind and solar power output prediction according to claim 1 or 4, characterized in that: The spiral search strategy is introduced to improve the global search capability and prevent the spider bee from falling into the local optimum when selecting the search space. The improved formula of step 3) is as follows: In the formula, is the latest position of the improved spider bee, and β is the defined search step size.

Images