Flexible control optimization method for multi-network interconnected novel energy system

Abstract

The invention discloses a flexible control optimization method for a multi-network interconnected novel energy system, and the method comprises the steps: collecting flexible control data and operation data of the novel energy system, and carrying out the preprocessing of the flexible control data and the operation data; performing load prediction on the flexible control data to obtain building energy consumption data, and performing time sequence deviation analysis on the building energy consumption data and the working data to obtain difference data; performing influence analysis on the associated data according to the difference data to obtain an adjusting and optimizing coefficient, and adjusting the flexible control data according to the adjusting and optimizing coefficient to obtain decision data; and constructing an energy system flexible control model according to the decision data, optimizing the energy system flexible control model, inputting to-be-optimized data into the energy system flexible control model, and outputting an optimization result. The method not only can improve the precision of flexible control optimization of the novel energy system, but also has good interpretability, and can be directly applied to a flexible control system.

Description

Technical Field

[0001] The present invention relates to the field of control, and in particular to a flexible control optimization method for a new type of energy system interconnected by multiple networks. Background Art

[0002] With the continuous growth of global energy demand and the diversification of energy structure, new energy systems with multi-network interconnection have gradually become an important trend in future energy supply. Such systems integrate multiple energy forms (such as solar energy, wind energy, geothermal energy, etc.) and multiple energy networks, aiming to improve energy efficiency and enhance system stability and reliability. However, the complexity of multi-network interconnection makes the control and management of the system face huge challenges, especially how to achieve flexible control of the energy system according to the real-time changing energy supply and demand, so as to optimize energy distribution, reduce waste and improve the overall performance of the system, which has become an urgent problem to be solved.

[0003] In the existing technology, although there are some energy management systems that can collect and analyze system operation data, most of them focus on the control optimization of a single energy form or a single network, and lack a comprehensive optimization strategy for multi-network interconnected systems. In addition, the in-depth mining and utilization of flexible control data is insufficient, making it difficult to accurately predict and respond to system load changes, resulting in control strategy lags and affecting system efficiency. Summary of the invention

[0004] The purpose of the present invention is to provide a flexible control optimization method for a new type of energy system interconnected by multiple networks.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Collecting flexible control data and operation data of the new energy system, and preprocessing the flexible control data and the operation data; the operation data includes working data and associated data;

[0008] Performing load forecasting on the flexible control data to obtain building energy consumption data, and performing time series deviation analysis on the building energy consumption data and the working data to obtain difference data;

[0009] Performing an impact analysis on the associated data according to the difference data to obtain a tuning coefficient, and adjusting the flexible control data according to the tuning coefficient to obtain decision data;

[0010] An energy system flexible control model is constructed according to the decision data, the energy system flexible control model is optimized, the data to be optimized is input into the energy system flexible control model, and the optimization result is output.

[0011] Furthermore, the method of performing load forecasting on the flexible control data to obtain building energy consumption data includes:

[0012] The flexible control data is input into the deep energy consumption prediction model, a white noise series is added to the flexible control data to obtain a time series, and the series is decomposed to obtain multimodal components. The expression is:

[0013]

[0014] The tth iteration The time series of moments is The i-th multimodal component is V i The upper limit of running time is The tth iteration The residual at time is The number of time series is N, and the qth sub-signal is The i-th multimodal component at time The number of sub-signals is The tth iteration The i-th multimodal component at time

[0015] The convolutional layer expression is given in terms of multimodal components:

[0016]

[0017] The expansion coefficient is α, the length of the multimodal component is L, the kth convolution filter is μ(k), and the The convolutional layer at this moment is The number of convolution filters is S, the i-th multimodal component of the L-α*ζ-th convolution filter;

[0018] Calculate the residual of the residual module:

[0019] V rd =Act(V i +F(V i ))

[0020] The mapping function of the multimodal component is F(V i ), the activation function is Act(·), and the residual is V rd ;

[0021] A fully connected layer is set at the end of the residual module to calculate the terminal output:

[0022]

[0023] The weight vector is The bias vector is γ, the dot product function is dot(·,·), and the predicted value is

[0024] Calculate the loss function of the fully connected layer, the expression is:

[0025]

[0026] The regularization parameter is η, and the loss function of the weight vector is The norm function is ‖·‖, and the actual value is

[0027] The linear fitting of the predicted energy consumption is expressed as follows:

[0028] U=f0+f1U1+f2U2

[0029] The predicted value is U, the predicted result of outdoor factors is U1, the predicted value of indoor factors is U2, and the linear parameters are f0, f1, and f2 respectively;

[0030] Calculate the cost function:

[0031]

[0032] The cost function of the linear parameter is C(f), and the maximum number of iterations is t max , the i-th predicted energy consumption is The actual energy consumption of the i-th The multimodal components at the input of the tth iteration are The multimodal components at the output of the tth iteration are

[0033] Iterate continuously until the cost function reaches the minimum value, and output the predicted energy consumption as building energy consumption data.

[0034] Furthermore, the method of performing time series deviation analysis on the building energy consumption data and the working data to obtain difference data includes:

[0035] A long short-term memory network is used to learn the long-term dependencies between building energy consumption data and work data through a gating mechanism, and the timing deviations in the time series are captured based on the long-term dependencies.

[0036] Furthermore, the method for obtaining the tuning coefficient includes:

[0037] Taking the difference data as the variation degree of the associated data, multivariate variance analysis is used to perform variation control analysis on the variation degree according to the relationship between multiple dependent variables and multiple independent variables to obtain the analysis results;

[0038] Calculate the tuning coefficient based on the analysis results:

[0039]

[0040] The jth associated data is λ j , the jth degree of change is Φ j , the degree of change Φ j and associated data λ j The correlation degree is β(Φ j ,λ j ), the control coefficient is κ, the number of associated data is Z, and the j+1th associated data is λ j+1 , the j+1th degree of change is Φ j+1 , associated data λ j and associated data λ j+1 The degree of change is m(λ j ,λ j+1 ).

[0041] Furthermore, the method for obtaining decision data by adjusting the flexible control data according to the tuning coefficient includes:

[0042] Calculate the fitness value of the particle:

[0043]

[0044] The penalty term for the cth particle at the sth temperature is β c (s), the upper temperature limit is The comfort function of the cth particle is The energy consumption load is Q, and the minimum value of the energy consumption load is Q min , the maximum value of energy consumption load is Q max , the thermal comfort is H, and the minimum value of thermal comfort is H min The maximum value of thermal comfort is H max ;

[0045] Take the particle with the largest comfort value as the attacker and calculate the average Euclidean distance between the particle and the attacker:

[0046]

[0047] The position of the cth particle is D c , the attacker's position is D ak , the average Euclidean distance between the particle and the attacker is The number of particles is

[0048] Select particles whose Euclidean distance to the attacker is greater than the average Euclidean distance, and update the particle position. The expression is:

[0049]

[0050] The updated position of the cth particle in the tth iteration is The upper bound of the search space is The lower bound of the search space is The position of the cth particle at the tth iteration is D c (t);

[0051] Calculate the fitness value of the updated particles and divide the attackers into two groups according to the fitness value;

[0052] Calculate the distance and step size between the particle and the searcher:

[0053]

[0054] ω=2r2·vv

[0055] The chaotic mutation is Cv, the random numbers from 0 to 1 are r1 and r2, the linear attenuation factor is v, and the distance between the particle and the searcher is The optimal position of the tth iteration is D bs (t), the position vector of the particle at the tth iteration is D cp (t), the moving step is ω;

[0056] Update the particle's position to obtain an adaptive position. The expression is:

[0057]

[0058] The adaptive position of the cth particle in the tth iteration is The maximum number of iterations is t max , randomly select the search agent as M;

[0059] Until the maximum number of iterations is reached to obtain the optimal solution, otherwise the fitness value of the particle is recalculated;

[0060] Calculate decision data based on the optimal solution:

[0061]

[0062] The optimal flexible control data is The tuning coefficient is χ and the deviation is σ.

[0063] Furthermore, a method for constructing a flexible control model of an energy system according to the decision data includes:

[0064] Energy system flexible control models include random forest algorithm, Hungarian algorithm, genetic algorithm, and special long short-term memory network;

[0065] The random forest algorithm divides the decision data into training data and test data in a ratio of 6:1;

[0066] The Hungarian algorithm constructs a cost matrix through training data, and uses row subtraction, column subtraction, and optimal matching selection to iteratively find key data in the maximum similarity interval;

[0067] Genetic algorithms iteratively evolve in the population of key data through operations such as selection, crossover, and mutation to obtain the optimal solution;

[0068] The special long short-term memory network combines the distributed decision-making ability of the multi-agent system and the long-term dependency learning of LSTM, and realizes the prediction and optimal control of the complex dynamic behavior of the energy system through the collaboration of the agent and the optimal solution;

[0069] The test data is input into the energy system flexible control model to obtain the predicted control data. The learning rate of the energy system flexible control model is adjusted according to the mean square error of energy consumption between the predicted control data and the actual control data. The adjustment is stopped when the mean square error of energy consumption is less than 0.014.

[0070] Furthermore, the method for optimizing the flexible control model of the energy system includes:

[0071] Update the position of the searcher, the expression is:

[0072]

[0073] The position of the searcher in the wth dimension is The updated position of the searcher in the wth dimension is The maximum bound of the wth dimension is b w , the minimum boundary of the wth dimension is z w , the proportional relationship between the searcher and the image is x;

[0074] The dynamic pheromone update rule is used to adjust the pheromone formula, and the expression is:

[0075]

[0076] The information density from path a to path y in the tth iteration is The information density from path a to path y in the t+1th iteration is The pheromone evaporation coefficient of the tth iteration is The sum of pheromones left by the searcher on path a to path y during the iteration is The maximum number of iterations is t max , the maximum value of the pheromone evaporation coefficient is The minimum value of the pheromone evaporation coefficient is

[0077] Calculate the probability density function of the t-distribution curve:

[0078]

[0079] The random variable is The probability density function of the t distribution is Non-negative integers are Pi is π, and the natural constant is e;

[0080] The probability density function is used to perturb the pheromone concentration, and the variation expression is:

[0081]

[0082] The t distribution operator with the degree of freedom parameter it is The information density of the t+1th iteration after mutation is

[0083]

[0084] According to the information concentration after mutation, the search sub-position is updated to obtain the optimized position. The expression is:

[0085]

[0086] The learning factor is ρ, and the optimal position of the i-th searcher in the t+1th iteration is The updated position of the i-th searcher in the t-th iteration is

[0087] In a second aspect, an embodiment of the present application further provides an electronic device, including:

[0088] A processor; and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the method steps described in the first aspect.

[0089] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0090] The beneficial effects of the present invention are:

[0091] The present invention is a flexible control optimization method for a new type of energy system interconnected by multiple networks. Compared with the prior art, the present invention has the following technical effects:

[0092] The present invention can improve the accuracy of flexible control of multi-network interconnected new energy systems through preprocessing, load forecasting, timing deviation analysis, impact analysis, acquisition of decision data, model construction and model optimization steps, thereby improving the precision of flexible control of multi-network interconnected new energy systems, optimizing the flexible control of multi-network interconnected new energy systems, greatly saving resources, and improving work efficiency, and can realize flexible automatic control of multi-network interconnected new energy systems, and perform multi-data correction and scheme optimization on the flexible control of multi-network interconnected new energy systems in real time, which is of great significance to the flexible control of multi-network interconnected new energy systems, can adapt to flexible control optimization of multi-network interconnected new energy systems of different standards and flexible control optimization requirements of different multi-network interconnected new energy systems, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 A flowchart of the steps of a flexible control optimization method for a multi-network interconnected new energy system according to the present invention;

[0094] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0095] The present invention is further described below by means of specific embodiments. The illustrative embodiments and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0096] The present invention provides a flexible control optimization method for a new type of energy system interconnected by multiple networks, comprising the following steps:

[0097] like Figure 1 As shown, in this embodiment, the following steps are included:

[0098] Collecting flexible control data and operation data of the new energy system, and preprocessing the flexible control data and the operation data; the operation data includes working data and associated data;

[0099] In the actual evaluation, XX building is taken as the research object, and the working data includes water usage, electricity usage, gas consumption, operating time, operating status, fault alarm, indoor temperature, humidity, air quality, smoke alarm, intrusion alarm, and video surveillance; the associated data includes temperature, humidity, pressure, electromagnetic interference, noise, user needs, accuracy, and response speed;

[0100] Performing load forecasting on the flexible control data to obtain building energy consumption data, and performing time series deviation analysis on the building energy consumption data and the working data to obtain difference data;

[0101] In the actual evaluation, the total energy consumption in the building energy consumption data in July was 1.3011 million KWh, and the difference data included water usage, electricity usage, gas consumption, operating time, operating status, indoor temperature, humidity, and air quality;

[0102] Performing an impact analysis on the associated data according to the difference data to obtain a tuning coefficient, and adjusting the flexible control data according to the tuning coefficient to obtain decision data;

[0103] In the actual evaluation, the tuning factor is 0.179;

[0104] Constructing an energy system flexible control model according to the decision data, optimizing the energy system flexible control model, inputting the data to be optimized into the energy system flexible control model, and outputting the optimization result;

[0105] In the actual evaluation, the comprehensive energy saving rate in July after optimization was 4.65%.

[0106] In this embodiment, the method for performing load forecasting on flexible control data to obtain building energy consumption data includes:

[0107] The flexible control data is input into the deep energy consumption prediction model, a white noise series is added to the flexible control data to obtain a time series, and the series is decomposed to obtain multimodal components. The expression is:

[0108]

[0109] The tth iteration The time series of moments is The i-th multimodal component is V i The upper limit of running time is The tth iteration The residual at time is The number of time series is N, and the qth sub-signal is The i-th multimodal component at time The number of sub-signals is The tth iteration The i-th multimodal component at time

[0110] The convolutional layer expression is given in terms of multimodal components:

[0111]

[0112] The expansion coefficient is α, the length of the multimodal component is L, the kth convolution filter is μ(k), and the The convolutional layer at this moment is The number of convolution filters is S, the i-th multimodal component of the L-α*ζ-th convolution filter;

[0113] Calculate the residual of the residual module:

[0114] V rd =Act(V i +F(V i ))

[0115] The mapping function of the multimodal component is F(V i ), the activation function is Act(·), and the residual is V rd ;

[0116] A fully connected layer is set at the end of the residual module to calculate the terminal output:

[0117]

[0118] The weight vector is The bias vector is γ, the dot product function is dot(·,·), and the predicted value is

[0119] Calculate the loss function of the fully connected layer, the expression is:

[0120]

[0121] The regularization parameter is η, and the loss function of the weight vector is The norm function is ‖·‖, and the actual value is

[0122] The linear fitting of the predicted energy consumption is expressed as follows:

[0123] U=f0+f1U1+f2U2

[0124] The predicted value is U, the predicted result of outdoor factors is U1, the predicted value of indoor factors is U2, and the linear parameters are f0, f1, and f2 respectively;

[0125] Calculate the cost function:

[0126]

[0127] The cost function of the linear parameter is C(f), and the maximum number of iterations is t max , the i-th predicted energy consumption is The actual energy consumption of the i-th The multimodal components at the input of the tth iteration are The multimodal components at the output of the tth iteration are

[0128] Iterate continuously until the cost function reaches the minimum value, and output the predicted energy consumption as building energy consumption data.

[0129] In this embodiment, the method of performing time series deviation analysis on the building energy consumption data and the work data to obtain difference data includes:

[0130] A long short-term memory network is used to learn the long-term dependencies between building energy consumption data and work data through a gating mechanism, and the timing deviations in the time series are captured based on the long-term dependencies.

[0131] In this embodiment, the method for obtaining the tuning coefficient includes:

[0132] Taking the difference data as the variation degree of the associated data, multivariate variance analysis is used to perform variation control analysis on the variation degree according to the relationship between multiple dependent variables and multiple independent variables to obtain the analysis results;

[0133] Calculate the tuning coefficient based on the analysis results:

[0134]

[0135] The jth associated data is λ j , the jth degree of change is φ j , the degree of change φ j and associated data λ j The correlation degree is β(φ j ,λ j ), the control coefficient is κ, the number of associated data is Z, and the j+1th associated data is λ j+1 , the j+1th degree of change is φ j+1 , associated data λ j and associated data λ j+1 The degree of change is m(λ j ,λ j+1 ).

[0136] In this embodiment, the method of adjusting the flexible control data according to the tuning coefficient to obtain decision data includes:

[0137] Calculate the fitness value of the particle:

[0138]

[0139] The penalty term for the cth particle at the sth temperature is β c (s), the upper temperature limit is The comfort function of the cth particle is The energy consumption load is Q, and the minimum value of the energy consumption load is Q min , the maximum value of energy consumption load is Q max, the thermal comfort is H, and the minimum value of thermal comfort is H min The maximum value of thermal comfort is H max ;

[0140] Take the particle with the largest comfort value as the attacker and calculate the average Euclidean distance between the particle and the attacker:

[0141]

[0142] The position of the cth particle is D c , the attacker's position is D ak , the average Euclidean distance between the particle and the attacker is The number of particles is

[0143] Select particles whose Euclidean distance to the attacker is greater than the average Euclidean distance, and update the particle position. The expression is:

[0144]

[0145] The updated position of the cth particle in the tth iteration is The upper bound of the search space is The lower bound of the search space is The position of the cth particle at the tth iteration is D c (t);

[0146] Calculate the fitness value of the updated particles and divide the attackers into two groups according to the fitness value;

[0147] Calculate the distance and step size between the particle and the searcher:

[0148]

[0149] ω=2r2·vv

[0150] The chaotic mutation is Cv, the random numbers from 0 to 1 are r1 and r2, the linear attenuation factor is v, and the distance between the particle and the searcher is The optimal position of the tth iteration is D bs (t), the position vector of the particle at the tth iteration is D cp (t), the moving step is ω;

[0151] Update the particle's position to obtain an adaptive position. The expression is:

[0152]

[0153] The adaptive position of the cth particle in the tth iteration is The maximum number of iterations is t max , randomly select the search agent as M;

[0154] Until the maximum number of iterations is reached to obtain the optimal solution, otherwise the fitness value of the particle is recalculated;

[0155] Calculate decision data based on the optimal solution:

[0156]

[0157] The optimal flexible control data is The tuning coefficient is χ and the deviation is σ.

[0158] In this embodiment, the method for constructing an energy system flexible control model according to the decision data includes:

[0159] Energy system flexible control models include random forest algorithm, Hungarian algorithm, genetic algorithm, and special long short-term memory network;

[0160] The random forest algorithm divides the decision data into training data and test data in a ratio of 6:1;

[0161] The Hungarian algorithm constructs a cost matrix through training data, and uses row subtraction, column subtraction, and optimal matching selection to iteratively find key data in the maximum similarity interval;

[0162] Genetic algorithms iteratively evolve in the population of key data through operations such as selection, crossover, and mutation to obtain the optimal solution;

[0163] The special long short-term memory network combines the distributed decision-making ability of the multi-agent system and the long-term dependency learning of LSTM, and realizes the prediction and optimal control of the complex dynamic behavior of the energy system through the collaboration of the agent and the optimal solution;

[0164] The test data is input into the energy system flexible control model to obtain the predicted control data. The learning rate of the energy system flexible control model is adjusted according to the mean square error of energy consumption between the predicted control data and the actual control data. The adjustment is stopped when the mean square error of energy consumption is less than 0.014.

[0165] In this embodiment, the method for optimizing the energy system flexible control model includes:

[0166] Update the position of the searcher, the expression is:

[0167]

[0168] The position of the searcher in the wth dimension is The updated position of the searcher in the wth dimension is The maximum bound of the wth dimension is b w , the minimum boundary of the wth dimension is z w, the proportional relationship between the searcher and the image is x;

[0169] The dynamic pheromone update rule is used to adjust the pheromone formula, and the expression is:

[0170]

[0171] The information density from path a to path y in the tth iteration is The information density from path a to path y in the t+1th iteration is The pheromone evaporation coefficient of the tth iteration is The sum of pheromones left by the searcher on path a to path y during the iteration is The maximum number of iterations is t max , the maximum value of the pheromone evaporation coefficient is The minimum value of the pheromone evaporation coefficient is

[0172] Calculate the probability density function of the t-distribution curve:

[0173]

[0174] The random variable is The probability density function of the t distribution is Non-negative integers are Pi is π, and the natural constant is e;

[0175] The probability density function is used to perturb the pheromone concentration, and the variation expression is:

[0176]

[0177] The t distribution operator with the degree of freedom parameter it is The information density of the t+1th iteration after mutation is

[0178]

[0179] According to the information concentration after mutation, the search sub-position is updated to obtain the optimized position. The expression is:

[0180]

[0181] The learning factor is ρ, and the optimal position of the i-th searcher in the t+1th iteration is The updated position of the i-th searcher in the t-th iteration is

[0182] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0183] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0184] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0185] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a flexible control optimization device for a new type of energy system with multiple interconnected networks at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the above-mentioned flexible control optimization methods for a new type of energy system with multiple interconnected networks.

[0186] The above application Figure 1The flexible control optimization method of a new energy system with multi-network interconnection disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in a decoding processor. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0187] The electronic device may also perform Figure 1 A flexible control optimization method for a new type of energy system with multi-network interconnection is proposed and implemented Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0188] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple application programs, any one of the aforementioned flexible control optimization methods for the new energy system with multi-network interconnection is executed.

[0189] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0190] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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 generate 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.

[0191] 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.

[0192] 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 in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0193] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0194] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0195] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0196] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0197] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt 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.) that contain computer-usable program code.

[0198] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A flexible control optimization method for a new type of energy system with multiple interconnected networks, characterized in that: The following steps are involved: Collecting flexible control data and operation data of the new energy system, and preprocessing the flexible control data and the operation data; the operation data includes working data and associated data; Performing load forecasting on the flexible control data to obtain building energy consumption data, and performing time series deviation analysis on the building energy consumption data and the working data to obtain difference data; Performing an impact analysis on the associated data according to the difference data to obtain a tuning coefficient, and adjusting the flexible control data according to the tuning coefficient to obtain decision data; An energy system flexible control model is constructed according to the decision data, the energy system flexible control model is optimized, the data to be optimized is input into the energy system flexible control model, and the optimization result is output.

2. According to the flexible control optimization method of a multi-network interconnected new energy system according to claim 1, it is characterized in that: The method for performing load forecasting on flexible control data to obtain building energy consumption data includes: The flexible control data is input into the deep energy consumption prediction model, a white noise series is added to the flexible control data to obtain a time series, and the series is decomposed to obtain multimodal components. The expression is: The tth iteration The time series of moments is The i-th multimodal component is V i The upper limit of running time is The tth iteration The residual at time is The number of time series is N, and the qth sub-signal is The i-th multimodal component at time The number of sub-signals is The tth iteration The i-th multimodal component at time The convolutional layer expression is given in terms of multimodal components: The expansion coefficient is α, the length of the multimodal component is L, the kth convolution filter is μ(k), and the The convolutional layer at this moment is The number of convolution filters is S, the i-th multimodal component of the L-α*ζ-th convolution filter; Calculate the residual of the residual module: V rd =Act(V i +F(V i )) The mapping function of the multimodal component is F(V i ), the activation function is Act(·), and the residual is V rd ; A fully connected layer is set at the end of the residual module to calculate the terminal output: The weight vector is The bias vector is γ, the dot product function is dot(·,·), and the predicted value is Calculate the loss function of the fully connected layer, the expression is: The regularization parameter is η, and the loss function of the weight vector is The norm function is ‖·‖, and the actual value is The linear fitting of the predicted energy consumption is expressed as follows: U=f0+f1U1+f2U2 The predicted value is U, the predicted result of outdoor factors is U1, the predicted value of indoor factors is U2, and the linear parameters are f0, f1, and f2 respectively; Calculate the cost function: The cost function of the linear parameter is C(f), and the maximum number of iterations is t max , the i-th predicted energy consumption is The actual energy consumption of the ith The multimodal components at the input of the tth iteration are The multimodal components at the output of the tth iteration are Iterate continuously until the cost function reaches the minimum value, and output the predicted energy consumption as building energy consumption data.

3. According to the flexible control optimization method of a multi-network interconnected new energy system according to claim 1, it is characterized in that: The method for performing time series deviation analysis on the building energy consumption data and the work data to obtain difference data comprises: A long short-term memory network is used to learn the long-term dependencies between building energy consumption data and work data through a gating mechanism, and the timing deviations in the time series are captured based on the long-term dependencies.

4. According to the flexible control optimization method of a multi-network interconnected new energy system as described in claim 1, it is characterized in that: The method for obtaining the tuning coefficient includes: Taking the difference data as the variation degree of the associated data, multivariate variance analysis is used to perform variation control analysis on the variation degree according to the relationship between multiple dependent variables and multiple independent variables to obtain the analysis results; Calculate the tuning coefficient based on the analysis results: The jth associated data is λ j , the jth degree of change is φ j , the degree of change φ j and associated data λ j The correlation degree is β(φ j ,λ j ), the control coefficient is κ, the number of associated data is Z, and the j+1th associated data is λ j+1 , the j+1th degree of change is φ j+1 , associated data λ j and associated data λ j+1 The degree of change is m(λ j ,λ j+1 ).

5. According to the flexible control optimization method of a multi-network interconnected new energy system as described in claim 1, it is characterized in that: The method for obtaining decision data by adjusting the flexible control data according to the tuning coefficient includes: Calculate the fitness value of the particle: The penalty term for the cth particle at the sth temperature is β c (s), the upper temperature limit is The comfort function of the cth particle is The energy consumption load is Q, and the minimum value of the energy consumption load is Q min , the maximum value of energy consumption load is Q max , the thermal comfort is H, and the minimum value of thermal comfort is H min The maximum value of thermal comfort is H max ; Take the particle with the largest comfort value as the attacker and calculate the average Euclidean distance between the particle and the attacker: The position of the cth particle is D c , the attacker's position is D ak , the average Euclidean distance between the particle and the attacker is The number of particles is Select particles whose Euclidean distance to the attacker is greater than the average Euclidean distance, and update the particle position. The expression is: The updated position of the cth particle in the tth iteration is The upper bound of the search space is The lower bound of the search space is The position of the cth particle at the tth iteration is D c (t); Calculate the fitness value of the updated particles and divide the attackers into two groups according to the fitness value; Calculate the distance and step size between the particle and the searcher: ω=2r2·vv The chaotic mutation is Cv, the random numbers from 0 to 1 are r1 and r2, the linear attenuation factor is v, and the distance between the particle and the searcher is The optimal position of the tth iteration is D bs (t), the position vector of the particle at the tth iteration is D cp (t), the moving step is ω; Update the particle's position to obtain an adaptive position. The expression is: The adaptive position of the cth particle in the tth iteration is The maximum number of iterations is t max , randomly select the search agent as M; Until the maximum number of iterations is reached to obtain the optimal solution, otherwise the fitness value of the particle is recalculated; Calculate decision data based on the optimal solution: The optimal flexible control data is The tuning coefficient is χ and the deviation is σ.

6. The flexible control optimization method for a multi-network interconnected new energy system according to claim 1 is characterized in that: The method for constructing an energy system flexible control model according to the decision data comprises: Energy system flexible control models include random forest algorithm, Hungarian algorithm, genetic algorithm, and special long short-term memory network; The random forest algorithm divides the decision data into training data and test data in a ratio of 6:1; The Hungarian algorithm constructs a cost matrix through training data, and uses row subtraction, column subtraction, and optimal matching selection to iteratively find key data in the maximum similarity interval; Genetic algorithms iteratively evolve in the population of key data through operations such as selection, crossover, and mutation to obtain the optimal solution; The special long short-term memory network combines the distributed decision-making ability of the multi-agent system and the long-term dependency learning of LSTM, and realizes the prediction and optimal control of the complex dynamic behavior of the energy system through the collaboration of the agent and the optimal solution; The test data is input into the energy system flexible control model to obtain the predicted control data. The learning rate of the energy system flexible control model is adjusted according to the mean square error of energy consumption between the predicted control data and the actual control data. The adjustment is stopped when the mean square error of energy consumption is less than 0.

014.

7. The flexible control optimization method for a multi-network interconnected new energy system according to claim 1 is characterized in that: The method for optimizing the flexible control model of the energy system includes: Update the position of the searcher, the expression is: The position of the searcher in the wth dimension is The updated position of the searcher in the wth dimension is The maximum bound of the wth dimension is b w , the minimum boundary of the wth dimension is z w , the proportional relationship between the searcher and the image is x; The dynamic pheromone update rule is used to adjust the pheromone formula, and the expression is: The information density from path a to path y in the tth iteration is l ay (t), the information density from path a to path y in the t+1th iteration is l ay (t+1), the pheromone evaporation coefficient of the tth iteration is The total amount of pheromones left by the searcher on path a to path y during the iteration is Δl ay (t), the maximum number of iterations is t max , the maximum value of the pheromone evaporation coefficient is The minimum value of the pheromone evaporation coefficient is Calculate the probability density function of the t-distribution curve: The random variable is The probability density function of the t distribution is Non-negative integers are Pi is π, and the natural constant is e; The probability density function is used to perturb the pheromone concentration, and the variation expression is: The t distribution operator with the degree of freedom parameter it is The information density of the t+1th iteration after mutation is According to the information concentration after mutation, the search sub-position is updated to obtain the optimized position. The expression is: The learning factor is ρ, and the optimal position of the i-th searcher in the t+1th iteration is The updated position of the i-th searcher in the t-th iteration is 8. An electronic device comprising: processor; as well as A memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute the method according to any one of claims 1 to 6.

Images