Method and device for optimizing cogeneration system based on data driving

Abstract

The embodiment of the invention provides a heat and power cogeneration system optimization method and device based on data driving, which can be used in the technical field of artificial intelligence, and the method comprises the steps: obtaining energy data in a heat and power cogeneration system; on the basis of multi-task learning, the energy data are predicted, and predicted power generation and heat supply data are generated; through the electric heating load characteristic curve of each heat supply unit, a system income objective function and constraint conditions are constructed according to the power generation heat supply data and the energy data; through an improved optimization algorithm and according to constraint conditions, optimization solution is carried out on a system income objective function, scheduling optimization of the cogeneration system is completed, electric heating load characteristic analysis is carried out through multi-task learning and the improved optimization algorithm, decoupling distributed optimization decomposition is carried out, therefore, a global optimal solution is sought, the calculation efficiency is improved, and the scheduling optimization of the cogeneration system is completed. The accuracy requirement of load prediction is met, calculation time and calculation resources are saved, and the calculation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, in particular to the field of artificial intelligence technology, and more particularly to a data-driven combined heat and power system optimization method and device. Background Art

[0002] With the increase in energy demand and the severity of environmental problems, how to improve energy efficiency and reduce pollution emissions has become an important research topic. The cogeneration system has received widespread attention and application due to its high efficiency and environmental protection. However, as a complex and highly coupled heating and power supply system, it has many problems in its operation. The electrical load and thermal load demand of the cogeneration system are often affected by the external environment and user needs, and have great volatility. This volatility poses a challenge to the stable operation and optimal control of the system. At the same time, the cogeneration system involves the conversion and transmission of multiple energy forms, the system structure is complex, and the control is difficult.

[0003] In the related technologies, system optimization is mainly achieved through machine learning algorithms such as support vector machine neural network, convolutional neural network, recurrent neural network, long short-term memory neural network, and optimization algorithms such as gradient descent method, Newton method, genetic algorithm, and particle swarm optimization algorithm. However, there are many factors affecting load forecasting, which is greatly affected by user autonomy. In addition, the user's load forecasting includes power load forecasting and heat load forecasting, and their coupling characteristics are too strong. Simple machine learning methods cannot meet the accuracy requirements of load forecasting. In addition, due to the complex coupling characteristics between the various devices in the cogeneration system, the traditional optimization algorithm is time-consuming and easy to fall into local optimality, which also wastes too many computing resources. Summary of the invention

[0004] One object of the present invention is to provide a data-driven combined heat and power system optimization method, which analyzes the characteristics of electric and thermal loads and performs decoupled distributed optimization decomposition through multi-task learning and improved optimization algorithms, thereby achieving the search for a global optimal solution, improving computational efficiency, meeting the accuracy requirements of load forecasting, saving computational time and computing resources, and improving computational efficiency. Another object of the present invention is to provide a data-driven combined heat and power system optimization device. Another object of the present invention is to provide a computer-readable medium. Still another object of the present invention is to provide a computer device.

[0005] In order to achieve the above objectives, the present invention discloses, on one hand, a data-driven combined heat and power system optimization method, comprising:

[0006] Obtain energy data in combined heat and power systems;

[0007] Based on multi-task learning, energy data is predicted to generate predicted power generation and heating data;

[0008] Through the electric and thermal load characteristic curves of each heating unit, according to the power generation and heating data and energy data, the system benefit objective function and constraint conditions are constructed;

[0009] Through the improved optimization algorithm, according to the constraints, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system.

[0010] Preferably, based on multi-task learning, energy data is predicted to generate predicted power generation and heating data, including:

[0011] Normalize the energy data to generate normalized energy features;

[0012] Generate a first output result according to the normalized energy characteristics through a long short-term memory network;

[0013] The first output result is calculated through the attention mechanism to generate the second output result;

[0014] The second output result is subjected to denormalization processing to generate power generation and heating data.

[0015] Preferably, the system benefit objective function and constraint conditions are constructed through the electric and thermal load characteristic curves of each heating unit according to the power generation and heating data and energy data, including:

[0016] Construct the system benefit objective function based on power generation and heating data, energy data and preset influencing factor coefficients;

[0017] Construct constraints based on the electric and thermal load characteristic curve and power generation and heating data.

[0018] Preferably, the power generation and heating data include wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation power, and cogeneration unit power generation power;

[0019] Energy data include wind power abandonment, photovoltaic power abandonment, grid-connected electricity price, coal-fired unit power, combined heat and power unit power, combined heat and power unit thermal power, number of coal-fired units, number of combined heat and power units, carbon emissions of coal-fired units, carbon emissions of combined heat and power units, peak load time, heating power of combined heat and power units, and heating power of electric boilers;

[0020] The influencing factor coefficients include wind abandonment penalty coefficient, solar abandonment penalty coefficient, coal consumption coefficient of coal-fired units, coal consumption coefficient of combined heat and power units, carbon emission coefficient of coal-fired units, carbon emission coefficient of combined heat and power units, carbon emission quota, multi-speed peak-shaving compensation boundary of the unit, heating price, and multi-speed peak-shaving compensation quotation of the unit;

[0021] According to the power generation and heating data, energy data and the preset influencing factor coefficients, the system benefit objective function is constructed, including:

[0022] Generate wind abandonment penalty fees based on abandoned wind power, on-grid electricity price and wind abandonment penalty coefficient;

[0023] Generate abandoned solar power penalty fees based on abandoned solar power quantity and abandoned solar power penalty coefficient;

[0024] Generate coal consumption costs based on the electrical power of coal-fired units, the electrical power of combined heat and power units, the thermal power of combined heat and power units, the number of coal-fired units, the number of combined heat and power units, the coal consumption coefficient of coal-fired units and the coal consumption coefficient of combined heat and power units;

[0025] Generate carbon trading fees for heating units based on carbon emissions of coal-fired units, carbon emissions of combined heat and power units, carbon emission coefficients of coal-fired units, carbon emission coefficients of combined heat and power units and carbon emission quotas;

[0026] Generate peak load compensation according to the unit's electrical load, peak load time, the unit's multi-speed peak load compensation boundary and the unit's multi-speed peak load compensation quotation;

[0027] Generate electricity price revenue based on the on-grid electricity price, wind power generation, photovoltaic power generation, coal-fired unit power generation and cogeneration unit power generation;

[0028] Generate heating revenue based on heating price, heating power of cogeneration unit and heating power of electric boiler;

[0029] The system profit objective function is generated based on the penalty for wind curtailment, solar curtailment, coal consumption costs, carbon trading costs of heating units, peak-shaving compensation, electricity price benefits and heating benefits.

[0030] Preferably, the constraints include system electric power balance constraints, system thermal power balance constraints and operation boundary constraints of each cogeneration unit;

[0031] According to the electric and thermal load characteristic curve and power generation and heating data, the constraints are constructed, including:

[0032] Generate system power balance constraints based on wind power generation, photovoltaic power generation, coal-fired unit power generation and combined heat and power unit power generation;

[0033] Generate system thermal power balance constraints based on the heating power of the cogeneration unit and the heating power of the electric boiler;

[0034] According to the electric and thermal load characteristic curve, the operating boundary constraints of each cogeneration unit are generated.

[0035] Preferably, the improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm;

[0036] Through the improved optimization algorithm, according to the constraints, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system, including:

[0037] Through the alternating direction multiplier method, decoupling is performed according to the system benefit objective function and constraint conditions to generate multiple sub-function equations;

[0038] Through the quantum particle swarm optimization algorithm, each sub-function equation is optimized and solved to generate the optimal solution, thus completing the scheduling optimization of the cogeneration system.

[0039] The present invention also discloses a data-driven combined heat and power system optimization device, comprising:

[0040] An acquisition unit, used for acquiring energy data in the combined heat and power system;

[0041] A prediction unit, which is used to predict energy data based on multi-task learning and generate predicted power generation and heating data;

[0042] A function construction unit is used to construct a system benefit objective function and constraint conditions according to the electric and thermal load characteristic curves of each heating unit, power generation and heating data, and energy data;

[0043] The optimization solving unit is used to optimize and solve the system benefit objective function according to the constraint conditions through the improved optimization algorithm, so as to complete the scheduling optimization of the cogeneration system.

[0044] Preferably, the prediction unit is specifically used to normalize the energy data to generate normalized energy characteristics; generate a first output result based on the normalized energy characteristics through a long short-term memory network; calculate the first output result through an attention mechanism to generate a second output result; and denormalize the second output result to generate power generation and heating data.

[0045] Preferably, the function construction unit is specifically used to construct a system benefit objective function based on power generation and heating data, energy data and preset influencing factor coefficients; and to construct constraint conditions based on the electric and thermal load characteristic curve and power generation and heating data.

[0046] Preferably, the power generation and heating data include wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation power, and cogeneration unit power generation power;

[0047] Energy data include wind power abandonment, photovoltaic power abandonment, grid-connected electricity price, coal-fired unit power, combined heat and power unit power, combined heat and power unit thermal power, number of coal-fired units, number of combined heat and power units, carbon emissions of coal-fired units, carbon emissions of combined heat and power units, peak load time, heating power of combined heat and power units, and heating power of electric boilers;

[0048] The influencing factor coefficients include wind abandonment penalty coefficient, solar abandonment penalty coefficient, coal consumption coefficient of coal-fired units, coal consumption coefficient of combined heat and power units, carbon emission coefficient of coal-fired units, carbon emission coefficient of combined heat and power units, carbon emission quota, multi-speed peak-shaving compensation boundary of the unit, heating price, and multi-speed peak-shaving compensation quotation of the unit;

[0049] The function construction unit is specifically used to generate wind abandonment penalty costs according to the abandoned wind power, on-grid electricity price and wind abandonment penalty coefficient; generate solar abandonment penalty costs according to the abandoned photovoltaic power and the abandoned photovoltaic penalty coefficient; generate coal consumption costs according to the electric power of coal-fired units, the electric power of cogeneration units, the thermal power of cogeneration units, the number of coal-fired units, the number of cogeneration units, the coal consumption coefficient of coal-fired units and the coal consumption coefficient of cogeneration units; generate heating unit carbon emissions according to the carbon emissions of coal-fired units, the carbon emissions of cogeneration units, the carbon emission coefficient of coal-fired units, the carbon emission coefficient of cogeneration units and carbon emission quotas. Generate carbon trading fees for the unit; generate peak-shaving compensation according to the unit's power load, peak-shaving time, the unit's multi-speed peak-shaving compensation boundary and the unit's multi-speed peak-shaving compensation quotation; generate electricity price benefits according to the on-grid electricity price, wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; generate heating benefits according to the heating price, cogeneration unit heating power and electric boiler heating power; generate system benefit objective function according to the wind abandonment penalty fee, solar abandonment penalty fee, coal consumption fee, heating unit carbon trading fees, peak-shaving compensation, electricity price benefits and heating benefits.

[0050] Preferably, the constraints include system electric power balance constraints, system thermal power balance constraints and operation boundary constraints of each cogeneration unit;

[0051] The function construction unit is specifically used to generate the system electric power balance constraints according to the wind power generation power, photovoltaic power generation power, coal-fired unit power generation and cogeneration unit power generation; generate the system thermal power balance constraints according to the cogeneration unit heating power and electric boiler heating power; generate the operating boundary constraints of each cogeneration unit according to the electric and thermal load characteristic curve.

[0052] Preferably, the improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm;

[0053] The optimization solving unit is specifically used to generate multiple sub-function equations by decoupling the system benefit objective function and constraint conditions through the alternating direction multiplier method; each sub-function equation is optimized and solved through the quantum particle swarm optimization algorithm to generate the optimal solution and complete the scheduling optimization of the cogeneration system.

[0054] The present invention also discloses a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the method described above is implemented.

[0055] The present invention also discloses a computer device, including a memory and a processor, wherein the memory is used to store information including program instructions, the processor is used to control the execution of program instructions, and the processor implements the above method when executing the program.

[0056] The present invention also discloses a computer program product, including a computer program / instruction, and the method described above is implemented when the computer program / instruction is executed by a processor.

[0057] The present invention obtains energy data in a cogeneration system; based on multi-task learning, the energy data is predicted to generate predicted power generation and heating data; through the electric and thermal load characteristic curves of each heating unit, the system benefit objective function and constraint conditions are constructed according to the power generation and heating data and the energy data; through an improved optimization algorithm, the system benefit objective function is optimized and solved according to the constraint conditions, and the scheduling optimization of the cogeneration system is completed; through multi-task learning and an improved optimization algorithm, the electric and thermal load characteristics are analyzed and decoupled distributed optimization decomposition is performed, thereby achieving the search for a global optimal solution, improving computing efficiency, meeting the accuracy requirements of load prediction, saving computing time and computing resources, and improving computing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0059] Figure 1 A flow chart of a data-driven combined heat and power system optimization method provided in an embodiment of the present invention;

[0060] Figure 2 A flow chart of another data-driven combined heat and power system optimization method provided in an embodiment of the present invention;

[0061] Figure 3 A schematic diagram of the structure of an LSTM network provided in an embodiment of the present invention;

[0062] Figure 4 A curve diagram of the heat supply electric-thermal characteristics of a condensing unit provided in an embodiment of the present invention;

[0063] Figure 5 A curve diagram of the electric-thermal characteristics of a high back pressure heating system provided by an embodiment of the present invention;

[0064] Figure 6A curve diagram of the electric-thermal characteristics of a low-pressure cylinder removal heating supply provided by an embodiment of the present invention;

[0065] Figure 7 A curve diagram of the electric-thermal characteristics of an electric boiler-condensation heating system provided in an embodiment of the present invention;

[0066] Figure 8 A schematic diagram of the structure of a data-driven cogeneration system optimization device provided by an embodiment of the present invention;

[0067] Fig. 9 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] It should be noted that the data-driven combined heat and power system optimization method and device disclosed in the present application can be used in the field of artificial intelligence technology, and can also be used in any field outside the field of artificial intelligence technology. The application field of the data-driven combined heat and power system optimization method and device disclosed in the present application is not limited.

[0070] In order to facilitate the understanding of the technical solution provided by this application, the relevant contents of the technical solution of this application are first explained below. The cogeneration system is a complex, highly coupled heating and power supply system, which is mainly used to meet the city's electricity and heat needs. Its electricity sources mainly include wind power generation, photovoltaic power generation, coal-fired unit power generation, and cogeneration unit power generation. Its heat sources mainly include cogeneration unit heating and electric boiler heating. The optimization decision of the cogeneration system mainly involves the prediction and scheduling of wind and photovoltaic power generation equipment, the prediction of user-side electric and thermal loads, the combination mode of cogeneration unit heating and power supply, the emission characteristics of the cogeneration unit, the operating cost and benefits of the overall system, etc.

[0071] This application takes power supply capacity and heating capacity as the starting point, and predicts wind power generation, photovoltaic power generation, user electric load and user thermal load through multi-task learning to reduce the impact of volatility; by analyzing the electric and thermal load characteristics of the cogeneration unit, the feasible domain of the electric and thermal load of various heating methods is obtained to provide guidance for the selection of heating methods; finally, the complex objective function is decomposed into relatively independent sub-problems through distributed optimization, and an optimization algorithm is introduced to achieve optimization and improve computing efficiency.

[0072] The following takes a data-driven combined heat and power system optimization device as an example of an implementation subject to illustrate the implementation process of the data-driven combined heat and power system optimization method provided by an embodiment of the present invention. It can be understood that the implementation subject of the data-driven combined heat and power system optimization method provided by an embodiment of the present invention includes but is not limited to a data-driven combined heat and power system optimization device.

[0073] Figure 1 A flow chart of a data-driven combined heat and power system optimization method provided in an embodiment of the present invention, such as Figure 1 As shown, the method includes:

[0074] Step 101: Obtain energy data in the cogeneration system.

[0075] Step 102: Based on multi-task learning (MTL), energy data is predicted to generate predicted power generation and heating data.

[0076] In the embodiment of the present invention, the underlying network of MTL uses a long short-term memory (LSTM) network. The LSTM network is improved from a recurrent neural network (RNN) and has the ability of RNN to mine time series data and remember part of the time series.

[0077] Step 103: construct the system benefit objective function and constraint conditions through the electric and thermal load characteristic curves of each heating unit, according to the power generation and heating data and energy data.

[0078] Step 104: By using the improved optimization algorithm and according to the constraint conditions, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system.

[0079] In the embodiment of the present invention, the improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm.

[0080] In the technical solution provided by the embodiment of the present invention, energy data in the cogeneration system is obtained; based on multi-task learning, the energy data is predicted to generate predicted power generation and heating data; through the electric and thermal load characteristic curves of each heating unit, the system benefit objective function and constraints are constructed according to the power generation and heating data and the energy data; through the improved optimization algorithm, according to the constraints, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system, and the electric and thermal load characteristics are analyzed through multi-task learning and the improved optimization algorithm, and the decoupled distributed optimization decomposition is performed, so as to achieve the search for the global optimal solution, improve the computing efficiency, meet the accuracy requirements of load prediction, save computing time and computing resources, and improve computing efficiency.

[0081] Figure 2 A flow chart of another data-driven combined heat and power system optimization method provided in an embodiment of the present invention is as follows: Figure 2 As shown, the method includes:

[0082] Step 201: Obtain energy data in the cogeneration system.

[0083] In the embodiment of the present invention, each step is performed by a data-driven cogeneration system optimization device.

[0084] In an embodiment of the present invention, energy data is relevant data of wind energy, photovoltaic energy, user electrical load and user thermal load, and the energy data includes but is not limited to wind speed, wind power, temperature, abandoned wind power, abandoned photovoltaic power, on-grid electricity price, coal-fired unit power, cogeneration unit power, cogeneration unit thermal power, number of coal-fired units, number of cogeneration units, carbon emissions of coal-fired units, carbon emissions of cogeneration units, peak shaving time, heating power of cogeneration units, and heating power of electric boilers.

[0085] Step 202: normalize the energy data to generate normalized energy characteristics.

[0086] In the embodiment of the present invention, taking wind power prediction as an example, energy data includes but is not limited to wind speed, wind power, and temperature; wind speed, wind power, and temperature are normalized to generate normalized wind speed characteristics, wind power characteristics, and temperature characteristics.

[0087] Step 203: Generate a first output result according to the normalized energy characteristics through a long short-term memory (LSTM) network.

[0088] Figure 3 A schematic diagram of the structure of an LSTM network provided in an embodiment of the present invention is shown in FIG. Figure 3 As shown, x t is the normalized energy feature of the input, h t-1 is the output of LSTM at the previous moment, h t is the first output result of the current unit, c t is the current state of the storage unit, c t-1 is the state of the storage unit at the last moment, f t is the forget gate, σ is the sigmoid function, i t is the input gate, is the new status information, o t is the output gate.

[0089] Forget Gate t Determines how much information of the previous unit the current unit retains. The calculation formula is:

[0090] f t =σ(w f ·[h t-1 ,x t ]+b f )

[0091] Among them, f t is the forget gate, w f is the forget gate weight matrix, b f is the forget gate bias, σ is the sigmoid function, h t-1 is the output of LSTM at the previous moment, x t is the normalized energy feature of the input.

[0092] Input Gate i t Update the state value of the input data to the storage unit, the calculation formula is:

[0093] i t =σ(w i ·[h t-1 ,x t ]+b i )

[0094] Among them, i t is the input gate, w i is the input gate i t The weight matrix, b i is the input gate i t Bias, σ is the sigmoid function, h t-1 is the output of LSTM at the previous moment, x t is the normalized energy feature of the input.

[0095] New status information The calculation formula is:

[0096]

[0097] in, is the new state information, w c is the new state information weight matrix, b c is the new state information bias, h t-1 is the output of LSTM at the previous moment, x t is the normalized energy feature of the input.

[0098] The current state c of the storage unit is obtained from the above formula t The calculation formula is:

[0099]

[0100] Among them, ct is the current state of the storage unit, c t-1 is the state of the storage unit at the last moment, f t For the forget gate, i t is the input gate, New status information.

[0101] Output gate o t The output value of the control storage unit status value is calculated as follows:

[0102] o t =σ(w o ·[h t-1 ,x t ]+b o )

[0103] Among them, t is the output gate, w o is the output gate weight matrix, b o is the output gate bias, σ is the sigmoid function, h t-1 is the output of LSTM at the previous moment, x t is the normalized energy feature of the input.

[0104] The first output result h of the current unit t The expression is:

[0105] h t =o t ·tan(c t )

[0106] Among them, h t is the first output result of the current unit, o t is the output gate, c t The current state of the storage unit.

[0107] In the embodiment of the present invention, an input gate, a forget gate and an output gate are introduced on the basis of the LSTM network, and different parameters are set through a gating mechanism to selectively forget historical data and retain useful information for the next neuron, thereby solving the problems of gradient vanishing and gradient exploding during back propagation and realizing the control of long-term memory of data.

[0108] Step 204: Calculate the first output result through the attention mechanism to generate a second output result.

[0109] In the embodiment of the present invention, considering the individual characteristics of each input information such as wind power output and the impact of each input energy characteristic on the result, an attention mechanism is also introduced in the LSTM network part of the underlying network. The scoring function of the attention mechanism is expressed as follows:

[0110] e t =u·tanh(w·h t +b)

[0111]

[0112]

[0113] Among them, e t Represents the first output result h of the LSTM network layer at time t t The determined attention probability distribution value; u and w are weight coefficients; h t is the first output result of the current unit; b is the bias coefficient; s t is the second output result of the attention layer at time t; α i is the intermediate parameter.

[0114] Step 205: Denormalize the second output result to generate power generation and heating data.

[0115] In the embodiment of the present invention, the final power generation and heating data are obtained by anti-normalization processing to maintain data consistency. Taking wind power generation power prediction as an example, the power generation and heating data are the wind power generation power prediction results.

[0116] In view of the autonomy of user electricity consumption and the impact of heating load demand caused by environmental changes, wind, light and user load forecasting based on multi-task learning is carried out. Multi-task learning (MTL) is a machine learning method that utilizes the correlation between multiple related tasks such as wind power generation, photovoltaic power generation, user electricity load and user heat load, and jointly trains a model or part of the model through information sharing to improve the generalization ability, thereby improving the learning effect of each task. A hard sharing architecture is adopted between tasks. Different tasks share the underlying network layer, and an attention mechanism is added to the underlying network to dynamically select and weight different parts of the underlying network output. At the same time, each exclusive network or simple output layer is set at the top level to reduce the risk of overfitting and improve the prediction ability.

[0117] The embodiments of the present invention can achieve accurate prediction of wind power output and photovoltaic output and optimize power dispatch. Wind power generation and photovoltaic power generation are important power supply equipment in the combined heat and power system, and are also important clean energy in the energy transformation process. Accurate prediction of wind power output and photovoltaic output helps to optimize power dispatch plans, maximize the use of wind power and photovoltaic power, reduce dependence on fossil fuels, and reduce power generation costs. At the same time, accurate prediction of wind power output and photovoltaic output can also reduce the phenomenon of wind and solar abandonment and improve the utilization rate of renewable energy.

[0118] In the embodiment of the present invention, the user's electricity consumption has a large degree of autonomy, which poses a great challenge to the dispatch of electricity. At the same time, environmental changes also affect the demand for heating load, so accurate load forecasting is of great significance. Accurate load forecasting can achieve the balance of power supply and demand in different time periods of the power system, and prevent the instability of the power grid caused by power shortage or surplus; at the same time, accurate load forecasting can help the dispatch center formulate more economical power generation and heating plans, and reduce the cost of power generation and heating.

[0119] Step 206: Construct a system benefit objective function based on the power generation and heating data, energy data and preset influencing factor coefficients.

[0120] In the embodiment of the present invention, the power generation and heating data include but are not limited to wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation power, and cogeneration unit power generation power.

[0121] In an embodiment of the present invention, the influencing factor coefficients include but are not limited to a wind abandonment penalty coefficient, a solar abandonment penalty coefficient, a coal consumption coefficient of a coal-fired unit, a coal consumption coefficient of a cogeneration unit, a carbon emission coefficient of a coal-fired unit, a carbon emission coefficient of a cogeneration unit, a carbon emission quota, a multi-speed peak-shaving compensation boundary of the unit, a heating price, and a multi-speed peak-shaving compensation quota of the unit.

[0122] It is worth noting that the power generation and heating data can be obtained based on predictions, the energy data can be obtained through actual acquisition, and the influencing factor coefficients can be set according to actual needs.

[0123] In an embodiment of the present invention, the maximum benefit of the system is taken as the objective function, which has multiple influencing factors, including but not limited to wind abandonment penalty fees, solar abandonment penalty fees, coal consumption costs of coal-fired units, coal consumption costs of heating units, carbon trading costs or income of heating units, peak-shaving compensation, electricity price benefits and heating benefits.

[0124] In the embodiment of the present invention, step 206 specifically includes:

[0125] Step 2061: Generate wind abandonment penalty fees based on the amount of abandoned wind power, the on-grid electricity price and the wind abandonment penalty coefficient.

[0126] Specifically, through The wind power curtailment, grid-connected electricity price and wind curtailment penalty coefficient are calculated to generate the wind curtailment penalty fee. Among them, M1 is the wind curtailment penalty fee, d1 is the grid-connected electricity price, ρ w is the wind abandonment penalty coefficient, Δp w (i,t) is the amount of wind power abandoned.

[0127] As an optional option, the wind curtailment penalty coefficient is between 0.1-0.5 yuan / KWh; the wind curtailment volume is the actual difference; and the on-grid electricity price is between 0.7-0.85 yuan / KWh.

[0128] Step 2062: Generate a penalty fee for abandoned solar power according to the amount of abandoned solar power and the penalty coefficient for abandoned solar power.

[0129] Specifically, through The abandoned photovoltaic power and the abandoned photovoltaic penalty coefficient are calculated to generate the abandoned photovoltaic penalty fee. Among them, M2 is the abandoned photovoltaic penalty fee, d1 is the on-grid electricity price, ρ s is the light abandonment penalty coefficient, Δp s (i,t) is the amount of wind power abandoned.

[0130] Step 2063: Generate coal consumption costs according to the electric power of the coal-fired units, the electric power of the cogeneration units, the thermal power of the cogeneration units, the number of coal-fired units, the number of cogeneration units, the coal consumption coefficient of the coal-fired units and the coal consumption coefficient of the cogeneration units.

[0131] Specifically, through The coal consumption cost is generated by calculating the electric power of coal-fired units, the electric power of combined heat and power units, the thermal power of combined heat and power units, the number of coal-fired units, the number of combined heat and power units, the coal consumption coefficient of coal-fired units and the coal consumption coefficient of combined heat and power units. Among them, M3 is the coal consumption cost, is the coal consumption cost of coal-fired units, is the coal consumption cost of the cogeneration unit, N g is the number of coal-fired units, N chp is the number of CHP units, C g is the coal consumption coefficient of the coal-fired unit, C chp is the coal consumption coefficient of the cogeneration unit, p g is the electrical power of the coal-fired unit, p chp is the electrical power of the combined heat and power unit, h chp is the thermal power of the cogeneration unit.

[0132] Step 2064: Generate the carbon trading fee of the heating unit according to the carbon emissions of the coal-fired unit, the carbon emissions of the cogeneration unit, the carbon emission coefficient of the coal-fired unit, the carbon emission coefficient of the cogeneration unit and the carbon emission quota.

[0133] Specifically, through The carbon emissions of coal-fired units, carbon emissions of combined heat and power units, carbon emission coefficients of coal-fired units, carbon emission coefficients of combined heat and power units and carbon emission quotas are calculated to generate the carbon trading costs of heating units. Among them, M4 is the carbon trading cost of heating units, N g is the number of coal-fired units, N chp is the number of CHP units, C' g is the carbon emission coefficient of coal-fired units, C' chp is the carbon emission coefficient of the cogeneration unit, A g is the carbon emission of coal-fired units, A chpare their respective carbon emissions, and A is the carbon emission quota.

[0134] Step 2065: Generate peak-shaving compensation according to the unit's electrical load, peak-shaving time, the unit's multi-speed peak-shaving compensation boundary, and the unit's multi-speed peak-shaving compensation quotation.

[0135] In the embodiment of the present invention, taking the three-level peak load compensation boundary and the three-level peak load compensation quotation as an example, specifically, by The peak load, peak load time, peak load compensation boundary and peak load compensation quotation of the unit are calculated to generate peak load compensation. Among them, M5 is peak load compensation, P is unit load, t is peak load time, C p1 is the first-tier peak load compensation quotation, C p2 The second-tier peak load compensation quotation, C p3 The three-level peak load compensation quotation, P 01 is the first-level peak-shaving compensation boundary, P 02 is the second-gear peak-shaving compensation boundary, P 03 It is the three-speed peak regulation compensation boundary.

[0136] Step 2066: Generate electricity price revenue based on the on-grid electricity price, wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power.

[0137] Specifically, through The grid-connected electricity price, wind power generation, photovoltaic power generation, coal-fired unit power generation and combined heat and power unit power generation are calculated to generate electricity price benefits. Among them, M6 is the electricity price benefit, p w is the wind power generation power, p s is the photovoltaic power generation power, p g is the power generation capacity of the coal-fired unit, p chp is the power generation capacity of the cogeneration unit, and d1 is the on-grid electricity price.

[0138] Step 2067: Generate heating revenue according to the heating price, the heating power of the cogeneration unit and the heating power of the electric boiler.

[0139] Specifically, through The heating price, heating power of the cogeneration unit and heating power of the electric boiler are calculated to generate the heating revenue. Among them, M7 is the heating revenue, d2 is the heating price, h chp is the heating power of the combined heat and power unit, h c Provides heating power for electric boilers.

[0140] Step 2068: Generate a system revenue objective function based on the wind abandonment penalty fee, solar abandonment penalty fee, coal consumption fee, carbon trading fee of heating units, peak load compensation, electricity price revenue and heating revenue.

[0141] In the embodiment of the present invention, the penalty for wind abandonment, the penalty for solar abandonment, the coal consumption cost, the carbon trading cost of the heating unit, the peak load compensation, the electricity price income and the heating income are added together to generate the system income objective function, which is as follows:

[0142] M=M1+M2+M3+M4+M5+M6+M7

[0143] Among them, M is the total system revenue, M1 is the penalty for wind abandonment, M2 is the penalty for solar abandonment, M3 is the coal consumption cost, M4 is the carbon trading fee of the heating unit, M5 is the peak-shaving compensation, M6 is the electricity price revenue, and M7 is the heating revenue.

[0144] In the embodiment of the present invention, the main heating equipment of the cogeneration system includes coal-fired heating units, cogeneration units and electric boilers. Among them, the coal-fired heating unit is a traditional heating unit that mainly relies on the combustion of fossil fuels; the cogeneration unit is a high-efficiency energy system that can produce electricity and heat energy at the same time. The main heating methods include extraction condensation heating, high back pressure heating, high and low bypass heating, and low pressure cylinder removal heating. Each heating method has a large electric-thermal coupling characteristic; electric boiler heating is to heat by converting the power generated by the cogeneration unit or other equipment into electric heat. By studying the different equipment combinations between the cogeneration systems, the power supply and heating requirements can be maximized and the energy utilization efficiency can be improved.

[0145] Step 207: Construct constraint conditions according to the electric and thermal load characteristic curve and the power generation and heating data.

[0146] In the embodiment of the present invention, the electric-heat load characteristic curve is a pre-generated feasible domain of electric-heat load for different heating modes of the cogeneration unit. In view of the electric-heat coupling characteristics of the cogeneration unit, the coefficients are first identified by the traditional least square method, and then the feasible domain of electric-heat load for different heating modes is obtained by curve fitting, and on this basis, the feasible domain of electric-heat load for different heating modes of the cogeneration unit is obtained. In addition, since coal consumption, CO2 emissions, etc. are also important parts of the objective function, it is also necessary to analyze the coal consumption characteristics and emission characteristics under different heating modes.

[0147] The heating methods of cogeneration units mainly include: extraction condensing unit heating, high back pressure heating, low pressure cylinder removal heating, and electric boiler auxiliary heating. Extraction condensing unit heating refers to extracting part of the steam in the medium and low pressure connecting pipes and flowing it into the heat network heater for heating. High back pressure heating refers to increasing the back pressure of the steam turbine, and by increasing the exhaust temperature of the low pressure cylinder, the heat network water absorbs the exhaust heat and heats up for heating. Low pressure cylinder removal heating is also called low pressure cylinder zero output heating, which means that most of the exhaust steam of the medium pressure cylinder of the steam turbine is used for external heating under low back pressure conditions, leaving only part of the steam to enter the low pressure cylinder to cool the low pressure rotor. Electric boiler heating is a kind of integrated heat source heating, which uses electric boiler heating as a compensating heat source to increase the heating capacity of the unit. Its essence is to convert the generated power of the unit into heat energy by consuming it. In actual applications, electric boiler heating is not used as a separate heating method, but is often used in combination with extraction condensing heating units to form an electric boiler-extraction condensing heating method.

[0148] Figure 4 A curve diagram of the heat supply electric-thermal characteristics of a condensing unit provided in an embodiment of the present invention, such as Figure 4 As shown, the horizontal axis is the heating load, in MW, and the vertical axis is the power generation load, in MW. In the curve diagram, line segment AB is the maximum load curve of the boiler, line segment CD is the minimum load curve of the boiler, BC is the minimum condensing condition curve, and AD is the pure condensing condition curve.

[0149] Figure 5 A curve diagram of a high back pressure heating electric-thermal characteristic provided by an embodiment of the present invention, such as Figure 5 As shown, the horizontal axis is the heating load, in MW, and the vertical axis is the power generation load, in MW. Point B in the curve is the minimum heating load point, and point A is the maximum heating load point.

[0150] Figure 6 A curve diagram of the electric-thermal characteristics of a low-pressure cylinder removal heating supply provided by an embodiment of the present invention, such as Figure 6 As shown, the horizontal axis is the heating load, in MW, and the vertical axis is the power generation load, in MW. In the curve diagram, segment B'C' is the minimum condensing gas condition curve after the low-pressure cylinder is removed, segment AB is the maximum load curve of the boiler, segment CD is the minimum load curve of the boiler, BC is the minimum condensing gas condition curve, AD is the pure condensing condition curve, segment BB' and segment CC' both indicate the mutation process, and the slopes of segment BC and segment B'C' are the same.

[0151] Figure 7 A curve diagram of the electric-thermal characteristics of an electric boiler-condensation heating system provided in an embodiment of the present invention, such as Figure 7As shown, the horizontal axis is the heating load, in MW, and the vertical axis is the power generation load, in MW. In the curve, segment B'C' is the minimum condensing gas condition curve after adding electric boiler coupled heating, C'D' is the boiler minimum load curve after adding electric boiler coupled heating, segment AB is the boiler maximum load curve, segment CD is the boiler minimum load curve, BC is the minimum condensing gas condition curve, AD is the pure condensing condition curve, segment BB' and segment DD' both indicate the mutation process, and the slopes of segment BC and segment B'C' are the same.

[0152] In the embodiment of the present invention, the constraint conditions include, but are not limited to, system electric power balance constraints, system thermal power balance constraints, and operation boundary constraints of each cogeneration unit.

[0153] In the embodiment of the present invention, step 207 specifically includes:

[0154] Step 2071: Generate system power balance constraints based on wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power.

[0155] Specifically, through The wind power generation, photovoltaic power generation, coal-fired unit power generation and combined heat and power unit power generation are calculated to generate the system power balance constraints. l (t) is the system electrical power, p w (i, t) is the wind power generation power, p s (i, t) is the photovoltaic power generation power, p g (i, t) is the power generation capacity of the coal-fired unit, p chg (i, t) is the power generation of the cogeneration unit, N g is the number of coal-fired units, N chp is the number of CHP units, N w is the number of wind turbines, N s is the number of photovoltaic units.

[0156] Step 2072: Generate a system thermal power balance constraint based on the heating power of the cogeneration unit and the heating power of the electric boiler.

[0157] Specifically, through The heating power of the combined heat and power unit and the heating power of the electric boiler are calculated to generate the system thermal power balance constraints. l (t) is the system thermal power, h chp (i,t) is the heating power of the cogeneration unit, h c (i,t) is the heating power of the electric boiler, N chp is the number of cogeneration units.

[0158] Step 2073: Generate operating boundary constraints for each cogeneration unit based on the electric and thermal load characteristic curve.

[0159] Specifically, according to the operation mode of each cogeneration unit, the electric and thermal load characteristic curves corresponding to each operation mode are calculated respectively, and the operation boundary constraints of each cogeneration unit are obtained as follows:

[0160]

[0161] Among them, y EH is the operating boundary of each cogeneration unit, y C is the minimum power generation load point, x C is the minimum heating load point, y AB is the maximum boiler load value of each cogeneration unit, y BC is the minimum condensing load value, x B The point of maximum heating load.

[0162] It is worth mentioning that the operating modes of each cogeneration unit include: extraction condensing unit heating and electricity-thermal characteristics, high back pressure heating and electricity-thermal characteristics, low-pressure cylinder cut-off heating and electricity-thermal characteristics, and electric boiler-extraction condensing heating and electricity-thermal characteristics.

[0163] Step 208: Decoupling is performed according to the system benefit objective function and constraint conditions by using the Alternating Direction Method of Multipliers (ADMM) to generate multiple sub-function equations.

[0164] In the embodiment of the present invention, the overall benefit objective function of the complex and coupled cogeneration system is decoupled by the ADMM algorithm and decomposed into multiple relatively independent sub-problems. The steps of the decoupling part of ADMM are: constructing an augmented Lagrangian function; decomposing the original problem into several sub-problems; and updating the Lagrangian multiplier through the dual gradient. The principle of ADMM is as follows:

[0165] Assume that the optimization objective function is min x,z f(x)+g(z), with the constraint stAx+Bz=b, construct the expression of the augmented Lagrangian function:

[0166] L(x,z,y)=f(x)+g(z)+y T (Ax+Bz-b)+(ρ / 2)||Ax+Bz-b||

[0167] On the basis of the augmented Lagrangian function, the joint optimization of the main variables (x, z) is changed to a separate alternating iteration, and the iterative equation is:

[0168] xk+1 =argmin x L β (x,z k ,y k )

[0169] z k+1 =argmin z L β (x k+1 ,z,y k )

[0170] y k+1 =y k -β(Ax k+1 +Bz k+1 -b)

[0171] Through the above iterative equations, the joint optimization problem in augmented Lagrangian can be completely divided into multiple relatively independent sub-problems.

[0172] Specifically, based on the principle of the above-mentioned ADMM, the system revenue objective function M, the system electric power balance constraint, the system thermal power balance constraint and the operating boundary constraints of each cogeneration unit are decoupled, and the expression of the augmented Lagrangian function is constructed. On the basis of the augmented Lagrangian function, the joint optimization of the main variables is changed to a separate alternating iteration, and the multiple iterative equations obtained are multiple sub-function equations.

[0173] Step 209: Optimize and solve each sub-function equation through the Quantum Particle Swarm Optimization (QPSO) algorithm to generate an optimal solution, thereby completing the scheduling optimization of the cogeneration system.

[0174] The QPSO algorithm is a particle swarm optimization algorithm that is improved by combining the principles of quantum mechanics. It improves the algorithm's global search capability and convergence speed through the probability characteristics of quantum states. The QPSO algorithm combines particle swarm optimization with the wave function and probability density function in quantum mechanics, and has stronger global search capabilities and the ability to adapt to complex optimization problems.

[0175] In the embodiment of the present invention, the QPSO algorithm is used to optimize the optimal power generation and heating capacity of each subsystem such as each coal-fired unit and cogeneration unit according to each sub-function equation, that is, each sub-problem is optimized and solved separately.

[0176] The basic idea of ​​the classic particle swarm optimization algorithm is to get the optimal solution by moving a group of particles in the search space, and each particle continuously adjusts its position according to its own experience and the optimal solution in the group. The classic PSO mainly includes speed update and position update, and its formula is:

[0177]

[0178] x i (t+1)=x i (t)+v i (t+1)

[0179] Among them, v i (t+1) is the velocity of the ith particle at time t+1, v i (t) is the velocity of the ith particle at time t, x i (t) is the position of the ith particle at time t, P i best is the best position found by the i-th particle itself, g best is the best position found in the population, ω is the inertia weight, c1 and c2 are learning factors, and r1 and r2 are random numbers.

[0180] Compared with the classical PSO, the QPSO algorithm does not directly update the velocity and position of the particle, but describes the movement of the particle through the quantum state. The formula for quantum position update is:

[0181] x i (t+1)=p i (t)±β·|p i (t)-x i (t)|·ln(1 / u)

[0182] Among them, x i (t+1) is the position of the i-th particle at time t+1; x i (t) is the position of the ith particle at time t; p i (t) is the historical optimal position of the ith particle; β is a control parameter, usually between [0,1]; u is a random number between [0,1].

[0183] Specifically, first, initialize the velocity and position of the particles and the individual optimal position p of each particle best , and determine the global optimal position g best ; Then calculate the fitness value of each particle, that is, the objective function value, and perform quantum update of the individual optimal position p best and the global optimal position g best ; Then calculate the new fitness value to determine whether it meets the fitness requirement. If it does not meet the conditions, repeatedly update the particle position until the conditions are met.

[0184] The present invention proposes a decoupled distributed optimization algorithm for algorithm optimization of the system benefit objective function. The decoupled distributed optimization algorithm is a type of method used to solve large-scale complex optimization problems with high coupling. The problem is decomposed into several smaller, relatively independent sub-problems, and the problems are solved on different computing nodes to improve the efficiency and scalability of the solution. Since the objective function contains a variety of complex coupling variables, this scheme uses ADMM for decomposition processing. ADMM is a classic decoupled distributed optimization algorithm. It introduces Lagrangian multipliers and augmented Lagrangian functions to decompose the problem into several sub-problems that are solved alternately, and the multipliers are updated in each iteration. The present invention obtains several relatively independent sub-problems through ADMM decomposition, and then uses the QPSO algorithm for optimization processing.

[0185] The combined heat and power system has a strong coupling property. It is impossible to simply rely on manual operation experience to establish the optimal combined heat and power mode. The commonly used intelligent optimization algorithm is easy to fall into the local optimal solution and cannot obtain reliable optimization results. By decoupling the distributed optimization algorithm to decouple the complex problem, the optimization algorithm is optimized on each sub-problem to obtain the optimal combined heat and power heating mode and improve the economic benefits.

[0186] By introducing a multi-task learning method, the present invention can achieve predictions about wind power output, photovoltaic output and user load with higher efficiency and higher accuracy, reduce scheduling difficulty, reduce wind and solar power abandonment, improve renewable energy utilization, and contribute to efficient energy utilization. By analyzing the electric-thermal coupling characteristics of the cogeneration unit, the feasible domain of electric-thermal loads with different heating modes and combinations is determined, providing guidance for the load distribution of the cogeneration unit. By applying the decoupled distributed optimization algorithm, the highly coupled large-scale complex optimization problem is decomposed, reducing the difficulty of solving the problem and improving the optimization efficiency. The optimization algorithm takes into account multiple factors, realizes the maximum benefit of the cogeneration system, and improves production efficiency.

[0187] It is worth noting that the acquisition, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of laws and regulations. The user information in the embodiments of this application is obtained through legal and compliant channels, and the acquisition, storage, use, and processing of user information are authorized and agreed by the customer.

[0188] It is worth noting that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0189] It is worth noting that the technical solution provided in this application provides users with corresponding operation entrances for them to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered.

[0190] In the technical solution of the data-driven combined heat and power system optimization method provided by the embodiment of the present invention, energy data in the combined heat and power system is obtained; based on multi-task learning, the energy data is predicted to generate predicted power generation and heating data; through the electric and thermal load characteristic curves of each heating unit, the system benefit objective function and constraints are constructed according to the power generation and heating data and the energy data; through the improved optimization algorithm, according to the constraints, the system benefit objective function is optimized and solved to complete the scheduling optimization of the combined heat and power system, and the electric and thermal load characteristics are analyzed through multi-task learning and the improved optimization algorithm, and the decoupled distributed optimization decomposition is performed, so as to achieve the search for the global optimal solution, improve the computing efficiency, meet the accuracy requirements of load prediction, save computing time and computing resources, and improve computing efficiency.

[0191] Figure 8 A schematic diagram of the structure of a data-driven cogeneration system optimization device provided in an embodiment of the present invention, wherein the device is used to execute the data-driven cogeneration system optimization method, such as Figure 8 As shown, the device includes: an acquisition unit 11, a prediction unit 12, a function construction unit 13 and an optimization solution unit 14.

[0192] The acquisition unit 11 is used to acquire energy data in the combined heat and power system.

[0193] The prediction unit 12 is used to predict energy data based on multi-task learning and generate predicted power generation and heating data.

[0194] The function construction unit 13 is used to construct the system benefit objective function and constraint conditions according to the electric and thermal load characteristic curves of each heating unit, the power generation and heating data and the energy data.

[0195] The optimization solving unit 14 is used to optimize and solve the system benefit objective function according to the constraint conditions through an improved optimization algorithm, so as to complete the scheduling optimization of the cogeneration system.

[0196] In an embodiment of the present invention, the prediction unit 12 is specifically used to normalize the energy data to generate normalized energy characteristics; generate a first output result based on the normalized energy characteristics through a long short-term memory network; calculate the first output result through an attention mechanism to generate a second output result; and denormalize the second output result to generate power generation and heating data.

[0197] In the embodiment of the present invention, the function construction unit 13 is specifically used to construct a system benefit objective function according to power generation and heating data, energy data and preset influencing factor coefficients; and to construct constraint conditions according to the electric and thermal load characteristic curve and power generation and heating data.

[0198] In the embodiment of the present invention, the power generation and heating data include wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation power, and cogeneration unit power generation power; the energy data include abandoned wind power, abandoned photovoltaic power, grid-connected electricity price, coal-fired unit power, cogeneration unit power, cogeneration unit thermal power, number of coal-fired units, number of cogeneration units, carbon emissions of coal-fired units, carbon emissions of cogeneration units, peak load time, cogeneration unit heating power, and electric boiler heating power; the influencing factor coefficients include wind abandonment penalty coefficient, solar abandonment penalty coefficient, coal consumption coefficient of coal-fired units, coal consumption coefficient of cogeneration units, carbon emission coefficient of coal-fired units, carbon emission coefficient of cogeneration units, carbon emission quota, multi-speed peak-shaving compensation boundary of the unit, heating price, and multi-speed peak-shaving compensation quota of the unit; the function construction unit 13 is specifically used to generate wind abandonment penalty fee according to the wind abandonment power, on-grid electricity price and wind abandonment penalty coefficient; generate solar abandonment penalty fee according to the solar abandonment power and solar abandonment penalty coefficient. Penalty fees; Generate coal consumption fees based on the electric power of coal-fired units, the electric power of cogeneration units, the thermal power of cogeneration units, the number of coal-fired units, the number of cogeneration units, the coal consumption coefficient of coal-fired units and the coal consumption coefficient of cogeneration units; Generate carbon trading fees for heating units based on the carbon emissions of coal-fired units, the carbon emissions of cogeneration units, the carbon emission coefficient of coal-fired units, the carbon emission coefficient of cogeneration units and carbon emission quotas; Generate carbon trading fees for heating units based on the unit's electrical load, peak-shaving time, and the unit's multi-speed peak-shaving compensation Generate peak-shaving compensation based on the multi-level peak-shaving compensation quotations of boundaries and units; generate electricity price benefits based on the on-grid electricity price, wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; generate heating benefits based on the heating price, cogeneration unit heating power and electric boiler heating power; generate the system benefit objective function based on the wind abandonment penalty fee, solar abandonment penalty fee, coal consumption fee, carbon trading fee of heating units, peak-shaving compensation, electricity price benefits and heating benefits.

[0199] In the embodiment of the present invention, the constraint conditions include system electric power balance constraint, system thermal power balance constraint and operation boundary constraint of each cogeneration unit; the function construction unit 13 is specifically used to generate system electric power balance constraint according to wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; generate system thermal power balance constraint according to the heating power of the cogeneration unit and the heating power of the electric boiler; generate operation boundary constraint of each cogeneration unit according to the electric and thermal load characteristic curve.

[0200] In an embodiment of the present invention, the improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm; the optimization solution unit 14 is specifically used to perform decoupling processing according to the system benefit objective function and constraint conditions through the alternating direction multiplier method to generate multiple sub-function equations; through the quantum particle swarm optimization algorithm, each sub-function equation is optimized and solved to generate an optimal solution, thereby completing the scheduling optimization of the cogeneration system.

[0201] In the scheme of the embodiment of the present invention, energy data in the cogeneration system is obtained; based on multi-task learning, the energy data is predicted to generate predicted power generation and heating data; through the electric and thermal load characteristic curves of each heating unit, the system benefit objective function and constraints are constructed according to the power generation and heating data and the energy data; through the improved optimization algorithm, according to the constraints, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system, and the electric and thermal load characteristics are analyzed through multi-task learning and the improved optimization algorithm, and the decoupled distributed optimization decomposition is performed, so as to achieve the search for the global optimal solution, improve the computing efficiency, meet the accuracy requirements of load prediction, save computing time and computing resources, and improve computing efficiency.

[0202] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer device, and specifically, the computer device may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0203] An embodiment of the present invention provides a computer device, including a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions. When the program instructions are loaded and executed by the processor, the steps of the embodiment of the above-mentioned data-driven combined heat and power system optimization method are implemented. For a specific description, please refer to the embodiment of the above-mentioned data-driven combined heat and power system optimization method.

[0204] Reference below Fig. 9 , which shows a schematic diagram of the structure of a computer device 600 suitable for implementing an embodiment of the present application.

[0205] like Fig. 9As shown, the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate operations and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the computer device 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0206] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal feedback device (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed as needed as the storage section 608.

[0207] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program tangibly contained on a machine-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 609, and / or installed from the removable medium 611.

[0208] 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 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.

[0209] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0210] 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.

[0211] 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.

[0212] 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.

[0213] 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.

[0214] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0215] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0216] 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.

[0217] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0218] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0219] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A data-driven combined heat and power system optimization method, characterized in that: The method comprises: Obtain energy data in combined heat and power systems; Based on multi-task learning, the energy data is predicted to generate predicted power generation and heating data; By using the electric and thermal load characteristic curves of each heating unit, according to the power generation and heating data and energy data, the system benefit objective function and constraint conditions are constructed; By using an improved optimization algorithm and according to the constraint conditions, the system benefit objective function is optimized and solved to complete the scheduling optimization of the cogeneration system.

2. The data-driven combined heat and power system optimization method according to claim 1, characterized in that: The method of predicting the energy data based on multi-task learning to generate predicted power generation and heating data includes: Normalizing the energy data to generate normalized energy characteristics; Generate a first output result according to the normalized energy characteristics through a long short-term memory network; By using an attention mechanism, the first output result is calculated to generate a second output result; The second output result is subjected to denormalization processing to generate the power generation and heat supply data.

3. The data-driven combined heat and power system optimization method according to claim 1, characterized in that: The electric heat load characteristic curves of each heating unit are used to construct the system benefit objective function and constraint conditions according to the power generation and heating data and energy data, including: Constructing a system benefit objective function based on the power generation and heating data, energy data and preset influencing factor coefficients; Constraint conditions are constructed based on the electric heating load characteristic curve and the power generation and heating data.

4. The data-driven combined heat and power system optimization method according to claim 3, characterized in that: The power generation and heating data include wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation, and cogeneration unit power generation; The energy data include abandoned wind power, abandoned photovoltaic power, grid-connected electricity price, coal-fired unit power, combined heat and power unit power, combined heat and power unit thermal power, number of coal-fired units, number of combined heat and power units, carbon emissions of coal-fired units, carbon emissions of combined heat and power units, peak load time, heating power of combined heat and power units, and heating power of electric boilers; The influencing factor coefficients include wind abandonment penalty coefficient, solar abandonment penalty coefficient, coal consumption coefficient of coal-fired units, coal consumption coefficient of cogeneration units, carbon emission coefficient of coal-fired units, carbon emission coefficient of cogeneration units, carbon emission quota, multi-speed peak-shaving compensation boundary of the unit, heating price, and multi-speed peak-shaving compensation quotation of the unit; The system benefit objective function is constructed according to the power generation and heating data, energy data and preset influencing factor coefficients, including: Generate a wind abandonment penalty fee according to the abandoned wind power, the on-grid electricity price and the wind abandonment penalty coefficient; Generate a penalty fee for abandoned solar power according to the abandoned solar power quantity and the abandoned solar power penalty coefficient; Generate coal consumption costs according to the electrical power of the coal-fired unit, the electrical power of the cogeneration unit, the thermal power of the cogeneration unit, the number of coal-fired units, the number of cogeneration units, the coal consumption coefficient of the coal-fired unit and the coal consumption coefficient of the cogeneration unit; Generate carbon trading fees for the heating unit according to the carbon emissions of the coal-fired unit, the carbon emissions of the combined heat and power unit, the carbon emission coefficient of the coal-fired unit, the carbon emission coefficient of the combined heat and power unit and the carbon emission quota; Generate peak-shaving compensation according to the unit's electrical load, peak-shaving time, the unit's multi-speed peak-shaving compensation boundary, and the unit's multi-speed peak-shaving compensation quotation; Generate electricity price income according to the on-grid electricity price, wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; Generate heating revenue according to the heating price, the heating power of the cogeneration unit and the heating power of the electric boiler; A system benefit objective function is generated based on the penalty for wind abandonment, penalty for solar abandonment, coal consumption costs, carbon trading costs of heating units, peak-shaving compensation, electricity price benefits and heating benefits.

5. The data-driven combined heat and power system optimization method according to claim 4, characterized in that: The constraints include system electric power balance constraints, system thermal power balance constraints and operation boundary constraints of each cogeneration unit; The constraining conditions are constructed according to the electric heat load characteristic curve and the power generation and heating data, including: Generate a system power balance constraint according to the wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; Generate a system thermal power balance constraint according to the heating power of the cogeneration unit and the heating power of the electric boiler; According to the electric and thermal load characteristic curve, the operating boundary constraints of each cogeneration unit are generated.

6. The data-driven combined heat and power system optimization method according to claim 1, characterized in that: The improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm; The improved optimization algorithm is used to optimize and solve the system benefit objective function according to the constraint conditions to complete the scheduling optimization of the cogeneration system, including: By using the alternating direction multiplier method, decoupling is performed according to the system benefit objective function and constraint conditions to generate multiple sub-function equations; Through the quantum particle swarm optimization algorithm, each sub-function equation is optimized and solved to generate the optimal solution, thereby completing the scheduling optimization of the cogeneration system.

7. A data-driven combined heat and power system optimization device, characterized in that: The device comprises: An acquisition unit, used for acquiring energy data in the combined heat and power system; A prediction unit, used to predict the energy data based on multi-task learning and generate predicted power generation and heating data; A function construction unit, used to construct a system benefit objective function and constraint conditions according to the power generation and heating data and energy data through the electric and thermal load characteristic curves of each heating unit; The optimization solving unit is used to optimize and solve the system benefit objective function according to the constraint conditions through an improved optimization algorithm to complete the scheduling optimization of the cogeneration system.

8. The data-driven cogeneration system optimization device according to claim 7, characterized in that: The prediction unit is specifically used to normalize the energy data to generate normalized energy characteristics; and generate a first output result according to the normalized energy characteristics through a long short-term memory network; By using an attention mechanism, the first output result is calculated to generate a second output result; The second output result is subjected to denormalization processing to generate the power generation and heat supply data.

9. The data-driven cogeneration system optimization device according to claim 7, characterized in that: The function construction unit is specifically used to construct a system benefit objective function according to the power generation and heating data, energy data and preset influencing factor coefficients; and to construct constraint conditions according to the electric and thermal load characteristic curve and the power generation and heating data.

10. The data-driven cogeneration system optimization device according to claim 9, characterized in that: The power generation and heating data include wind power generation power, photovoltaic power generation power, unit electrical load, coal-fired unit power generation, and cogeneration unit power generation; The energy data include abandoned wind power, abandoned photovoltaic power, grid-connected electricity price, coal-fired unit power, combined heat and power unit power, combined heat and power unit thermal power, number of coal-fired units, number of combined heat and power units, carbon emissions of coal-fired units, carbon emissions of combined heat and power units, peak load time, heating power of combined heat and power units, and heating power of electric boilers; The influencing factor coefficients include wind abandonment penalty coefficient, solar abandonment penalty coefficient, coal consumption coefficient of coal-fired units, coal consumption coefficient of cogeneration units, carbon emission coefficient of coal-fired units, carbon emission coefficient of cogeneration units, carbon emission quota, multi-speed peak-shaving compensation boundary of the unit, heating price, and multi-speed peak-shaving compensation quotation of the unit; The function construction unit is specifically used to generate wind abandonment penalty fees according to the abandoned wind power, grid-connected electricity price and wind abandonment penalty coefficient; generate solar abandonment penalty fees according to the abandoned solar power and solar abandonment penalty coefficient; generate coal consumption fees according to the coal-fired unit electric power, the cogeneration unit electric power, the cogeneration unit thermal power, the number of coal-fired units, the number of cogeneration units, the coal consumption coefficient of the coal-fired unit and the coal consumption coefficient of the cogeneration unit; generate heat supply costs according to the carbon emissions of the coal-fired unit, the carbon emissions of the cogeneration unit, the carbon emission coefficient of the coal-fired unit, the carbon emission coefficient of the cogeneration unit and the carbon emission quota The carbon trading fee of the unit; generate peak-shaving compensation according to the unit's power load, peak-shaving time, the unit's multi-speed peak-shaving compensation boundary and the unit's multi-speed peak-shaving compensation quotation; generate electricity price income according to the on-grid electricity price, wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; generate heating income according to the heating price, cogeneration unit heating power and electric boiler heating power; generate system income objective function according to the wind abandonment penalty fee, solar abandonment penalty fee, coal consumption fee, heating unit carbon trading fee, peak-shaving compensation, electricity price income and heating income.

11. The data-driven cogeneration system optimization device according to claim 10, characterized in that: The constraints include system electric power balance constraints, system thermal power balance constraints and operation boundary constraints of each cogeneration unit; The function construction unit is specifically used to generate system electric power balance constraints according to the wind power generation power, photovoltaic power generation power, coal-fired unit power generation power and cogeneration unit power generation power; generate system thermal power balance constraints according to the cogeneration unit heating power and electric boiler heating power; and generate operating boundary constraints for each cogeneration unit according to the electric and thermal load characteristic curve.

12. The data-driven cogeneration system optimization device according to claim 7, characterized in that: The improved optimization algorithm includes an alternating direction multiplier method and a quantum particle swarm optimization algorithm; The optimization solving unit is specifically used to perform decoupling processing according to the system benefit objective function and constraint conditions through the alternating direction multiplier method to generate multiple sub-function equations; optimize and solve each sub-function equation through the quantum particle swarm optimization algorithm to generate the optimal solution, thereby completing the scheduling optimization of the cogeneration system.

13. A computer readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the data-driven combined heat and power system optimization method described in any one of claims 1 to 6 is implemented.

14. A computer device comprising a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, characterized in that: When the program instructions are loaded and executed by the processor, the data-driven combined heat and power system optimization method according to any one of claims 1 to 6 is implemented.

15. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the data-driven combined heat and power system optimization method according to any one of claims 1 to 6 is implemented.

Images