Park multi-energy complementary system day-ahead low-carbon economic production optimization scheduling method and system

Abstract

The invention provides a day-ahead low-carbon economic production optimization scheduling method and system for a park multi-energy complementary system, and relates to the technical field of industrial park production energy optimization scheduling. A deep neural network regression model is constructed by using the nonlinear fitting regression capability of a deep convolutional neural network; according to the model, the total output predicted value of the time-sharing multi-energy complementary system under the energy supply control strategy of each energy conversion device of the multi-energy complementary system in at least one day in the future is predicted, and nonlinear factors of input, conversion and output links in the actual energy conversion process of the multi-energy complementary system are comprehensively considered. Establishing an optimization objective function for production optimization scheduling by taking minimization of the sum of the total operation cost and the carbon transaction cost as an objective, and obtaining an optimal hourly energy supply control strategy of each energy conversion device of the multi-energy complementary system in at least one day in the future after optimization; accurate day-ahead low-carbon economic production optimization scheduling is carried out on an actual existing industrial park comprehensive energy system.

Description

Technical Field

[0001] This invention relates to the field of energy optimization scheduling technology for industrial parks, and in particular to a day-ahead low-carbon economic production optimization scheduling method and system for a multi-energy complementary system in an industrial park. Background Technology

[0002] Current research on low-carbon economic operation optimization methods for multi-energy complementary systems in industrial parks mainly relies on idealized matrix modeling of energy conversion equipment (i.e., energy hub models). This is combined with methods such as linear programming, mixed-integer linear programming (MILP), swarm intelligence optimization algorithms, and deep reinforcement learning to optimize the control strategies of each energy conversion device in the multi-energy complementary system. The energy hub model can represent the conversion and coupling between various energy sources, and is an abstract model of various actual devices and their combinations. However, because the matrix model of energy conversion equipment cannot fully consider the changes in operating efficiency of each energy conversion device under varying operating conditions, the changes in ramp rate during actual load increases and decreases, and the nonlinear factors in each stage of the actual energy conversion process, such as irreversible losses during energy flow transfer, these existing technologies are only suitable for the initial design phase of planning and construction of multi-energy complementary systems in industrial parks. They cannot be used for accurate low-carbon economic operation optimization scheduling of existing multi-energy complementary systems in industrial parks, let alone for accurate day-ahead low-carbon economic production optimization scheduling of existing integrated energy systems in industrial parks. Summary of the Invention

[0003] To address the shortcomings of existing technologies that fail to fully consider the nonlinear factors in the energy input, conversion, and output stages of multi-energy complementary systems in industrial parks, thus hindering accurate day-ahead low-carbon economic production optimization scheduling for existing multi-energy complementary systems in industrial parks, this invention proposes a day-ahead low-carbon economic production optimization scheduling method and system for multi-energy complementary systems in industrial parks.

[0004] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a method for optimizing and scheduling low-carbon economic production in a multi-energy complementary system in an industrial park, comprising: constructing a deep neural network regression model of the multi-energy complementary system using a deep convolutional neural network, taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output; acquiring historical production operation data of the multi-energy complementary system in the industrial park to be optimized and scheduled to train the deep neural network regression model of the multi-energy complementary system; wherein, the historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system; and calculating the low-carbon economic production schedule within at least one day in the future using the trained deep neural network regression model of the multi-energy complementary system. The overall output forecast of the multi-energy complementary system is calculated based on the time-sharing power supply control strategies of each energy conversion device in the multi-energy complementary system. The energy load required for park production in the next at least one day is generated based on the time-sharing external energy purchase price, time-sharing carbon trading price, and production plan. With the objective of minimizing the sum of the overall operating cost and the overall carbon trading cost in the next at least one day, and with constraints set based on the energy load and the overall output forecast, an optimization objective function for production optimization scheduling is established. The optimization objective function is optimized to obtain the best hourly power supply control strategy for each energy conversion device in the multi-energy complementary system in the next at least one day, which serves as the day-ahead low-carbon economic production optimization scheduling scheme for the park's multi-energy complementary system.

[0005] To improve the accuracy of model prediction, the present invention provides a preferred solution in the first aspect, wherein the step of acquiring historical production operation data of the multi-energy complementary system of the industrial park to be optimized for production scheduling to train the deep neural network regression model of the multi-energy complementary system includes, after acquiring the historical production operation data, performing data preprocessing including data type conversion, filling missing data, data smoothing and outlier detection, and using the preprocessed historical production operation data to train the deep neural network regression model of the multi-energy complementary system.

[0006] To improve the robustness of outlier identification and removal, the present invention provides a preferred solution in the first aspect, wherein the outlier detection employs a Hampel filter, which identifies and removes outliers in the data based on the median and median absolute deviation (MAD).

[0007] Furthermore, the power supply control strategy data for each energy conversion device includes: the electrical power supplied to the electric refrigeration equipment, the natural gas power supplied to the gas boiler, the thermal power supplied to the absorption refrigeration equipment, the natural gas power supplied to the combined heat and power equipment, and the power of electricity purchased from outside the park; the overall output data of the multi-energy complementary system includes: the overall available power supply of the multi-energy complementary system, the overall available heat supply of the multi-energy complementary system, and the overall available cooling supply of the multi-energy complementary system.

[0008] Furthermore, the purchased energy price includes: purchased electricity price and purchased natural gas price; the energy load required for production in the park includes: electrical load, heat load and cooling load.

[0009] Furthermore, the overall operating cost is calculated using the purchased energy price for each period and the energy consumed under the energy conversion equipment power supply control strategy for each period; the overall carbon trading cost is calculated using the carbon emission factor of the purchased natural gas power per unit of power, the carbon emission quota of the purchased natural gas power per unit of power, the carbon emission factor of the purchased electricity power per unit of power, the carbon emission quota of the purchased electricity power per unit of power, the carbon dioxide trading unit price for each period, the amount of purchased natural gas for each period, and the amount of purchased electricity for each period.

[0010] The present invention provides a preferred embodiment in a first aspect, wherein the step of establishing an optimization objective function for production optimization scheduling, with the objective of minimizing the sum of total operating costs and carbon trading costs over at least one day in the future, and setting constraints based on the energy load and the total output forecast, specifically involves introducing a penalty function to set the constraints, and the resulting optimization objective function is as follows: ; The penalty function is: m is the penalty coefficient, and a refers to the first parameter in each penalty term, which actually corresponds to... , or ; b refers to the second parameter in each penalty term, which actually corresponds to , or , This represents the operating cost of the multi-energy complementary system in the i-th time period. This represents the carbon trading cost of a multi-energy complementary system in time period i. This represents the cooling load required for production in the park during the i-th time period. This represents the heat load required for production in the park during the i-th time period. This represents the electrical load required for production in the park during the i-th time period. Let i be the total available cooling power of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period. Let i be the total available thermal power of the multi-energy complementary system under the control strategy of the energy conversion equipment in the i-th time period. Let represent the total power supply of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period.

[0011] In a first aspect, the present invention provides a preferred embodiment in which the gray wolf optimization algorithm is used to optimize the objective function and obtain the best hourly energy supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future as the day-ahead low-carbon economic production optimization scheduling scheme of the multi-energy complementary system in the park.

[0012] The present invention provides a preferred embodiment in the first aspect, wherein the position of the gray wolf in the optimization objective function is the hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, including the electric power supplied to the electric refrigeration equipment, the natural gas power supplied to the gas boiler, the natural gas power supplied to the cogeneration equipment, the thermal power supplied to the absorption refrigeration equipment, and the power purchased from outside the park for each period of the future day.

[0013] In a second aspect, this invention provides a day-ahead low-carbon economic production optimization scheduling system for a multi-energy complementary system in an industrial park, used to execute the method, comprising: a model construction and training module, used to construct a deep neural network regression model of the multi-energy complementary system using a deep convolutional neural network, taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output; acquiring historical production operation data of the multi-energy complementary system of the industrial park to be optimized and scheduled to train the deep neural network regression model of the multi-energy complementary system; wherein, the historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system; and an output prediction module, used to calculate the output prediction for at least one day in the future using the trained deep neural network regression model of the multi-energy complementary system. The system comprises: a time-sharing overall output forecast of the multi-energy complementary system under the energy supply control strategies of each energy conversion device; a load forecasting module, used to generate the energy load required for park production for at least one day based on the time-sharing external energy price, time-sharing carbon trading price, and production plan for at least one day in the future; an optimization objective function establishment module, used to minimize the sum of the overall operating cost and the overall carbon trading cost for at least one day in the future, and set constraints based on the energy load and the overall output forecast to establish an optimization objective function for production optimization scheduling; and an optimization module, used to optimize the optimization objective function to obtain the best hourly energy supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future as the day-ahead low-carbon economic production optimization scheduling scheme for the park's multi-energy complementary system.

[0014] Compared with the prior art, the present invention has the following beneficial technical effects: First, this invention, based on historical production and operation data of industrial parks, utilizes the nonlinear fitting and regression capabilities of deep convolutional neural networks. Taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output, it constructs a deep neural network regression model for the industrial park's multi-energy complementary system. This establishes a high-precision nonlinear mapping model from device control strategies to the system's total output. This model predicts the time-sharing total output of the multi-energy complementary system under the energy supply control strategies of each energy conversion device within at least one day. This model comprehensively considers the nonlinear factors at each stage of the actual energy conversion process in the industrial park's multi-energy complementary system, including input, conversion, and output. Therefore, it can more accurately improve the accuracy of system output prediction, providing a reliable model foundation for subsequent optimized scheduling and avoiding scheduling deviations caused by model distortion.

[0015] Then, this invention aims to minimize the sum of total operating costs and total carbon trading costs over at least one day. Under the constraints of meeting the energy load required for production and the predicted total output of the multi-energy complementary system, an optimization objective function for production optimization scheduling is established. Through optimization, the optimal hourly energy supply control strategy for each energy conversion device of the multi-energy complementary system over at least one day is obtained as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the industrial park. This enables accurate day-ahead low-carbon economic production optimization scheduling of the existing integrated energy system of the industrial park, providing a scientific basis and technical support for formulating appropriate and reasonable day-ahead low-carbon economic production optimization scheduling strategies for the multi-energy complementary system in the park, and realizing cost reduction, efficiency improvement, and low-carbon high-efficiency operation of the park. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the steps of a day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park, provided as a specific embodiment of the present invention. Figure 2 A schematic diagram of a low-carbon economic production optimization scheduling system for a multi-energy complementary system in a park, provided for a specific embodiment of the invention; Figure 3 A flowchart of a day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park, provided as a specific embodiment of the present invention; Figure 4A schematic diagram of an integrated energy system for combined cooling, heating and power (CCHP) in an industrial park, which is the application of a method for optimizing low-carbon economic production scheduling in a multi-energy complementary system for industrial parks according to a specific embodiment of the present invention. Figure 5 The present invention provides a time-sharing curve of the park's electricity load, heat load, and cooling load for a future day in a method for optimizing low-carbon economic production scheduling of a multi-energy complementary system in a park, as described in a specific embodiment of the present invention. Figure 6 A line graph showing the time-sharing energy price changes for the next day (1 to 24 hours) in a day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park provided by a specific embodiment of the present invention. Figure 7 This invention provides a method for optimizing low-carbon economic production scheduling in a multi-energy complementary system for industrial parks, based on a specific embodiment of the present invention. This method optimizes the hourly power supply control strategy for electric refrigeration equipment before and after optimization (the power supplied to the electric refrigeration equipment at various times during the next day). Comparison curves; Figure 8 This invention provides a method for optimizing low-carbon economic production scheduling in a multi-energy complementary system for industrial parks, based on a specific embodiment of the present invention. This method optimizes the hourly energy supply control strategy of the gas-fired boilers before and after optimization (the natural gas power supplied to the gas-fired boilers at various times during the next day). Comparison curves; Figure 9 This invention provides a method for optimizing low-carbon economic production scheduling in a multi-energy complementary system for industrial parks, based on a specific embodiment of the present invention. This method optimizes the hourly power supply control strategy for combined heat and power (CHP) equipment before and after optimization (the natural gas power supplied to CHP equipment at various times during the next day). Comparison curves; Figure 10 This invention provides a method for optimizing low-carbon economic production scheduling in a multi-energy complementary system for industrial parks, based on a specific embodiment of the present invention. This method optimizes the hourly energy supply control strategy for absorption chillers before and after optimization (the heat power supplied to the absorption chillers at various times during the next day). Comparison curves; Figure 11 This invention provides a specific embodiment of a method for optimizing the day-ahead low-carbon economic production scheduling of a multi-energy complementary system in a park, which optimizes the hourly power supply control strategy of the park's purchased electricity before and after optimization (power of purchased electricity in the park at various times during the next day). Comparison curves; Figure 12 This is a graph showing the hourly operating cost comparison of the multi-energy complementary system in a park before and after optimization by a low-carbon economic production optimization scheduling method for a park multi-energy complementary system, provided by a specific embodiment of the present invention, for the next day. Figure 13 This is a graph showing the hourly carbon trading cost comparison curves of the multi-energy complementary system in a park before and after optimization of the daytime low-carbon economic production optimization scheduling method provided by a specific embodiment of the present invention, for the next day. Figure 14 This is a comparison curve showing the sum of hourly operating costs and carbon trading costs of a multi-energy complementary system in a park before and after optimization by a day-ahead low-carbon economic production optimization scheduling method provided by a specific embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please refer to Figure 1 In a preferred embodiment, a day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park is provided, which is mainly implemented through the following steps: S1. Using a deep convolutional neural network, with the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output, a deep neural network regression model for the multi-energy complementary system is constructed; historical production operation data of the multi-energy complementary system in the industrial park to be optimized and scheduled are obtained to train the deep neural network regression model for the multi-energy complementary system; wherein, the historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system; S2. Using the trained deep neural network regression model of the multi-energy complementary system, calculate the time-sharing prediction of the overall output of the multi-energy complementary system under the energy supply control strategies of each energy conversion device in the multi-energy complementary system for at least one day in the future. S3. Based on the time-of-use energy purchase price, time-of-use carbon trading price and production plan for at least one day in the future, generate the energy load required for park production for at least one day in the future. S4. With the goal of minimizing the sum of total operating costs and total carbon trading costs over at least one day in the future, and based on the energy load and the total output forecast, establish an optimization objective function for production optimization scheduling; S5. Optimize the objective function to obtain the best hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park.

[0020] Please refer to Figure 2 Correspondingly, this embodiment provides a day-ahead low-carbon economic production optimization scheduling system for a multi-energy complementary system in a park, used to execute the above method, which mainly consists of the following modules: Model building and training module 1 is used to construct a deep neural network regression model for a multi-energy complementary system by using a deep convolutional neural network, taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output; and to obtain historical production operation data of the multi-energy complementary system of the industrial park to be optimized and scheduled for production in order to train the deep neural network regression model of the multi-energy complementary system; wherein, the historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system; The output prediction module 2 is used to calculate the time-sharing overall output prediction value of the multi-energy complementary system under the energy supply control strategy of each energy conversion device in the multi-energy complementary system through the trained deep neural network regression model of the multi-energy complementary system for at least one day in the future. Load forecasting module 3 is used to generate the energy load required for park production for at least one day in the future based on the time-sharing external energy price, time-sharing carbon trading price and production plan for at least one day in the future. The objective function establishment module 4 is used to establish an optimization objective function for production optimization scheduling with the goal of minimizing the sum of the total operating cost and the total carbon trading cost over at least one day in the future, and based on the energy load and the total output forecast, setting constraints. Optimization Module 5 is used to optimize the objective function and obtain the best hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park.

[0021] The above embodiments, based on historical production and operation data of the industrial park, utilize the nonlinear fitting and regression capabilities of deep convolutional neural networks. Taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output, a deep neural network regression model of the industrial park's multi-energy complementary system is constructed. This establishes a high-precision nonlinear mapping model from device control strategies to the total system output. This model predicts the time-sharing total output of the multi-energy complementary system under the energy supply control strategies of each energy conversion device within at least one day. This model comprehensively considers the nonlinear factors at each stage of the actual energy conversion process in the industrial park's multi-energy complementary system, including input, conversion, and output. Therefore, it can more accurately improve the accuracy of system output prediction, providing a reliable model foundation for subsequent optimized scheduling and avoiding scheduling deviations caused by model distortion.

[0022] The above embodiments aim to minimize the sum of total operating costs and total carbon trading costs over at least one day. Under the constraints of meeting the energy load required for production and the overall output forecast of the multi-energy complementary system, an optimization objective function for production optimization scheduling is established. Then, through optimization, the best hourly energy supply control strategy for each energy conversion device of the multi-energy complementary system over at least one day in the future is obtained as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park. This enables accurate day-ahead low-carbon economic production optimization scheduling of the existing industrial park's integrated energy system, providing a scientific basis and technical support for formulating appropriate and reasonable day-ahead low-carbon economic production optimization scheduling strategies for the park's multi-energy complementary system, and achieving cost reduction, efficiency improvement, and low-carbon, high-efficiency operation of the park.

[0023] The following will provide a more detailed explanation of each step in the above embodiments, and offer a more preferred and detailed implementation method, such as... Figure 3 As shown, the complete implementation process of the day-ahead low-carbon economic production optimization scheduling method for the multi-energy complementary system in the industrial park is presented in this specific embodiment. Meanwhile, to test the effectiveness of the optimization scheduling method proposed in this invention for accurately optimizing day-ahead low-carbon economic production scheduling of existing integrated energy systems in industrial parks, in a preferred embodiment, as follows... Figure 4 The example shown is a combined cooling, heating, and power (CCHP) integrated energy system in a park. Figure 4 In this system, AB is a gas-fired boiler with a maximum heating capacity of 400kW; CERG is an electric refrigeration unit with a maximum cooling capacity of 300kW; WRG is an absorption refrigeration unit with a maximum cooling capacity of 300kW; and CHP is a combined heat and power (CHP) unit with a maximum power supply of 120kW and a maximum heating capacity of 160kW. The energy conversion equipment control strategy of this park's integrated energy system is based on the power supplied to the electric refrigeration unit. Natural gas power supplied to gas boilers The heat power supplied to the absorption refrigeration equipment Natural gas power supplied to cogeneration equipment and the power purchased from outside the park Composed of.

[0024] S1. Using a deep convolutional neural network, a deep neural network regression model for a multi-energy complementary system is constructed, with the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output. Historical production operation data of the multi-energy complementary system in the industrial park to be optimized and scheduled are obtained to train the deep neural network regression model of the multi-energy complementary system. Among them, the historical production operation data includes the energy supply control strategy data of each energy conversion device under various variable operating conditions and the overall output data of the multi-energy complementary system.

[0025] In step S1, historical production operation data is obtained as follows: Historical production operation data of the multi-energy complementary system of the industrial park to be optimized and scheduled is collected, including power supply control strategy data for each energy conversion device under various variable operating conditions (such as the electrical power supplied to the electric refrigeration equipment). Natural gas power supplied to gas boilers The heat power supplied to the absorption refrigeration equipment Natural gas power supplied to cogeneration equipment and the power purchased from outside the park The data includes data on the output of energy conversion equipment, i.e., the overall output data of the multi-energy complementary system (such as the total available power supply, total available heat supply, and total available cooling power of the multi-energy complementary system in an industrial park). In a preferred embodiment, the acquired historical production operation data undergoes data cleaning preprocessing, including data type conversion, missing data filling, data smoothing, and outlier detection. Gaussian smoothing is used for data smoothing with a data window of 20; an outlier detection uses a Hampel filter. The Hampel filter, as a powerful time series outlier handling tool, plays an important role in data cleaning and analysis. The Hampel filter is a filter based on the median and median absolute deviation (MAD) designed to identify and remove outliers in time series data. Compared to traditional mean and standard deviation methods, the Hampel filter is more robust to outliers.

[0026] The criteria for identifying outliers in the Hampel filter are as follows: ; Where X is a time-series vector of historical production operation data consisting of multiple (e.g., 20) data points. Let be the i-th data point in the time series vector of this data segment. The median absolute deviation (MAD) is: .

[0027] In step S1, a deep neural network regression model of the multi-energy complementary system is established using a deep convolutional neural network. The model takes the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output. Essentially, this model uses a deep convolutional neural network to deeply regress and learn the nonlinear factors of each link in the energy input-conversion-output process of the multi-energy complementary system.

[0028] The structure of the deep neural network regression model for multi-energy complementary systems is shown in Table 1.

[0029] Table 1: Structure of Deep Neural Network Regression Model for Multi-Energy Complementary Systems ; In Table 1, the input layer parameter 'n' represents the number of controllable parameters for the power supply control strategy of each energy conversion device in the multi-energy complementary system of the industrial park to be optimized for production scheduling. In a preferred embodiment, the controllable parameters of the power supply control strategy of each energy conversion device in the multi-energy complementary system include: the electrical power supplied to the electric refrigeration equipment. Natural gas power supplied to gas boilers The heat power supplied to the absorption refrigeration equipment Natural gas power supplied to cogeneration equipment and the power purchased from outside the park At this point, n=5.

[0030] S2. Using the trained deep neural network regression model of the multi-energy complementary system, calculate the time-sharing prediction value of the overall output of the multi-energy complementary system under the energy supply control strategies of each energy conversion device in the multi-energy complementary system for at least one day in the future.

[0031] In step S2, the deep neural network regression model of the multi-energy complementary system is used to calculate in real time the predicted overall output of the multi-energy complementary system when operating under different energy conversion equipment power supply control strategies. In a preferred embodiment, the predicted overall output of the multi-energy complementary system includes the predicted total available power, heat, and cooling power of the multi-energy complementary system. Therefore, this model can also be called: a deep regression model of the power supply control strategies of each energy conversion equipment in the multi-energy complementary system and the total available power, heat, and cooling power of the system.

[0032] S3. Generate the time-of-use energy load required for park production for at least one day in the future, based on the time-of-use purchased energy price, time-of-use carbon trading price, and production plan. In this step, the purchased energy price mainly includes the purchased electricity price and the purchased natural gas price; the energy load required for park production includes: electricity load, heat load, and cooling load. Taking data from the next day as an example, specifically, this step generates the time-of-use park electricity load, heat load, and cooling load curves for the next day based on the time-of-use park electricity price, purchased natural gas price, carbon trading price, and park production plan for the next day, as shown below. Figure 5 As shown. For the hourly energy prices of the next day, such as... Figure 6 As shown in the figure, "electricity price" refers to the price of purchased electricity, that is, the price per unit of purchased electricity (yuan / kWh), and "gas price" refers to the price of purchased natural gas, that is, the price per unit of purchased natural gas power (yuan / kWh). Meanwhile, in this embodiment, it is assumed that the carbon trading price is fixed at 0.3 yuan / kg.

[0033] S4. An optimization objective function for production optimization scheduling is established, with the objective of minimizing the sum of total operating costs and total carbon trading costs for at least one day in the future, and based on constraints set according to the energy load and the total power output forecast. The total operating cost is calculated using the purchased energy price for each time period and the energy consumed under the energy conversion equipment power supply control strategy for each time period. The total carbon trading cost is calculated using the carbon emission factor per unit of purchased natural gas power, the carbon emission quota per unit of purchased natural gas power, the carbon emission factor per unit of purchased electricity power, the carbon emission quota per unit of purchased electricity power, the carbon dioxide trading price per unit (i.e., the price per kilogram of carbon dioxide), the amount of purchased natural gas, and the amount of purchased electricity for each time period.

[0034] Specifically, the design optimization objective is to minimize the sum of the total operating cost and carbon trading cost of the park's multi-energy complementary system for the next day, while meeting the constraints of the time-of-use electricity, heat, and cooling loads required for production in the park. This constraint is achieved by introducing a penalty function into the optimization objective. The optimization objective function is: The penalty function is: 'm' is the penalty coefficient, and 'a' refers to the first parameter in each penalty term of the optimization objective function, which actually corresponds to... , or ; b refers to the second parameter in each penalty term, which actually corresponds to , or ; This represents the operating cost of the park's multi-energy complementary system in the i-th time period (i.e., the price of purchased energy in the i-th time period multiplied by the energy consumed under the current energy conversion equipment power supply control strategy). Let be the price (yuan / kWh) of the unit of natural gas power purchased in the i-th time period. Let be the amount of natural gas purchased in the i-th time period (kWh). Let be the price per unit of purchased electrical power in the i-th time period (yuan / kWh). Let i be the amount of electricity purchased from outside the system during the i-th time period (kWh). This represents the carbon trading cost of the multi-energy complementary system in the park during the i-th time period. Let the carbon emission factor (kg / (kWh)) be the unit of power generated from purchased natural gas in time period i. The carbon emission allowance per unit of purchased natural gas power in time period i is kg / kWh. The carbon emission factor (kg / (kWh)) for a unit of purchased electricity in time period i. The carbon emission allowance per unit of purchased electricity in time period i is kg / kWh. Let represent the price per kilogram of carbon dioxide traded in the i-th time period (in yuan / kg). This represents the cooling load required for production in the park during the i-th time period. This represents the heat load required for production in the park during the i-th time period. This represents the electrical load required for production in the park during the i-th time period. Let i be the total available cooling power of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period. Let i be the total available thermal power of the multi-energy complementary system under the control strategy of the energy conversion equipment in the i-th time period. Let represent the total power supply of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period.

[0035] S5. Optimize the objective function to obtain the best hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park.

[0036] In a preferred embodiment, the Grey Wolf optimization algorithm is used for optimization. Specifically, the Grey Wolf optimization algorithm, based on the time-of-use energy prices, carbon trading prices, and the time-of-use electricity, heat, and cooling loads required for production in the park for the next day, aims to minimize the sum of the total operating cost and carbon trading cost of the multi-energy complementary system for the next day, while satisfying the constraints of the time-of-use electricity, heat, and cooling loads required for production in the park. The algorithm optimizes the hourly energy supply control strategy for each energy conversion device in the multi-energy complementary system, such as the power supply to the cooling equipment at different times of the next day. Natural gas power supplied to gas boilers The heat power supplied to the absorption refrigeration equipment Natural gas power supplied to cogeneration equipment and the power purchased from outside the park wait.

[0037] The Grey Wolf Optimizer (GWO) is a metaheuristic algorithm that simulates the hunting behavior of grey wolf packs in nature. Based on the social hierarchy and hunting behavior of grey wolves, the algorithm seeks optimal solutions by simulating their hierarchy and hunting actions. In the algorithm, grey wolves are divided into four types: Alpha (α), Beta (β), Delta (δ), and Omega (ω). Alpha represents the optimal solution, Beta and Delta represent the second-best and third-best solutions, respectively, while Omega is responsible for making decisions and exploring the search space. Grey wolves update their positions based on their distance from prey. The positions of the grey wolves determine the hourly control strategies for each energy conversion device in the multi-energy complementary system of the park for the next day. ,like Figure 3As shown, the optimal hourly control strategy for each energy conversion device in the multi-energy complementary system of the park is obtained through the iterative process of the Grey Wolf optimization algorithm. This includes the optimal power supply for electric refrigeration equipment, natural gas supply for gas boilers, natural gas supply for combined heat and power (CHP) equipment, heat supply for absorption chillers, and purchased electricity from outside the park. Through continuous iterative updates, the wolf gradually moves towards the prey's location (optimal solution) in the solution space, ultimately obtaining the globally optimal solution. That is, the optimization objective function is obtained. Hourly control strategy for each energy conversion device in the multi-energy complementary system of the park, minimizing the optimal level of control over the next day. This includes the optimal power supply for electric refrigeration equipment, natural gas supply for gas boilers, natural gas supply for combined heat and power (CHP) equipment, thermal power supply for absorption chillers, and purchased electricity for the park, for each time period of the future day, thereby obtaining the day-ahead low-carbon economic production optimization scheduling scheme for the park's multi-energy complementary system. In this embodiment, the wolf pack size is taken as 500, and the maximum number of iterations is taken as 200.

[0038] The hourly control strategies for each energy conversion device in a multi-energy complementary system in a park before and after optimization, based on the above embodiments of the present invention and the day-ahead low-carbon economic production optimization scheduling method provided by the present invention, are as follows: Figures 7 to 11 As shown. Figure 7 To optimize the hourly power supply control strategy for the pre- and post-heating refrigeration equipment (the electrical power supplied to the refrigeration equipment at various times during the next day). Compare the curves. Figure 8 To optimize the hourly energy supply control strategy for the upstream and downstream gas-fired boilers (the natural gas power supplied to the gas-fired boilers at various times during the next day) Compare the curves. Figure 9 To optimize the hourly power supply control strategy for the combined heat and power (CHP) equipment (the natural gas power supplied to the CHP equipment at various times during the next day) Compare the curves. Figure 10 To optimize the hourly power supply control strategy for pre- and post-absorption refrigeration equipment (the heat power supplied to the absorption refrigeration equipment at various times during the next day). Compare the curves. Figure 11 To optimize the hourly power supply control strategy for the park's externally purchased electricity (power purchased from external sources at various times during the next day) Comparison curves.

[0039] Meanwhile, the hourly operating costs and carbon trading costs of a multi-energy complementary system in a park before and after optimization using the day-ahead low-carbon economic production optimization scheduling method provided in the above embodiments of the present invention, as well as the comparison results of the sum of operating costs and carbon trading costs, for the multi-energy complementary system in a park for the next day are as follows: Figures 12 to 14 As shown. Figure 12 To optimize the hourly operating cost comparison curve of the multi-energy complementary system in the park before and after, for the next day, Figure 13 To optimize the hourly carbon trading cost comparison curve for the multi-energy complementary system in the park before and after the optimization, Figure 14 This is a comparison curve showing the hourly operating costs and carbon trading costs of the multi-energy complementary system in the park before and after optimization for the next day. Figures 12 to 14 It is evident that, under the constraint of meeting the time-sharing electricity, heat, and cooling loads required for production in the park, the sum of the overall operating cost and carbon trading cost of the multi-energy complementary system in the park for the next day after optimization is significantly reduced compared to before optimization with the Grey Wolf optimization iteration. That is, before and after optimization, the 24-hour operating cost, carbon trading cost, and the sum of operating cost and carbon trading cost are all significantly reduced. Furthermore, the overall energy available in the park after optimization (total power supply, heat, and cooling power of the park) and the overall electricity, heat, and cooling load required for production in the park achieve a supply-demand balance. Moreover, the computation time of the day-ahead optimization scheduling proposed in this invention is shorter than that of intelligent optimization algorithms such as particle swarm optimization and genetic algorithms, greatly reducing the computation time of optimization and demonstrating the superiority of the Grey Wolf optimization algorithm while ensuring the optimization effect.

[0040] As can be seen from the above specific implementation cases, the method proposed in this invention can achieve the goal of accurately optimizing the day-ahead low-carbon economic production of existing multi-energy complementary systems in industrial parks, and provide a scientific basis and technical support for formulating appropriate and reasonable day-ahead low-carbon economic production optimization scheduling strategies for multi-energy complementary systems in industrial parks.

[0041] Specifically, based on the above embodiments, the present invention can achieve the following beneficial technical effects: This invention aims to minimize the sum of total system operating costs and carbon trading costs. Under the constraint of meeting electricity, heat, and cooling load demands, it utilizes the Grey Wolf optimization algorithm to optimize the control strategies of various energy devices hourly. This algorithm features strong global search capabilities and fast convergence speed. The multi-objective low-carbon economic scheduling method based on the Grey Wolf optimization algorithm can efficiently find the optimal solution under complex nonlinear constraints, achieving synergistic optimization of the economy and low-carbon performance of multi-energy complementary systems. Compared to traditional optimization algorithms such as genetic algorithms and particle swarm optimization, it significantly reduces computation time while maintaining optimization effectiveness, making it more suitable for practical engineering applications.

[0042] This invention introduces a penalty function into the optimization objective, transforming the load balancing constraint (electricity, heat, and cooling supply and demand balance) into a soft constraint. This ensures that actual production needs are always met during the optimization process, avoiding infeasible solutions. It enhances the robustness and practicality of the optimization model, ensuring the scheduling scheme is feasible and reliable in actual operation and avoiding system operation risks caused by constraint violations.

[0043] This invention integrates historical data cleaning steps such as Hampel filtering outlier detection and Gaussian smoothing, utilizes deep convolutional neural network modeling, load forecasting, and gray wolf optimization to construct a complete day-ahead scheduling framework. This framework can generate optimal hourly energy conversion equipment power supply control strategies based on future electricity prices, gas prices, carbon prices, and production plans. This method is applicable not only to newly built systems but also to real-time optimization scheduling in existing industrial parks, demonstrating strong engineering applicability and promotional value.

[0044] This invention achieves precise, low-carbon, and economical management of day-ahead scheduling for multi-energy complementary systems in industrial parks by combining deep learning modeling, intelligent optimization algorithms, and a data-driven framework. This method not only overcomes the shortcomings of traditional models in characterizing nonlinear factors but also achieves a good balance between computational efficiency and optimization effectiveness, providing reliable technical support for the green and low-carbon transformation of industrial parks.

[0045] In summary, the method of this invention can efficiently and accurately establish a deep regression model of the control strategies of each energy conversion device in a multi-energy complementary system of an industrial park, along with the overall available power, heat, and cooling capacity of the system, by utilizing the nonlinear fitting and regression capabilities of deep convolutional neural networks. Then, the Grey Wolf optimization algorithm is used to find the optimal hourly power supply control strategy for each energy conversion device in the multi-energy complementary system of the industrial park, with the goal of minimizing the sum of the overall operating cost and carbon trading cost of the multi-energy complementary system of the industrial park for the next day, while meeting the constraints of the time-sharing power, heat, and cooling load required for the park's production in the next day. This achieves the goal of accurately optimizing and scheduling day-ahead low-carbon economic production in existing industrial park integrated energy systems, providing a scientific basis and technical support for formulating appropriate and reasonable day-ahead low-carbon economic production optimization and scheduling strategies for multi-energy complementary systems in industrial parks, and realizing cost reduction, efficiency improvement, and low-carbon, high-efficiency operation of the park.

[0046] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0047] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification. Furthermore, the above embodiments only illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. For those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A method for optimizing the low-carbon economic production scheduling of a multi-energy complementary system in a park, characterized in that: include: Using a deep convolutional neural network, a deep neural network regression model for a multi-energy complementary system is constructed, taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output. Historical production operation data of the multi-energy complementary system in the industrial park to be optimized and scheduled are obtained to train the deep neural network regression model. The historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system. Using a trained deep neural network regression model of a multi-energy complementary system, the total output of the multi-energy complementary system is predicted to be time-sharing under the energy supply control strategies of each energy conversion device in the multi-energy complementary system for at least one day in the future. The energy load required for park production in the time-sharing of at least one day in the future is generated based on the time-sharing external energy price, time-sharing carbon trading price, and production plan. With the goal of minimizing the sum of total operating costs and total carbon trading costs over at least one day in the future, and based on the energy load and the total output forecast, an optimization objective function for production optimization scheduling is established. The objective function is optimized to obtain the best hourly power supply control strategy for each energy conversion device in the multi-energy complementary system for at least one day in the future, which serves as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park.

2. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 1, characterized in that, The step of acquiring historical production operation data of the multi-energy complementary system of the industrial park to be optimized for production scheduling in order to train the deep neural network regression model of the multi-energy complementary system includes, after acquiring the historical production operation data, performing data preprocessing including data type conversion, filling missing data, data smoothing and outlier detection, and using the preprocessed historical production operation data to train the deep neural network regression model of the multi-energy complementary system.

3. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 2, characterized in that, The outlier detection employs a Hampel filter, which identifies and removes outliers from the data based on the median and median absolute deviation (MAD).

4. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 1, characterized in that, The power supply control strategy data for each energy conversion device includes: the electrical power supplied to the electric refrigeration equipment, the natural gas power supplied to the gas boiler, the thermal power supplied to the absorption refrigeration equipment, the natural gas power supplied to the combined heat and power equipment, and the power of electricity purchased from outside the park; the overall output data of the multi-energy complementary system includes: the overall available power supply of the multi-energy complementary system, the overall available heat supply of the multi-energy complementary system, and the overall available cooling supply of the multi-energy complementary system.

5. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 4, characterized in that, The purchased energy price includes: purchased electricity price and purchased natural gas price; the energy load required for production in the park includes: electrical load, heat load and cooling load.

6. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 5, characterized in that, The overall operating cost is calculated using the purchased energy price for each period and the energy consumed under the energy conversion equipment power supply control strategy for each period. The overall carbon trading cost is calculated using the carbon emission factor of the purchased natural gas power per unit of power, the carbon emission quota of the purchased natural gas power per unit of power, the carbon emission factor of the purchased electricity power per unit of power, the carbon emission quota of the purchased electricity power per unit of power, the carbon dioxide trading unit price for each period, the amount of purchased natural gas for each period, and the amount of purchased electricity for each period.

7. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 5, characterized in that, The step of establishing an optimization objective function for production optimization scheduling, with the objective of minimizing the sum of total operating costs and carbon trading costs over at least one day in the future, and setting constraints based on the energy load and the total output forecast, specifically involves introducing a penalty function to set the constraints. The resulting optimization objective function is as follows: ； The penalty function is: m is the penalty coefficient, and a refers to the first parameter in each penalty term, which actually corresponds to... , or ; b refers to the second parameter in each penalty term, which actually corresponds to , or , This represents the operating cost of the multi-energy complementary system in the i-th time period. This represents the carbon trading cost of a multi-energy complementary system in time period i. This represents the cooling load required for production in the park during the i-th time period. This represents the heat load required for production in the park during the i-th time period. This represents the electrical load required for production in the park during the i-th time period. Let i be the total available cooling power of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period. Let i be the total available thermal power of the multi-energy complementary system under the control strategy of the energy conversion equipment in the i-th time period. Let represent the total power supply of the multi-energy complementary system under the energy conversion equipment control strategy in the i-th time period.

8. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 1, characterized in that, In the step of optimizing the objective function to obtain the best hourly energy supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future as the day-ahead low-carbon economic production optimization scheduling scheme of the multi-energy complementary system in the park, the gray wolf optimization algorithm is used for optimization.

9. The day-ahead low-carbon economic production optimization scheduling method for a multi-energy complementary system in a park according to claim 8, characterized in that, In the optimization objective function, the position of the gray wolf is the hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, including the electric power supplied to the electric refrigeration equipment, the natural gas power supplied to the gas boiler, the natural gas power supplied to the cogeneration equipment, the heat power supplied to the absorption refrigeration equipment, and the power purchased from outside the park for each time period of the future day.

10. A low-carbon economic production optimization scheduling system for a multi-energy complementary system in a park, used to execute the method described in any one of claims 1 to 9, characterized in that, include: The model building and training module is used to construct a deep neural network regression model for a multi-energy complementary system by using a deep convolutional neural network, taking the energy supply control strategies of each energy conversion device in the multi-energy complementary system as input and the overall output of the multi-energy complementary system as output; and to obtain historical production operation data of the multi-energy complementary system of the industrial park to be optimized and scheduled for production in order to train the deep neural network regression model of the multi-energy complementary system; wherein, the historical production operation data includes energy supply control strategy data of each energy conversion device under various variable operating conditions and overall output data of the multi-energy complementary system; The output prediction module is used to calculate the time-sharing overall output prediction value of the multi-energy complementary system under the energy supply control strategy of each energy conversion device in the multi-energy complementary system, through a trained deep neural network regression model of the multi-energy complementary system for at least one day in the future. The load forecasting module is used to generate the energy load required for the park's production for at least one day in the future, based on the time-of-use external energy price, time-of-use carbon trading price, and production plan for at least one day in the future. The objective function establishment module is used to establish an optimization objective function for production optimization scheduling with the goal of minimizing the sum of the total operating cost and the total carbon trading cost over at least one day in the future, and based on the energy load and the total output forecast, setting constraints. The optimization module is used to optimize the objective function and obtain the best hourly power supply control strategy for each energy conversion device of the multi-energy complementary system for at least one day in the future, which serves as the day-ahead low-carbon economic production optimization scheduling scheme for the multi-energy complementary system in the park.

Images