Key parameter data mining and long-term configuration prediction method and system for comprehensive energy

Abstract

The present invention discloses a method and system for mining key parameter data and long-term configuration prediction for comprehensive energy, adopts a key parameter information extraction method to obtain the latest parameters, and uses a deep clustering algorithm to cluster the parameters of comprehensive energy equipment, constructs a comprehensive energy equipment library and an internal optimization configuration model for quantitative economic efficiency of energy supply hubs; obtains typical energy consumption characteristics in different seasons of the base year; uses a source-load matching selection modeling method to optimize and combine four typical energy hubs, and obtains a configuration scheme that meets various energy needs in different seasons; predicts the parameter change trend of typical equipment, and then uses a variety of models to perform life cycle curve fitting methods to perform long-term prediction of energy demand; finally, the obtained prediction results are optimized and combined using a source-load matching selection modeling method to obtain a long-term prediction configuration scheme. The present invention improves the prediction accuracy, has automatic parameter recognition and automatic correction, and makes the prediction results more in line with the actual situation.

Description

Technical Field

[0001] The present invention belongs to the technical field of data mining and long-term configuration prediction, and specifically relates to a method and system for key parameter data mining and long-term configuration prediction for comprehensive energy. Background Art

[0002] In recent years, integrated energy systems have become a new research hotspot and frontier. In integrated energy systems, energy hubs are defined as an input-output port model used to describe the exchange and coupling relationship between energy, loads, and networks in multi-energy systems. The coupling matrix describing the input energy and output load ports can briefly represent the various coupling relationships such as conversion, storage, and transmission between various forms of energy such as electricity, heat, and gas, and plays an important role in the planning and operation research of multi-energy systems. In the process of modeling multi-energy systems using energy hubs, there are many equipment parameters and the energy input and output relationship is complex. How to optimize the modeling is a current concern; at the same time, in the long-term forecast of integrated energy, how to improve the forecast accuracy is also a relatively important research point. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for key parameter data mining and long-term configuration prediction for integrated energy in view of the deficiencies in the above-mentioned prior art. By combining data mining technology, the multi-dimensional input parameters of the model are automatically extracted, clustered and predicted, and the modeling is optimized. A quantitative internal optimization configuration model and a source-load matching model for the economic efficiency of the energy supply hub are established, thereby realizing the long-term configuration prediction of integrated energy.

[0004] The present invention adopts the following technical solutions:

[0005] The key parameter data mining and long-term configuration prediction method for comprehensive energy includes the following steps:

[0006] S1. Use convolutional neural network to classify network information, then use recurrent neural network to extract key parameter information, use deep clustering method to cluster key parameter information, and build a comprehensive energy equipment library;

[0007] S2. Couple different integrated energy equipment from the integrated energy equipment library constructed in step S1, construct an economic quantitative internal optimization configuration model for the energy supply hub, solve the economic quantitative internal optimization configuration model for the energy supply hub under the condition of maximum profit, and obtain the optimal configuration schemes for four energy hubs including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas type energy hub, the electricity-cold / heat type energy hub, and the electricity-gas-hydrogen type energy hub;

[0008] S3. Use clustering feature extraction technology to pre-process the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtain the typical energy consumption characteristics of different parks at different times in the M year through clustering;

[0009] S4. Adopt the selection modeling method of park source-load matching, according to the typical energy consumption characteristics of different parks at different times in the Mth year obtained in step S3, select the optimal selection scheme of the energy hub obtained in step S2 to optimize the energy supply, solve the park source-load matching model under the condition of the lowest annual cost of the park, and obtain the quantity ratio of various energy supply hubs that meet various energy consumption needs in different seasons; add up the configurations of all parks to obtain the comprehensive energy configuration plan for the corresponding region in the Mth year;

[0010] S5. Using the key parameter information of the equipment in step S1, a time series prediction method is used to convert the time series into a curve shape, and a convolutional neural network model is used for analysis to predict the change trend of key parameter information of typical equipment from the M+1 year to the end of the Nth year;

[0011] S6. Use the life cycle curve fitting method to predict the energy demand of the park, and predict the development demand of various types of energy in the park from the M+1 year to the end of the Nth year;

[0012] S7. Taking the typical equipment parameter change trend predicted in step S5 from the M+1th year to the end of the Nth year and the various energy development demands of the park from the M+1th year to the end of the Nth year predicted in step S6 as input parameters, the park source-load matching selection modeling method of step S4 is used to predict the long-term comprehensive energy configuration plan for the region from the M+1th year to the end of the Nth year.

[0013] Specifically, in step S1, the key parameter information is extracted by the recurrent neural network as follows:

[0014] The web page text data is preprocessed to obtain word vectors, which are input into the convolutional neural network for model training. A text classifier is built, and the data to be extracted is labeled. The BIO sequence and text word vectors are input into the recurrent neural network for training to obtain an information extraction model, which can extract the key unstructured data in the text into structured data and obtain key equipment parameters including capacity, power, energy conversion efficiency, life, investment cost, maintenance cost and expected net residual value.

[0015] Specifically, in step S2, the annual cost of the electricity-cold / heat-gas-hydrogen type energy hub includes:

[0016] The production energy cost is:

[0017]

[0018] The fixed construction cost of the energy hub is:

[0019]

[0020] The annual revenue of the electricity-cold / heat-gas-hydrogen energy hub includes:

[0021] The revenue of energy supply service of energy hub is:

[0022]

[0023] The revenue from energy hub backup services is:

[0024]

[0025]

[0026]

[0027] The revenue from the electric vehicle charging pile power supply service is:

[0028]

[0029] The annual cost of the electricity-cold / heat-gas energy hub includes: The production and purchase cost is:

[0030]

[0031] The fixed construction cost of the energy hub is:

[0032]

[0033] The annual revenue of the electricity-cold / heat-gas type energy hub includes: The energy hub energy supply service revenue is:

[0034]

[0035] The revenue from energy hub backup services is:

[0036]

[0037]

[0038]

[0039] The revenue from the electric vehicle charging pile power supply service is:

[0040]

[0041] The annual cost of the electricity-cold / heat energy hub includes: The production and purchase cost of energy is:

[0042]

[0043] The fixed construction cost of the energy hub is:

[0044]

[0045] The annual revenue of the electric-cold / heat energy hub includes: The energy hub energy supply service revenue is:

[0046]

[0047] The revenue from energy hub backup services is:

[0048]

[0049]

[0050] The revenue from the electric vehicle charging pile power supply service is:

[0051]

[0052] The annual cost of the electricity-gas-hydrogen energy hub includes: The production and purchase cost is:

[0053]

[0054] The fixed construction cost of the energy hub is:

[0055]

[0056] The annual revenue of the Electricity-Gas-Hydrogen Energy Hub includes:

[0057] The revenue of energy supply service of energy hub is:

[0058]

[0059] The revenue from energy hub backup services is:

[0060]

[0061]

[0062]

[0063] The revenue from the electric vehicle charging pile power supply service is:

[0064]

[0065] Among them, EH1, EH2, EH3, and EH4 are respectively the energy hubs of electricity-cold / heat-gas-hydrogen type, electricity-cold / heat-gas type, electricity-cold / heat type, and electricity-gas-hydrogen type; D is the total number of days in a year; N wis the number of typical scenes in different seasons of the year; w is the typical scenes in different seasons of the year, which are summer typical scenes, transitional season typical scenes and winter typical scenes; T is the total time period in a day; t is the specific time of the day; p w is the probability of occurrence of scenes in different seasons; c e 、c g prices for external electricity and natural gas purchases; is the amount of electricity purchased from the external main grid by each energy hub during period t in scenario w, is the natural gas purchase amount of each energy hub in time period t under scenario w, I is the set of energy equipment types included in the electricity-cold / heat-gas-hydrogen type energy hub, GSHP-ground source heat pump, EB-electric boiler, ISAC-ice storage air conditioner, PV-photovoltaic, EVSE-electric vehicle charging station; i is the energy equipment type included in the electricity-cold / heat-gas-hydrogen type energy hub; is the annualized initial investment cost per unit capacity and annual fixed maintenance cost per unit capacity corresponding to each energy equipment; V i The standard capacity for each type of equipment; is the number of devices configured in this type of energy hub; e , c , h , g , The corresponding energy supply prices for electricity, cold energy, heat and cold energy, natural gas and hydrogen energy; The price of electricity for charging stations; It is the annual unit capacity reserve service price of electricity and the annual unit capacity reserve service price of natural gas. Revenue from electricity supply services for electric vehicle charging stations; Revenue from power reserve services for each energy hub; Revenue from natural gas energy backup services for various energy hubs; The amount of electricity provided by each energy hub during period t in scenario w; The cold energy provided by each energy hub to the outside during period t in scenario w; The thermal energy provided by each energy hub to the outside during period t under scenario w; is the amount of natural gas provided by each energy hub during period t under scenario w; The hydrogen energy provided by each energy hub during period t under scenario w; is the power supply of electric vehicle charging piles in each energy hub w scenario during period t.

[0066] Furthermore, the demand and supply constraints of electricity, cold energy, heat energy, natural gas and hydrogen energy in the electricity-cold / heat-gas-hydrogen energy hub are as follows:

[0067] Electricity-cold / heat-gas-hydrogen type energy hub power load supply constraints:

[0068]

[0069] Cooling load supply constraints of electricity-cooling / heat-gas-hydrogen type energy hub:

[0070]

[0071] Heat load supply constraints of electricity-cold / heat-gas-hydrogen energy hub:

[0072]

[0073] Natural gas load supply constraints for power-cooling / heat-gas-hydrogen energy hubs:

[0074]

[0075] Hydrogen load supply constraints for electricity-cold / heat-gas-hydrogen type energy hubs:

[0076]

[0077] For the power balance constraints of electricity and natural gas in the power-cold / heat-gas-hydrogen type energy hub, the following relationship exists:

[0078] Power balance constraints:

[0079]

[0080]

[0081] Natural gas balance constraints:

[0082]

[0083]

[0084] The demand and supply constraints for electricity, cooling energy, heat energy and natural gas in the electricity-cooling / heat-gas energy hub are as follows:

[0085] Electricity-cold / heat-gas type energy hub power load supply constraints:

[0086]

[0087] Cooling load supply constraints of electricity-cooling / heat-gas type energy hub:

[0088]

[0089] Heat load supply constraints of electricity-cold / heat-gas type energy hub:

[0090]

[0091] Natural gas load supply constraints for power-cooling / heat-gas type energy hubs:

[0092]

[0093] For the power-cold / heat-gas type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0094] Power balance constraints:

[0095]

[0096]

[0097] Natural gas balance constraints:

[0098]

[0099]

[0100] The supply constraints of electricity, cooling and heat energy in the electricity-cooling / heating energy hub are as follows:

[0101] Electricity-cold / heat energy hub power load supply constraints:

[0102]

[0103] Cooling load supply constraints of electricity-cooling / heating energy hub:

[0104]

[0105] Heat load supply constraints of electricity-cold / heat type energy hub:

[0106]

[0107] For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0108] Power balance constraints:

[0109]

[0110]

[0111] Natural gas balance constraints:

[0112]

[0113]

[0114] The supply constraints for electricity, natural gas and hydrogen in the electricity-gas-hydrogen energy hub are as follows:

[0115] Electricity-gas-hydrogen energy hub power load supply constraints:

[0116]

[0117] Natural gas load supply constraints for electricity-gas-hydrogen energy hubs:

[0118]

[0119] Hydrogen load supply constraints for electricity-gas-hydrogen energy hubs:

[0120]

[0121] For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0122] Power balance constraints:

[0123]

[0124]

[0125] Natural gas balance constraints:

[0126]

[0127]

[0128] Specifically, in step S4, the annual cost of the park includes:

[0129] The production and purchase costs of energy for all types of energy hubs in the region:

[0130]

[0131] The investment and construction costs of all types of energy hubs in the region:

[0132]

[0133] Among them, Park is four types of parks, including school parks, industrial parks, residential parks and commercial parks; The corresponding amount of electricity purchased from the external main grid by each type of energy hub during period t; is the corresponding natural gas purchase volume of each type of energy hub during period t; The optimal number of four energy hubs to be built in different parks; The total annual investment and construction costs of the four types of energy hubs; The load demand corresponding to the electricity in each park period; The load demand corresponding to the electric vehicle charging piles in each park period; The upper limit of hydrogen supply ratio for each park.

[0134] Furthermore, the corresponding load demand constraints for electricity, cooling energy, heat energy, natural gas, hydrogen energy and electric vehicle charging piles in the park in each season are as follows:

[0135] Park power load supply constraints:

[0136]

[0137] Park cooling load supply constraints:

[0138]

[0139] Park heat load supply constraints:

[0140]

[0141] Park gas load supply constraints:

[0142]

[0143] Hydrogen load supply constraints in the park:

[0144]

[0145] The power supply of electric vehicle charging piles in the park meets the demand constraints:

[0146]

[0147] Specifically, in step S5, in the time series prediction method, a convolutional neural network model is used to predict the changing trend of equipment parameters.

[0148] Specifically, in step S6, multiple models are used to fit the life cycle curve:

[0149] Three life cycle functions are selected for fitting, and the historical energy demand data of each park are input into the Bass model, Gomperz model and polynomial model, and the prediction results of different models are output respectively.

[0150] Another technical solution of the present invention is a comprehensive energy key parameter data mining and long-term configuration prediction system, comprising:

[0151] The data module uses a convolutional neural network to classify network information, then extracts key parameter information through a recurrent neural network, clusters the key parameter information using a deep clustering method, and builds a comprehensive energy equipment library;

[0152] The solution module couples different integrated energy equipment from the integrated energy equipment library constructed by the data module, constructs an economic quantitative internal optimization configuration model for the energy supply hub, solves the economic quantitative internal optimization configuration model for the energy supply hub under the condition of maximum profit, and obtains the optimal configuration schemes for four energy hubs including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas type energy hub, the electricity-cold / heat type energy hub, and the electricity-gas-hydrogen type energy hub;

[0153] The preprocessing module uses clustering feature extraction technology to preprocess the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtains the typical energy consumption characteristics of different parks at different times in the M year through clustering;

[0154] The configuration module adopts the selection modeling method of park source-load matching. According to the typical energy consumption characteristics of different parks at different times in the Mth year obtained by the preprocessing module, the optimal configuration scheme of the energy hub obtained by the solution module is selected to optimize the energy supply. The park source-load matching model is solved under the condition of the lowest annual cost of the park, and the quantity ratio of various energy supply hubs that meet various energy demands in different seasons is obtained; the configuration of all parks is added up to obtain the comprehensive energy configuration plan of the corresponding region in the Mth year;

[0155] The conversion module uses the key parameter information extracted from the data module, uses the time series prediction method to convert the time series into a curve shape, and uses the convolutional neural network model for analysis to predict the change trend of key parameter information of typical equipment from the M+1 year to the end of the Nth year;

[0156] The fitting module uses the life cycle curve fitting method to predict the energy demand of the park and predict the development demand of various types of energy in the park from the M+1 year to the end of the Nth year;

[0157] The prediction module takes the changing trend of key parameter information of typical equipment predicted by the conversion module from the M+1 year to the end of the N year and the various energy development demands of the park predicted by the fitting module from the M+1 year to the end of the N year as input parameters, and uses the park source-load matching selection modeling method of the configuration module to predict the long-term comprehensive energy configuration plan for the region from the M+1 year to the end of the N year.

[0158] Compared with the prior art, the present invention has at least the following beneficial effects:

[0159] The key parameter data mining and long-term configuration prediction method for comprehensive energy of the present invention adopts a key parameter information extraction method to obtain the latest parameters, and uses a deep clustering algorithm to cluster the comprehensive energy equipment parameters, thereby standardizing the input parameters of the model and improving the accuracy of the calculation results; then the model is optimized to construct an internal optimization configuration model for the economic quantification of the energy supply hub and a source-load matching selection model, so that the calculated results are more practical; finally, the time series prediction method and the life cycle curve fitting method are used to respectively make long-term predictions of equipment parameters and energy demand, and the prediction results are input into the above-mentioned selection model, which provides a credible reference for equipment selection and energy supply plan selection, thereby improving business profits.

[0160] Furthermore, a real-time key parameter information extraction model is proposed. By assigning different weights to the three parts of data, namely title, keyword, and content, and inputting them into convolutional neural networks of different scales, a multi-input text classifier model is constructed, which makes the accuracy of parameter information extraction higher.

[0161] Furthermore, in the constructed internal optimization configuration model for the economic quantification of energy hubs, the cost structure of the energy hubs is fully considered when calculating the annual costs of the four energy hubs. The annual costs are divided into two parts, including the production energy purchase costs and the fixed investment and construction costs of the energy hubs. At the same time, in each part of the cost, the costs are broken down to each type of equipment, so that the model can be more in line with the actual situation.

[0162] Furthermore, within the constraints of the internal optimization configuration model for the economic quantification of the energy hub, the sum of the electricity, cooling, heat, gas, and hydrogen energy outputs of all equipment in the energy hub model must not be less than the energy demand of the four energy hubs. When this constraint is met, the optimal configuration scheme obtained is more reasonable.

[0163] Furthermore, in the selection modeling method of park source-load matching, the annual cost is also fully taken into consideration. The annual cost is divided into two parts, including the production and purchase cost of all types of energy hubs in the region and the fixed investment and construction cost, so that the calculated structure is more in line with the actual situation.

[0164] Furthermore, in the model constraints of the selection modeling of source-load matching in the park, it is necessary to ensure that the sum of the electricity, cooling, heat, gas and hydrogen energy output of all energy hubs in the model is not less than the actual energy demand of the park. When this constraint is met, the optimal solution is more reasonable.

[0165] Furthermore, the time series prediction method is used to predict the changing trends of typical equipment parameters. The time series is converted into a curve shape, and then the convolutional neural network-based model is applied for analysis to predict the changing trends of typical equipment parameters from the M+1 year to the end of the Nth year.

[0166] Furthermore, by fitting three common life cycle curves, the optimal modeling results under different life cycle curves can be obtained, so as to compare and analyze the results and draw more comprehensive conclusions.

[0167] In summary, the present invention optimizes multiple links based on data mining and prediction algorithms, which improves the prediction accuracy compared to traditional methods; at the same time, with the increase of information and the passage of time, the model has automatic identification and automatic correction of parameters, so that the prediction results can be more in line with the actual situation.

[0168] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0169] Figure 1 It is a general flow chart of the key parameter data mining and long-term configuration prediction method for integrated energy;

[0170] Figure 2 It is an energy hub diagram (taking the electricity-cold / heat-gas-hydrogen energy hub as an example);

[0171] Figure 3 Detailed diagram of the multi-input text classifier model in the key parameter information extraction method;

[0172] Figure 4 This is a summary diagram of the energy hub benefits under the optimal investment and construction quantity in each region in Example 1;

[0173] Figure 5 It is a curve chart of income changes in each region during the 20-year life cycle of each region in Example 1;

[0174] Figure 6 The cost change curve of each region in Example 1 during the 20-year life cycle of each region. DETAILED DESCRIPTION

[0175] The present invention provides a method for data mining and long-term configuration prediction of key parameters for comprehensive energy. The method adopts a key parameter information extraction method to obtain the latest parameters, and uses a deep clustering algorithm to cluster the parameters of comprehensive energy equipment to construct a comprehensive energy equipment library; secondly, a quantitative internal optimization configuration model for the economic efficiency of energy supply hubs is constructed; then, a clustering feature extraction technology is used to obtain the typical energy consumption characteristics of different seasons in the base year; then, a source-load matching selection modeling method is used to optimize and combine four typical energy hubs to obtain a configuration scheme that meets various energy consumption requirements in different seasons; then, the parameter change trend of typical equipment is predicted, and then a variety of models are used to perform life cycle curve fitting methods to perform long-term prediction of energy demand; finally, the obtained prediction results are optimized and combined using a source-load matching selection modeling method to obtain a long-term prediction configuration scheme. Based on data mining and prediction algorithms, the present invention optimizes multiple links, and improves the prediction accuracy compared to traditional methods. At the same time, with the increase of information and the development of time, the model has automatic parameter recognition and automatic correction, so that the prediction results can be more in line with the actual situation.

[0176] See also Figure 1 , the present invention provides a method for data mining and long-term configuration prediction of key parameters for comprehensive energy, optimizes the configuration of internal comprehensive energy equipment for various types of standardized energy hubs in the park, and determines the optimal configuration quantity of internal energy equipment for various types of standardized energy hubs; then, based on the former, optimizes the number of four types of standardized energy hubs in each region as a whole, and determines the optimal number of four types of standardized energy hubs in school parks, industrial parks, residential parks and commercial parks. Long-term predictions are made for various equipment parameters and energy development needs, and the optimal configuration plan of the energy hub and the long-term prediction parameter results are selected and modeled using the park source-load matching selection modeling method to obtain the predicted configuration results. The specific steps are as follows:

[0177] S1. A key parameter information extraction method based on real-time network information is proposed. A convolutional neural network is used to classify network information, and then a recurrent neural network is used to extract key parameter information. A deep clustering method is used to cluster a large number of equipment parameters to build a comprehensive energy equipment library.

[0178] The key parameter information extraction method based on real-time network information is as follows: preprocessing the web page text data to obtain word vectors, inputting them into the convolutional neural network for model training, building a text classifier, and then manually annotating the data to be extracted. The BIO sequence and text word vectors are input into the recurrent neural network for training to obtain the information extraction model, and finally the key unstructured data in the text is extracted into structured data to obtain various equipment model parameters.

[0179] During clustering, the VaDE algorithm is used to cluster the parameters of various models of equipment to obtain representative standardized equipment parameters.

[0180] Key equipment parameters include: capacity, power, energy conversion efficiency, life span, investment cost, maintenance cost and estimated net residual value.

[0181] S2. Construct an internal optimization configuration model for the economic quantification of the energy supply hub, and couple different integrated energy equipment from the integrated energy equipment library obtained in step S1. Figure 3 , construct a model to obtain the optimal configuration schemes for four typical energy hubs, including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas energy hub, the electricity-cold / heat energy hub, and the electricity-gas-hydrogen energy hub; solve the model under the condition of maximum profit to obtain the optimal equipment configuration scheme for the energy hub;

[0182] In the internal optimization configuration model of energy supply hub economic quantification, for four standardized energy hubs, namely "electricity-cold / heat-gas-hydrogen type", "electricity-cold / heat-gas type", "electricity-cold / heat type" and "electricity-gas-hydrogen type", the economic quantification analysis modeling of each is as follows:

[0183] (1) Electricity-cold / heat-gas-hydrogen energy hub

[0184] See also Figure 2 The annual cost of the electricity-cold / heat-gas-hydrogen energy hub includes the production and purchase costs and the fixed investment and construction costs of the energy hub, which are as follows:

[0185] 1) The production energy purchase cost is:

[0186]

[0187] 2) The fixed investment and construction cost of the energy hub is:

[0188]

[0189] The annual revenue of the electricity-cold / heat-gas-hydrogen type energy hub includes the energy hub supply service revenue, energy hub backup service revenue and electric vehicle charging pile power supply service revenue, specifically:

[0190] 1) The revenue from energy hub supply services is:

[0191]

[0192] 2) Energy hub backup service income is:

[0193]

[0194] 3) The revenue from the electric vehicle charging station power supply service is:

[0195]

[0196] The demand and supply constraints of electricity, cold energy, heat energy, natural gas and hydrogen energy in the electricity-cold / heat-gas-hydrogen energy hub are as follows:

[0197] 1) Electricity-cold / heat-gas-hydrogen type energy hub power load supply constraints:

[0198]

[0199] 2) Cooling load supply constraints of electricity-cooling / heat-gas-hydrogen type energy hub:

[0200]

[0201] 3) Heat load supply constraints of electricity-cold / heat-gas-hydrogen type energy hub:

[0202]

[0203] 4) Natural gas load supply constraints for power-cooling / heat-gas-hydrogen energy hubs:

[0204]

[0205] 5) Hydrogen load supply constraints for electricity-cold / heat-gas-hydrogen type energy hubs:

[0206]

[0207] For the power balance constraints of electricity and natural gas in the power-cold / heat-gas-hydrogen type energy hub, the following relationship exists:

[0208] 1) Power balance constraints:

[0209]

[0210]

[0211] 2) Natural gas balance constraints:

[0212]

[0213]

[0214] (2) Electricity-cold / heat-gas energy hub

[0215] The annual cost of an electricity-cold / heat-gas energy hub includes the production and purchase costs and the fixed investment and construction costs of the energy hub, which are as follows:

[0216] 1) The production energy purchase cost is:

[0217]

[0218] 2) The fixed investment and construction cost of the energy hub is:

[0219]

[0220] The annual revenue of the electricity-cold / heat-gas type energy hub includes the energy hub supply service revenue, energy hub backup service revenue and electric vehicle charging pile power supply service revenue, specifically:

[0221] 1) The revenue from energy hub supply services is:

[0222]

[0223] 2) Energy hub backup service income is:

[0224]

[0225] 3) The revenue from the electric vehicle charging station power supply service is:

[0226]

[0227] The demand and supply constraints for electricity, cooling energy, heat energy and natural gas in the electricity-cooling / heat-gas energy hub are as follows:

[0228] 1) Electricity-cold / heat-gas type energy hub power load supply constraints:

[0229]

[0230] 2) Cooling load supply constraints of electricity-cold / heat-gas type energy hubs:

[0231]

[0232] 3) Heat load supply constraints of electricity-cold / heat-gas type energy hubs:

[0233]

[0234] 4) Natural gas load supply constraints for power-cooling / heat-gas type energy hubs:

[0235]

[0236] For the power-cold / heat-gas type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0237] 1) Power balance constraints:

[0238]

[0239]

[0240] 2) Natural gas balance constraints:

[0241]

[0242]

[0243] (3) Electricity-cold / heat energy hub

[0244] The annual cost of an electricity-cold / heat energy hub includes the production and purchase costs and the fixed investment and construction costs of the energy hub, which are as follows:

[0245] 1) The production energy purchase cost is:

[0246]

[0247] 2) The fixed investment and construction cost of the energy hub is:

[0248]

[0249] The annual revenue of the electric-cold / heat energy hub includes the energy hub supply service revenue, energy hub backup service revenue and electric vehicle charging pile power supply service revenue, specifically:

[0250] 1) The revenue from energy hub supply services is:

[0251]

[0252] 2) Energy hub backup service income is:

[0253]

[0254] 3) The revenue from the electric vehicle charging station power supply service is:

[0255]

[0256] The supply constraints of electricity, cooling and heat energy in the electricity-cooling / heating energy hub are as follows:

[0257] 1) Electricity-cold / heat energy hub power load supply constraints:

[0258]

[0259] 2) Cooling load supply constraints of electricity-cooling / heating energy hubs:

[0260]

[0261] 3) Heat load supply constraints of electricity-cold / heat type energy hub:

[0262]

[0263] For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0264] 1) Power balance constraints:

[0265]

[0266]

[0267] 2) Natural gas balance constraints:

[0268]

[0269]

[0270] (4) Electricity-gas-hydrogen energy hub

[0271] The annual cost of an electricity-gas-hydrogen energy hub includes the production and purchase costs and the fixed investment and construction costs of the energy hub, which are as follows:

[0272] 1) The production energy purchase cost is:

[0273]

[0274] 2) The fixed investment and construction cost of the energy hub is:

[0275]

[0276] The annual revenue of the electricity-gas-hydrogen energy hub includes the energy hub supply service revenue, energy hub backup service revenue and electric vehicle charging pile power supply service revenue, specifically:

[0277] 1) The revenue from energy hub supply services is:

[0278]

[0279] 2) Energy hub backup service income is:

[0280]

[0281] 3) The revenue from the electric vehicle charging station power supply service is:

[0282]

[0283] The supply constraints for electricity, natural gas and hydrogen in the electricity-gas-hydrogen energy hub are as follows:

[0284] 1) Electricity-gas-hydrogen energy hub power load supply constraints:

[0285]

[0286] 2) Natural gas load supply constraints for electricity-gas-hydrogen energy hubs:

[0287]

[0288] 3) Hydrogen load supply constraints of electricity-gas-hydrogen energy hubs:

[0289]

[0290] For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship:

[0291] 1) Power balance constraints:

[0292]

[0293]

[0294] 2) Natural gas balance constraints:

[0295]

[0296]

[0297] Among them, EH1, EH2, EH3, and EH4 are respectively the energy hubs of electricity-cold / heat-gas-hydrogen type, electricity-cold / heat-gas type, electricity-cold / heat type, and electricity-gas-hydrogen type; D is the total number of days in a year; N w is the number of typical scenes in different seasons of the year; w is the typical scenes in different seasons of the year, which are summer typical scenes, transitional season typical scenes (spring and autumn) and winter typical scenes; T is the total time period in a day; t is the specific time of the day; p w is the probability of occurrence of scenes in different seasons; c e 、c g prices for external electricity and natural gas purchases; The amount of electricity purchased by each energy hub from the external main grid during period t under scenario w. is the natural gas purchase amount of each energy hub in time period t under scenario w. I is the set of energy equipment types included in the electricity-cold / heat-gas-hydrogen type energy hub, specifically: GSHP-ground source heat pump, CCHP-CCHP, EB-electric boiler, ISA C-ice storage air conditioner, P2G-P2G, PV-photovoltaic, EVSE-electric vehicle charging pile; i is the energy equipment type included in the electricity-cold / heat-gas-hydrogen type energy hub, and its value is: GSHP, CCHP, EB, ISAC, P2G, PV, EVSE; is the annualized initial investment cost per unit capacity and annual fixed maintenance cost per unit capacity corresponding to each energy equipment; V i The standard capacity for each type of equipment; is the number of devices configured in this type of energy hub; e , c , h , g , The corresponding energy supply prices for electricity, cold energy, heat and cold energy, natural gas and hydrogen energy; The price of electricity for charging stations; It is the annual unit capacity reserve service price of electricity and the annual unit capacity reserve service price of natural gas. Revenue from electricity supply services for electric vehicle charging stations; Revenue from power reserve services for each energy hub; Revenue from natural gas energy backup services for various energy hubs; The amount of electricity provided by each energy hub during period t in scenario w; The cold energy provided by each energy hub to the outside during period t in scenario w; The thermal energy provided by each energy hub to the outside during period t under scenario w; is the amount of natural gas provided by each energy hub during period t under scenario w; The hydrogen energy provided by each energy hub during period t under scenario w; is the power supply of electric vehicle charging piles in each energy hub w scenario during period t.

[0298] S3. Use clustering feature extraction technology to pre-process the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtain the typical energy consumption characteristics of different parks at different times in the four seasons through clustering;

[0299] The load data is clustered through a deep clustering algorithm, and finally the load characteristics of four types of parks, namely schools, commercial, residential and industrial parks, at different times in the four seasons of spring, summer, autumn and winter are obtained.

[0300] S4. A selection modeling method for park source-load matching is proposed. According to the typical energy consumption characteristics of different parks at different times in the Mth year obtained in step S3, a typical energy hub solution in step S2 is selected to optimize energy supply. The park source-load matching model is solved under the condition of the lowest annual cost of the park, and the quantity ratio and minimum equipment cost of various energy supply hubs that meet various energy demands in different seasons are obtained; the configurations of all parks in the region are further added up to obtain the comprehensive energy configuration plan for the corresponding region in the Mth year;

[0301] The annual cost of the park includes the production and purchase costs of all types of energy hubs in the region and the investment and construction costs of all types of energy hubs in the region, specifically:

[0302] 1) Production and purchase costs of energy at all types of energy hubs in the region:

[0303]

[0304] 2) Investment and construction costs of all types of energy hubs in the region:

[0305]

[0306] The corresponding load demand constraints for electricity, cooling energy, heat energy, natural gas, hydrogen energy and electric vehicle charging piles in the park in each season are as follows:

[0307] 1) The power load supply constraints of the park are as follows:

[0308]

[0309] 2) The cooling load supply constraints of the park are as follows:

[0310]

[0311] 3) The heat load supply constraints of the park are as follows:

[0312]

[0313] 4) The gas load supply constraints of the park are as follows:

[0314]

[0315] 5) The hydrogen load supply constraints of the park are as follows:

[0316]

[0317] 6) The power supply of electric vehicle charging piles in the park meets the following demand constraints:

[0318]

[0319] Among them, Park is four types of parks, including school parks, industrial parks, residential parks and commercial parks; The corresponding amount of electricity purchased from the external main grid by each type of energy hub during period t; is the corresponding natural gas purchase volume of each type of energy hub during period t; The optimal number of four energy hubs to be built in different parks; The total annual investment and construction costs of the four types of energy hubs; The load demand corresponding to the electricity in each park period; The load demand corresponding to the electric vehicle charging piles in each park period; The upper limit of hydrogen supply ratio for each park.

[0320] S5. Use the time series prediction method to predict the change trend of equipment parameters, obtain the key parameter information of various types of equipment in step S1, use the time series prediction method to convert the time series into a curve shape, and then use the convolutional neural network-based model for analysis to predict the change trend of typical equipment parameters from the M+1 year to the end of the Nth year;

[0321] In the time series prediction method, the convolutional neural network model is used to predict the changing trend of equipment parameters.

[0322] S6. Use multiple models to perform life cycle curve fitting methods to make long-term forecasts of energy demand in each park, and predict the development demand of various types of energy demand in the park from year M+1 to the end of year N;

[0323] A variety of models are used for life cycle curve fitting, and three life cycle functions are selected for fitting. The historical energy demand data of each park is input into the three models, and then the prediction results of different models are output respectively. The model selects functions commonly used for life cycle curve fitting, including: Bass model, Gomperz model, and polynomial model.

[0324] S7. Taking the typical equipment parameter change trend predicted in step S5 from the M+1th year to the end of the Nth year and the various energy development demands of the park from the M+1th year to the end of the Nth year predicted in step S6 as input parameters, the park source-load matching selection modeling method of step S4 is used to predict the long-term comprehensive energy configuration plan for the region from the M+1th year to the end of the Nth year.

[0325] In another embodiment of the present invention, a system for key parameter data mining and long-term configuration prediction for integrated energy is provided. The system can be used to implement the above-mentioned key parameter data mining and long-term configuration prediction method for integrated energy. Specifically, the system for key parameter data mining and long-term configuration prediction for integrated energy includes a data module, a solution module, a preprocessing module, a configuration module, a conversion module, a fitting module and a prediction module.

[0326] Among them, the data module uses convolutional neural network to classify network information, then extracts key parameter information through recurrent neural network, and uses deep clustering method to cluster key parameter information to build a comprehensive energy equipment library;

[0327] The solution module couples different integrated energy equipment from the integrated energy equipment library constructed by the data module, constructs an economic quantitative internal optimization configuration model for the energy supply hub, solves the economic quantitative internal optimization configuration model for the energy supply hub under the condition of maximum profit, and obtains the optimal configuration schemes for four energy hubs including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas type energy hub, the electricity-cold / heat type energy hub, and the electricity-gas-hydrogen type energy hub;

[0328] The preprocessing module uses clustering feature extraction technology to preprocess the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtains the typical energy consumption characteristics of different parks at different times in the M year through clustering;

[0329] The configuration module adopts the selection modeling method of park source-load matching. According to the typical energy consumption characteristics of different parks at different times in the Mth year obtained by the preprocessing module, the optimal configuration scheme of the energy hub obtained by the solution module is selected to optimize the energy supply. The park source-load matching model is solved under the condition of the lowest annual cost of the park, and the quantity ratio of various energy supply hubs that meet various energy demands in different seasons is obtained; the configuration of all parks is added up to obtain the comprehensive energy configuration plan of the corresponding region in the Mth year;

[0330] The conversion module uses the key parameter information extracted from the data module, uses the time series prediction method to convert the time series into a curve shape, and uses the convolutional neural network model for analysis to predict the change trend of key parameter information of typical equipment from the M+1 year to the end of the Nth year;

[0331] The fitting module uses the life cycle curve fitting method to predict the energy demand of the park and predict the development demand of various types of energy in the park from the M+1 year to the end of the Nth year;

[0332] The prediction module takes the changing trend of key parameter information of typical equipment predicted by the conversion module from the M+1 year to the end of the N year and the various energy development demands of the park predicted by the fitting module from the M+1 year to the end of the N year as input parameters, and uses the park source-load matching selection modeling method of the configuration module to predict the long-term comprehensive energy configuration plan for the region from the M+1 year to the end of the N year.

[0333] Example 1

[0334] In order to verify the long-term configuration prediction effect of the method of the present invention, Shaanxi Province was selected as the source range of experimental data, 2020 was taken as the base year, the deduction cycle was 20 years, and the deduction model was the Bass model. The equipment parameters were automatically extracted and long-term predicted. At the same time, the energy hub and the energy load demand of the park were feature extracted and long-term predicted. Finally, by solving the economic quantitative internal optimization configuration model and source-load matching model of the energy supply hub, a 20-year configuration selection plan was obtained, thereby realizing the comprehensive energy long-term configuration prediction. The specific results are as follows:

[0335] According to the model, the objective function is to minimize the cost of building energy hubs, combined with constraints such as energy supply and demand balance, to obtain the optimal number of investments in each region, the income of each region, and the cost of each region under the 20-year life cycle. The predicted results of energy hub benefits under the optimal number of investments in each region in 2040 are as follows: Figure 4 shown.

[0336] The revenue and cost of each region in the 20-year life cycle are as follows: Figure 5 , Figure 6 As shown. Figure 5 It can be seen that the income changes in various regions in Shaanxi Province show an S-shaped change pattern: slowly climbing in the early stage of the deduction, rising rapidly in the middle stage of the deduction, and the market profit tends to be saturated at the end of the deduction; the inflection point of the curve appears around 2030, when the market profit grows fastest and the profit scale is half of the steady-state scale. Specifically, in the steady state, the profit in northern Shaanxi is about 448.054 million yuan, the Guanzhong region is about 594.498 million yuan, and the southern Shaanxi region is about 214.032 million yuan. The profit scale of each region in the steady state is about twice that of the initial situation. According to the marginal effect of economic principles, as investment pours into the integrated energy market, the market economy scale will continue to tend to saturation, which shows that the deduction results have a certain rationality.

[0337] In summary, the present invention provides a method and system for key parameter data mining and long-term configuration prediction for integrated energy, which improves prediction accuracy by automatically extracting model input parameters and optimizing the modeling process. In the integrated energy business, it provides a reliable reference for equipment selection and energy supply solution selection, thereby increasing business profits.

[0338] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for mining key parameter data and long-term configuration prediction for comprehensive energy, characterized in that: The following steps are involved: S1. Use convolutional neural network to classify network information, then use recurrent neural network to extract key parameter information, use deep clustering method to cluster key parameter information, and build a comprehensive energy equipment library; S2. Couple different integrated energy equipment from the integrated energy equipment library constructed in step S1, construct an economic quantitative internal optimization configuration model for the energy supply hub, solve the economic quantitative internal optimization configuration model for the energy supply hub under the condition of maximum profit, and obtain the optimal configuration schemes for four energy hubs including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas type energy hub, the electricity-cold / heat type energy hub, and the electricity-gas-hydrogen type energy hub; S3. Use clustering feature extraction technology to pre-process the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtain the typical energy consumption characteristics of different parks at different times in the M year through clustering; S4. Adopt the selection modeling method of park source-load matching, according to the typical energy consumption characteristics of different parks at different times in the Mth year obtained in step S3, select the optimal selection scheme of the energy hub obtained in step S2 to optimize the energy supply, solve the park source-load matching model under the condition of the lowest annual cost of the park, and obtain the quantity ratio of various energy supply hubs that meet various energy consumption needs in different seasons; add up the configurations of all parks to obtain the comprehensive energy configuration plan for the corresponding region in the Mth year; S5. Using the key parameter information of the equipment in step S1, a time series prediction method is used to convert the time series into a curve shape, and a convolutional neural network model is used for analysis to predict the change trend of key parameter information of typical equipment from the M+1 year to the end of the Nth year; S6. Use the life cycle curve fitting method to predict the energy demand of the park, and predict the development demand of various types of energy in the park from the M+1 year to the end of the Nth year; S7. Taking the typical equipment parameter change trend predicted in step S5 from the M+1th year to the end of the Nth year and the various energy development demands of the park from the M+1th year to the end of the Nth year predicted in step S6 as input parameters, the park source-load matching selection modeling method of step S4 is used to predict the long-term comprehensive energy configuration plan for the region from the M+1th year to the end of the Nth year.

2. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 1 is characterized in that: In step S1, the key parameter information is extracted by the recurrent neural network as follows: The web page text data is preprocessed to obtain word vectors, which are input into the convolutional neural network for model training. A text classifier is built, and the data to be extracted is labeled. The BIO sequence and text word vectors are input into the recurrent neural network for training to obtain an information extraction model, which can extract the key unstructured data in the text into structured data and obtain key equipment parameters including capacity, power, energy conversion efficiency, life, investment cost, maintenance cost and expected net residual value.

3. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 1, characterized in that: In step S2, the annual cost of the electricity-cold / heat-gas-hydrogen type energy hub includes: The production energy cost is: The fixed construction cost of the energy hub is: The annual revenue of the electricity-cold / heat-gas-hydrogen energy hub includes: The revenue of energy supply service of energy hub is: The revenue from energy hub backup services is: The revenue from the electric vehicle charging pile power supply service is: The annual costs of an electricity-cold / heat-gas energy hub include: The production energy cost is: The fixed construction cost of the energy hub is: The annual revenue of the electricity-cold / heat-gas energy hub includes: The revenue of energy supply service of energy hub is: The revenue from energy hub backup services is: The revenue from the electric vehicle charging pile power supply service is: The annual costs of an electricity-cooling / heating energy hub include: The production energy cost is: The fixed construction cost of the energy hub is: The annual revenue of the electric-cold / heat energy hub includes: The revenue of energy supply service of energy hub is: The revenue from energy hub backup services is: The revenue from the electric vehicle charging pile power supply service is: The annual costs of an electricity-gas-hydrogen energy hub include: The production energy cost is: The fixed construction cost of the energy hub is: The annual revenue of the Electricity-Gas-Hydrogen Energy Hub includes: The revenue of energy supply service of energy hub is: The revenue from energy hub backup services is: The revenue from the electric vehicle charging pile power supply service is: Among them, EH1, EH2, EH3, and EH4 are respectively the energy hubs of electricity-cold / heat-gas-hydrogen type, electricity-cold / heat-gas type, electricity-cold / heat type, and electricity-gas-hydrogen type; D is the total number of days in a year; N w is the number of typical scenes in different seasons of the year; w is the typical scenes in different seasons of the year, which are summer typical scenes, transitional season typical scenes and winter typical scenes; T is the total time period in a day; t is the specific time of the day; p w is the probability of occurrence of scenes in different seasons; c e 、c g prices for external electricity and natural gas purchases; is the amount of electricity purchased from the external main grid by each energy hub during period t in scenario w, is the natural gas purchase volume of each energy hub in period t under scenario w, I is the set of energy equipment types included in the electricity-cold / heat-gas-hydrogen type energy hub, GSHP-ground source heat pump, EB-electric boiler, ISAC-ice storage air conditioner, PV-photovoltaic, EVSE-electric vehicle charging station; i is the energy equipment type included in the electricity-cold / heat-gas-hydrogen type energy hub; is the annualized initial investment cost per unit capacity and annual fixed maintenance cost per unit capacity corresponding to each energy equipment; V i The standard capacity for each type of equipment; is the number of devices configured in this type of energy hub; e , c , h , g , The corresponding energy supply prices for electricity, cold energy, heat and cold energy, natural gas and hydrogen energy; The price of electricity for charging stations; The annual unit capacity reserve service price for electric energy and the annual unit capacity reserve service price for natural gas; Revenue from electricity supply services for electric vehicle charging stations; Revenue from power reserve services for each energy hub; Revenue from natural gas energy backup services for various energy hubs; The amount of electricity provided by each energy hub during period t under scenario w; The cold energy provided by each energy hub to the outside during period t in scenario w; The thermal energy provided by each energy hub to the outside during period t under scenario w; is the amount of natural gas provided by each energy hub during period t under scenario w; The hydrogen energy provided by each energy hub during period t under scenario w; It is the power supply of electric vehicle charging piles in each energy hub w scenario during period t.

4. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 3 is characterized in that: The demand and supply constraints of electricity, cold energy, heat energy, natural gas and hydrogen energy in the electricity-cold / heat-gas-hydrogen energy hub are as follows: Electricity-cold / heat-gas-hydrogen type energy hub power load supply constraints: Cooling load supply constraints of electricity-cooling / heat-gas-hydrogen type energy hub: Heat load supply constraints of electricity-cold / heat-gas-hydrogen energy hub: Natural gas load supply constraints for power-cooling / heat-gas-hydrogen energy hubs: Hydrogen load supply constraints for electricity-cold / heat-gas-hydrogen type energy hubs: For the power balance constraints of electricity and natural gas in the power-cold / heat-gas-hydrogen type energy hub, the following relationship exists: Power balance constraints: Natural gas balance constraints: The demand and supply constraints for electricity, cooling energy, heat energy and natural gas in the electricity-cooling / heat-gas energy hub are as follows: Electricity-cold / heat-gas type energy hub power load supply constraints: Cooling load supply constraints of electricity-cooling / heat-gas type energy hub: Heat load supply constraints of electricity-cold / heat-gas type energy hub: Natural gas load supply constraints for power-cooling / heat-gas type energy hubs: For the power-cold / heat-gas type energy hub, the power balance constraints of electricity and natural gas have the following relationship: Power balance constraints: Natural gas balance constraints: The supply constraints of electricity, cooling and heat energy in the electricity-cooling / heating energy hub are as follows: Electricity-cold / heat energy hub power load supply constraints: Cooling load supply constraints of electricity-cooling / heating energy hub: Heat load supply constraints of electricity-cold / heat type energy hub: For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship: Power balance constraints: Natural gas balance constraints: The supply constraints for electricity, natural gas and hydrogen in the electricity-gas-hydrogen energy hub are as follows: Electricity-gas-hydrogen energy hub power load supply constraints: Natural gas load supply constraints for electricity-gas-hydrogen energy hubs: Hydrogen load supply constraints for electricity-gas-hydrogen energy hubs: For the power-cold / heat type energy hub, the power balance constraints of electricity and natural gas have the following relationship: Power balance constraints: Natural gas balance constraints:

5. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 1, characterized in that: In step S4, the annual cost of the park includes: The production and purchase costs of energy for all types of energy hubs in the region: The investment and construction costs of all types of energy hubs in the region: Among them, Park is four types of parks, including school parks, industrial parks, residential parks and commercial parks; The corresponding amount of electricity purchased from the external main grid by each type of energy hub during period t; is the corresponding natural gas purchase volume of each type of energy hub during period t; The optimal number of four energy hubs to be built in different parks; The total annual investment and construction costs of the four types of energy hubs; The load demand corresponding to the electricity in each park period; The load demand corresponding to the electric vehicle charging piles in each park period; The upper limit of hydrogen supply ratio for each park.

6. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 5, characterized in that: The corresponding load demand constraints for electricity, cooling energy, heat energy, natural gas, hydrogen energy and electric vehicle charging piles in the park in each season are as follows: Park power load supply constraints: Park cooling load supply constraints: Park heat load supply constraints: Park gas load supply constraints: Hydrogen load supply constraints in the park: The power supply of electric vehicle charging piles in the park meets the demand constraints:

7. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 1, characterized in that: In step S5, in the time series prediction method, a convolutional neural network model is used to predict the changing trend of equipment parameters.

8. The method for comprehensive energy key parameter data mining and long-term configuration prediction according to claim 1, characterized in that: In step S6, multiple models are used to fit the life cycle curve: Three life cycle functions are selected for fitting, and the historical energy demand data of each park are input into the Bass model, Gomperz model and polynomial model, and the prediction results of different models are output respectively.

9. A comprehensive energy key parameter data mining and long-term configuration prediction system, characterized in that: include: The data module uses a convolutional neural network to classify network information, then extracts key parameter information through a recurrent neural network, clusters the key parameter information using a deep clustering method, and builds a comprehensive energy equipment library; The solution module couples different integrated energy equipment from the integrated energy equipment library constructed by the data module, constructs an economic quantitative internal optimization configuration model for the energy supply hub, solves the economic quantitative internal optimization configuration model for the energy supply hub under the condition of maximum profit, and obtains the optimal configuration schemes for four energy hubs including the electricity-cold / heat-gas-hydrogen energy hub, the electricity-cold / heat-gas type energy hub, the electricity-cold / heat type energy hub, and the electricity-gas-hydrogen type energy hub; The preprocessing module uses clustering feature extraction technology to preprocess the electricity, cooling, heating, gas, and hydrogen load data of the base year M, and obtains the typical energy consumption characteristics of different parks at different times in the M year through clustering; The configuration module adopts the selection modeling method of park source-load matching. According to the typical energy consumption characteristics of different parks at different times in the Mth year obtained by the preprocessing module, the optimal configuration scheme of the energy hub obtained by the solution module is selected to optimize the energy supply. The park source-load matching model is solved under the condition of the lowest annual cost of the park, and the quantity ratio of various energy supply hubs that meet various energy demands in different seasons is obtained; the configuration of all parks is added up to obtain the comprehensive energy configuration plan of the corresponding region in the Mth year; The conversion module uses the key parameter information extracted from the data module, uses the time series prediction method to convert the time series into a curve shape, and uses the convolutional neural network model for analysis to predict the change trend of key parameter information of typical equipment from the M+1 year to the end of the Nth year; The fitting module uses the life cycle curve fitting method to predict the energy demand of the park and predict the development demand of various types of energy in the park from the M+1 year to the end of the Nth year; The prediction module takes the changing trend of key parameter information of typical equipment predicted by the conversion module from the M+1 year to the end of the N year and the various energy development demands of the park predicted by the fitting module from the M+1 year to the end of the N year as input parameters, and uses the park source-load matching selection modeling method of the configuration module to predict the long-term comprehensive energy configuration plan for the region from the M+1 year to the end of the N year.

Images