Multi-source data driven development zone comprehensive energy intelligent evaluation method and system

Abstract

This application discloses a multi-source data-driven intelligent evaluation method and system for integrated energy in development zones, relating to the field of industrial big data technology. The method includes: calculating the load fluctuation synchronization index within each unit time window based on historical electrical load time series, historical heat load time series, and historical cold load time series; identifying load coupling groups; extracting typical fluctuation cycles; correcting the load amplitude within the typical fluctuation cycles; generating dynamic predicted load curves; performing hourly superposition; reducing the superimposed peak value; obtaining an equivalent integrated load sequence; calculating its capacity demand value; and iteratively searching for the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and annual shortage capacity of the equipment. This solves the technical problems of existing technologies that fail to consider the synchronous coupling characteristics of load fluctuations between energy-consuming units and fail to combine actual production plans and waste heat recovery characteristics to correct the load, resulting in large comprehensive load prediction errors and unreasonable power supply facility configurations.

Description

Technical Field

[0001] This application relates to the field of industrial big data technology, specifically to a multi-source data-driven intelligent evaluation method and system for comprehensive energy in development zones. Background Technology

[0002] As the core carrier of industrial agglomeration and development, development zones gather a large number of energy-consuming units of different types. The rationality of the planning and configuration of their integrated energy systems directly affects the regional energy utilization efficiency and operational economy.

[0003] However, most existing energy assessment methods for development zones directly predict the total load statistically without considering the synchronous coupling characteristics of load fluctuations between different energy-consuming units. They easily overlook the load peak superposition effect caused by the synchronicity of production scheduling, resulting in a large error in the predicted comprehensive load.

[0004] Meanwhile, most existing methods do not take into account the actual production plans of enterprises in the development zone and the characteristics of equipment waste heat recovery to adjust the load, nor do they carry out multi-objective optimization configuration for capacity requirements under different operating conditions. This can easily lead to problems such as excessive redundancy or insufficient capacity of power supply facilities, making it difficult to meet the actual needs of accurate evaluation and optimization configuration of the integrated energy system in the development zone. Summary of the Invention

[0005] This application provides a multi-source data-driven intelligent evaluation method and system for comprehensive energy in development zones, which solves the technical problems of existing technologies that fail to consider the synchronous coupling characteristics of load fluctuations between energy-consuming units and fail to combine actual production plans and waste heat recovery characteristics to correct loads, resulting in large comprehensive load prediction errors and unreasonable power supply facility configurations.

[0006] The technical solution to the above-mentioned technical problems in this application is as follows:

[0007] Firstly, this application provides a multi-source data-driven intelligent evaluation method for comprehensive energy in development zones, the method comprising: Collect historical electricity load time series, historical heat load time series and historical cooling load time series of each energy-consuming unit in the development zone, and obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit; Based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, calculate the load fluctuation synchronization index within each unit time window, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation period of each coupling group. For each load coupling group, the load amplitude within a typical fluctuation period is corrected based on the planned production scheduling data to generate a dynamic predicted load curve; The dynamic predicted load curve is superimposed hourly. By converting the waste heat recovery electricity substitution and the cooling load electricity equivalent, the superimposed peak value is reduced to obtain the equivalent comprehensive load sequence. The capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles is then calculated. Using the capacity demand value as the initial population, the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment is used as the recommended configuration capacity of the power supply facilities in the development zone.

[0008] Secondly, this application provides a multi-source data-driven integrated energy intelligent evaluation system for development zones, including: The historical data acquisition module is used to collect historical electricity load time series, historical heat load time series and historical cooling load time series of various energy-consuming units in the development zone, and to obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit. The fluctuation cycle extraction module is used to calculate the load fluctuation synchronization index within each unit time window based on the historical electrical load time sequence, historical heat load time sequence, and historical cooling load time sequence, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation cycle of each coupling group. The load amplitude correction module is used to correct the load amplitude within a typical fluctuation period for each load coupling group based on the planned production schedule data, and generate a dynamic predicted load curve. The load curve overlay module is used to overlay the dynamic predicted load curve hourly, reduce the overlay peak value by replacing waste heat recovery electricity with cold load electricity equivalent, obtain the equivalent comprehensive load sequence, and calculate the capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles. The configuration capacity recommendation module is used to iteratively search for the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment, using the capacity demand value as the initial population, and then uses this as the recommended configuration capacity of the power supply facilities in the development zone.

[0009] This application provides one or more technical solutions, which have at least the following technical effects or advantages: This application provides a multi-source data-driven intelligent evaluation method and system for comprehensive energy in development zones. First, it identifies load coupling groups with synchronous fluctuation characteristics based on multi-source historical load data, fully considering the load peak superposition effect caused by the synchronicity of production schedules of different energy-consuming units, effectively reducing the prediction error of comprehensive load. Second, this method combines the actual planned production schedule data of each energy-consuming unit in the development zone to correct the load amplitude within typical fluctuation cycles. It also introduces waste heat recovery electricity substitution and cold load electricity equivalent conversion processing, taking into account the energy recycling characteristics within the development zone. Finally, using the capacity demand of multiple quantiles as the initial population, and minimizing the weighted sum of annual idle capacity and annual shortage capacity as the optimization objective, iteratively searches to obtain the optimal recommended configuration capacity of power supply facilities, balancing power supply reliability and operational economy.

[0010] Through the above technical solution, this application effectively avoids the problem of excessive redundancy or insufficient capacity in the configuration, and can provide reliable support for the accurate evaluation and optimized configuration of the integrated energy system in the development zone. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the multi-source data-driven intelligent energy evaluation method for development zones provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of the multi-source data-driven intelligent energy evaluation system for development zones provided in this application embodiment.

[0013] The components represented by each number in the attached diagram are explained below: Historical data acquisition module 11, fluctuation cycle extraction module 12, load amplitude correction module 13, load curve overlay module 14, and configuration capacity recommendation module 15. Detailed Implementation

[0014] This application provides a multi-source data-driven intelligent evaluation method and system for comprehensive energy in development zones. It addresses the technical problems of existing technologies that fail to consider the synchronous coupling characteristics of load fluctuations between energy-consuming units and fail to adjust loads based on actual production plans and waste heat recovery characteristics, resulting in large comprehensive load prediction errors and unreasonable power supply facility configurations.

[0015] Example 1, as Figure 1As shown in the embodiments of this application, a multi-source data-driven intelligent evaluation method for comprehensive energy in development zones is provided, including: S10: Collect historical electricity load time series, historical heat load time series and historical cooling load time series of each energy-consuming unit in the development zone, and obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit; The planned production scheduling data includes at least the start time of production orders, the end time of production orders, and the target output of products corresponding to the orders for each energy-consuming unit within the target planning period. The equipment start-up and shutdown characteristic parameters include at least the rated output thermal power of each energy-consuming equipment and the maximum recovery power of the waste heat recovery device.

[0016] In this embodiment of the application, the historical electricity load time series, historical heat load time series and historical cooling load time series of each energy-consuming unit in the development zone are first collected. Specifically, this includes collecting hourly load operation data for at least one year at preset time intervals. During the collection process, missing data points are supplemented by the load average value of adjacent time points, and abnormal extreme values ​​exceeding three times the standard deviation are replaced and corrected by the corresponding load value of the same type of working day in the same period, so as to ensure the integrity and accuracy of the input load time series data.

[0017] Based on this, the planned production schedule data of each energy-consuming unit within the target planning period is retrieved simultaneously from the enterprise production management system of the development zone, and the start-up and shutdown characteristic parameters of each core energy-consuming equipment are extracted from the energy equipment ledger. Among them, the planned production schedule data includes the start time of production orders, the end time of production orders, and the corresponding product output targets of each energy-consuming unit within the target planning period, reflecting the fluctuation trend of the enterprise's production load within the planning period; the equipment start-up and shutdown characteristic parameters include the rated output thermal power of each energy-consuming equipment and the maximum recovery power of the waste heat recovery device, reflecting the waste heat recovery and utilization potential of the equipment itself.

[0018] S20: Based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, calculate the load fluctuation synchronization index within each unit time window, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation period of each coupling group; In this embodiment of the application, the load fluctuation synchronization index within each unit time window is calculated by dividing the historical load time series of all energy-consuming units into a single unit time window, standardizing the load sequence of each energy-consuming unit within each time window, calculating the Pearson correlation coefficient between each pair of energy-consuming units, and taking the average of all pairwise correlation coefficients as the load fluctuation synchronization index of that time window.

[0019] Furthermore, when the load fluctuation synchronization index of a pair of energy-consuming units is greater than the preset synchronization index threshold, the pair of energy-consuming units is marked as having a synchronous fluctuation relationship. Based on this, all interconnected energy-consuming units are divided into the same load coupling group, and the main period of load sequence fluctuation of each group is extracted by spectrum analysis. This main period is used as the typical fluctuation period of the corresponding coupling group.

[0020] Specifically, based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, the load fluctuation synchronization index within each unit time window is calculated, including: The historical electrical load time series, historical heat load time series, and historical cooling load time series are divided into multiple consecutive unit time windows according to the same time interval, and each unit time window contains several consecutive time points. Within each unit time window, the changes in electrical load, heat load, and cooling load of each energy-consuming unit relative to the first time point within the window are calculated to obtain the electrical load change sequence, heat load change sequence, and cooling load change sequence. For any two energy-consuming units, calculate the Pearson correlation coefficient between their electrical load change sequences, the Pearson correlation coefficient between their heat load change sequences, and the Pearson correlation coefficient between their cooling load change sequences. The arithmetic mean of the three Pearson correlation coefficients is taken as the load fluctuation synchronization index of the two energy-consuming units within the current unit time window.

[0021] In this embodiment of the application, the collected historical electrical load time series, historical heat load time series and historical cold load time series are first divided into multiple consecutive unit time windows according to the same time interval, for example, each 7 days is divided into a unit time window, and each unit time window contains 168 hourly time points, covering the complete production cycle fluctuation cycle.

[0022] Subsequently, within each unit time window, the changes in electrical load, heat load, and cooling load of each energy-consuming unit relative to the first time point within the window are calculated to obtain standardized electrical load change sequences, heat load change sequences, and cooling load change sequences. Specifically, the changes are standardized by dividing the changes by the historical standard deviation of the corresponding load type of the energy-consuming unit to eliminate the impact of load magnitude differences between different energy-consuming units on the synchronization calculation.

[0023] Secondly, for any two energy-consuming units, calculate the Pearson correlation coefficient between their electrical load change sequences, the Pearson correlation coefficient between their heat load change sequences, and the Pearson correlation coefficient between their cooling load change sequences. Take the arithmetic mean of these three Pearson correlation coefficients as the load fluctuation synchronization index of the two energy-consuming units within the current unit time window.

[0024] The calculation of the Pearson correlation coefficient is an existing technology that can measure the degree of linear correlation between two variables. The value range is [-1, 1]. The larger the absolute value, the stronger the linear correlation between the two variables. When the changing trends of the two variables are completely consistent, the correlation coefficient is 1; when the changing trends are completely opposite, it is -1; and when there is no obvious linear correlation, it approaches 0. The specific calculation process is as follows: divide the covariance of the corresponding sequences of the two variables by the product of the standard deviations of the two sequences, and finally obtain the Pearson correlation coefficient between the two.

[0025] Furthermore, load coupling groups with synchronous fluctuation characteristics are identified, and the typical fluctuation period of each coupling group is extracted, including: Arrange the load fluctuation synchronization index between any two energy-consuming units calculated within each unit time window in chronological order to form a synchronization index sequence. Traverse all energy-consuming unit pairs, and filter out energy-consuming unit pairs in the synchronization index sequence whose load fluctuation synchronization index is greater than a preset synchronization index threshold and whose number of consecutive time windows greater than the preset synchronization index threshold exceeds a preset consecutive time window threshold, and mark them as strong synchronization pairs. By establishing connections between each energy-consuming unit as nodes and the strong synchronization pairs as connecting edges, a synchronization association graph is constructed. Extract the connected groups in the synchronous association graph, and take the set of energy-consuming units in each connected group as a load coupling group; For each load coupling group, extract the load data of all energy-consuming units in the group within the same time window from the historical electrical load time series, historical thermal load time series and historical cooling load time series. Calculate the peak time of the total load of the load coupling group within each unit time window, count the time interval between adjacent peak times in all time windows, and take the mode of the time interval as the typical fluctuation period.

[0026] In this embodiment of the application, firstly, the synchronization indexes of all pairs of energy-consuming units within each unit time window are arranged according to the time dimension to obtain the synchronization index change sequence of each pair of energy-consuming units.

[0027] Subsequently, all energy-consuming unit pairs are traversed, and strong synchronous association pairs with a synchronization index that is consistently higher than the preset synchronization index threshold are selected. A synchronization association graph is constructed with energy-consuming units as nodes and strong synchronization associations as edges. Then, each independent load coupling group is obtained through connected component analysis to ensure that the load fluctuations of energy-consuming units within the same group have stable synchronization characteristics. The preset synchronization index threshold is usually determined based on the lower quartile of the historical synchronization index distribution of known energy-consuming unit pairs with collaborative production relationships within the development zone. Since a Pearson correlation coefficient greater than 0.8 is considered a high positive correlation, in this embodiment, the preset synchronization index threshold can be taken as 0.8.

[0028] Specifically, the synchronous correlation graph is constructed by using each energy-consuming unit as a vertex. Only when two energy-consuming units are marked as a strong synchronous pair is an undirected connection edge added between the two vertices. This results in a synchronous correlation graph that reflects the load synchronous fluctuation relationship between energy-consuming units. Then, a breadth-first search algorithm is used to traverse the entire graph and divide all interconnected vertices into the same connected group. Each connected group corresponds to a load coupling group. Energy-consuming units within the same group maintain stable synchronous load fluctuation characteristics in the long term, while the load fluctuation synchronization between different groups is weak.

[0029] Secondly, after obtaining all load coupling groups, the typical fluctuation cycle corresponding to each group is extracted: first, the load data of all energy-consuming units in the same time window within the group is extracted from the historical load time series, and the total load time series of the group is obtained by summing them. Then, the peak occurrence time of the total load in each time window is counted, the time interval between adjacent peaks is calculated, and the mode of all time intervals is taken as the typical fluctuation cycle of the load coupling group. This cycle can accurately reflect the overall load fluctuation pattern of the energy-consuming units in the group.

[0030] S30: For each load coupling group, adjust the load amplitude within a typical fluctuation period based on the planned production scheduling data, and generate a dynamic predicted load curve; In this embodiment of the application, for each load coupling group, the planned production scheduling data of the corresponding date is matched according to the date covered by each energy-consuming unit in the group within the typical fluctuation period, and the electrical load amplitude correction coefficient is calculated to complete the preliminary correction of the electrical load amplitude. Finally, the dynamic predicted load curve of the load coupling group after scheduling correction is obtained.

[0031] Specifically, step S30 in the method includes: Extract the start time of production orders, end time of production orders, and product output targets for each energy-consuming unit within the load coupling group during the target planning period from the planned production scheduling data. Align the typical fluctuation cycle with the target planning period according to the time axis, and in the time interval that does not cover the production order, call the load amplitude corresponding to the typical fluctuation cycle in the historical power load time series as the base amplitude; Within the time interval covering production orders, extract the historical average electrical load amplitude corresponding to the production interval with the same product output target from the historical electrical load time series, and calculate the electrical load amplitude correction coefficient by combining it with the original electrical load amplitude within the typical fluctuation cycle. Based on the electrical load amplitude correction coefficient and the original electrical load amplitude, the corrected electrical load amplitude is calculated. The electrical load amplitude correction coefficient is applied to both the heat load amplitude and the cooling load amplitude to calculate the corrected heat load amplitude and the corrected cooling load amplitude. The corrected electrical load amplitude, corrected heat load amplitude, and corrected cooling load amplitude are filled into the time points corresponding to the typical fluctuation cycle in chronological order to form the dynamic predicted load curve of the load coupling group.

[0032] In this embodiment, firstly, the production arrangement information of all energy-consuming units in the load coupling group within the target planning period is extracted from the planned production scheduling data. After aligning the typical fluctuation cycle to the target planning period according to the time axis, the time intervals covered by production orders are processed separately. For non-production periods not covered by production orders, the load amplitude of the corresponding typical fluctuation cycle in the historical period of the coupling group is directly called as the base amplitude.

[0033] Secondly, for the production period covering the production order, the historical average load amplitude of the same production range as the current order production target is first matched from the historical load database. Then, the average amplitude is divided by the corresponding average amplitude of the original typical fluctuation cycle to obtain the power load amplitude correction coefficient, that is, power load amplitude correction coefficient = historical average power load amplitude / original power load amplitude.

[0034] Then, the correction factor is multiplied by the electrical load amplitude at each time point of the original typical fluctuation period to obtain the corrected electrical load amplitude, that is, the corrected electrical load amplitude = the original electrical load amplitude × the electrical load amplitude correction factor.

[0035] Within a load coupling group, changes in production order output simultaneously drive fluctuations in the same direction among electrical, heating, and cooling loads. For example, as output increases, the power consumption of electrical equipment rises, and the demand for equipment heating and cooling also increases accordingly. Furthermore, the proportional relationship between these three loads and output is approximately constant under typical production conditions. Therefore, using the same correction factor to scale the heating and cooling loads proportionally can reflect the synchronous changes of these three loads with engineering-acceptable accuracy, while avoiding the additional data requirements and model complexity that would be incurred by separately calculating correction factors for heating and cooling loads.

[0036] Specifically, the corrected heat load amplitude = original heat load amplitude × electrical load amplitude correction factor, and the corrected cooling load amplitude = original cooling load amplitude × electrical load amplitude correction factor. Finally, the three types of corrected load amplitudes are filled into the corresponding time points of the typical fluctuation cycle in chronological order to form the dynamic predicted load curve of the load coupling group.

[0037] S40: The dynamic predicted load curve is superimposed hourly. By converting the waste heat recovery electricity substitution and the cooling load electricity equivalent, the superimposed peak value is reduced to obtain the equivalent comprehensive load sequence. The capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles is calculated. The values ​​of the cumulative probability quantiles are selected based on the frequency distribution of peak loads in the historical electrical load time series, corresponding to normal operating conditions, peak operating conditions, and extreme operating conditions, respectively, with the cumulative probabilities being 0.5, 0.85, and 0.95, respectively.

[0038] In this embodiment, the dynamic predicted load curves of all load coupling groups are first superimposed hourly to obtain the total electrical load, total heat load, and total cooling load for each hour within the target planning period. Then, the heat load is reduced hourly by using waste heat recovery for electrical substitution: combining the start-up and shutdown status of the operating equipment in the current period, and based on the rated output heat power of the equipment and the maximum recovery power of the waste heat recovery device, the equivalent electrical power corresponding to the recoverable waste heat in the current period is calculated. This equivalent electrical power is subtracted from the total electrical load to complete the peak reduction of the electrical load by the heat load. Subsequently, the cooling load is converted into electrical equivalents. Based on the comprehensive performance coefficient (COP) of the chiller units in the current period, the cooling load demand is converted into the corresponding electrical load power and added to the total electrical load. Finally, the hourly equivalent comprehensive electrical load sequence is obtained.

[0039] Furthermore, after obtaining the equivalent comprehensive load sequence, all load values ​​in the sequence are sorted from smallest to largest, and the cumulative probability corresponding to each load value after sorting is calculated. Load values ​​at the quantiles of 0.5, 0.85, and 0.95 are selected as the comprehensive energy capacity demand values ​​of the development zone under normal operating conditions, peak operating conditions, and extreme operating conditions.

[0040] Specifically, the 0.5 quantile (median) of the frequency distribution of peak loads in the historical electrical load time series corresponds to the normal operating condition, indicating that 50% of the peak loads are below this value; the 0.85 quantile corresponds to the peak operating condition, indicating that 85% of the peak loads are below this value; and the 0.95 quantile corresponds to the extreme operating condition, indicating that 95% of the peak loads are below this value.

[0041] The dynamic predicted load curves are superimposed hourly, and the superimposed peak value is reduced by converting waste heat recovery electricity substitution and cooling load electricity equivalent, resulting in an equivalent comprehensive load sequence, including: The electrical load, heat load, and cooling load at the same time point in the dynamic predicted load curves of each load coupling group are summed to obtain the hourly total electrical load sequence, hourly total heat load sequence, and hourly total cooling load sequence of the development area. For each time point, the maximum recoverable waste heat power is calculated based on the rated output heat power and the maximum recovery power of the waste heat recovery device in the equipment start-stop characteristic parameters, and the smaller value between the maximum recoverable waste heat power and the corresponding heat load value in the hourly total heat load sequence is taken as the actual waste heat recovery amount. The waste heat power generation is obtained by multiplying the actual waste heat recovery amount by the thermoelectric conversion efficiency, wherein the thermoelectric conversion efficiency is determined based on the actual operating performance parameters of the waste heat recovery device. The waste heat power generation is subtracted from the corresponding electrical load value of the hourly total electrical load sequence to obtain the preliminary equivalent electrical load; Based on the cooling load value in the hourly total cooling load sequence and the energy efficiency ratio of the refrigeration unit, the electrical power required for cooling is calculated, wherein the energy efficiency ratio of the refrigeration unit is determined based on the actual operating performance parameters of the refrigeration unit. Based on the electrical power required for cooling and the preliminary equivalent electrical load, calculate the hourly equivalent electrical load; The hourly equivalent electrical load is used as the equivalent comprehensive load value at the corresponding time point. The equivalent comprehensive load values ​​at all time points are arranged in chronological order to obtain the equivalent comprehensive load sequence.

[0042] In this embodiment of the application, firstly, the loads of the same type in all coupled groups at the same time point are added together to obtain the hourly sequence of total electricity, total heat, and total cooling loads for the entire target planning period of the development zone.

[0043] Then, the equivalent power generation corresponding to the recoverable waste heat is calculated hourly: For each time point, combined with the current equipment start-up and shutdown status, the maximum recoverable waste heat power at that time point is obtained based on the rated output heat power of the equipment and the maximum recovery power of the waste heat recovery device, that is, the maximum recoverable waste heat power = min(rated output heat power, maximum recovery power of the waste heat recovery device). At the same time, the smaller value between the maximum recoverable waste heat power and the actual heat load demand at the current time point is taken as the actual recoverable waste heat, that is, the actual waste heat recovery amount = min(rated output heat power, maximum recovery power of the waste heat recovery device, total heat load value at each hour).

[0044] Secondly, the actual amount of waste heat recovered is multiplied by the thermoelectric conversion efficiency of waste heat power generation to obtain the waste heat power generation power that can reduce the power supply demand of the grid. Then, this waste heat power generation power is subtracted from the total electrical load to obtain the preliminary equivalent electrical load, that is, the preliminary equivalent electrical load = the corresponding electrical load value of the total electrical load sequence - the waste heat power generation power. The waste heat power generation power = the actual amount of waste heat recovered × the thermoelectric conversion efficiency, thus completing the reduction of the peak electrical load due to the heat-side waste heat recovery. The thermoelectric conversion efficiency is usually taken as 0.25 to 0.45.

[0045] Next, the cooling load is converted to electrical equivalent. The cooling load demand at the current time point is divided by the comprehensive energy efficiency ratio (COP) of the chiller under the current operating conditions to obtain the electrical power required for cooling. That is, the electrical power required for cooling = total cooling load per hour / chiller energy efficiency ratio. The chiller energy efficiency ratio is determined based on the actual operating performance parameters of the chiller, and is usually taken as 3.0 to 5.0.

[0046] Furthermore, the electrical power required for cooling is added to the preliminary equivalent electrical load that has already completed waste heat reduction, and finally the equivalent comprehensive electrical load at that time point is obtained. That is, the hourly equivalent electrical load = preliminary equivalent electrical load + electrical power required for cooling. The equivalent comprehensive electrical loads at all time points are arranged in chronological order to obtain a complete equivalent comprehensive load sequence.

[0047] For example, if the target planned daily peak electricity load of a development zone is 12MW, the corresponding total heat load is 8MW, the rated output heat power of the equipment is 7MW, the maximum recoverable power of the waste heat recovery device is 6MW, and the thermoelectric conversion efficiency is taken as 0.3, then the maximum recoverable waste heat power is: The actual waste heat recovery amount is Waste heat power generation capacity is The preliminary equivalent electrical load is If the total cooling load during this period is 12MW and the energy efficiency ratio of the chiller is taken as 4.0, then the electrical power required for cooling is: The final equivalent comprehensive electrical load for this period is This completes the equivalent load calculation for that point in time.

[0048] Further, the capacity demand values ​​of the equivalent composite load sequence at multiple cumulative probability quantiles are calculated, including: Arrange the load values ​​in the equivalent comprehensive load sequence from smallest to largest to obtain the sorted load sequence, and record the sorting number of each load value in the sorted load sequence; The total number of load values ​​contained in the equivalent comprehensive load sequence is counted. For each preset cumulative probability quantile, the corresponding cumulative probability is multiplied by the total number of load values ​​to obtain the theoretical position value. If the theoretical position value is an integer, then the load value corresponding to the sorting number that is equal to the theoretical position value is extracted from the sorted load sequence and used as the capacity demand value at the cumulative probability quantile. If the theoretical position value is not an integer, then the integer value taken down from the theoretical position value is taken as the first sorting number and the integer value taken up is taken as the second sorting number. Extract the first load value corresponding to the first sorting number and the second load value corresponding to the second sorting number from the sorted load sequence; Calculate the decimal part of the theoretical position value, use the decimal part as a weight, and perform a weighted summation on the first load value and the second load value according to the weight. Use the weighted summation result as the capacity demand value at the cumulative probability quantile.

[0049] In this embodiment, firstly, all load values ​​in the equivalent comprehensive load sequence are sorted in ascending order to obtain the sorted load sequence. Then, the sorting number of each load value in the sorted load sequence is recorded. Next, the total number of all load values ​​is counted. For each preset cumulative probability quantile, the cumulative probability corresponding to the quantile is multiplied by the total number of loads to obtain the theoretical position value corresponding to the quantile. That is, the theoretical position value = the cumulative probability corresponding to the quantile × the total number of loads. The cumulative probabilities corresponding to the quantiles are 0.5, 0.85, and 0.95, respectively. Therefore, the theoretical position values ​​corresponding to the three quantiles can be obtained respectively.

[0050] Specifically, if the theoretical position value is an integer, the load value corresponding to the sorting number in the sorting sequence is directly extracted as the comprehensive energy capacity demand value at that quantile. If the theoretical position value is not an integer, the value is rounded down to obtain the first sorting number and rounded up to obtain the second sorting number. The first load value and the second load value corresponding to the two numbers are extracted respectively. The decimal part of the theoretical position value is then used as the weight to perform a weighted sum of the two load values. The result is used as the capacity demand value at that cumulative probability quantile.

[0051] For example, if the target planning day contains 24 hourly equivalent composite load values, corresponding to a total of 24 load values, the theoretical position value of the cumulative probability 0.5 quantile is calculated as follows: Since the value is an integer, the load value of the 12th element after sorting is directly extracted as the capacity requirement value under normal operating conditions.

[0052] If the theoretical position value of the cumulative probability 0.85 quantile is calculated as follows: Since the integer is not a whole number, rounding down gives the first sorting number 20, and rounding up gives the second sorting number 21, with a decimal part of 0.4. If the first load value is 12.6MW and the second load value is 13.1MW, then the weighted sum gives the capacity demand value for that quantile. This is used to calculate the capacity demand under peak operating conditions.

[0053] S50: Using the capacity demand value as the initial population, iteratively search for the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment, and use this as the recommended configuration capacity of the power supply facilities in the development zone.

[0054] In this embodiment of the application, the capacity demand values ​​under three operating conditions—normal, peak, and extreme—are used as the initial population. The weighted sum of the annual idle capacity and the annual shortage capacity of the equipment is set as the optimization objective function. The weight of the idle capacity corresponds to the annual investment cost per unit capacity, and the weight of the shortage capacity corresponds to the annual power shortage loss cost per unit capacity.

[0055] Subsequently, the particle swarm optimization algorithm is used for iterative search, continuously selecting, crossing over, and mutating configuration schemes in the population. Disadvantageous configuration schemes with larger objective function values ​​are eliminated, while advantageous configuration schemes with smaller objective function values ​​are retained, until the number of iterations reaches a preset upper limit or the objective function value converges and no longer changes significantly. Finally, the minimum weighted sum and the corresponding power supply facility configuration capacity are used as the recommended configuration capacity of the power supply facilities in the development zone.

[0056] Among them, the particle swarm optimization algorithm is an existing technology. By continuously updating the position and velocity of each particle in the swarm, the particles track their current optimal position and the global optimal position, iteratively searching towards a smaller objective function value. Each particle corresponds to a set of power supply facility configuration capacity schemes. In each iteration, the particle adjusts its own configuration capacity parameters based on its historical optimal configuration and the current optimal configuration found by the entire swarm, recalculates the objective function value, and updates the individual optimal and global optimal results. When the iteration termination condition is met, the configuration capacity corresponding to the final global optimal result is the recommended configuration capacity of the power supply facilities in the development zone obtained from this optimization.

[0057] Specifically, step S50 in the method includes: The capacity demand values ​​at multiple cumulative probability quantiles are used as the initial positions of multiple search individuals, and the position coordinates of each search individual are used as candidate values ​​for the power supply facility configuration capacity. For each search individual, the capacity margin is calculated point-in-time by comparing the candidate values ​​of the power supply facility configuration capacity with the equivalent comprehensive load sequence. The annual idle capacity of equipment is calculated by summing up the capacity margin values ​​at the points in the year when the capacity margin is greater than zero, and the annual shortage capacity of equipment is calculated by summing up the absolute values ​​of the shortages at the points in the year when the capacity margin is less than zero. The fitness value of each search individual is obtained by multiplying the annual shortage capacity of the equipment by the shortage penalty coefficient and then adding it to the annual idle capacity of the equipment, wherein the shortage penalty coefficient is a constant greater than 1. Select the individual with the smallest fitness value from all searched individuals as the current optimal individual, and record the position coordinates of the current optimal individual; Calculate the difference between the current position of each search individual and the current optimal individual position, and randomly generate a random number between 0 and 1 as the step size coefficient; The difference is multiplied by the step size coefficient and then added to the current position of each search individual to obtain the updated position coordinates; Repeat the fitness value calculation, optimal individual selection and position update operations until the change in the fitness value of the current optimal individual is less than the preset convergence limit in multiple consecutive iterations; The location coordinates of the current best individual obtained last time are used as the recommended configuration capacity of the power supply facilities.

[0058] In this embodiment of the application, firstly, the capacity demand values ​​of the three operating conditions (normal, peak, and extreme) obtained at the three quantile points are used as the initial positions of the three search individuals, and the position coordinates of each search individual are a set of candidate values ​​for the configuration capacity of the power supply facilities to be evaluated.

[0059] Then, for each search individual, the difference between the candidate configuration capacity value of the individual and the equivalent comprehensive electrical load at the corresponding time point is calculated to obtain the capacity margin at each time point, that is, capacity margin = configuration capacity - equivalent comprehensive load value.

[0060] Specifically, when the configured capacity is greater than the current equivalent load, the margin is positive, indicating that there is surplus and idle capacity; when the configured capacity is less than the current equivalent load, the margin is negative, indicating that there is insufficient capacity.

[0061] Secondly, statistical analysis is conducted at all points in time throughout the year. All positive margins are summed up to obtain the annual idle capacity of the equipment corresponding to the individual, and the absolute values ​​of all negative margins are summed up to obtain the annual shortage capacity of the equipment. Then, the fitness value of the individual is calculated according to the fitness function. Fitness value = annual idle capacity of equipment + shortage penalty coefficient × annual shortage capacity of equipment. The shortage penalty coefficient is a constant greater than 1, which is used to highlight the loss impact caused by capacity shortage.

[0062] Next, the individual with the smallest fitness value is found among all the search individuals, and its position coordinates are recorded as the current global optimal position. Then, for each search individual, the difference between its current position and the current optimal position is calculated, and a step size coefficient between 0 and 1 is randomly generated. The difference is multiplied by the step size coefficient and added to the individual's current position to complete the position coordinate update and obtain a new configuration capacity candidate value.

[0063] Repeat the iterative process of fitness value calculation, current best individual selection, and individual position update until the fitness value change of the current best individual is less than the preset convergence limit in a set number of consecutive iterations, indicating that the optimization result has stabilized. At this point, the position coordinates of the current best individual obtained in the last iteration are the recommended configuration capacity of the power supply facilities in the development zone obtained by the final search.

[0064] For example, if the initial candidate configuration capacity values ​​for the three search individuals are 12.8MW, 13.5MW, and 14.2MW, respectively, the corresponding annual idle capacity obtained from the equivalent comprehensive load sequence throughout the year is 150MW, 320MW, and 480MW, and the annual shortage capacity is 120MW, 45MW, and 12MW, respectively, and the shortage penalty coefficient is 2, then the fitness values ​​of the three individuals are respectively , , At this point, the current optimal individual is the 12.8MW configuration scheme corresponding to a fitness value of 390. For each subsequent search individual, the position coordinates are updated using the same method, and the fitness value is recalculated, continuously updating the current optimal individual. When the fitness value change of the current optimal individual is less than 0.1MW in five consecutive iterations, the iteration stops, and the final output is the position coordinates of the current optimal individual, which is the recommended configuration capacity of the power supply facilities in the development zone obtained from this optimization.

[0065] In summary, compared with existing technologies, this application can fully consider the characteristics of multi-energy coupling, waste heat recovery, and cold load electrical equivalent conversion in the development zone, uniformly convert different types of energy loads into equivalent comprehensive electrical loads, and then carry out optimized configuration based on the capacity requirements of different probability quantiles. The recommended configuration capacity of the power supply facilities obtained is more in line with the actual operation scenario of the integrated energy system in the development zone, avoiding the capacity redundancy problem caused by the traditional configuration based on the superposition of various load peaks, and effectively improving the rationality and economy of the configuration results.

[0066] On the other hand, this method balances the dual risks of capacity idleness and capacity shortage through a weighted objective function, optimizes the objective to fit the investment and loss logic of actual projects, and has a simple and clear calculation process that is easy to implement in engineering.

[0067] In summary, the embodiments of this application have at least the following technical effects: This application provides a multi-source data-driven intelligent evaluation method for comprehensive energy in development zones. First, it identifies load coupling groups with synchronous fluctuation characteristics based on multi-source historical load data, fully considering the load peak superposition effect caused by the synchronicity of production schedules of different energy-consuming units, effectively reducing the prediction error of comprehensive load. Second, this method combines the actual planned production schedule data of each energy-consuming unit in the development zone to correct the load amplitude within typical fluctuation cycles. It also introduces waste heat recovery electricity substitution and cold load electricity equivalent conversion processing, taking into account the energy recycling characteristics within the development zone. Finally, using the capacity demand of multiple quantiles as the initial population, and with the minimum weighted sum of annual idle capacity and annual shortage capacity as the optimization objective, iteratively searches to obtain the optimal recommended configuration capacity of power supply facilities, balancing power supply reliability and operational economy.

[0068] Through the above technical solution, this application effectively avoids the problem of excessive redundancy or insufficient capacity in the configuration, and can provide reliable support for the accurate evaluation and optimized configuration of the integrated energy system in the development zone.

[0069] Example 2, as Figure 2 As shown, based on the same inventive concept as the multi-source data-driven intelligent energy evaluation method for development zones provided in Embodiment 1, this application also provides a multi-source data-driven intelligent energy evaluation system for development zones, including: The historical data acquisition module 11 is used to collect the historical electricity load time sequence, historical heat load time sequence and historical cold load time sequence of each energy-consuming unit in the development zone, and to obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit. The fluctuation cycle extraction module 12 is used to calculate the load fluctuation synchronization index within each unit time window based on the historical electrical load time sequence, historical heat load time sequence and historical cold load time sequence, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation cycle of each coupling group. The load amplitude correction module 13 is used to correct the load amplitude within a typical fluctuation period for each load coupling group based on the planned production schedule data, and generate a dynamic predicted load curve. The load curve overlay module 14 is used to overlay the dynamic predicted load curve hourly, reduce the overlay peak value by replacing waste heat recovery electricity with cold load electricity equivalent, obtain the equivalent comprehensive load sequence, and calculate the capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles. The capacity recommendation module 15 is used to iteratively search for the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment, using the capacity demand value as the initial population, and then use this as the recommended configuration capacity of the power supply facilities in the development zone.

[0070] Furthermore, in one embodiment of the application, the planned production scheduling data includes at least the start time of the production order, the end time of the production order, and the target output of the corresponding product for each energy-consuming unit within the target planning period. The equipment start-up and shutdown characteristic parameters include at least the rated output thermal power of each energy-consuming device and the maximum recovery power of the waste heat recovery device.

[0071] Further, in one embodiment of the application, the load fluctuation synchronization index within each unit time window is calculated based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, including: The historical electrical load time series, historical heat load time series, and historical cooling load time series are divided into multiple consecutive unit time windows according to the same time interval, and each unit time window contains several consecutive time points. Within each unit time window, the changes in electrical load, heat load, and cooling load of each energy-consuming unit relative to the first time point within the window are calculated to obtain the electrical load change sequence, heat load change sequence, and cooling load change sequence. For any two energy-consuming units, calculate the Pearson correlation coefficient between their electrical load change sequences, the Pearson correlation coefficient between their heat load change sequences, and the Pearson correlation coefficient between their cooling load change sequences. The arithmetic mean of the three Pearson correlation coefficients is taken as the load fluctuation synchronization index of the two energy-consuming units within the current unit time window.

[0072] Furthermore, load coupling groups with synchronous fluctuation characteristics are identified, and the typical fluctuation period of each coupling group is extracted, including: Arrange the load fluctuation synchronization index between any two energy-consuming units calculated within each unit time window in chronological order to form a synchronization index sequence. Traverse all energy-consuming unit pairs, and filter out energy-consuming unit pairs in the synchronization index sequence whose load fluctuation synchronization index is greater than a preset synchronization index threshold and whose number of consecutive time windows greater than the preset synchronization index threshold exceeds a preset consecutive time window threshold, and mark them as strong synchronization pairs. By establishing connections between each energy-consuming unit as nodes and the strong synchronization pairs as connecting edges, a synchronization association graph is constructed. Extract the connected groups in the synchronous association graph, and take the set of energy-consuming units in each connected group as a load coupling group; For each load coupling group, extract the load data of all energy-consuming units in the group within the same time window from the historical electrical load time series, historical thermal load time series and historical cooling load time series. Calculate the peak time of the total load of the load coupling group within each unit time window, count the time interval between adjacent peak times in all time windows, and take the mode of the time interval as the typical fluctuation period.

[0073] In one embodiment, the load amplitude correction module 13 is specifically used for: Extract the start time of production orders, end time of production orders, and product output targets for each energy-consuming unit within the load coupling group during the target planning period from the planned production scheduling data. Align the typical fluctuation cycle with the target planning period according to the time axis, and in the time interval that does not cover the production order, call the load amplitude corresponding to the typical fluctuation cycle in the historical power load time series as the base amplitude; Within the time interval covering production orders, extract the historical average electrical load amplitude corresponding to the production interval with the same product output target from the historical electrical load time series, and calculate the electrical load amplitude correction coefficient by combining it with the original electrical load amplitude within the typical fluctuation cycle. Based on the electrical load amplitude correction coefficient and the original electrical load amplitude, the corrected electrical load amplitude is calculated. The electrical load amplitude correction coefficient is applied to both the heat load amplitude and the cooling load amplitude to calculate the corrected heat load amplitude and the corrected cooling load amplitude. The corrected electrical load amplitude, corrected heat load amplitude, and corrected cooling load amplitude are filled into the time points corresponding to the typical fluctuation cycle in chronological order to form the dynamic predicted load curve of the load coupling group.

[0074] Furthermore, in one embodiment, the dynamically predicted load curve is overlaid hourly, and the overlaid peak value is reduced by converting waste heat recovery electricity substitution with cooling load electrical equivalent, to obtain an equivalent comprehensive load sequence, including: The electrical load, heat load, and cooling load at the same time point in the dynamic predicted load curves of each load coupling group are summed to obtain the hourly total electrical load sequence, hourly total heat load sequence, and hourly total cooling load sequence of the development area. For each time point, the maximum recoverable waste heat power is calculated based on the rated output heat power and the maximum recovery power of the waste heat recovery device in the equipment start-stop characteristic parameters, and the smaller value between the maximum recoverable waste heat power and the corresponding heat load value in the hourly total heat load sequence is taken as the actual waste heat recovery amount. The waste heat power generation is obtained by multiplying the actual waste heat recovery amount by the thermoelectric conversion efficiency, wherein the thermoelectric conversion efficiency is determined based on the actual operating performance parameters of the waste heat recovery device. The waste heat power generation is subtracted from the corresponding electrical load value of the hourly total electrical load sequence to obtain the preliminary equivalent electrical load; Based on the cooling load value in the hourly total cooling load sequence and the energy efficiency ratio of the refrigeration unit, the electrical power required for cooling is calculated, wherein the energy efficiency ratio of the refrigeration unit is determined based on the actual operating performance parameters of the refrigeration unit. Based on the electrical power required for cooling and the preliminary equivalent electrical load, calculate the hourly equivalent electrical load; The hourly equivalent electrical load is used as the equivalent comprehensive load value at the corresponding time point. The equivalent comprehensive load values ​​at all time points are arranged in chronological order to obtain the equivalent comprehensive load sequence.

[0075] Further, the capacity demand values ​​of the equivalent composite load sequence at multiple cumulative probability quantiles are calculated, including: Arrange the load values ​​in the equivalent comprehensive load sequence from smallest to largest to obtain the sorted load sequence, and record the sorting number of each load value in the sorted load sequence; The total number of load values ​​contained in the equivalent comprehensive load sequence is counted. For each preset cumulative probability quantile, the corresponding cumulative probability is multiplied by the total number of load values ​​to obtain the theoretical position value. If the theoretical position value is an integer, then the load value corresponding to the sorting number that is equal to the theoretical position value is extracted from the sorted load sequence and used as the capacity demand value at the cumulative probability quantile. If the theoretical position value is not an integer, then the integer value taken down from the theoretical position value is taken as the first sorting number and the integer value taken up is taken as the second sorting number. Extract the first load value corresponding to the first sorting number and the second load value corresponding to the second sorting number from the sorted load sequence; Calculate the decimal part of the theoretical position value, use the decimal part as a weight, and perform a weighted summation on the first load value and the second load value according to the weight. Use the weighted summation result as the capacity demand value at the cumulative probability quantile.

[0076] Furthermore, the values ​​of the cumulative probability quantiles are selected based on the frequency distribution of peak load occurrence in the historical electrical load time series, corresponding to normal operating conditions, peak operating conditions, and extreme operating conditions, respectively, with the cumulative probabilities being 0.5, 0.85, and 0.95, respectively.

[0077] In one embodiment, the configuration capacity recommendation module 15 is specifically used for: The capacity demand values ​​at multiple cumulative probability quantiles are used as the initial positions of multiple search individuals, and the position coordinates of each search individual are used as candidate values ​​for the power supply facility configuration capacity. For each search individual, the capacity margin is calculated point-in-time by comparing the candidate values ​​of the power supply facility configuration capacity with the equivalent comprehensive load sequence. The annual idle capacity of equipment is calculated by summing up the capacity margin values ​​at the points in the year when the capacity margin is greater than zero, and the annual shortage capacity of equipment is calculated by summing up the absolute values ​​of the shortages at the points in the year when the capacity margin is less than zero. The fitness value of each search individual is obtained by multiplying the annual shortage capacity of the equipment by the shortage penalty coefficient and then adding it to the annual idle capacity of the equipment, wherein the shortage penalty coefficient is a constant greater than 1. Select the individual with the smallest fitness value from all searched individuals as the current optimal individual, and record the position coordinates of the current optimal individual; Calculate the difference between the current position of each search individual and the current optimal individual position, and randomly generate a random number between 0 and 1 as the step size coefficient; The difference is multiplied by the step size coefficient and then added to the current position of each search individual to obtain the updated position coordinates; Repeat the fitness value calculation, optimal individual selection and position update operations until the change in the fitness value of the current optimal individual is less than the preset convergence limit in multiple consecutive iterations; The location coordinates of the current best individual obtained last time are used as the recommended configuration capacity of the power supply facilities.

Claims

1. A multi-source data-driven intelligent evaluation method for comprehensive energy in development zones, characterized in that: The method includes: Collect historical electricity load time series, historical heat load time series and historical cooling load time series of each energy-consuming unit in the development zone, and obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit; Based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, calculate the load fluctuation synchronization index within each unit time window, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation period of each coupling group. For each load coupling group, the load amplitude within a typical fluctuation period is corrected based on the planned production scheduling data to generate a dynamic predicted load curve; The dynamic predicted load curve is superimposed hourly. By converting the waste heat recovery electricity substitution and the cooling load electricity equivalent, the superimposed peak value is reduced to obtain the equivalent comprehensive load sequence. The capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles is then calculated. Using the capacity demand value as the initial population, the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment is used as the recommended configuration capacity of the power supply facilities in the development zone.

2. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, The planned production scheduling data includes at least the start time of production orders, the end time of production orders, and the target output of products corresponding to the orders for each energy-consuming unit within the target planning period. The equipment start-up and shutdown characteristic parameters include at least the rated output thermal power of each energy-consuming equipment and the maximum recovery power of the waste heat recovery device.

3. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, Based on the historical electrical load time series, historical heat load time series, and historical cooling load time series, calculate the load fluctuation synchronization index within each unit time window, including: The historical electrical load time series, historical heat load time series, and historical cooling load time series are divided into multiple consecutive unit time windows according to the same time interval, and each unit time window contains several consecutive time points. Within each unit time window, the changes in electrical load, heat load, and cooling load of each energy-consuming unit relative to the first time point within the window are calculated to obtain the electrical load change sequence, heat load change sequence, and cooling load change sequence. For any two energy-consuming units, calculate the Pearson correlation coefficient between their electrical load change sequences, the Pearson correlation coefficient between their heat load change sequences, and the Pearson correlation coefficient between their cooling load change sequences. The arithmetic mean of the three Pearson correlation coefficients is taken as the load fluctuation synchronization index of the two energy-consuming units within the current unit time window.

4. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, Identify load coupling groups with synchronous fluctuation characteristics and extract the typical fluctuation period of each coupling group, including: Arrange the load fluctuation synchronization index between any two energy-consuming units calculated within each unit time window in chronological order to form a synchronization index sequence. Traverse all energy-consuming unit pairs, and filter out energy-consuming unit pairs in the synchronization index sequence whose load fluctuation synchronization index is greater than a preset synchronization index threshold and whose number of consecutive time windows greater than the preset synchronization index threshold exceeds a preset consecutive time window threshold, and mark them as strong synchronization pairs. By establishing connections between each energy-consuming unit as nodes and the strong synchronization pairs as connecting edges, a synchronization association graph is constructed. Extract the connected groups in the synchronous association graph, and take the set of energy-consuming units in each connected group as a load coupling group; For each load coupling group, extract the load data of all energy-consuming units in the group within the same time window from the historical electrical load time series, historical thermal load time series and historical cooling load time series. Calculate the peak time of the total load of the load coupling group within each unit time window, count the time interval between adjacent peak times in all time windows, and take the mode of the time interval as the typical fluctuation period.

5. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, For each load coupling group, the load amplitude within a typical fluctuation period is corrected based on the planned production scheduling data to generate a dynamic predicted load curve, including: Extract the start time of production orders, end time of production orders, and product output targets for each energy-consuming unit within the load coupling group during the target planning period from the planned production scheduling data. Align the typical fluctuation cycle with the target planning period according to the time axis, and in the time interval that does not cover the production order, call the load amplitude corresponding to the typical fluctuation cycle in the historical power load time series as the base amplitude; Within the time interval covering production orders, extract the historical average electrical load amplitude corresponding to the production interval with the same product output target from the historical electrical load time series, and calculate the electrical load amplitude correction coefficient by combining it with the original electrical load amplitude within the typical fluctuation cycle. Based on the electrical load amplitude correction coefficient and the original electrical load amplitude, the corrected electrical load amplitude is calculated. The electrical load amplitude correction coefficient is applied to both the heat load amplitude and the cooling load amplitude to calculate the corrected heat load amplitude and the corrected cooling load amplitude. The corrected electrical load amplitude, corrected heat load amplitude, and corrected cooling load amplitude are filled into the time points corresponding to the typical fluctuation cycle in chronological order to form the dynamic predicted load curve of the load coupling group.

6. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, The dynamic predicted load curves are overlaid hourly. By converting waste heat recovery electricity substitution with cooling load electrical equivalents, the overlaid peak value is reduced to obtain an equivalent comprehensive load sequence, including: The electrical load, heat load, and cooling load at the same time point in the dynamic predicted load curves of each load coupling group are summed to obtain the hourly total electrical load sequence, hourly total heat load sequence, and hourly total cooling load sequence of the development area. For each time point, the maximum recoverable waste heat power is calculated based on the rated output heat power and the maximum recovery power of the waste heat recovery device in the equipment start-stop characteristic parameters, and the smaller value between the maximum recoverable waste heat power and the corresponding heat load value in the hourly total heat load sequence is taken as the actual waste heat recovery amount. The waste heat power generation is obtained by multiplying the actual waste heat recovery amount by the thermoelectric conversion efficiency, wherein the thermoelectric conversion efficiency is determined based on the actual operating performance parameters of the waste heat recovery device. The waste heat power generation is subtracted from the corresponding electrical load value of the hourly total electrical load sequence to obtain the preliminary equivalent electrical load; Based on the cooling load value in the hourly total cooling load sequence and the energy efficiency ratio of the refrigeration unit, the electrical power required for cooling is calculated, wherein the energy efficiency ratio of the refrigeration unit is determined based on the actual operating performance parameters of the refrigeration unit. Based on the electrical power required for cooling and the preliminary equivalent electrical load, calculate the hourly equivalent electrical load; The hourly equivalent electrical load is used as the equivalent comprehensive load value at the corresponding time point. The equivalent comprehensive load values ​​at all time points are arranged in chronological order to obtain the equivalent comprehensive load sequence.

7. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, Calculating the capacity demand values ​​of the equivalent composite load sequence at multiple cumulative probability quantiles includes: Arrange the load values ​​in the equivalent comprehensive load sequence from smallest to largest to obtain the sorted load sequence, and record the sorting number of each load value in the sorted load sequence; The total number of load values ​​contained in the equivalent comprehensive load sequence is counted. For each preset cumulative probability quantile, the corresponding cumulative probability is multiplied by the total number of load values ​​to obtain the theoretical position value. If the theoretical position value is an integer, then the load value corresponding to the sorting number that is equal to the theoretical position value is extracted from the sorted load sequence and used as the capacity demand value at the cumulative probability quantile. If the theoretical position value is not an integer, then the integer value taken down from the theoretical position value is taken as the first sorting number and the integer value taken up is taken as the second sorting number. Extract the first load value corresponding to the first sorting number and the second load value corresponding to the second sorting number from the sorted load sequence; Calculate the decimal part of the theoretical position value, use the decimal part as a weight, and perform a weighted summation on the first load value and the second load value according to the weight. Use the weighted summation result as the capacity demand value at the cumulative probability quantile.

8. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 7, characterized in that, The values ​​of the cumulative probability quantiles are selected based on the frequency distribution of peak loads in the historical electrical load time series, corresponding to normal operating conditions, peak operating conditions, and extreme operating conditions, respectively, with the cumulative probabilities being 0.5, 0.85, and 0.95, respectively.

9. The multi-source data-driven intelligent evaluation method for comprehensive energy in development zones according to claim 1, characterized in that, Using the capacity demand value as the initial population, an iterative search is performed to determine the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment, including: The capacity demand values ​​at multiple cumulative probability quantiles are used as the initial positions of multiple search individuals, and the position coordinates of each search individual are used as candidate values ​​for the power supply facility configuration capacity. For each search individual, the capacity margin is calculated point-in-time by comparing the candidate values ​​of the power supply facility configuration capacity with the equivalent comprehensive load sequence. The annual idle capacity of equipment is calculated by summing up the capacity margin values ​​at the points in the year when the capacity margin is greater than zero, and the annual shortage capacity of equipment is calculated by summing up the absolute values ​​of the shortages at the points in the year when the capacity margin is less than zero. The fitness value of each search individual is obtained by multiplying the annual shortage capacity of the equipment by the shortage penalty coefficient and then adding it to the annual idle capacity of the equipment, wherein the shortage penalty coefficient is a constant greater than 1. Select the individual with the smallest fitness value from all searched individuals as the current optimal individual, and record the position coordinates of the current optimal individual; Calculate the difference between the current position of each search individual and the current optimal individual position, and randomly generate a random number between 0 and 1 as the step size coefficient; The difference is multiplied by the step size coefficient and then added to the current position of each search individual to obtain the updated position coordinates; Repeat the fitness value calculation, optimal individual selection and position update operations until the change in the fitness value of the current optimal individual is less than the preset convergence limit in multiple consecutive iterations; The location coordinates of the current best individual obtained last time are used as the recommended configuration capacity of the power supply facilities.

10. A multi-source data-driven intelligent energy evaluation system for development zones, characterized in that: The method for implementing the multi-source data-driven intelligent energy evaluation method for development zones according to any one of claims 1-9 includes: The historical data acquisition module is used to collect historical electricity load time series, historical heat load time series and historical cooling load time series of various energy-consuming units in the development zone, and to obtain the planned production scheduling data and equipment start-up and shutdown characteristic parameters of each energy-consuming unit. The fluctuation cycle extraction module is used to calculate the load fluctuation synchronization index within each unit time window based on the historical electrical load time sequence, historical heat load time sequence, and historical cooling load time sequence, identify load coupling groups with synchronous fluctuation characteristics, and extract the typical fluctuation cycle of each coupling group. The load amplitude correction module is used to correct the load amplitude within a typical fluctuation period for each load coupling group based on the planned production schedule data, and generate a dynamic predicted load curve. The load curve overlay module is used to overlay the dynamic predicted load curve hourly, reduce the overlay peak value by replacing waste heat recovery electricity with cold load electricity equivalent, obtain the equivalent comprehensive load sequence, and calculate the capacity demand value of the equivalent comprehensive load sequence at multiple cumulative probability quantiles. The configuration capacity recommendation module is used to iteratively search for the power supply facility configuration capacity that minimizes the weighted sum of the annual idle capacity and the annual shortage capacity of the equipment, using the capacity demand value as the initial population, and then uses this as the recommended configuration capacity of the power supply facilities in the development zone.

Images