A method and device for reducing temperature load considering the influence of random load

Abstract

The present invention relates to the technical field of power system load measurement and prediction, and specifically provides a method and device for cooling load that takes into account the influence of random loads, including: using a variational modal decomposition algorithm to divide the cooling load in the current time period into K modal components, and obtaining the first k low-frequency components of the K modal components; using the k low-frequency components and meteorological characteristic data in the current time period as input data of a pre-trained long short-term memory neural network model, and obtaining the cooling load in the predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model. The technical solution provided by the present invention eliminates the influence of random factors in the cooling load, can improve the correlation between the cooling load and meteorological parameters, and at the same time accurately measures and predicts the cooling load, further improving the prediction accuracy of the total power load, which is of great significance to the load measurement and prediction business.

Description

Technical Field

[0001] The present invention relates to the technical field of power system load measurement and prediction, and in particular to a method and device for cooling loads taking into account the influence of random loads. Background Art

[0002] Load fluctuations are influenced by numerous factors, including economic development level, industrial structure, electricity pricing structure, meteorological factors, and electricity usage habits, making accurate predictions extremely challenging. During the summer, cooling loads, primarily air conditioning, account for an increasing proportion of the grid load, becoming the primary reason for record-breaking summer grid loads. Meteorological factors are the primary driver of cooling load fluctuations, while random factors such as emergencies and policy decisions are also increasingly impacting cooling loads. Measuring and forecasting cooling loads has become a key component of supporting power supply security, and accurately measuring and forecasting cooling loads is crucial for load analysis and forecasting.

[0003] However, the current cooling load calculation and prediction methods do not eliminate the influence of random loads, resulting in a low correlation between the cooling load and meteorological factors. Summary of the Invention

[0004] In order to overcome the above-mentioned defects, the present invention proposes a method and device for reducing the temperature load by considering the influence of random load.

[0005] In a first aspect, a method for reducing a temperature load by taking into account the influence of random loads is provided. The method for reducing a temperature load by taking into account the influence of random loads comprises:

[0006] Using a variational modal decomposition algorithm to divide the cooling load in the current time period into K modal components, and obtaining the first k low-frequency components of the K modal components;

[0007] Using the k low-frequency components and meteorological characteristic data within the current time period as input data of a pre-trained long short-term memory neural network model, obtaining a cooling load within a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model;

[0008] Wherein, K is a first preset value, k is a second preset value, and k≤K-1.

[0009] Preferably, the cooling load of the i-th industry in week r in the cooling load during the current time period is as follows:

[0010] P 降温,r,i =P 需求,r,i -P 基础,r,i -P 随机,r,i

[0011] In the above formula, P 降温,r,iis the cooling load of the i-th industry in week r in the current time period, P 需求,r,i is the demand load of the i-th industry in week r of the current time period, P 基础,r,i is the base load of the industry i in week r of the current time period, P 随机,r,i is the random load of the i-th industry in week r of the current time period.

[0012] Furthermore, the base load of the i-th industry in week r of the cooling load in the current time period is as follows:

[0013] P 基础,r,i =P 基线,r,i ×(1+α r ) y +P 新报装,r,i ×bP 销户,r,i

[0014] In the above formula, P 基线,r,i is the baseline load of the industry in week r of the baseline sample week, α r is the average weekly growth rate coefficient corresponding to week r in the baseline sample week, P 新报装,r,i is the newly reported electricity load of industry i in week r of the current time period, b is the increase coefficient of business expansion, P 销户,r,i is the account cancellation load of the ith industry in week r of the current time period, and y is the interval between the week to which week r belongs and the baseline sample week in the current time period.

[0015] Furthermore, the baseline load of the ith industry in week r of the baseline sample week is as follows:

[0016]

[0017] In the above formula, P 需求,n,r,i is the load demand of industry i in week r of the nth baseline sample week, and N is the total number of baseline sample weeks.

[0018] Furthermore, the average weekly growth rate coefficient corresponding to week r within the baseline sample week is as follows:

[0019]

[0020] In the above formula, P r,N is the 96-point average load of week r within the Nth baseline sample week, P r,1 is the 96-point average load of week r within the first baseline sample week.

[0021] Furthermore, the demand load of the i-th industry in week r during the current time period is as follows:

[0022] P 需求,r,i =P 实际,r,i +P管理负荷,r,i +P 分布式光伏,r,i

[0023] In the above formula, P 实际,r,i is the actual electricity load of the i-th industry in week r of the current time period, P 管理负荷,r,i is the management load of the industry i in week r of the current time period, P 分布式光伏,r,i It is the self-generated and self-used load of distributed photovoltaic power generation in the i-th industry in week r of the current time period.

[0024] Furthermore, the actual electricity load of the i-th industry in week r during the current time period is as follows:

[0025] P 实际,r,i =P r ' ,i +Q r,i Q r,全行业 ×△P 偏差

[0026] In the above formula, P r ' ,i Q is the electricity load of industry i in week r of the current time period, excluding line loss. r,i is the daily electricity consumption of industry i in week r of the current time period, Q r,全行业 is the total daily electricity consumption of the entire industry in week r during the current time period, △P 偏差 For power consumption deviation.

[0027] Furthermore, the power consumption deviation is as follows:

[0028] △P 偏差 =P 调度 -P 全行业 -△P 厂用电 -△P 自备电厂

[0029] In the above formula, P 调度 is the dispatching load, P 全行业 is the electricity load of the entire industry, △P 厂用电 is the plant power load, △P 自备电厂 It is the electricity load of self-generated electricity in the captive power plant.

[0030] Preferably, the process of determining the first preset value and the second preset value includes:

[0031] Step a. Initialize K to be equal to a first preset starting value;

[0032] Step b. Using the variational mode decomposition algorithm to divide the cooling load in the historical period into K modal components;

[0033] Step c. Initialize k=1;

[0034] Step d. Calculating the correlation coefficient between the sum of the first k low-frequency components of the K modal components and the meteorological characteristic data within the historical period;

[0035] Step e. Determine whether k is equal to K-1. If so, output the correlation coefficient and execute step f. Otherwise, set k=k+1 and return to step d.

[0036] Step f. Determine whether K is equal to the first preset termination value. If so, use the values of K and k corresponding to the maximum correlation coefficient as the first preset value and the second preset value respectively. Otherwise, set K=K+1 and return to step a.

[0037] Furthermore, the correlation coefficient is as follows:

[0038]

[0039] In the above formula, T m is the mth element value in the sum of the first k low-frequency components in the K modal components, is the mean value of each element in the sum of the first k low-frequency components in the K modal components, P m is the mth element value in the meteorological characteristic data during the historical period, is the mean value of each element in the meteorological characteristic data within the historical period, r is the correlation coefficient, and M is the preset number of elements.

[0040] Preferably, the training process of the pre-trained long short-term memory neural network model includes:

[0041] The cooling load in the historical time period is divided into K modal components using a variational modal decomposition algorithm, and the first k low-frequency components of the K modal components are obtained;

[0042] Constructing training data using the k low-frequency components, meteorological characteristic data within a historical time period, and cooling load;

[0043] The training data is used to train the initial long short-term memory neural network model.

[0044] In a second aspect, a temperature reduction load device that considers the influence of random loads is provided, wherein the temperature reduction load device that considers the influence of random loads comprises:

[0045] A decomposition module is used to divide the cooling load in the current time period into K modal components using a variational modal decomposition algorithm, and obtain the first k low-frequency components of the K modal components;

[0046] An analysis module is configured to use the k low-frequency components and the meteorological characteristic data within the current time period as input data of a pre-trained long short-term memory neural network model, and obtain a cooling load within a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model;

[0047] Wherein, K is a first preset value, k is a second preset value, and k≤K-1.

[0048] In a third aspect, a computer device is provided, comprising: one or more processors;

[0049] The processor is configured to execute one or more programs;

[0050] When the one or more programs are executed by the one or more processors, the method for reducing the cooling load by considering the influence of random load is implemented.

[0051] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed, the method for reducing the cooling load by considering the influence of random load is implemented.

[0052] The above one or more technical solutions of the present invention have at least one or more of the following beneficial effects:

[0053] The present invention provides a method and device for cooling loads that considers the influence of random loads. The method comprises: using a variational mode decomposition algorithm to divide the cooling load within a current time period into K modal components, and obtaining the first k low-frequency components of the K modal components; using the k low-frequency components and meteorological characteristic data within the current time period as input data for a pre-trained long short-term memory neural network model, and obtaining the cooling load within a predicted time period corresponding to the current time period, as output by the pre-trained long short-term memory neural network model. The technical solution provided by the present invention uses a variational mode decomposition method to eliminate random loads from the cooling load, thereby improving the correlation between the cooling load and meteorological parameters, reducing the impact of random load fluctuations on the cooling load calculation results, and further improving the accuracy of cooling load prediction through accurate calculation of the cooling load. The method aims to improve the accuracy of cooling load measurement and prediction in the power grid, facilitate the power grid to accurately grasp the changing characteristics of the cooling load, provide support for power grid load analysis and prediction, support power supply during peak load periods, and ensure the safe and stable operation of the power grid during the summer under the new power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flow chart of the main steps of a method for reducing the temperature load considering the influence of random loads according to an embodiment of the present invention;

[0055] Figure 2This is a full-year load curve diagram of an embodiment of the present invention;

[0056] Figure 3 This is a load baseline diagram of different day types according to an embodiment of the present invention;

[0057] Figure 4 Schematic diagram of the correlation between cooling-random load and temperature index according to an embodiment of the present invention;

[0058] Figure 5 Schematic diagram of various load components when K=4 in an embodiment of the present invention;

[0059] Figure 6 This is a comparison chart of the cooling-random load and cooling load during the first ten days of summer according to an embodiment of the present invention;

[0060] Figure 7 This is a comparison chart of the power load prediction value and the actual value without eliminating the influence of random factors in an embodiment of the present invention;

[0061] Figure 8 This is a comparison chart between the predicted value of power load and the actual value after eliminating the influence of random factors in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0064] Example 1

[0065] See attached Figure 1 , Figure 1 FIG. 1 is a flow chart showing the main steps of a method for reducing the temperature load by considering the influence of random loads according to an embodiment of the present invention. Figure 1 As shown, the method for reducing the temperature load considering the influence of random load in the embodiment of the present invention mainly includes the following steps:

[0066] Step S101: using a variational modal decomposition algorithm to divide the cooling load in the current time period into K modal components, and obtaining the first k low-frequency components of the K modal components;

[0067] Step S102: using the k low-frequency components and the meteorological characteristic data in the current time period as input data of a pre-trained long short-term memory neural network model, and obtaining a cooling load in a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model;

[0068] Wherein, K is a first preset value, k is a second preset value, and k≤K-1.

[0069] In this embodiment, the calculation process of the cooling load is often affected by random factors. If the influence of the random load on the cooling load is not eliminated, its proportion in the cooling load will further increase, affecting the correlation with meteorological factors, resulting in a large deviation between the final calculation result of the cooling load and the actual result. Therefore, it is necessary to further eliminate the random load from the cooling load. The cooling load of each industry is obtained by subtracting the basic load from the electricity load demand of each industry, and then subtracting the random load. The cooling load of the i-th industry in week r of the cooling load in the current time period is as follows:

[0070] P 降温,r,i =P 需求,r,i -P 基础,r,i -P 随机,r,i

[0071] In the above formula, P 降温,r,i is the cooling load of the i-th industry in week r in the current time period, P 需求,r,i is the demand load of the i-th industry in week r of the current time period, P 基础,r,i is the base load of the industry i in week r of the current time period, P 随机,r,i is the random load of the i-th industry in week r of the current time period.

[0072] In one embodiment, the base load of the i-th industry in week r of the cooling load in the current time period is as follows:

[0073] P 基础,r,i =P 基线,r,i ×(1+α r ) y +P 新报装,r,i ×bP 销户,r,i

[0074] In the above formula, P 基线,r,i is the baseline load of the industry in week r of the baseline sample week, α r is the average weekly growth rate coefficient corresponding to week r in the baseline sample week, P 新报装,r,i is the newly reported electricity load of industry i in week r of the current time period, b is the increase coefficient of business expansion, P 销户,r,iis the account cancellation load of the ith industry in week r of the current time period, and y is the interval between the week to which week r belongs and the baseline sample week in the current time period.

[0075] In one embodiment, a total of N weeks (calculated by natural weeks) from March to May in spring are selected as the baseline sample period. When the temperature is around 20°C and there is no large-scale rainfall or other extreme weather conditions, the load demand on the rth day (r = 1, 2, ..., 7) of the nth week (n = 1, 2, ..., N) is recorded and identified as the rth typical day without cooling load in the region. The load data is averaged according to the typical day type to obtain the baseline load of the corresponding industry at the corresponding time. The baseline load of the i-th industry in the week r of the baseline sample week is as follows:

[0076]

[0077] In the above formula, P 需求,n,r,i is the load demand of industry i in week r of the nth baseline sample week, and N is the total number of baseline sample weeks.

[0078] In one embodiment, the average weekly growth rate coefficient corresponding to the week r within the baseline sample week is as follows:

[0079]

[0080] In the above formula, P r,N is the 96-point average load of week r within the Nth baseline sample week, P r,1 is the 96-point average load of week r within the first baseline sample week.

[0081] In one embodiment, the demand load of the i-th industry in week r of the current time period is as follows:

[0082] P 需求,r,i =P 实际,r,i +P 管理负荷,r,i +P 分布式光伏,r,i

[0083] In the above formula, P 实际,r,i is the actual electricity load of the i-th industry in week r of the current time period, P 管理负荷,r,i is the management load of the industry i in week r of the current time period, P 分布式光伏,r,i It is the self-generated and self-used load of distributed photovoltaic power generation in the i-th industry in week r of the current time period.

[0084] In one embodiment, the actual electricity load of the ith industry in week r during the current time period is as follows:

[0085] P 实际,r,i =P r ' ,i +Q r,iQ r,全行业 ×△P 偏差

[0086] In the above formula, P r ' ,i Q is the electricity load of industry i in week r of the current time period, excluding line loss. r,i is the daily electricity consumption of industry i in week r of the current time period, Q r,全行业 is the total daily electricity consumption of the entire industry in week r during the current time period, △P 偏差 For power consumption deviation.

[0087] In one embodiment, the power usage deviation is as follows:

[0088] △P 偏差 =P 调度 -P 全行业 -△P 厂用电 -△P 自备电厂

[0089] In the above formula, P 调度 is the dispatching load, P 全行业 is the electricity load of the entire industry, △P 厂用电 is the plant power load, △P 自备电厂 It is the electricity load of self-generated electricity in the captive power plant.

[0090] In this embodiment, a variational modal decomposition algorithm is used to transform the cooling load calculation results from the time domain to the frequency domain, and the frequency domain is automatically decomposed into a defined number of frequency segments, ultimately obtaining K modal components with a certain bandwidth. The low-frequency component has a clearer time domain waveform, and its variation trend is the same as that of the input load, exhibiting a certain load periodic variation pattern, and is less affected by random factors. The high-frequency component, on the other hand, has a more complex waveform, with large amplitude fluctuations and unclear variation patterns, which are primarily caused by random factors. Therefore, the decomposed high-frequency component is removed as a random load, while the low-frequency component is retained as the cooling load.

[0091] At the same time, setting a reasonable K value can further improve the decomposition effect of the variational mode decomposition algorithm. If the K value is too large, the original signal will be over-decomposed, resulting in mode overlap. If the K value is too small, the original signal will be under-decomposed, causing some information in the signal to be lost. The Pearson correlation analysis method is introduced to calculate the correlation between the low-frequency component and the meteorological condition. Specifically, the process of determining the first preset value and the second preset value includes:

[0092] Step a. Initialize K to be equal to a first preset starting value;

[0093] Step b. Using the variational mode decomposition algorithm to divide the cooling load in the historical period into K modal components;

[0094] Step c. Initialize k=1;

[0095] Step d. Calculating the correlation coefficient between the sum of the first k low-frequency components of the K modal components and the meteorological characteristic data within the historical period;

[0096] Step e. Determine whether k is equal to K-1. If so, output the correlation coefficient and execute step f. Otherwise, set k=k+1 and return to step d.

[0097] Step f. Determine whether K is equal to the first preset termination value. If so, use the values of K and k corresponding to the maximum correlation coefficient as the first preset value and the second preset value respectively. Otherwise, set K=K+1 and return to step a.

[0098] The value range of K is generally 3-9, and the correlation coefficient is as follows:

[0099]

[0100] In the above formula, T m is the mth element value in the sum of the first k low-frequency components in the K modal components, is the mean value of each element in the sum of the first k low-frequency components in the K modal components, P m is the mth element value in the meteorological characteristic data during the historical period, is the mean value of each element in the meteorological characteristic data within the historical period, r is the correlation coefficient, and M is the preset number of elements.

[0101] In this embodiment, the training process of the pre-trained long short-term memory neural network model includes:

[0102] The cooling load in the historical time period is divided into K modal components using a variational modal decomposition algorithm, and the first k low-frequency components of the K modal components are obtained;

[0103] Constructing training data using the k low-frequency components, meteorological characteristic data within a historical time period, and cooling load;

[0104] The training data is used to train the initial long short-term memory neural network model.

[0105] In a specific embodiment, the present invention uses electricity consumption data of a certain area to calculate the temperature control load. The area is hot in summer and there is no centralized heating in winter. Air conditioning, heat pumps and other equipment are basically used for cooling and heating. Figure 2The figure shows the region's daily maximum and average loads for 2022. The figure shows that the maximum and average loads between April 10th and May 20th are the lowest for the year. This is because the region enjoys warm weather from April to May, with an average temperature around 18°C and little rainfall, eliminating the need for temperature-controlled loads. Starting in late May and early June, the maximum and average loads begin to rise steadily, reaching a peak around mid-August. At this time, the region's average temperature remains around 32°C, creating high temperatures and high humidity, creating a muggy and humid climate. Both residential and commercial temperature-controlled loads operate at high power, leading to a peak in the region's daily maximum and average loads. Temperatures then gradually cool, and the region's daily maximum and average loads gradually decline, reaching their lowest point around October. Cooler temperatures then trigger a further increase in load due to the use of heating equipment. Furthermore, the average daily load around October in the second half of the year is slightly higher than the average daily load from April 10th to May 20th in the first half of the year. This is because the region's base load has increased with economic development.

[0106] Therefore, the present invention selects the load of the area from April 10 to May 20 (excluding the May Day holiday) as the load day without temperature control, and calculates 7 baseline loads according to formula 4, such as Figure 3 As shown. According to formula 5, the weekly growth rate of different baseline loads is calculated, and the weekly growth rate matrix of different day types is obtained as a = [0.0144 0.0069 0.0173 0.0151 0.0231 0.0153 0.0115]T. Then, according to formula 6, the basic load of different day types is obtained. The original load is subtracted from the basic load to obtain the cooling-random load. The Pearson correlation coefficient of the cooling-random daily maximum load and the daily maximum temperature, daily minimum temperature, and daily average temperature is calculated by Figure 4 It can be seen that the maximum average temperature correlation coefficient is 0.876. Subsequently, the correlation between cooling load and average temperature is used as a screening index for variational mode decomposition.

[0107] Set K to 3 to 9 respectively, select the sum of several low-frequency load components as the cooling load, and calculate the correlation with the average temperature. The results are shown in Table 1. When the sum of the first two load components is used as the cooling load when K = 4, the correlation with the average temperature is the highest, 0.8943, which is 2.09% higher than the correlation when the random load is not deleted. The load decomposition is shown in Table 1. Figure 5 As shown, in order to intuitively display the cooling load results, the cooling load of the first ten days is intercepted and compared with the cooling-random load, as shown in Figure 6 As shown in Figure 2, it can be clearly seen that the cooling load is smoother than before the random load is deleted.

[0108] Table 1

[0109] Sum of load components K＝3 K＝4 K＝5 K＝6 K＝7 K＝8 K＝9 First load component 0.8618 0.8610 0.8577 0.8572 0.8572 0.8567 0.8567 First two load components 0.8937 0.8943 0.8933 0.8937 0.8937 0.8940 0.8940 The first three load components / 0.8743 0.8694 0.8693 0.8693 0.8686 0.8686 The first 4 load components / / 0.8744 0.8743 0.8746 0.8725 0.8725 First 5 load components / / / 0.8748 0.8744 0.8741 0.8745 First 6 load components / / / / 0.8748 0.8746 0.8741 First 7 load components / / / / / 0.8749 0.8748 First 8 load components / / / / / / 0.8749

[0110] Based on the calculated cooling-random load and cooling load, the LSTM prediction model is used to predict them, and the basic cooling prediction accumulation method is used to predict the total power load. The power load forecasts from August 25th to 31st are intercepted for comparison. Figure 7 、 8 As shown in the figure, it can be seen that the power load forecast accuracy is higher after eliminating the influence of random factors, and the average power load forecast accuracy is improved by 0.7% compared with the power load forecast without eliminating the influence of random factors.

[0111] Example 2

[0112] Based on the same inventive concept, the present invention further provides a temperature reduction load device that takes into account the influence of random loads, and the temperature reduction load device that takes into account the influence of random loads comprises:

[0113] A decomposition module is used to divide the cooling load in the current time period into K modal components using a variational modal decomposition algorithm, and obtain the first k low-frequency components of the K modal components;

[0114] An analysis module is configured to use the k low-frequency components and the meteorological characteristic data within the current time period as input data of a pre-trained long short-term memory neural network model, and obtain a cooling load within a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model;

[0115] Wherein, K is a first preset value, k is a second preset value, and k≤K-1.

[0116] Preferably, the cooling load of the i-th industry in week r in the cooling load during the current time period is as follows:

[0117] P 降温,r,i =P 需求,r,i -P 基础,r,i -P 随机,r,i

[0118] In the above formula, P 降温,r,i is the cooling load of the i-th industry in week r in the current time period, P 需求,r,i is the demand load of the i-th industry in week r of the current time period, P 基础,r,i is the base load of the industry i in week r of the current time period, P 随机,r,i is the random load of the i-th industry in week r of the current time period.

[0119] Furthermore, the base load of the i-th industry in week r of the cooling load in the current time period is as follows:

[0120] P 基础,r,i =P 基线,r,i ×(1+αr ) y +P 新报装,r,i ×bP 销户,r,i

[0121] In the above formula, P 基线,r,i is the baseline load of the industry in week r of the baseline sample week, α r is the average weekly growth rate coefficient corresponding to week r in the baseline sample week, P 新报装,r,i is the newly reported electricity load of industry i in week r of the current time period, b is the increase coefficient of business expansion, P 销户,r,i is the account cancellation load of the ith industry in week r of the current time period, and y is the interval between the week to which week r belongs and the baseline sample week in the current time period.

[0122] Furthermore, the baseline load of the ith industry in week r of the baseline sample week is as follows:

[0123]

[0124] In the above formula, P 需求,n,r,i is the load demand of industry i in week r of the nth baseline sample week, and N is the total number of baseline sample weeks.

[0125] Furthermore, the average weekly growth rate coefficient corresponding to week r within the baseline sample week is as follows:

[0126]

[0127] In the above formula, P r,N is the 96-point average load of week r within the Nth baseline sample week, P r,1 is the 96-point average load of week r within the first baseline sample week.

[0128] Furthermore, the demand load of the i-th industry in week r during the current time period is as follows:

[0129] P 需求,r,i =P 实际,r,i +P 管理负荷,r,i +P 分布式光伏,r,i

[0130] In the above formula, P 实际,r,i is the actual electricity load of the i-th industry in week r of the current time period, P 管理负荷,r,i is the management load of the industry i in week r of the current time period, P 分布式光伏,r,i It is the self-generated and self-used load of distributed photovoltaic power generation in the i-th industry in week r of the current time period.

[0131] Furthermore, the actual electricity load of the i-th industry in week r during the current time period is as follows:

[0132] P实际,r,i =P r ' ,i +Q r,i Q r,全行业 ×△P 偏差

[0133] In the above formula, P r ' ,i Q is the electricity load of industry i in week r of the current time period, excluding line loss. r,i is the daily electricity consumption of industry i in week r of the current time period, Q r,全行业 is the total daily electricity consumption of the entire industry in week r during the current time period, △P 偏差 For power consumption deviation.

[0134] Furthermore, the power consumption deviation is as follows:

[0135] △P 偏差 =P 调度 -P 全行业 -△P 厂用电 -△P 自备电厂

[0136] In the above formula, P 调度 is the dispatching load, P 全行业 is the electricity load of the entire industry, △P 厂用电 is the plant power load, △P 自备电厂 It is the electricity load of self-generated electricity in the captive power plant.

[0137] Preferably, the process of determining the first preset value and the second preset value includes:

[0138] Step a. Initialize K to be equal to a first preset starting value;

[0139] Step b. Using the variational mode decomposition algorithm to divide the cooling load in the historical period into K modal components;

[0140] Step c. Initialize k=1;

[0141] Step d. Calculating the correlation coefficient between the sum of the first k low-frequency components of the K modal components and the meteorological characteristic data within the historical period;

[0142] Step e. Determine whether k is equal to K-1. If so, output the correlation coefficient and execute step f. Otherwise, set k=k+1 and return to step d.

[0143] Step f. Determine whether K is equal to the first preset termination value. If so, use the values of K and k corresponding to the maximum correlation coefficient as the first preset value and the second preset value respectively. Otherwise, set K=K+1 and return to step a.

[0144] Furthermore, the correlation coefficient is as follows:

[0145]

[0146] In the above formula, T m is the mth element value in the sum of the first k low-frequency components in the K modal components, is the mean value of each element in the sum of the first k low-frequency components in the K modal components, P m is the mth element value in the meteorological characteristic data during the historical period, is the mean value of each element in the meteorological characteristic data within the historical period, r is the correlation coefficient, and M is the preset number of elements.

[0147] Preferably, the training process of the pre-trained long short-term memory neural network model includes:

[0148] The cooling load in the historical time period is divided into K modal components using a variational modal decomposition algorithm, and the first k low-frequency components of the K modal components are obtained;

[0149] Constructing training data using the k low-frequency components, meteorological characteristic data within a historical time period, and cooling load;

[0150] The training data is used to train the initial long short-term memory neural network model.

[0151] Example 3

[0152] Based on the same inventive concept, the present invention also provides a computer device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function, so as to implement the steps of a cooling load method considering the influence of random load in the above embodiment.

[0153] Example 4

[0154] Based on the same inventive concept, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It can be understood that the computer-readable storage medium here can include both built-in storage media in the computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of a method for reducing the temperature load that takes into account the influence of random loads in the above embodiment.

[0155] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0156] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0157] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0158] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for reducing temperature load considering the influence of random load, characterized in that: The method comprises: Using a variational modal decomposition algorithm to divide the cooling load in the current time period into K modal components, and obtaining the first k low-frequency components of the K modal components; Using the k low-frequency components and meteorological characteristic data within the current time period as input data of a pre-trained long short-term memory neural network model, obtaining a cooling load within a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model; Wherein, K is a first preset value, k is a second preset value, and k≤K-1; The process of determining the first preset value and the second preset value includes: Step a. Initialize K to be equal to a first preset starting value; Step b. Using the variational mode decomposition algorithm to divide the cooling load in the historical period into K modal components; Step c. Initialize k=1; Step d. Calculating the correlation coefficient between the sum of the first k low-frequency components of the K modal components and the meteorological characteristic data within the historical period; Step e. Determine whether k is equal to K-1. If so, output the correlation coefficient and execute step f. Otherwise, set k=k+1 and return to step d. Step f. Determine whether K is equal to the first preset end value. If so, set the values of K and k corresponding to the maximum correlation coefficient as the first preset value and the second preset value, respectively. Otherwise, set K = K + 1 and return to step a. The correlation coefficient is as follows: In the above formula, T m is the mth element value in the sum of the first k low-frequency components in the K modal components, is the mean value of each element in the sum of the first k low-frequency components in the K modal components, P m is the mth element value in the meteorological characteristic data during the historical period, is the mean value of each element in the meteorological characteristic data within the historical period, r is the correlation coefficient, and M is the preset number of elements.

2. The method according to claim 1, wherein The cooling load of the i-th industry in week r during the current time period is as follows: P 降温,r,i =P 需求,r,i -P 基础,r,i -P 随机,r,i In the above formula, P 降温,r,i is the cooling load of the i-th industry in week r in the current time period, P 需求,r,i is the demand load of the i-th industry in week r of the current time period, P 基础,r,i is the base load of the industry i in week r of the current time period, P 随机,r,i is the random load of the i-th industry in week r of the current time period.

3. The method according to claim 2, wherein The basic load of the cooling load in the current time period in the i-th industry in week r is as follows: P 基础,r,i =P 基线,r,i ×(1+α r ) y +P 新报装,r,i ×bP 销户,r,i In the above formula, P 基线,r,i is the baseline load of the industry in week r of the baseline sample week, α r is the average weekly growth rate coefficient corresponding to week r in the baseline sample week, P 新报装,r,i is the newly reported electricity load of industry i in week r of the current time period, b is the increase coefficient of business expansion, P 销户,r,i is the account cancellation load of the ith industry in week r of the current time period, and y is the interval between the week to which week r belongs and the baseline sample week in the current time period.

4. The method according to claim 3, wherein The baseline load of the industry in week r of the baseline sample week is as follows: In the above formula, P 需求,n,r,i is the load demand of industry i in week r of the nth baseline sample week, and N is the total number of baseline sample weeks.

5. The method according to claim 4, wherein The average weekly growth rate coefficient corresponding to week r within the baseline sample week is as follows: In the above formula, P r,N is the 96-point average load of week r within the Nth baseline sample week, P r,1 is the 96-point average load of week r within the first baseline sample week.

6. The method according to claim 5, wherein The demand load of the i-th industry in week r of the current time period is as follows: P 需求,r,i =P 实际,r,i +P 管理负荷,r,i +P 分布式光伏,r,i In the above formula, P 实际,r,i is the actual electricity load of the i-th industry in week r of the current time period, P 管理负荷,r,i is the management load of the industry i in week r of the current time period, P 分布式光伏,r,i It is the self-generated and self-used load of distributed photovoltaic power generation in the i-th industry in week r of the current time period.

7. The method according to claim 6, wherein The actual power load of the i-th industry in week r during the current time period is as follows: P 实际,r,i =P′ r,i +Q r,i / Q r,全行业 ×ΔP 偏差 In the above formula, P′ r,i Q is the electricity load of industry i in week r of the current time period, excluding line loss. r,i is the daily electricity consumption of industry i in week r of the current time period, Q r,全行业 is the total daily electricity consumption of the entire industry in week r during the current time period, ΔP 偏差 For power consumption deviation.

8. The method according to claim 7, wherein The power consumption deviation is as follows: ΔP 偏差 =P 调度 -P 全行业 -ΔP 厂用电 -ΔP 自备电厂 In the above formula, P 调度 is the dispatching load, P 全行业 is the electricity load of the entire industry, ΔP 厂用电 is the plant power load, ΔP 自备电厂 It is the electricity load of self-generated electricity in the captive power plant.

9. The method according to claim 1, wherein The training process of the pre-trained long short-term memory neural network model includes: The cooling load in the historical time period is divided into K modal components using a variational modal decomposition algorithm, and the first k low-frequency components of the K modal components are obtained; Constructing training data using the k low-frequency components, meteorological characteristic data within a historical time period, and cooling load; The training data is used to train the initial long short-term memory neural network model.

10. A device for reducing the temperature load by considering the random load effect according to any one of claims 1 to 9, characterized in that: The device comprises: A decomposition module is used to divide the cooling load in the current time period into K modal components using a variational modal decomposition algorithm, and obtain the first k low-frequency components of the K modal components; An analysis module is configured to use the k low-frequency components and the meteorological characteristic data within the current time period as input data of a pre-trained long short-term memory neural network model, and obtain a cooling load within a predicted time period corresponding to the current time period output by the pre-trained long short-term memory neural network model; Wherein, K is a first preset value, k is a second preset value, and k≤K-1.

11. A computer device, characterized in that: include: one or more processors; The processor is configured to execute one or more programs; When the one or more programs are executed by the one or more processors, the method for reducing the cooling load considering the influence of random loads as claimed in any one of claims 1 to 9 is implemented.

12. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed, the method for reducing the temperature load by considering the influence of random loads as claimed in any one of claims 1 to 9 is implemented.

Images