Industrial interruptible load identification and quantification method, system and equipment based on power load data driving and medium

Abstract

The invention discloses an industrial interruptible load identification and quantification method, system and device based on power load data driving and a medium. The method comprises the following steps: acquiring power load data of a user and preprocessing the power load data; performing primary AP clustering by taking daily power load data of the user as an index according to the power load data of the user on workdays in a certain period; calculating a characteristic index of the user power load data of each clustering center obtained by the primary clustering; performing secondary AP clustering on the primary clustering result according to the characteristic index of the user power load data to obtain all power consumption modes of the user; identifying an interruptible load power utilization mode of the user; and interruptible load resources of the user are quantified according to the load of the user under the normal production condition and the load in the interruption period. According to the method, the most possible interruptible load resources are mined through two times of clustering; based on power load data driving, the identification speed of the industrial interruptible load resources is improved, and the investigation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method, system, equipment and medium for identifying and quantifying industrial interruptible loads driven by electric power load data. Background Art

[0002] Power grid data shows that in recent years, there has been significant pressure to reduce peak loads and fill valleys, and the utilization rate of grid investment and construction has been low. To alleviate this pressure, the power sector is gradually promoting the construction of new power systems and expanding the types and volume of flexible resources to address the increasingly tight power supply and demand situation. Among the various types of flexible resources, load-side resources, with their low development and utilization costs and high interactivity, are a key resource for mitigating grid peaks.

[0003] Industrial production is characterized by high power consumption, high loads, and multiple process flows. These characteristics provide a certain amount of interruptibility for industrial loads and provide a good peak-shaving effect. However, in the actual work of power grid marketing and dispatching departments, industrial interruptible loads are still obtained through surveys or contracts, which are subject to underreporting and concealment by enterprises. This results in a limited number of available industrial interruptible load resources in actual load regulation, or the regulation of industrial loads can only be carried out in a "one-size-fits-all" manner, which greatly harms regional interests. Therefore, it is urgent to carry out the identification and quantification of industrial interruptible load resources based on power data. Based on existing research, the most common method for mining industrial load demand potential is to design interruptible potential indicators based on load factors such as load rate and peak-to-valley difference. However, electricity usage patterns vary among different users, and relying solely on general indicators does not reflect the actual production conditions of enterprises and is subject to significant errors. At the same time, some studies have quantified industrial demand potential by combining power data with equipment physical characteristics. However, these studies are based on a clear list of industrial interruptible users, which is not in line with actual work conditions. Furthermore, calculating the demand potential of industrial interruptible users based on physical parameters requires a large number of equipment characteristics, which is not universally applicable. Summary of the Invention

[0004] In response to the above-mentioned deficiencies in the existing technology, the present invention provides a method, system, equipment and medium for identifying and quantifying industrial interruptible loads based on power load data, which solves two technical problems: one is how to mine common power consumption patterns from load data and extract interruptible features; the other is how to quantify the interruptible capacity only through load data based on the extraction of interruptible features.

[0005] In a first aspect, a method for identifying and quantifying industrial interruptible loads based on power load data is provided, comprising the following steps:

[0006] S1: Obtain the user's power load data and pre-process it;

[0007] S2: Based on the power load data of users on weekdays within a certain period, the initial AP clustering (Affinity Propagation) is performed using the user's daily power load data as an indicator;

[0008] S3: Calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering;

[0009] S4: Perform secondary AP clustering on the cluster centers obtained from the initial AP clustering based on the characteristic indicators of the user's power load data to obtain all the user's power consumption patterns;

[0010] S5: Identify the user's interruptible load power consumption pattern, specifically: for each user's power consumption pattern, identify it according to the following standards: calculate the load change rate at each time node of the power consumption pattern, define the load change at the time node when the change rate is greater than the decrease amplitude threshold q as 1, and the load change at the time node when the change rate is less than q as 0; traverse each time node, and define the transition point from 0 to 1 as the pre-interruption inflection point; count the load decrease at the subsequent moments before the interruption inflection point, and define the point where the load stops decreasing as the post-interruption inflection point, and within a period of time after the post-interruption inflection point, the load fluctuation at adjacent moments must not exceed the preset fluctuation threshold v; when both the pre-interruption inflection point and the post-interruption inflection point appear in the user's power consumption pattern, the power consumption pattern is considered to be an interruptible load power consumption pattern;

[0011] S6: Quantify the user's interruptible load resources based on the user's load under normal production conditions and the load during the interruption period.

[0012] Furthermore, the preprocessing process includes removing outliers and filling missing values.

[0013] Furthermore, the characteristic indicators of the user power load data include daily load rate, daily peak-to-valley difference and daily maximum load utilization hours.

[0014] Furthermore, the AP clustering method can find all representative curves of users without pre-setting clusters, and is very suitable for finding all power usage patterns of users, including the following steps:

[0015] S21: Set the attenuation coefficient λ and the maximum number of iterations M;

[0016] S22: Initialize the attraction matrix R and the attribution matrix A;

[0017] S23: Update the attraction matrix R. The update formula is:

[0018]

[0019] Where r m+1(i,k) represents the degree to which the load data of the kth day in the m+1th iteration is suitable as the cluster center of the load data of the ith day; S(i,k) represents the similarity between the load data of the ith day and the load data of the kth day; a m (i, j) represents the suitability of the load data of day i in the mth iteration to select the load data of day k as its cluster center;

[0020] S24: Update the attribution matrix A. The update formula is:

[0021]

[0022] S25: Attenuate the attraction matrix R and the attribution matrix A according to the attenuation coefficient λ. The attenuation formula is:

[0023]

[0024] S26: Iterate steps S23-S25 until the iteration termination condition is met;

[0025] S27: Output the largest element k in the R+A matrix as the clustering center of the data.

[0026] Furthermore, in the initial AP clustering process, S(i,k) is obtained by taking the negative of the square of the Euclidean distance between the node load vectors of the load data on the i-th day and the load data on the k-th day; in the secondary AP clustering process, S(i,k) is obtained by taking the negative of the square of the Euclidean distance between the characteristic index vectors of the load data on the i-th day and the load data on the k-th day.

[0027] Furthermore, the identifying of the user's interruptible load power usage pattern specifically includes:

[0028] S51: Set the drop threshold q, q>0;

[0029] S52: Calculate the multi-order moving average of each time node of each power consumption mode of the user as the average load of the time node;

[0030] S53: Calculate the average load change rate p at each time node t , specifically: change rate = (current time node load - next time node load) / current time node load;

[0031] S54: Determine load changes When the average load change rate p t When it is greater than the falling threshold q, Recorded as 1, otherwise recorded as 0;

[0032] S55: Determine the turning point before the interruption At a certain time point When it changes from 0 to 1, the point is recorded as the turning point before the interruption;

[0033] S56: Set the fluctuation threshold v, v>0;

[0034] S57: Determine load fluctuations When the average load change rate p t When the absolute value of is greater than the fluctuation threshold v, Recorded as 1, otherwise recorded as 0;

[0035] S58: Determine the turning point after the interruption At a certain time point When it changes from 1 to 0, and multiple time nodes specified after this time node and If the values ​​of are all 0, then the point is recorded as the turning point after the interruption;

[0036] S59: Identify an interruptible power usage pattern, and when at least one pair of interruption inflection points appears in the user's power usage pattern, identify the power usage pattern as an interruptible load power usage pattern;

[0037] S510: Output the inflection point pair with the largest difference

[0038] Furthermore, the user's interruptible load resource is calculated by the following formula:

[0039]

[0040] Where L is the user's interruptible load resource, is the average load of the user under normal production conditions, The average load of users during the outage.

[0041] Secondly, a system for identifying and quantifying industrial interruptible loads driven by power load data is provided, including:

[0042] Data preprocessing module, used to obtain the user's power load data and perform preprocessing;

[0043] The initial clustering module is used to perform initial AP clustering based on the power load data of users on weekdays within a certain period of time, taking the user's daily power load data as an indicator;

[0044] The characteristic index extraction module is used to calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering;

[0045] The secondary clustering module is used to perform secondary AP clustering on the cluster centers obtained by the initial AP clustering according to the characteristic indicators of the user's power load data to obtain all the power consumption patterns of the user;

[0046] The interruptible load power consumption pattern identification module is used to identify the user's interruptible load power consumption pattern. Specifically, for each power consumption pattern of the user, identification is performed according to the following standards: the load change rate at each time node of the power consumption pattern is calculated, and the load change at the time node when the change rate is greater than the decrease amplitude threshold q is defined as 1, and the load change at the time node when the change rate is less than q is defined as 0; each time node is traversed, and the conversion point from 0 to 1 is defined as the pre-interruption inflection point; the load decrease at the subsequent moments of the pre-interruption inflection point is counted, and the point where the load stops decreasing is defined as the post-interruption inflection point, and the load fluctuation at adjacent moments must not exceed the preset fluctuation threshold v within a period of time after the post-interruption inflection point; when both the pre-interruption inflection point and the post-interruption inflection point appear in the user's power consumption pattern, the power consumption pattern is considered to be an interruptible load power consumption pattern;

[0047] The interruptible load resource quantification module is used to quantify the user's interruptible load resources according to the user's load under normal production conditions and the load during the interruption period.

[0048] According to a third aspect, an electronic device is provided, including:

[0049] Memory on which computer programs or instructions are stored;

[0050] The processor is used to load and execute the computer program or instructions to implement the industrial interruptible load identification and quantification method driven by power load data as described above.

[0051] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the method for identifying and quantifying industrial interruptible loads driven by power load data as described above is implemented.

[0052] The present invention proposes a method, system, device and medium for identifying and quantifying industrial interruptible loads based on power load data. Compared with the existing technology, the method has the following advantages: First, the method fully considers the various power consumption patterns of users, deeply mines the power consumption characteristics of users, and mines the most likely interruptible load resources through two clusterings, which is helpful for load management during peak power consumption periods; Second, the method is driven by power load data, and the results are obtained more quickly and objectively. On the one hand, this avoids the false reporting or omission of interruptible load resources by large industrial users to improve enterprise production, and on the other hand, it improves the identification speed of industrial interruptible load resources, reduces the investigation cost, and is conducive to better promotion and application; Third, the identification and quantification of industrial interruptible load resources of the present invention are more accurate and precise. In the process of identifying and quantifying industrial interruptible load resources, the present invention can accurately identify the inflection points before and after the interruption of a specific industrial interruptible load resource and quantify the size of the interruptible load resource, which helps the power grid system to accurately dispatch, implement peak shaving and valley filling policies, and reduce power losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 This is a flow chart of an industrial interruptible load identification and quantification method based on power load data drive provided by an embodiment of the present invention;

[0055] Figure 2 This is a load diagram of the 96 time nodes of the xx company in 2023 provided by an embodiment of the present invention;

[0056] Figure 3 This is a cluster center load diagram of a certain city's xx company after initial AP clustering provided by an embodiment of the present invention;

[0057] Figure 4 This is a diagram of electricity consumption patterns of users of a company in a certain city in 2023, provided by an embodiment of the present invention;

[0058] Figure 5 This is a graph of the interruption inflection point identification results of a company in a certain city provided by an embodiment of the present invention;

[0059] Figure 6 This is an interruptible load diagram of a company in a certain city provided by an embodiment of the present invention;

[0060] Figure 7This is a load diagram of the 96 time nodes of a certain province xx company in 2023 provided by an embodiment of the present invention;

[0061] Figure 8 This is a diagram of the initial AP cluster center load of a company in a certain province, provided by an embodiment of the present invention;

[0062] Figure 9 This is a diagram of the electricity consumption pattern of a company in a certain province in 2023, provided by an embodiment of the present invention;

[0063] Figure 10 This is a category 1 interruptible load diagram of a company in a certain province provided by an embodiment of the present invention;

[0064] Figure 11 This is the category 3 interruptible load of a certain province xx company provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0065] To make the objectives, technical solutions, and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other implementations obtained by those of ordinary skill in the art without inventive effort are within the scope of protection of the present invention.

[0066] Example 1

[0067] In this embodiment, the power data used is the daily 96-node load data of a certain city xx company in 2023. The power load situation of the whole year is shown in the attached Figure 2 .

[0068] Specifically, this embodiment provides an industrial interruptible load identification and quantification method based on power load data drive, such as Figure 1 As shown, the following steps are included:

[0069] S1: Obtain the user's power load data and preprocess it. The preprocessing process includes removing outliers and filling missing values.

[0070] Specifically, they are: the Lagrange difference method to fill missing values ​​and the 6 times standard deviation principle to find outliers.

[0071] In this embodiment, the preprocessing operation is as follows: first, the number of missing values ​​in the user load data is counted, and users with a missing value data ratio exceeding 20% ​​are eliminated; for the remaining users, their 6 times standard deviation data is calculated, and the data exceeding 6 times the standard deviation is marked as an outlier, and then the missing values ​​and outliers are filled using the Lagrange interpolation method.

[0072] S2: Based on the power load data of the users on working days within a certain period (one year in this embodiment), the initial AP clustering is performed using the user's daily power load data as an indicator.

[0073] Specifically, the initial AP clustering includes the following steps:

[0074] S21: Set the attenuation coefficient λ and the maximum number of iterations M;

[0075] S22: Initialize the attraction matrix R and the attribution matrix A;

[0076] S23: Update the attraction matrix R. The update formula is:

[0077]

[0078] Where r m+1 (i, k) represents the degree to which the load data of the kth day in the m+1th iteration is suitable as the cluster center of the load data of the ith day; S(i, k) represents the similarity between the load data of the ith day and the load data of the kth day, and S(i, k) is obtained by taking the negative of the square of the Euclidean distance between the node load vectors (in this embodiment, the vectors composed of 96 node loads) of the load data of the ith day and the load data of the kth day; a m (i, j) represents the suitability of the load data of day i in the mth iteration to select the load data of day k as its cluster center;

[0079] S24: Update the attribution matrix A. The update formula is:

[0080]

[0081] S25: Attenuate the attraction matrix R and the attribution matrix A according to the attenuation coefficient λ. The attenuation formula is:

[0082]

[0083] S26: Iterate steps S23-S25 until the iteration termination condition is met;

[0084] S27: Output the largest element k in the R+A matrix as the cluster center of the data; specifically: assume that the k value that maximizes a(i,k)+r(i,k) is k', if i=k', then the load data of the i-th day is the cluster center, if, then the load data of the i-th day belongs to the cluster center k' (load data of the k-th day).

[0085] In this embodiment, there are 22 cluster centers in the initial clustering, and the power load situation of the cluster centers is as follows: Figure 3As shown in the figure, the horizontal axis represents the time node of 24 hours a day, and the vertical axis represents the user's load level. It can be found that compared with the original data, the clustered power load situation can more clearly reflect the various production power consumption modes of a certain city's xx company.

[0086] S3: Calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering.

[0087] In this embodiment, the characteristic indicators of the user power load data include daily load rate, daily peak-to-valley difference and daily maximum load utilization hours. The calculation formulas for each characteristic indicator are: Daily load rate: Among them, lv represents the daily load rate, l mean represents the average daily load, l max Indicates the daily maximum load. Daily peak-to-valley difference: Among them, ld represents the daily peak-to-valley difference, Indicates the peak load of the daily load curve, Indicates the valley load of the daily load curve. Daily maximum load utilization hours: Among them, lh is the daily maximum load utilization hours, and ∑l is the total daily load.

[0088] S4: Perform secondary AP clustering on the cluster centers obtained from the initial AP clustering according to the characteristic indicators of the user's power load data to obtain all the power usage patterns of the user.

[0089] Secondary AP clustering follows the same principles as primary AP clustering. The difference is that the characteristic indicators for primary AP clustering are node load data, while those for secondary AP clustering are daily load factor, daily peak-to-valley difference, and daily maximum load utilization hours. Therefore, in secondary AP clustering, S(i,k) is calculated by taking the negative of the square of the Euclidean distance between the characteristic indicator vectors for the load data on day i and day k.

[0090] In this embodiment, the results after the initial clustering (see Appendix Figure 3 ) obviously have similar electricity consumption patterns, so we consider clustering by the characteristic indicators of power load data. The characteristic indicators used for clustering are daily load rate, daily peak-to-valley difference and daily maximum load utilization hours. After clustering, there are four types of cluster centers (see Appendix Figure 4 ), category 0 represents the all-day production electricity consumption mode, category 1 represents the electricity consumption mode with slow production cessation, category 2 represents the no-production electricity consumption mode, and category 3 represents the production electricity consumption mode with a production reduction period, that is, the production mode with interruptible characteristics.

[0091] S5: Identify the user's interruptible load power consumption pattern, specifically: for each power consumption pattern of the user, identify it according to the following standards: calculate the change rate of the load at each time node of the power consumption pattern, define the load change at the time node when the change rate is greater than the decrease amplitude threshold q as 1, and the load change at the time node when the change rate is less than q as 0; traverse each time node, and define the conversion point from 0 to 1 as the pre-interruption inflection point; count the load drop at the subsequent moments before the interruption inflection point, and define the point where the load stops dropping as the post-interruption inflection point, and within a period of time after the post-interruption inflection point, the load fluctuation at adjacent moments must not exceed the preset fluctuation threshold v; when the pre-interruption inflection point and the post-interruption inflection point appear in the user's power consumption pattern at the same time, it is considered that the power consumption pattern has interruptible characteristics, and the power consumption pattern is an interruptible load power consumption pattern.

[0092] In this embodiment, identifying the user's interruptible load power usage pattern specifically includes the following steps:

[0093] S51: Set the drop threshold q, q>0;

[0094] S52: Calculate the multi-order moving average (hereinafter, the fifth-order moving average) of each time node of each power consumption mode of the user as the average load of the time node. The calculation formula is as follows:

[0095]

[0096] Where, represents the average load at time node t, l i represents the load at time node i;

[0097] S53: Calculate the average load change rate p at each time node t Specifically, the change rate = (current time node load - next time node load) / current time node load. The calculation formula is as follows:

[0098]

[0099] S54: Determine load changes When the average load change rate p at a certain time node t When it is greater than the falling threshold q, the time node Recorded as 1, otherwise recorded as 0;

[0100] S55: Determine the turning point before the interruption At a certain time point When it changes from 0 to 1, the point is recorded as the inflection point before the interruption. like Figure 5 As shown;

[0101] S56: Set the fluctuation threshold v, v>0;

[0102] S57: Determine load fluctuations When the average load change rate p at a certain time node t When the absolute value of is greater than the fluctuation threshold v, the Recorded as 1, otherwise recorded as 0;

[0103] S58: Determine the turning point after the interruption At a certain time point When the value changes from 1 to 0, and there are at least 5 time nodes within the 6 time nodes after the time node (at least 5 is to avoid the influence of a single load jump value on the identification) and If the values ​​are all 0, then the point is recorded as the inflection point after the interruption and the inflection point before the interruption. like Figure 5 As shown;

[0104] S59: Identify an interruptible power usage pattern, and when at least one pair of interruption inflection points appears in the user's power usage pattern, identify the power usage pattern as an interruptible load power usage pattern;

[0105] S510: Output the inflection point pair with the largest difference

[0106] In this embodiment, the four power usage patterns obtained in step S4 are analyzed, wherein only the power usage pattern belonging to category 3 is identified as an interruptible load power usage pattern, and the corresponding interruption inflection point pair is (44, 59).

[0107] S6: Quantify the user's interruptible load resources based on the user's load under normal production conditions and the load during the interruption period. The user's interruptible load resources are calculated using the following formula:

[0108]

[0109] Where L is the user's interruptible load resource, Provide users with a turning point before interruption (i.e. the average load under normal user production), Help users find a turning point after an interruption The average load during the outage.

[0110] In this implementation case, as shown in the attached Figure 6 As shown, the average load under normal production conditions is 5182 kW, the average load during the interruption period is 1691 kW, and the resulting interruptible capacity load is 3490 kW.

[0111] Example 2

[0112] In this embodiment, the power data used is the daily 96-node load data of a certain province xx company in 2023. The power load situation of the whole year is shown in the attached Figure 7 .

[0113] Specifically, this embodiment provides an industrial interruptible load identification and quantification method based on power load data drive, such as Figure 1 As shown, the following steps are included:

[0114] S1: Obtain the user's power load data and preprocess it. The preprocessing process includes removing outliers and filling missing values.

[0115] In this embodiment, the preprocessing operation is as follows: first, the number of missing values ​​in the user load data is counted, and users with a missing value data ratio exceeding 20% ​​are eliminated; for the remaining users, their 6 times standard deviation data is calculated, and the data exceeding 6 times the standard deviation is marked as an outlier, and then the missing values ​​and outliers are filled using the Lagrange interpolation method.

[0116] S2: Based on the power load data of users on weekdays throughout the year, the initial AP clustering is performed using the user's daily power load data as an indicator.

[0117] In this embodiment, there are 22 cluster centers in the initial clustering, and the power load situation of the cluster centers is as follows: Figure 8 As shown in the figure, the horizontal axis represents the time node of 24 hours a day, and the vertical axis represents the user's load level. It can be found that compared with the original data, the clustered power load situation can more clearly reflect the various production power consumption modes of xx company in a certain province.

[0118] S3: Calculate characteristic indicators of user power load data corresponding to each cluster center obtained by the initial AP clustering. In this embodiment, the characteristic indicators of user power load data include daily load rate, daily peak-to-valley difference, and daily maximum load utilization hours.

[0119] S4: Perform secondary AP clustering on the cluster centers obtained from the initial AP clustering according to the characteristic indicators of the user's power load data to obtain all the power usage patterns of the user.

[0120] In this embodiment, the results after the initial clustering (see Appendix Figure 8 ) obviously have similar electricity consumption patterns, so we consider clustering by the characteristic indicators of power load data. The characteristic indicators used for clustering are daily load rate, daily peak-to-valley difference and daily maximum load utilization hours. There are five types of cluster centers after clustering (see Appendix Figure 9 ), category 0 and category 4 represent two all-day production electricity consumption modes, category 2 represents the nighttime production electricity consumption mode, and category 1 and category 3 represent two different production electricity consumption modes that can reduce production, that is, production modes with interruptible characteristics.

[0121] S5: Identify the user's interruptible load power consumption pattern, specifically: for each power consumption pattern of the user, identify it according to the following standards: calculate the change rate of the load at each time node of the power consumption pattern, define the load change at the time node when the change rate is greater than the decrease amplitude threshold q as 1, and the load change at the time node when the change rate is less than q as 0; traverse each time node, and define the conversion point from 0 to 1 as the pre-interruption inflection point; count the load decrease at the subsequent moments before the interruption inflection point, and define the point where the load stops decreasing as the post-interruption inflection point, and within a period of time after the post-interruption inflection point, the load fluctuation at adjacent moments must not exceed the preset fluctuation threshold v; when the pre-interruption inflection point and the post-interruption inflection point appear in the user's power consumption pattern at the same time, the power consumption pattern is considered to be an interruptible load power consumption pattern.

[0122] In this embodiment, the five power usage patterns obtained in step S4 are analyzed, where category 1 and category 3 are identified as interruptible load power usage patterns, and the corresponding inflection point pairs are (31, 41) and (37, 51).

[0123] S6: Quantify the user's interruptible load resources based on the user's load under normal production conditions and the load during the interruption period. The user's interruptible load resources are calculated using the following formula:

[0124]

[0125] Where L is the user's interruptible load resource, For users in normal production conditions (inflection point before interruption The average load of For users during the interruption period (inflection point after the interruption The average load of the following).

[0126] In this implementation case, the two interruptible power modes of a certain province's xx company are shown in the attached Figure 10 , Attachment Figure 11 The average load of the obtained interruptible load power consumption mode category 1 under normal production conditions is 12045 kilowatts, the average load during the interruption period is about 6318 kilowatts, and the load of the obtained interruptible capacity is about 5727 kilowatts; the average load of the obtained interruptible load power consumption mode category 3 under normal production conditions is about 16163 kilowatts, the average load during the interruption period is about 3929 kilowatts, and the load of the obtained interruptible capacity is about 12233 kilowatts.

[0127] Example 3

[0128] This embodiment also provides an industrial interruptible load identification and quantification system driven by power load data, including:

[0129] Data preprocessing module, used to obtain the user's power load data and perform preprocessing;

[0130] The initial clustering module is used to perform initial AP clustering based on the power load data of users on weekdays within a certain period of time, taking the user's daily power load data as an indicator;

[0131] The characteristic index extraction module is used to calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering;

[0132] The secondary clustering module is used to perform secondary AP clustering on the cluster centers obtained by the initial AP clustering according to the characteristic indicators of the user's power load data to obtain all the power consumption patterns of the user;

[0133] The interruptible load power consumption pattern identification module is used to identify the user's interruptible load power consumption pattern. Specifically, for each power consumption pattern of the user, identification is performed according to the following standards: the load change rate at each time node of the power consumption pattern is calculated, and the load change at the time node when the change rate is greater than the decrease amplitude threshold q is defined as 1, and the load change at the time node when the change rate is less than q is defined as 0; each time node is traversed, and the conversion point from 0 to 1 is defined as the pre-interruption inflection point; the load decrease at the subsequent moments of the pre-interruption inflection point is counted, and the point where the load stops decreasing is defined as the post-interruption inflection point, and the load fluctuation at adjacent moments must not exceed the preset fluctuation threshold v within a period of time after the post-interruption inflection point; when both the pre-interruption inflection point and the post-interruption inflection point appear in the user's power consumption pattern, the power consumption pattern is considered to be an interruptible load power consumption pattern;

[0134] The interruptible load resource quantification module is used to quantify the user's interruptible load resources according to the user's load under normal production conditions and the load during the interruption period.

[0135] It should be understood that the functional unit modules in various embodiments of the present invention can be concentrated in one processing unit, or each unit module can exist physically separately, or two or more unit modules can be integrated into one unit module, and can be implemented in the form of hardware or software.

[0136] Example 4

[0137] This embodiment provides an electronic device, including:

[0138] Memory on which computer programs or instructions are stored;

[0139] The processor is used to load and execute the computer program or instructions to implement the industrial interruptible load identification and quantification method driven by power load data as described above.

[0140] Example 5

[0141] In a fourth aspect of this embodiment, a computer-readable storage medium is provided, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the industrial interruptible load identification and quantification method driven by power load data as described above is implemented.

[0142] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0143] Compared with the existing technology, the present invention has the following beneficial effects: First, the present invention fully considers the various electricity consumption patterns of users, deeply mines the electricity consumption characteristics of users, and through two clustering, mines the most likely interruptible load resources, which is helpful for load management during peak hours; Second, the present invention is driven by power load data, and obtains results more quickly and objectively. On the one hand, this avoids the false reporting or omission of interruptible load resources by large industrial users in order to improve enterprise production, and on the other hand, it improves the identification speed of industrial interruptible load resources, reduces the investigation cost, and helps to better promote and apply; Third, the identification and quantification of industrial interruptible load resources of the present invention are more accurate and precise. In the process of identifying and quantifying industrial interruptible load resources, the present invention can accurately identify the inflection points before and after the interruption of a specific industrial interruptible load resource, and quantify the size of the interruptible load resource, which helps the power grid system to accurately dispatch, implement peak shaving and valley filling policies, and reduce power losses.

[0144] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt 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.) that contain computer-usable program code.

[0145] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. 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.

[0146] 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.

[0147] 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.

[0148] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A method for identifying and quantifying industrial interruptible loads based on power load data, characterized in that: The steps include: S1: Obtain the user's power load data and pre-process it; S2: Based on the power load data of users on weekdays within a certain period, the initial AP clustering is performed using the user's daily power load data as an indicator; S3: Calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering; S4: Perform secondary AP clustering on the cluster centers obtained from the initial AP clustering based on the characteristic indicators of the user's power load data to obtain all the user's power consumption patterns; S5: Identify the user's interruptible load power consumption pattern. Specifically, for each user's power consumption pattern, identify it according to the following criteria: calculate the load change rate at each time node of the power consumption pattern, define the load change at the time node where the change rate is greater than the decrease amplitude threshold q as 1, and define the load change at the time node where the change rate is less than q as 0; traverse each time node and define the transition point from 0 to 1 as the inflection point before interruption; The load drop after the pre-interruption inflection point is counted, and the point where the load stops decreasing is defined as the post-interruption inflection point. Furthermore, the load fluctuations at adjacent moments must not exceed a preset fluctuation threshold v for a period of time after the post-interruption inflection point. When both the pre-interruption inflection point and the post-interruption inflection point appear in a user's power usage pattern, the power usage pattern is considered to be an interruptible load power usage pattern. S6: Quantify the user's interruptible load resources based on the user's load under normal production conditions and the load during the interruption period.

2. The method for identifying and quantifying industrial interruptible loads based on power load data according to claim 1, characterized in that: The preprocessing process includes removing outliers and filling in missing values.

3. The method for identifying and quantifying industrial interruptible loads based on power load data according to claim 1, characterized in that: The characteristic indicators of the user power load data include daily load rate, daily peak-to-valley difference and daily maximum load utilization hours.

4. The method for identifying and quantifying industrial interruptible loads based on power load data according to claim 1, characterized in that: AP clustering includes the following steps: S21: Set the attenuation coefficient λ and the maximum number of iterations M; S22: Initialize the attraction matrix R and the attribution matrix A; S23: Update the attraction matrix R. The update formula is: Where r m+1 (i,k) represents the degree to which the load data of the kth day in the m+1th iteration is suitable as the cluster center of the load data of the ith day; S(i,k) represents the similarity between the load data of the ith day and the load data of the kth day; a m (i, j) represents the suitability of the load data of day i in the mth iteration to select the load data of day k as its cluster center; S24: Update the attribution matrix A. The update formula is: S25: Attenuate the attraction matrix R and the attribution matrix A according to the attenuation coefficient λ. The attenuation formula is: S26: Iterate steps S23-S25 until the iteration termination condition is met; S27: Output the largest element k in the R+A matrix as the clustering center of the data.

5. The method for identifying and quantifying industrial interruptible loads based on power load data drive according to claim 4 is characterized in that: In the initial AP clustering process, S(i,k) is obtained by taking the negative of the square of the Euclidean distance between the node load vectors of the load data on the i-th day and the load data on the k-th day; in the secondary AP clustering process, S(i,k) is obtained by taking the negative of the square of the Euclidean distance between the characteristic index vectors of the load data on the i-th day and the load data on the k-th day.

6. The method for identifying and quantifying industrial interruptible loads based on power load data according to claim 1, characterized in that: The identification of the user's interruptible load power consumption mode specifically includes: S51: Set the drop threshold q, q>0; S52: Calculate the multi-order moving average of each time node of each power consumption mode of the user as the average load of the time node; S53: Calculate the average load change rate p at each time node t ; S54: Determine load changes When the average load change rate p t When it is greater than the falling threshold q, Recorded as 1, otherwise recorded as 0; S55: Determine the turning point before the interruption At a certain time point When it changes from 0 to 1, the point is recorded as the turning point before the interruption; S56: Set the fluctuation threshold v, v>0; S57: Determine load fluctuations When the average load change rate p t When the absolute value of is greater than the fluctuation threshold v, Recorded as 1, otherwise recorded as 0; S58: Determine the turning point after the interruption At a certain time point When it changes from 1 to 0, and multiple time nodes specified after this time node and If the values ​​of are all 0, then the point is recorded as the turning point after the interruption; S59: Identify an interruptible power usage pattern, and when at least one pair of interruption inflection points appears in the user's power usage pattern, identify the power usage pattern as an interruptible load power usage pattern; S510: Output the inflection point pair with the largest difference 7. The method for identifying and quantifying industrial interruptible loads based on power load data drive according to claim 1, characterized in that: The user's interruptible load resource is calculated by the following formula: Where L is the user's interruptible load resource, is the average load of the user under normal production conditions, The average load of users during the outage.

8. An industrial interruptible load identification and quantification system based on power load data, characterized in that: include: Data preprocessing module, used to obtain the user's power load data and perform preprocessing; The initial clustering module is used to perform initial AP clustering based on the power load data of users on weekdays within a certain period of time, taking the user's daily power load data as an indicator; The characteristic index extraction module is used to calculate the characteristic index of the user power load data corresponding to each cluster center obtained by the initial AP clustering; The secondary clustering module is used to perform secondary AP clustering on the cluster centers obtained by the initial AP clustering according to the characteristic indicators of the user's power load data to obtain all the power consumption patterns of the user; The interruptible load power consumption pattern identification module is used to identify the user's interruptible load power consumption pattern. Specifically, for each user's power consumption pattern, identification is performed according to the following criteria: the load change rate at each time node of the power consumption pattern is calculated, and the load change at the time node where the change rate is greater than the decrease amplitude threshold q is defined as 1, and the load change at the time node where the change rate is less than q is defined as 0; each time node is traversed, and the transition point from 0 to 1 is defined as the inflection point before interruption; The load drop after the pre-interruption inflection point is counted, and the point where the load stops decreasing is defined as the post-interruption inflection point. Furthermore, the load fluctuations at adjacent moments must not exceed a preset fluctuation threshold v for a period of time after the post-interruption inflection point. When both the pre-interruption inflection point and the post-interruption inflection point appear in a user's power usage pattern, the power usage pattern is considered to be an interruptible load power usage pattern. The interruptible load resource quantification module is used to quantify the user's interruptible load resources according to the user's load under normal production conditions and the load during the interruption period.

9. An electronic device, characterized in that: include: Memory on which computer programs or instructions are stored; A processor is used to load and execute the computer program or instructions to implement the industrial interruptible load identification and quantification method based on power load data as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the method for identifying and quantifying industrial interruptible loads based on power load data as described in any one of claims 1 to 7 is implemented.

Images