A method, system and related equipment for evaluating power demand response potential

Abstract

The present invention discloses a method, system and related equipment for evaluating the potential of power demand response. The method includes: S1, constructing a load decomposition model of a special transformer user based on an STL algorithm, which is used to decompose the load of the special transformer user to obtain a load periodic component; S2, constructing a load curve platform power determination model based on an S-G filtering algorithm, determining the power that can represent the load curve platform, and each local minimum point in the load periodic component processed by the S-G filtering algorithm can represent the power of the load curve platform; S3, according to the specified start time of the demand response, using the special transformer user load decomposition model and the load curve platform power determination model to determine the real-time load power of the load periodic component at the response start time, and calculating the difference of all load curve platform powers less than the real-time load power, wherein the maximum value is the demand response potential power. The demand response potential of the special transformer user can be obtained more accurately, and the management efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power intelligent management, and in particular to an electric power demand response potential assessment method, system and related equipment. Background Art

[0002] As the load on the power grid continues to rise in recent years, the peak-to-valley difference in load continues to increase, and peak-to-frequency regulation has become a difficulty in building a highly reliable power grid. In the past, when facing short-term peak load demands, the approach was to increase the construction of generator sets and supporting transmission and distribution networks. These equipment have low utilization rates and low economic benefits.

[0003] Demand-side management is an important way of smart electricity consumption. By formulating effective and reasonable rules, it can guide user groups to use electricity in a direction that is beneficial to the operation of the power grid without affecting the basic electricity demand of users, thereby improving electricity efficiency and enhancing the reliability of the power grid.

[0004] Demand response is a market behavior. The power grid provides various price policies and incentive policies. When users choose to respond, they can gain benefits by changing their own electricity consumption methods. For the power grid, it can improve the load pressure of the power grid system. Demand response is implemented in the form of smart grid, helping users to actively participate in the grid regulation, improve the economic efficiency of electricity consumption, save resources and reduce energy consumption, accelerate the development of the power market mechanism, and improve the reliability and stability of the power grid.

[0005] Users can decide whether to participate in demand response, which determines that the effect of demand-side response depends on the user's electricity consumption behavior and response habits. Analysis of demand-side holographic data and in-depth mining of users' historical electricity consumption behavior, and selection of high-quality potential users are conducive to improving the implementation efficiency of demand response. Summary of the invention

[0006] The purpose of the present invention is to provide a method, system and related equipment for evaluating the potential of power demand response, which can more accurately obtain the demand response potential of special transformer users, provide scientific guidance for power companies to implement demand-side management, and be used to screen high-quality users who participate in demand response and to provide targeted incentives for users who have a negative attitude towards demand response.

[0007] In order to solve the above technical problems, an embodiment of the present invention provides a method for evaluating power demand response potential, comprising:

[0008] S1. Construct a dedicated transformer user load decomposition model based on the STL algorithm, which is used to decompose the load of the dedicated transformer user to obtain load cycle components. The input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date;

[0009] S2. Construct a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform.

[0010] S3, according to the specified start time of demand response, the real-time load power of the load periodic component at the response start time is determined by utilizing the dedicated transformer user load decomposition model and the load curve platform power determination model, and the difference is calculated for all load curve platform powers that are less than the real-time load power, and the maximum value among them is the demand response potential power.

[0011] Wherein, the S1 includes:

[0012] The load of the dedicated transformer user is used as an input load sequence to remove the load trend component, and the load subsequence is subjected to low-throughput filtering to obtain a load cycle component and a load residual component;

[0013] Among them, the load trend component represents the load that continues to run without being cut off during multiple scheduled sampling days in the production process of the dedicated transformer user, which is used to reflect the changes in the daytime production scale of the sampling day; the load cycle component represents the regular electricity load extracted from the sampling day, which is used to reflect the daily production or business plan and reflect the regularity of daily electricity load changes; the load residual component represents sudden load fluctuations outside the planned production.

[0014] Wherein, the S1 comprises:

[0015] Determine the internal circulation by load components;

[0016] Calculate the robustness weight term to control the data from generating abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop;

[0017] After the cycle is completed, the load cycle components are post-smoothed based on a local quadratic fit.

[0018] Wherein, determining the inner cycle by load component comprises:

[0019] S11, remove the trend of the last iteration for the load series sampled over multiple days

[0020] S11, LOESS regression processing is performed on each load subsequence, and one cycle is extended before and after, and the smoothing parameter is n (s) , the smoothed result is recorded as

[0021] S12, smoothing the result Make a length of n (p) 、n (p) , 3 sliding average, and then parameter n (l) LOESS regression, and get a sequence of length N

[0022] S13, obtain the periodic component of the multi-day load series,

[0023] S14, de-cycle,

[0024] S15, yes Use LOESS algorithm to smooth and obtain the load trend component judge Convergence, if converged, output the result, otherwise return to step S11,

[0025] in, Determine the load trend component and load cycle component at the end of the k-1th cycle in the inner cycle for the load component. The initial time n (i) is the number of inner loop layers, n (o) is the number of outer loop layers, n (p) is the number of period samples, n (s) 、n (l) 、n (t) are the LOESS smoothing parameters in S12, S13, and S14 respectively.

[0026] Wherein, the calculation robustness weight term includes:

[0027] The robustness weight term is calculated using the following formula:

[0028] δ v =6*f median (|R v |);

[0029]

[0030]

[0031] Where v is the position of the load point in the load sequence, δ v is the robustness weight.

[0032] Wherein, after S3, the following is also included:

[0033] Determining whether the demand response potential power is greater than the current maximum load supply capacity;

[0034] If so, the maximum load supply capacity is increased and an alarm message is output.

[0035] In addition, an embodiment of the present application further provides a power demand response potential assessment system, including:

[0036] A dedicated transformer user load decomposition model construction module is used to construct a dedicated transformer user load decomposition model based on the STL algorithm, decompose the dedicated transformer user's load to obtain load cycle components, and the input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date;

[0037] A platform power determination model building module is used to build a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform.

[0038] The demand response potential power calculation module is used to determine the real-time load power of the load periodic component at the response start time according to the specified demand response start time, using the dedicated transformer user load decomposition model and the load curve platform power determination model, and to find the difference between all load curve platform powers that are less than the real-time load power, and the maximum value among them is the demand response potential power.

[0039] The dedicated variable user load decomposition model construction module includes an inner loop determination unit, a robustness weight item calculation unit, and a smoothing processing unit; wherein,

[0040] The inner loop determination unit is used to determine the inner loop through the load component, remove the load trend component from the load of the special transformer user as the input load sequence, and obtain the load cycle component and the load residual component after low-throughput filtering of the load subsequence. The load trend component represents the load that is continuously operated and not cut off during multiple predetermined sampling days in the production process of the special transformer user, and is used to reflect the change of the daily production scale of the sampling day; the load cycle component represents the regular power load extracted from the sampling day, and is used to reflect the plan of production or business within the day, and reflects the law of the change of power load within the day; the load residual component represents the sudden load fluctuation outside the planned production;

[0041] The robustness weight item calculation module is used to calculate the robustness weight item to control the data to generate abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop;

[0042] The smoothing processing unit is used to perform post-smoothing on the load cycle component based on local quadratic fitting after the cycle ends.

[0043] In addition, an embodiment of the present application provides a power demand response potential assessment device, including:

[0044] A memory and a processor; wherein the memory is used to store a computer program, and the processor is used to implement the steps of the above-mentioned method for evaluating the potential for power demand response when executing the computer program.

[0045] In addition, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the power demand response potential assessment device method as described above are implemented.

[0046] The power demand response potential assessment method, system and related devices provided by the embodiments of the present invention have the following advantages compared with the prior art:

[0047] The power demand response potential assessment method, system and related equipment use the STL algorithm to decompose the load of the dedicated transformer user to obtain the load cycle component, and then process the load cycle component through the SG filtering algorithm, each of which can represent the power of the load curve platform. Finally, according to the specified start time of the demand response, the real-time load power of the load cycle component at the response start time is determined, and the difference of all load curve platform powers less than the real-time load power is calculated, and the maximum value is the demand response potential power. This method can more accurately obtain the demand response potential of the dedicated transformer user, provide scientific guidance for power companies to implement demand-side management, and be used to screen high-quality users participating in demand response and to provide targeted incentives for users with negative attitudes towards demand response. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A schematic diagram of a specific implementation of a method for evaluating the potential for power demand response provided by the present invention;

[0050] Figure 2 This is a schematic flow chart of the steps of inner loop determination in one embodiment of the power demand response potential assessment method provided by the present invention,

[0051] Figure 3A schematic diagram of the structure of a power demand response potential assessment system provided by an embodiment of the invention of this application;

[0052] Figure 4 A load decomposition result based on the STL algorithm and a schematic diagram of each component in one embodiment of the power demand response potential assessment system provided by the present invention;

[0053] Figure 5 A comparison diagram of load periodicity curves before and after filtering based on the SG algorithm in one embodiment of the power demand response potential assessment system provided by the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in 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 creative work are within the scope of protection of the present invention.

[0055] Please refer to Figures 1 to 5 , Figure 1 A schematic diagram of a specific implementation of a method for evaluating the potential for power demand response provided by the present invention; Figure 2 This is a schematic flow chart of the steps of inner loop determination in one embodiment of the power demand response potential assessment method provided by the present invention, Figure 3 A schematic diagram of the structure of a power demand response potential assessment system provided by an embodiment of the invention of this application; Figure 4 A load decomposition result based on the STL algorithm and a schematic diagram of each component in one embodiment of the power demand response potential assessment system provided by the present invention; Figure 5 A comparison diagram of load periodicity curves before and after filtering based on the SG algorithm in one embodiment of the power demand response potential assessment system provided by the present invention.

[0056] In a specific implementation, the power demand response potential assessment method includes:

[0057] S1. Construct a dedicated transformer user load decomposition model based on the STL algorithm, which is used to decompose the load of the dedicated transformer user to obtain load cycle components. The input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date;

[0058] S2. Construct a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform.

[0059] S3, according to the specified start time of demand response, the real-time load power of the load periodic component at the response start time is determined by utilizing the dedicated transformer user load decomposition model and the load curve platform power determination model, and the difference is calculated for all load curve platform powers that are less than the real-time load power, and the maximum value among them is the demand response potential power.

[0060] The load of the dedicated transformer user is decomposed by using the STL algorithm to obtain the load cycle components, and then the load cycle components are processed by the SG filtering algorithm. Each local minimum point can represent the power of the load curve platform. Finally, according to the specified start time of the demand response, the real-time load power of the load cycle component at the response start time is determined, and the difference of all load curve platform powers less than the real-time load power is calculated. The maximum value is the demand response potential power. This method can more accurately obtain the demand response potential of dedicated transformer users, provide scientific guidance for power companies to implement demand-side management, and be used to screen high-quality users participating in demand response and to provide targeted incentives for users with negative attitudes towards demand response.

[0061] The present application does not limit the decomposition and acquisition method of the load periodic components. In one embodiment, S1 includes:

[0062] The load of the dedicated transformer user is used as an input load sequence to remove the load trend component, and the load subsequence is subjected to low-throughput filtering to obtain a load cycle component and a load residual component;

[0063] Among them, the load trend component represents the load that continues to run without being cut off during multiple scheduled sampling days in the production process of the dedicated transformer user, which is used to reflect the changes in the daytime production scale of the sampling day; the load cycle component represents the regular electricity load extracted from the sampling day, which is used to reflect the daily production or business plan and reflect the regularity of daily electricity load changes; the load residual component represents sudden load fluctuations outside the planned production.

[0064] The present application does not limit the filter used for filtering, and the staff can select a suitable filter as needed.

[0065] Since the STL algorithm is mainly used in this application to decompose the load of the dedicated transformer user to obtain the load cycle components, the specific calculation process in this process is not limited. In one embodiment, S1 includes:

[0066] Determine the internal circulation by load components;

[0067] Calculate the robustness weight term to control the data from generating abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop;

[0068] After the cycle is completed, the load cycle components are post-smoothed based on a local quadratic fit.

[0069] Since the periodic component is mainly obtained through the inner loop, the specific calculation process is not limited. In a specific embodiment, determining the inner loop through the load component includes:

[0070] S11, remove the trend of the last iteration for the load series sampled over multiple days

[0071] S12, LOESS regression processing is performed on each load subsequence, and one cycle is extended before and after, and the smoothing parameter is n (s) , the smoothed result is recorded as

[0072] S13, smoothing the result Make a length of n (p) 、n (p) , 3 sliding average, and then parameter n (l) LOESS regression, and get a sequence of length N

[0073] S14, obtain the periodic component of the multi-day load series,

[0074] S15, de-cycle,

[0075] S16, yes Use LOESS algorithm to smooth and obtain the load trend component judge Convergence, if converged, output the result, otherwise return to step S11,

[0076] in, Determine the load trend component and load cycle component at the end of the k-1th cycle in the inner cycle for the load component. The initial time n (i) is the number of inner loop layers, n (o) is the number of outer loop layers, n (p) is the number of period samples, n (s) 、n (l) 、n (t) are the LOESS smoothing parameters in S12, S13, and S14 respectively.

[0077] The present application includes but is not limited to the above-mentioned calculation method, and the staff may also choose to use other types of calculation methods.

[0078] The calculation and operation process of the robustness weight item in the present application are not limited. In one embodiment, the calculation of the robustness weight item includes:

[0079] The robustness weight term is calculated using the following formula:

[0080] δ v =6*f median (|R v| );

[0081]

[0082]

[0083] Where v is the position of the load point in the load sequence, δ v is the robustness weight.

[0084] The main purpose of this application is to calculate the user's demand response potential, realize demand-side management, and improve management efficiency. Inevitably, there will be a situation where the power supply is insufficient. Therefore, it is necessary to increase the power supply responsibility appropriately, thereby improving the power reliability of the power user and meeting the power demand, thereby ensuring their normal production and life. Therefore, in one embodiment, after S3, it also includes:

[0085] Determining whether the demand response potential power is greater than the current maximum load supply capacity;

[0086] If so, the maximum load supply capacity is increased and an alarm message is output.

[0087] Of course, it should be pointed out that due to changes in management, production efficiency and production volume, different enterprises or units will not only experience an increase in load, but also a decrease in electricity load. In order to achieve efficient electricity management, dynamic load management of different users can be implemented to achieve more efficient electricity management.

[0088] The load curve platform is defined in this application as the power range of the equipment combination in use in the production of the dedicated transformer user. When new equipment is put into use or the equipment in use is removed, the load curve will rise or fall sharply and reach a new load curve platform. In the load periodic component, there is still a certain load fluctuation on a load curve platform, and the load platform with a duration that is too short is less than the time required for demand response, which is not suitable as a basis for evaluating the demand response potential of dedicated transformer users.

[0089] Therefore, the SG filtering algorithm is used to determine the power that can represent the load curve platform. The principle of the SG filtering algorithm, that is, the determination of the power that represents the load curve platform, is as follows.

[0090] Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power P of the load curve platform where it is located. Loc_min = {P 1,L_min ,P 2,L_min ,……,P i,L_min ,……,P k,L_min}.

[0091] According to the start time of demand response required by the power company, determine the real-time load power of the periodic component of the dedicated transformer user's load at the response start time, and calculate the difference for all load curve platform powers of the user that are less than its own real-time load power. The maximum value is the demand response potential power of the user.

[0092]

[0093] In addition, an embodiment of the present application further provides a power demand response potential assessment system, including:

[0094] The dedicated transformer user load decomposition model construction module 10 is used to construct a dedicated transformer user load decomposition model based on the STL algorithm, decompose the dedicated transformer user's load to obtain load cycle components, and the input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date;

[0095] The platform power determination model building module 20 is used to build a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform.

[0096] The demand response potential power calculation module 30 is used to determine the real-time load power of the load periodic component at the response start time according to the specified demand response start time, using the dedicated transformer user load decomposition model and the load curve platform power determination model, and to find the difference between all load curve platform powers that are less than the real-time load power, and the maximum value among them is the demand response potential power.

[0097] Since the power demand response potential assessment system is a system corresponding to the above-mentioned power demand response potential assessment method and has the same beneficial effects, this application will not elaborate on it.

[0098] The present application does not limit the process of using the STL algorithm to perform load decomposition. In one embodiment, the dedicated variable user load decomposition model construction module includes an inner loop determination unit, a robustness weight item calculation unit, and a smoothing processing unit; wherein,

[0099] The inner loop determination unit is used to determine the inner loop through the load component, remove the load trend component from the load of the special transformer user as the input load sequence, and obtain the load cycle component and the load residual component after low-throughput filtering of the load subsequence. The load trend component represents the load that is continuously operated and not cut off during multiple predetermined sampling days in the production process of the special transformer user, and is used to reflect the change of the daily production scale of the sampling day; the load cycle component represents the regular power load extracted from the sampling day, and is used to reflect the plan of production or business within the day, and reflects the law of the change of power load within the day; the load residual component represents the sudden load fluctuation outside the planned production;

[0100] The robustness weight item calculation module is used to calculate the robustness weight item to control the data to generate abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop;

[0101] The smoothing processing unit is used to perform post-smoothing on the load cycle component based on local quadratic fitting after the cycle ends.

[0102] In this application, the main steps of load decomposition using the STL algorithm include three parts: load component determination inner loop, robustness weight balancing outer loop and load period component post-smoothing.

[0103] (1) Load component determines the inner loop

[0104] In the load component determination inner loop, the load trend component T is removed from the input load sequence. v , low-throughput filtering is performed on the load subsequence to obtain the load cycle component S v and the load residual component R v .set up Determine the load trend component and load cycle component at the end of the k-1th cycle in the inner cycle for the load component. The initial time n (i) is the number of inner loop layers, n (o) is the number of outer loop layers, n (p) is the number of period samples, n (s) 、n (l) 、n (t) They are the LOESS smoothing parameters in ②, ③, and ④ respectively. The specific process is shown in the attached Figure 1 shown.

[0105] ① Remove the trend of the last iteration for the load series sampled over multiple days

[0106]

[0107] ② Perform LOESS regression processing on each load subsequence, extend one cycle before and after, and the smoothing parameter is n (s) , the smoothed result is recorded as

[0108] ③ Perform LOESS regression on the subsequence, that is, the smoothing result in ② Make a length of n (p) 、n (p) , 3 sliding average, and then parameter n (l) LOESS regression, and get a sequence of length N

[0109] ④ Obtain the periodic component of multi-day load series:

[0110]

[0111] ⑤De-cycle:

[0112]

[0113] ⑥Pair Use LOESS algorithm to smooth and get the load trend component judge Convergence, if converged, output the result, otherwise return to step ①:

[0114]

[0115] (2) Robust weight balancing outer loop

[0116] The function of the outer loop is to calculate the robustness weight term to control the abnormal value of the data in the process of load decomposition, and substitute the weight value into the inner loop for calculation. Assume:

[0117] δ v =6*f median (|R v |)

[0118]

[0119]

[0120] Where v is the position of the load point in the load sequence, δ v is the robustness weight.

[0121] (3) Load cycle component post-smoothing

[0122] After the loop is finished, there will be burrs in the load in the periodic component because the smoothing in the inner loop is only performed in each window. After the load series are integrated according to the load sampling time, the smoothness of the entire load sampling series cannot be guaranteed. The post-smoothing of the load periodic component is based on local quadratic fitting and does not require robust iteration in loess.

[0123] After STL decomposition, the three components are obtained, among which the load trend component T v The load that is continuously operated and not cut off during several sampling days in the production process of the dedicated transformer user can reflect the change of the daily production scale on the sampling day. The load cycle component S v Represents the regular electricity load extracted from several sampling days, which can reflect the production or business plan within the day and the law of changes in the electricity load within the day. v Represents sudden load fluctuations outside of planned production.

[0124] The load curve platform is defined as the power range of the state where the equipment combination used in the production of the dedicated transformer user remains unchanged. When new equipment is put into use or the equipment in use is removed, the load curve will rise or fall sharply and reach a new load curve platform. In the load periodic component, there is still a certain load fluctuation on a load curve platform, and the load platform with a duration that is too short is less than the time required for demand response, which is not suitable as a basis for evaluating the demand response potential of dedicated transformer users. Therefore, the SG filtering algorithm is used to determine the power that can represent the load curve platform.

[0125] Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power P of the load curve platform where it is located. Loc_min = {P 1,L_min ,P 2,L_min ,……,P i,L_min ,……,P k,L_min}.

[0126] According to the start time of demand response required by the power company, determine the real-time load power of the periodic component of the dedicated transformer user's load at the response start time, and calculate the difference for all load curve platform powers of the user that are less than its own real-time load power. The maximum value is the demand response potential power of the user.

[0127]

[0128] In addition, an embodiment of the present application provides a power demand response potential assessment device, including:

[0129] A memory and a processor; wherein the memory is used to store a computer program, and the processor is used to implement the steps of the above-mentioned method for evaluating the potential for power demand response when executing the computer program.

[0130] Since the processor of the power demand response potential assessment device is used to implement the steps of the power demand response potential assessment method as described above when executing the computer program, it has the same beneficial effects, and this application does not limit this.

[0131] This application does not elaborate on the types of power demand response potential assessment equipment.

[0132] In addition, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the power demand response potential assessment device method as described above are implemented.

[0133] Similarly, since the computer-readable storage medium stores a computer program, the computer program, when executed by the processor, implements the steps of the power demand response potential assessment device method as described above, and has the same beneficial effects, which will not be elaborated in this application.

[0134] The present application does not limit the type of computer-readable storage medium, which may be a CDROM, an EEPROM, a USB flash drive, a mobile hard disk or other storage media, and the present application does not limit this.

[0135] In one embodiment, the user's industry is foundry industry, and the load decomposition results 15 days before the response date are as follows: Figure 4 As shown. Analyzing the load decomposition results, the load periodic component S v It can be seen that the production mode of this user is production during the day, rest at night, and a slight peak avoidance phenomenon occurs at noon; according to the load trend component T v It can be seen that the user has arranged the production scale in a variable manner during these 15 days, and the overall trend is downward on the basis of certain fluctuations; there is a load residual component R v It can be seen that outside the production plan, there are certain sudden load fluctuations.

[0136] Extract the load periodicity component within a day and perform SG filtering. The load periodicity component of the user and the result after SG filtering are as follows: Figure 5 shown.

[0137] Comparing before and after filtering, it can be found that small load fluctuations have been smoothed, so as not to affect the load platform determination for demand response capabilities. The larger load fluctuations that occurred around 8:00 am were also smoothed, which is reasonable. Although the load difference caused by this fluctuation is large, it lasts for a short time and usually cannot meet the demand response requirement of load reduction for half an hour or more.

[0138] Finally, there are three load platforms obtained by using the objective demand response capability determination model based on the SG filtering algorithm, and the corresponding power values ​​are P 1,L_min =559.93kW, P 2,L_min =1816.73kW, P 3,L_min = 0, the response preload P corresponding to the response start time DR_p =2187.09kWThe objective response capacity is 2187.09kW, the response power participating in the demand response is 1772.4kW, and the response level is 81.04%. This can illustrate the rationality of the demand response objective capacity determination model and the accuracy of the determination results.

[0139] In summary, the power demand response potential assessment method, system and related equipment provided in the embodiments of the present invention use the STL algorithm to decompose the load of the dedicated transformer user to obtain the load cycle component, and then process the load cycle component through the SG filtering algorithm, where each local minimum point can represent the power of the load curve platform. Finally, according to the specified start time of the demand response, the real-time load power of the load cycle component at the response start time is determined, and the difference of all load curve platform powers less than the real-time load power is calculated, and the maximum value is the demand response potential power. This method can more accurately obtain the demand response potential of dedicated transformer users, provide scientific guidance for power companies to implement demand-side management, and be used to screen high-quality users participating in demand response and to provide targeted incentives for users with negative attitudes towards demand response.

[0140] The above is a detailed introduction to the power demand response potential assessment method, system and related equipment provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A method for evaluating power demand response potential, characterized in that: include: S1. Construct a dedicated transformer user load decomposition model based on the STL algorithm, which is used to decompose the load of the dedicated transformer user to obtain load cycle components. The input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date; S2. Construct a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform. S3, according to the specified start time of the demand response, the real-time load power of the load periodic component at the response start time is determined by using the dedicated transformer user load decomposition model and the load curve platform power determination model, and the difference of all load curve platform powers less than the real-time load power is calculated, and the maximum value among them is the demand response potential power; The S1 includes: The load of the dedicated transformer user is used as an input load sequence to remove the load trend component, and the load subsequence is subjected to low-throughput filtering to obtain a load cycle component and a load residual component; Among them, the load trend component represents the load that is continuously operated and not cut off during multiple scheduled sampling days in the production process of the dedicated transformer user, which is used to reflect the change of the daily production scale of the sampling day; the load cycle component represents the regular power load extracted from the sampling day, which is used to reflect the production or business plan within the day and reflect the law of the change of power load within the day; the load residual component represents the sudden load fluctuation outside the planned production; S1 includes: Determine the internal circulation by load components; Calculate the robustness weight term to control the data from generating abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop; After the cycle ends, post-smoothing the load cycle component based on local quadratic fitting; Determining the inner cycle by load component includes: S11, remove the trend of the last iteration for the load series sampled over multiple days S12, LOESS regression processing is performed on each load subsequence, and one cycle is extended before and after, and the smoothing parameter is n (s) , the smoothed result is recorded as S13, smoothing the result Sequentially make the length n (p) 、n (p) , 3 sliding average, and then parameter n (l) LOESS regression, and get a sequence of length N S14, obtain the periodic component of the multi-day load series, S15, de-cycle, S16, yes Use LOESS algorithm to smooth and obtain the load trend component judge Convergence, if converged, output the result, otherwise return to step S11, in, Determine the load trend component and load cycle component at the end of the k-1th cycle in the inner cycle for the load component. The initial time n (i) is the number of inner loop layers, n (o) is the number of outer loop layers, n (p) is the number of period samples, n (s) 、n (l) 、n (t) are the LOESS smoothing parameters in S12, S13, and S14 respectively; The calculation of the robustness weight term includes: The robustness weight term is calculated using the following formula: d v =6*f median (R v ); Where v is the position of the load point in the load sequence, δ v is the robustness weight; After S3, the following is also included: Determining whether the demand response potential power is greater than the current maximum load supply capacity; If yes, increase the maximum load supply capacity and output an alarm message Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform where it is located. P Loc_min ={P 1,L_min ,P 2,L_min ,……,P i,L_min ,……,P k,L_min }; According to the start time of demand response required by the power company, determine the real-time load power of the load periodic component of the dedicated transformer user at the start time of response. For all load curve platform powers of the user that are less than its own real-time load power, calculate the difference. The maximum value is the demand response potential power of the user.

2. The method for evaluating the potential of power demand response according to claim 1, characterized in that: After S3, the following is also included: Determining whether the demand response potential power is greater than the current maximum load supply capacity; If so, the maximum load supply capacity is increased and an alarm message is output.

3. A power demand response potential assessment system, characterized in that: The method for evaluating the potential of power demand response as claimed in claim 1 comprises: A dedicated transformer user load decomposition model construction module is used to construct a dedicated transformer user load decomposition model based on the STL algorithm, decompose the dedicated transformer user's load to obtain load cycle components, and the input quantity in the dedicated transformer user load decomposition model is the load observation quantity of the dedicated transformer user for a specified time period before the response date; A platform power determination model building module is used to build a load curve platform power determination model based on the SG filtering algorithm to determine the power that can represent the load curve platform. Each local minimum point in the load periodic component processed by the SG filtering algorithm can represent the power of the load curve platform where it is located; The demand response potential power calculation module is used to determine the real-time load power of the load periodic component at the response start time according to the specified demand response start time by using the dedicated transformer user load decomposition model and the load curve platform power determination model, and to calculate the difference of all load curve platform powers less than the real-time load power, and the maximum value among them is the demand response potential power; The dedicated variable user load decomposition model construction module includes an inner loop determination unit, a robustness weight item calculation unit, and a smoothing processing unit; wherein, The inner loop determination unit is used to determine the inner loop through the load component, remove the load trend component from the load of the special transformer user as the input load sequence, and obtain the load cycle component and the load residual component after low-throughput filtering of the load subsequence. The load trend component represents the load that is continuously operated and not cut off during multiple predetermined sampling days in the production process of the special transformer user, and is used to reflect the change of the daily production scale of the sampling day; the load cycle component represents the regular power load extracted from the sampling day, and is used to reflect the plan of production or business within the day, and reflects the law of the change of power load within the day; the load residual component represents the sudden load fluctuation outside the planned production; The robustness weight item calculation module is used to calculate the robustness weight item to control the data to generate abnormal values ​​during the load decomposition process, and substitute the weight value into the inner loop for calculation to realize the robustness weight balance outer loop; The smoothing processing unit is used to perform post-smoothing on the load cycle component based on local quadratic fitting after the cycle ends.

4. A power demand response potential assessment device, characterized in that: include: A memory and a processor; wherein the memory is used to store a computer program, and the processor is used to implement the steps of the method for evaluating the potential for power demand response as described in any one of claims 1 to 2 when executing the computer program.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for evaluating the potential of electric power demand response equipment are implemented as described in any one of claims 1 to 2.

Images