A power load identification method and system based on maximum likelihood method

Abstract

The application provides a non-invasive power load identification method and system based on a maximum likelihood method, total active power collected at a power consumption entrance is subjected to an event detection algorithm, power load switch events are quickly searched, and a load power difference value is calculated; the power difference value is subjected to an iteration method and a power range value to obtain a possible existence of a power load; the possible existence of the power load is subjected to a maximum likelihood method and a probability density function feature library of each power load operation power, operation power change, operation time and shutdown time to calculate a possible probability value of each power load; the probability value is compared and sorted to obtain a maximum value, which is a power load identification result. The event detection algorithm can quickly search for start and stop of the power load, the power range and the iteration method can effectively reduce the calculation dimension, and the maximum likelihood method is directly used to effectively identify each power load type probability value, and the power load identification has strong reliability.

Description

Technical Field

[0001] This invention relates to the field of electrical load identification technology, particularly to identification technology for everyday electrical equipment, and more specifically, to an electrical load identification method and system based on the maximum likelihood method. Background Technology

[0002] With the development of the times, today's society is facing urgent economic and environmental problems, such as the global energy crisis and the greenhouse effect. Improving energy efficiency and conserving energy are important ways to break through and solve these problems. Currently, industrial, commercial, and residential electricity users share the following common problems: weak awareness of energy conservation; insufficient understanding of the energy consumption of electrical equipment; failure to participate in grid peak-shaving and valley-filling measures according to demand-side response; and lack of reasonable energy-saving plans. Furthermore, the current electricity management model still follows traditional methods of energy conservation, promoting energy conservation awareness through slogans, without taking concrete measures to achieve controllable, manageable, and under-controllable electricity consumption.

[0003] The maximum likelihood-based load identification method identifies the types and operating states of each user's load by collecting total power values ​​at the main power inlet. Compared to intrusive load monitoring technology, this method has advantages such as low cost, simple communication network, and easy maintenance. It also obtains the operating status and energy consumption of each electrical device, allowing for effective energy-saving measures based on comprehensive information such as energy consumption, time-of-use pricing, and electricity metering. Furthermore, with increasingly strained energy relations, the maximum likelihood-based load identification method and system can provide efficient demand-side response and energy allocation, participating in grid regulation and significantly improving the public's satisfaction with electricity usage. Summary of the Invention

[0004] This invention addresses the problems of non-intrusive electrical load identification by proposing a method and system for electrical load identification based on the maximum likelihood method. The method monitors the total power of the electrical load at the user's power inlet; it uses an event detection algorithm to quickly search for the start / stop status of each load based on the collected active power value; then, iterative methods are used to initially identify the probability of load presence, reducing the dimensionality of load identification calculations; finally, the probability of each electrical load's existence is calculated using the maximum likelihood classification method, and the maximum probability value is the electrical load identification result. Experimental verification shows that this method can accurately identify loads and their operating status.

[0005] One objective of this invention is to propose a method for identifying electrical loads based on the maximum likelihood method, comprising: monitoring the total active power of the electrical load at the user's power inlet; using an event detection algorithm to determine electrical load switching events and calculate the load power difference; using an iterative method to obtain the probability of the electrical load's existence; calculating the probability of each electrical load's existence using the maximum likelihood method and based on a feature library of probability density functions for each electrical load's operating power, changes in operating power, operating time, and shutdown time; and comparing and ranking the probability values ​​to obtain the maximum value as the electrical load identification result.

[0006] One objective of this invention is to provide an electrical load identification system based on the maximum likelihood method, comprising: a monitoring device for monitoring the total active power at the user's power inlet; an event detection device for analyzing the collected total active power and the switching status of electrical loads, and calculating the power difference; a load combination determination device for determining the possible loads at a given moment by solving the switching status of the electrical loads using an iterative method; a maximum likelihood calculation device for calculating the possible probability value of each electrical load based on the operating power, changes in operating power, operating time, and shutdown time probability density function feature library; and an electrical load determination device for sorting the possible probability values ​​and determining the maximum value as the electrical load identification result.

[0007] The technical solution adopted in this invention is:

[0008] A method for identifying electrical loads based on the maximum likelihood method, the method comprising:

[0009] The total active power collected at the power inlet will be used to quickly retrieve power load switch events and calculate the load power difference through an event detection algorithm.

[0010] The power difference is used to determine the possibility of the existence of electrical load by using an iterative method and based on the power range value;

[0011] The probability of the existence of the electrical load is calculated using the maximum likelihood method and based on the feature library of probability density functions for the operating power, changes in operating power, operating time, and shutdown time of each electrical load.

[0012] The probability values ​​are compared and sorted to obtain the maximum value, which is the result of electricity load identification.

[0013] A power load identification system based on the maximum likelihood method, the system comprising:

[0014] The event detection device analyzes the total active power collected at the power inlet to obtain the power load switching status and calculate the power difference.

[0015] A load combination determination device is used to solve the electrical load switching situation using an iterative method to determine the possible loads at that moment.

[0016] The maximum likelihood calculation device calculates the probability value of the existence of each electrical load based on the feature library of probability density functions of the operating power, the change of operating power, operating time and shutdown time of each electrical load.

[0017] The power load determination device sorts and compares the possible probability values ​​and obtains the maximum value as the power load identification result.

[0018] The beneficial result of this invention lies in its proposed method that organically combines iterative dimensionality reduction and maximum likelihood classification to address the problems existing in the prior art. First, event detection is used to determine the on / off events of loads, enabling a rapid search for the start / stop status of each load. Then, an iterative method is used to initially identify the probability of a load's existence. Finally, based on the maximum likelihood method and combined with characteristics such as the operating power characteristics, power changes, and operating time of each load, the probability of each load's existence is calculated, and the maximum probability value is the load identification result. This method offers strong reliability for electrical load identification.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating an implementation method for an electrical load identification method based on the maximum likelihood method, provided in this embodiment of the invention;

[0022] Figure 2 The total active power variation diagram from 00:00 to 24:00 in a specific embodiment provided by the present invention;

[0023] Figure 3 Event detection algorithm flowchart;

[0024] Figure 4 Event detection algorithm power difference map;

[0025] Figure 5 Probability diagram of positive and negative power sets for a refrigerator;

[0026] Figure 6 Probability value graph of refrigerator opening and closing time sets;

[0027] Figure 7 Load identification result diagram;

[0028] Figure 8 This is a structural block diagram of an embodiment of an electrical load identification system based on the maximum likelihood method provided by the present invention.

[0029] Figure 9 This is a structural block diagram of a second embodiment of an electrical load identification system based on the maximum likelihood method provided in this invention. Detailed Implementation

[0030] Figure 1 This is a flowchart illustrating an implementation of a maximum likelihood-based method for identifying electrical loads, as provided in an embodiment of the present invention. Figure 1 As can be seen, in the implementation method, the method specifically includes:

[0031] S101: The total active power collected at the power inlet will be quickly retrieved using an event detection algorithm to determine the power load switch events and calculate the load power difference, such as... Figure 2 The diagram shows the total active power between 00:00 and 24:00. The event detection algorithm flow is as follows: Figure 3 As shown. First, calculate the mean and standard deviation of the four times preceding each time t, as shown in equations (1) and (2):

[0032]

[0033]

[0034] Then calculate the power difference ΔP(t), as shown in equation (3). However, the electrical load may be on or off, meaning ΔP(t) can be positive or negative. And determine σ. P (t) and |ΔP(t)| with threshold σ g The value σ is used to determine load fluctuations. If the load fluctuation exceeds a threshold, a switching event is considered to have occurred; otherwise, no switching event is detected. g The fluctuations in the activation of each load were obtained through experiments.

[0035] ΔP(t)=P(t)-P(t-1) (3)

[0036] Where: P(t) is the power value at time t, and P(t-1) is the power value at time t-1.

[0037] For example in Figure 2The elliptical area enclosed by the dashed line represents the main concentrated electricity consumption time range for residents on that day, namely 7:00-8:00 AM and 6:00-10:00 PM. This paper's algorithm mainly focuses on the 6:00-10:00 PM period. The algorithm uses an event detection method to determine load switching events and calculates the power load difference event. The results are as follows... Figure 4 As shown.

[0038] S102: Solve the active power difference using an iterative method, and determine the possible load at that moment based on the power range, specifically including:

[0039] During the use of each load, the power consumption changes differently due to the different types of loads started by the user; the power difference ΔP(t) is obtained through event detection, and the load start-stop vector that may exist when the appliance is used is X; according to the power balance, equation (4) can be obtained, and the entire component function that meets the conditions is calculated by computer programming iteratively according to equation (4).

[0040] ||ΔP(t)|-P i T ·X|≤Δmax (4)

[0041] Where X(i)∈{0,1}, the load start / stop vector X takes only two values, 0 and 1, where 0 indicates the appliance is off and 1 indicates the appliance is on; P i min ≤P i ≤P i max P i P represents the active power status of the i-th electrical load during operation. i min and P i max The minimum and maximum active power during load operation are shown in Table 1, which lists different types of electrical loads and their power ranges; Δmax represents the maximum allowable error in load power.

[0042] The calculation for this example is 15W, mainly considering factors such as standby power of the load and line losses. Using the above formula (4) and the iterative method, the possibility of the electrical load existing during the event detection can be determined.

[0043] Table 1

[0044]

[0045] Solving the problem using an iterative method, such as... Figure 4 The possible loads obtained from the power difference, for example, if the power difference is 53W at 18:37:10, the possible loads can be determined by the iterative method as five possibilities: energy-saving lamp 1, energy-saving lamp 2, laptop, desktop computer, and refrigerator.

[0046] S103: The probability of the existence of the electrical loads is calculated using the maximum likelihood classification method, based on the characteristic library of probability density functions for the operating power, changes in operating power, operating time, and shutdown time of each electrical load. The maximum likelihood method is used to identify load types. This method is applied to decisions involving two or more classes. It uses statistical methods to establish a nonlinear discriminant function set based on the maximum likelihood ratio Bayesian decision criterion. Assuming the distribution functions of each class are known, and selecting training areas, the probability of each sample area to be classified is calculated for classification.

[0047] The possibility of load existing when a switching event occurs is obtained through an iterative method, denoted as […]. S represents the set of possible loads, and i represents the number of loads. The probability of each load's existence can then be calculated using the maximum likelihood classification method. The maximum probability is the event detection load identification result, expressed by formula (5):

[0048]

[0049] Where X is the load sample

[0050] Maximum likelihood load identification calculates the probability of each load's existence based on the probability density function in the load feature database, and the maximum probability is the load identification result at that moment.

[0051] The load characteristic library mainly considers the positive power set of the load operation. Running negative power set Load operation variation power set S ΔP Load operating time set and load shutdown time set Where S ΔP It is only applicable to loads with multiple changing states.

[0052] According to various load types and S ΔP Images from each dataset, for and S ΔP The probability density function is approximated using a normal distribution, as shown in equation (6).

[0053]

[0054] Where P is the load power value at different times, and μ and σ are the mean and standard deviation of the normal distribution, respectively.

[0055] Similarly, for and The Weibull distribution is adopted, as shown in equation (7).

[0056]

[0057] For example, a refrigerator is a multi-state load and does not have a load operation variation power set S. ΔP Its positive power set and negative power set image like Figure 5 runtime set and closing time set like Figure 6 .

[0058] The probability of a load's existence is determined using an iterative load analysis method. Then, the load type needs to be identified based on its characteristics. For example, if the power difference is 53W at 18:37:10, the probabilities of each load's existence are W(energy-saving lamp 1) = 0, W(energy-saving lamp 2) = 0, W(laptop) = 0.2, W(computer) = 0.3, and W(refrigerator) = 0.5. Therefore, the final load identification result is the refrigerator. Similarly, the maximum likelihood method is used to identify the load for the power difference between 18:00 and 22:00. The identification results are as follows... Figure 7 .

[0059] This invention proposes a method for identifying electrical loads based on the maximum likelihood method, comprising: obtaining electrical load switching events by passing the collected total active power value through an event detection algorithm; determining the probability of the existence of electrical loads by passing the electrical load switching events through an iterative method and based on the load range value; calculating the probability value of the existence of each electrical load by passing the maximum likelihood classification method and based on the feature library of probability density functions of the operating power, changes in operating power, operating time, and shutdown time of each electrical load; comparing the probability values, and the maximum value is the electrical load identification result, which has strong reliability for electrical load identification.

[0060] Figure 8 This is a structural block diagram of a first embodiment of a household load identification system based on the maximum likelihood method provided by the present invention. Figure 8 It can be seen that, in implementation method one, the system specifically includes:

[0061] Event detection device 201 is used to analyze the total active power collected to obtain the power load switching status, calculate the fluctuation value at each moment using formula (1) and formula (2) and compare it with the threshold, quickly retrieve the power load switching event, and solve the power difference using formula (3).

[0062] The load combination determination device 202 is used to determine the possible loads at a given time based on the power difference using formula (4). For example, if the power difference is 53W at 18:37:10 and the sampling frequency of this monitoring device is 1s, then the event detection can be considered as only one load being either on or off. Through 0-1 programming, five possible loads can be obtained: energy-saving lamp 1, energy-saving lamp 2, laptop computer, desktop computer, and refrigerator.

[0063] The maximum likelihood calculation device 203 calculates the possible probability value of each electrical load based on the characteristic library of probability density functions of the operating power, the change of operating power, the operating time and the shutdown time of each electrical load.

[0064] The load determination device 204 sorts and compares various electrical loads, and determines the load with the highest probability as the identification result. For example, at 18:37:10, the power difference is 53W, and the possible probabilities of each load are W(energy-saving lamp 1) = 0, W(energy-saving lamp 2) = 0, W(laptop) = 0.2, W(computer) = 0.3, and W(refrigerator) = 0.5. Therefore, the final load identification result at this time is the refrigerator.

[0065] Figure 9 This is a structural block diagram of a second embodiment of a household electricity load identification system based on maximum power consumption provided by the present invention. Figure 9 It can be seen that, in implementation method two, the system further includes:

[0066] Pre-acquisition device 205 is used to pre-acquire the operating positive power set corresponding to various types of loads. Running negative power set Load operation variation power set S ΔP Load operating time set and load shutdown time set Where S ΔP This only applies to loads with multiple changing states, such as electric fans with different speed settings corresponding to different power outputs. The operating positive power set is the set of power changes when various electrical loads are turned on; the operating negative power set is the set of power changes when various electrical loads are turned off; the operating variable power set is the set of power changes when various electrical loads are running; the load operating time set refers to the time from when an electrical load is turned on to when it is turned off; and the load turning off time set refers to the time from when an electrical load is turned off to when it is turned on.

[0067] Characteristic function library computing device 206, based on various loads regarding and S ΔP Images from each dataset, for and S ΔP The probability density function is approximated using a normal distribution and calculated using formula (5). For... and The calculation is performed using the Weibull distribution and formula (6).

[0068] This invention proposes an electrical load identification system based on the maximum likelihood method, which organically combines iterative and maximum likelihood classification methods. First, it detects load switching events to quickly search for load start-up and shutdown statuses. Then, it iterates to determine the probability of load existence, reducing the dimensionality of load identification calculations. Finally, based on the maximum likelihood method and combined with characteristics such as operating power, power variation, and operating time of each load, it calculates the probability of each load's existence; the maximum probability value is the load identification result. This system demonstrates high reliability for electrical load identification.

[0069] To ensure the applicability of the above verification algorithm and load sampling time to each user, several residential users were selected for experimental verification. First, feature databases of each load of each household user were collected. Then, the load of residential users was identified based on event detection, integer programming and maximum likelihood classification. The load identification results of different residential users are shown in Table 2.

[0070] Table 2

[0071] family Accuracy (%) Average battery power error (%) Family 1 84 4.5 Family 2 91 2.3 Family 3 79 13.1 Family 4 82 6.8 Family 5 93 3.5 Family 6 85 7.8 average value 85.67 6.33

[0072] Accuracy represents the accuracy of the overall load identification results. This is a fairly stringent metric, requiring that the identified set of load start / stop states completely overlaps with the actual set. It is defined as follows:

[0073]

[0074] Where m is the number of samples for identifying the load, I() represents a mapping, a sign function, I(true) = 1, I(false) = 0.

[0075] Power consumption and error: In addition to determining the working status of the load, another purpose of this paper is to determine the power consumption of each load, thereby improving users' awareness of energy conservation. The power consumption calculation data for each type of load is obtained from the load identification results. The power consumption calculation formula is as shown in equation (9), and the power consumption error is calculated as shown in equation (10).

[0076]

[0077]

[0078] As shown in Table 2, the average accuracy of each electrical load identification reaches over 85%, indicating good electrical load identification performance.

[0079] In summary, the beneficial result of this invention lies in proposing a maximum likelihood classification method for identifying electrical loads. First, event detection is used to determine the on / off events of loads, quickly searching for the start / stop status of each load and obtaining the corresponding load power difference at that moment. Then, an iterative method is used to determine the probability of a load's existence. Finally, based on the maximum likelihood method and combined with the operating power characteristics, operating power changes, and operating time of each load, the probability of each load's existence is calculated, and the maximum probability value is the load identification result. This method demonstrates strong reliability for electrical load identification.

[0080] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A power load identification method based on maximum likelihood method, characterized in that, The method comprises: The total active power collected at the power consumption inlet is subjected to an event detection algorithm to quickly retrieve the power consumption load switch event, and the load power difference value is calculated; The power difference value is subjected to an iterative method and the power range value to obtain the possible existence of the power consumption load; The possible existence of the power consumption load is subjected to a maximum likelihood method and the probability density function characteristic library of the running power, the change of the running power, the running time and the shutdown time of each power consumption load to calculate the possible probability value of the existence of each power consumption load; The probability value is compared and sorted to obtain the maximum value, which is the power consumption load identification result; The possible existence of the power consumption load is subjected to a maximum likelihood classification method and the probability density function characteristic library of the running power, the change of the running power, the running time and the shutdown time of each power consumption load to calculate the possible probability value of the existence of each power consumption load; which specifically comprises: First, the characteristic library of each load probability density function is calculated, mainly considering the load operation positive power set Operation negative power set S P - Load operation change power set S ΔP Load operation time set And load shutdown time set Where S ΔP Only for multi-change state load; According to the various loads on S P - and S ΔP each data set image, for S P - and S ΔP The normal distribution is used to approximate the probability density function, and the formula (5) is used for calculation: Wherein, P is the load power value at different time, μ and σ are the mean and standard deviation of normal distribution respectively; By analogy, for and Using the Weibull distribution, as in equation (6): Then, the load existence possibility is obtained by the initial identification of the iteration method, denoted as S = (s1, s2, λs i ), S is the load existence possibility set, i is the number of load existence, and then the maximum likelihood method is used to solve the probability of each load existence. The maximum probability is the event detection load identification result, which is expressed by formula (7). Wherein, X is the load sample; the probability of the existence of each load is calculated according to the probability density function in the load characteristic library.

2. The maximum likelihood method-based power load identification method according to claim 1, characterized by, The total active power is subjected to an event detection algorithm to quickly retrieve the switch event of the load, and the power difference value of the power consumption load is calculated, which specifically comprises: First, the average value and the standard deviation of the previous four time points at each time t are calculated, as shown in formula (1) and (2): Then, the power difference value ΔP(t) is calculated, as shown in formula (3): ΔP(t) = P(t) - P(t-1) (3) But the power load exists opening and closing, namely ΔP(t) exists positive and negative; and judge σ P (t) and |ΔP(t)| and threshold σ g , judge the load fluctuation condition, when the load fluctuation condition is greater than the threshold, namely the load exists switching event, otherwise, there is no switching event; threshold σ g The fluctuation condition of each load opening is obtained through experiment verification.

3. The maximum likelihood method-based power load identification method according to claim 1, characterized by, The active power difference value is solved by using the iterative method, and the possible existing load at this time is determined according to the power range, which specifically comprises: During the use of each load, the use of different types of loads by the user leads to different changes in power consumption; the power difference value ΔP(t) is obtained through event detection, and the load start-stop vector that may exist at this time is X; according to the power balance, formula (4) is obtained, and all groups of functions that meet the conditions are solved by computer programming and formula (4); | ΔP(t) - P i T | X | ≤ Δmax (4) where X(i) e {0, 1}, the load start-stop vector X takes values limited to two values of 0 and 1, 0 represents that the electrical appliance is closed, and 1 represents that the electrical appliance is opened; P i min ≤ P i ≤ P i max , P i represents the running active power condition of the i-th electrical load, P i min and P i max are the minimum active power and the maximum active power when the load is running, and Δmax represents the maximum error allowed by the load power.

4. A power load identification system based on maximum likelihood method, characterized by, The system comprises: An event detection device analyzes the total active power collected at the power consumption inlet to obtain the power consumption load switch condition and calculate the power difference value; A load existing combination determination device is used to solve the power consumption load switch condition by the iterative method to determine the possible existing load at this time; A maximum likelihood calculation device calculates the possible probability value of the existence of each power consumption load according to the probability density function characteristic library of the running power, the change of the running power, the running time and the shutdown time of each power consumption load; A power consumption load determination device compares and sorts the possible probability value to obtain the maximum value, which is the power consumption load identification result; The probability value is compared and sorted to obtain the maximum value, which is the power consumption load identification result; The possible existence of the power consumption load is subjected to a maximum likelihood classification method and the probability density function characteristic library of the running power, the change of the running power, the running time and the shutdown time of each power consumption load to calculate the possible probability value of the existence of each power consumption load; which specifically comprises: First, the characteristic library of each load probability density function is calculated, mainly considering the load operation positive power set Operation negative power set S P - Load operation change power set S ΔP Load operation time set And load shutdown time set Where S ΔP Only for multi-change state load; According to the various loads on S P - and S ΔP Each data set image, for S P - and S ΔP The normal distribution is used to approximate the probability density function, and the formula (5) is used for calculation: where P is the load power value at different time, μ and σ are the mean and standard deviation of normal distribution, respectively; and for and The Weibull distribution is used, as in equation (6): Similarly, Then, the load existence possibility is obtained by the initial identification of the iteration method, denoted as S = (s1, s2, λs i ), S is the load existence possibility set, i is the number of load existence, and then the maximum likelihood method is used to solve the probability of each load existence. The maximum probability is the event detection load identification result, which is expressed by formula (7). Wherein, X is the load sample; the probability of the existence of each load is calculated according to the probability density function in the load characteristic library.

5. The system of claim 4, wherein, The system further comprises: A pre-acquisition device is used to pre-acquire the running positive power set corresponding to various types of loads The running negative power set S P - The load running change power set S ΔP The load running time set And the load shutdown time set The characteristic function library calculating device calculates the operating positive power set of each type of load through normal distribution The operating negative power set S P - And the load operating change power set S ΔP The probability density function of the operating time set And the load shutdown time set The characteristic function library of each type of load is determined.

Images