Load loss value random characteristic analysis method

Abstract

The invention provides a loss load value random characteristic analysis method, and belongs to the technical field of safe and stable operation analysis and control of a power system. According to the analysis method, the dynamic characteristic of the lost load value has a random characteristic, and the random digital characteristic of the lost load value is obtained by quantifying the influence of a property factor and a quantity factor; comprising the following steps: firstly, constructing a mathematical model of property factors and an extraction method of random numeric features; secondly, constructing a simulation model of quantity factors and an extraction method of random digital characteristics of the quantity factors; and finally, a third step, constructing a mathematical model of a loss load value, and further obtaining a final random number feature. The analysis method provided by the invention provides a more accurate basic analysis tool for safe and stable operation analysis and control of the power system, and can better realize the maximization of the overall benefits of the whole society.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system security and stability operation analysis and control, and particularly relates to a method for accurately determining the value of lost load in a power system, and provides a method for analyzing the stochastic characteristics of the lost load value. The lost load value is an underlying basic parameter (key parameter) for all research on power system frequency security and operation control. Background Art

[0002] The lost load value generally refers to the economic loss quantity of the power system for losing a unit quantity of load, and it is a basic element in the power security value system. In all work related to power system security and optimal operation, especially in analyses such as the formulation of power system planning schemes, the preparation of operation plans, and the design of control strategies, this parameter is indispensable and belongs to an underlying basic parameter. If there are large deviations in the determination of the lost load value quantity, the credibility of the work results based on this will be greatly reduced. Therefore, the accurate determination of the lost load value has always received wide attention, and many theories and methods have been proposed and applied in practice.

[0003] Currently, the practice of setting the lost load value as a fixed value essentially ignores its dynamic characteristics and assumes that this parameter is static and unchanged.

[0004] The research on the lost load value can be traced back at least to the 1950s. For the economic investment problem required to improve the reliability of the power system, theories and methods such as the impact of disturbances on the power system and cost calculation, and power reliability value assessment have been explored. With the continuous in-depth research, typical calculation methods for the lost load value have emerged, such as calculating the lost load value based on the input-output method in "Research on the Value of Power Lost Load Based on the Input-Output Method" proposed by scholars like He Yongxiu, and calculating the lost load value based on the production function method in "The Formation Mechanism and Model of the Time-of-Use Trading Price in the Medium- and Long-Term Power Market in China" proposed by scholars like Huang Shanshan, etc. Using these models, theories and methods, the lost load value quantities of different regions and different electricity-consuming industries can be estimated.

[0005] Used to measure the value given by users to the electric energy commodity, the lost load value can be expressed as the average price that users are willing to pay per megawatt-hour of electric energy, that is, per unit quantity of electric energy, to avoid sudden power outages. In our understanding, the average here has two meanings: one is the average among the prices of all users, and the other is the average of the prices at all times. Currently, setting the lost load value as the average price to carry out relevant work has become a common practice in the academic and engineering circles at home and abroad.

[0006] The magnitude of the value of lost load depends on the types of industries associated with the relevant load. For a load of a single industry type, due to factors such as production and life patterns, its value of lost load may vary over time; while for the entire power system, since the load components at different load nodes are not identical, the value of lost load between each node, that is, spatially, may also be different; the interaction of the two factors of time and space will cause the value of lost load to exhibit dynamic characteristics.

[0007] Using the average price in the relevant analysis, that is, setting the value of lost load as a constant value, essentially ignores the dynamic characteristics of the value of lost load and assumes that this parameter is static and unchanged.

[0008] Taking the optimal operation management and control of the power system as the object of discussion and analysis, if the stochastic characteristics of the system's value of lost load are relatively obvious, it may lead to insufficient or even shortage of the safety margin of the designed operation plan, and may cause major losses to the system or the operation plan to be too conservative, thereby resulting in a significant increase in operating costs and being uneconomical. Therefore, using a static value of lost load in a system with strong stochastic characteristics of the value of lost load cannot maximize the benefits.

[0009] Although the value of lost load has dynamic stochastic characteristics, due to the relatively sufficient frequency stability defense ability of the earlier power system, the frequency safety and stability situation is not tense; in addition, the system operation conditions are relatively stable, and the load composition does not change significantly. Therefore, the dynamic stochastic characteristics of the value of lost load have not attracted attention, and this characteristic is generally ignored in the system operation management work, and the average price obtained from the weighted average value of lost load model or marginal value of lost load model based on the electricity consumption industry is used for relevant research and analysis, and the actual application also meets the industry requirements.

[0010] With the gradual advancement of the construction of the new power system, on the one hand, as the proportion of new energy capacity continues to increase, conventional units are replaced and gradually withdrawn from operation, resulting in a shortage of frequency control resources, and the frequency safety and stability state will be more severe; at the same time, the grid operation conditions will be more complex, and the frequency spatio-temporal distribution characteristics will be more obvious, making the dynamic characteristics of the value of lost load more significant. On the other hand, with the continuous enrichment of frequency response regulation resources, more accurate cost accounting is required to conduct refined management of various types of expensive control resources for frequency regulation and better achieve optimal resource allocation. Obviously, if the existing static data of the value of lost load is still used for analysis, it is impossible to better maximize the overall benefits of the whole society.

[0011] In summary, as a basic element in the power security value system, the value of load loss has increasingly prominent dynamic stochastic characteristics in the new power system. If not considered in the system operation and management work, it is impossible to maximize the overall social benefits in the probabilistic sense. Analyzing the stochastic characteristics of the load loss value based on stochastic theory, revealing the influence mechanism of the nature and quantity factors that determine the nature and quantity of the load loss in the power system on the stochastic characteristics of the load loss value, can establish an analytical mapping relationship between multiple influencing factors and the stochastic characteristics of the load loss value. Constructing the node distribution mode of the load loss value in the power system and the stochastic process model of the load loss mode can analyze the influence of the stochastic factors of nature and quantity on the stochastic characteristics of the load loss value in the system. On this basis, proposing an analysis method for the stochastic characteristics of the load loss value under the superposition of multiple stochastic factors can achieve the analytical calculation of the stochastic characteristics of the load loss value, develop the representation form of the load loss value from a static parameter to a dynamic stochastic parameter, and then form a dynamic stochastic analysis theory of the load loss value, as the development and supplement of the theory for accurately determining the quantity of the load loss value, providing a more accurate basic analysis tool for the analysis and control of the safe and stable operation of the power system. Summary of the Invention

[0012] Aiming at the above problems existing in the prior art, based on the inherent dynamic stochastic characteristics of the load loss value, this invention revolves around the proposed concept of the stochastic characteristics of the load loss value, reveals and quantifies the influencing factors of its stochastic characteristics, and proposes a method for obtaining its stochastic digital characteristics, providing a basic and accurate analysis tool for the analysis and control of the safe and stable operation of the power system.

[0013] To achieve the above object, the technical solution adopted by this invention is as follows:

[0014] A method for analyzing the stochastic characteristics of the load loss value, in which the dynamic characteristics of the load loss value in the analysis method have stochastic characteristics. Due to the extremely complex stochastic characteristics caused by the superposition of the two influencing factors of nature and quantity, the premise for obtaining the stochastic digital characteristics of the load loss value is to quantify the influence of the nature factor and the quantity factor. The specific steps are as follows:

[0015] 1) Construct a mathematical model of the nature factor and a method for extracting stochastic digital characteristics

[0016] The nature of load loss is determined by the types of relevant loads. It is closely related to factors such as the production and living patterns of users. The inherent random characteristics of these factors will result in the load loss value having random characteristics in space-time distribution. Since these influencing factors determine the nature of load loss, they can be simply referred to as nature factors. In terms of time, for a single type of load, the random characteristics of the load quantity fluctuations caused by factors such as production and living patterns will cause the load loss value of a single type of load to possibly change dynamically over time and have random characteristics; for a load node composed of different types of electrical loads, its load components may also change dynamically over time and have random characteristics. The combined effect of this factor and the random characteristics of a single type of load will result in the quantity of the load loss value of the node having random characteristics; in terms of space, generally speaking, the load components of each node in the power system are different, which leads to different load loss values for each node.

[0017] Furthermore, the combination of the time distribution characteristics and the space distribution characteristics results in the distribution of the load loss value of the nodes in each section, that is, the load loss value node distribution pattern, which is also different and shows random characteristics. Therefore, through analysis, a mathematical model for calculating nature factors is constructed. The specific steps are as follows:

[0018] First, calculate the load proportion of the load connected to the load node through historical load data. The expression is:

[0019]

[0020] In formula (1), is the load proportion of load k in node i at time t, is the total load of node i at time t, is the load value of load type k in node i at time t, and K is the total number of loads connected to the node.

[0021] Subsequently, calculate the load loss value of each node, which is composed of the load proportions of each load in the node and the load loss values of each load connected to the node. The expression is:

[0022]

[0023] In formula (2), C i,t is the load loss value of node i at time t, and V k is the load loss value of load k.

[0024] From formula (2), the load proportions of each load component in each node at each moment can be obtained. Through kernel density estimation, the random digital characteristics of the load proportions of the node load components in different time periods can be obtained. In this step, the random digital characteristics are the first moment and the second moment of the load proportion.

[0025] Finally, it can be seen from Equation (2) that to obtain the value of load shedding at node i at time t, in addition to calculating the load ratio, data on the load shedding values of each load connected to this node are also required. Then, the data regarding C can be obtained through Equation (2), and then the first moment and second moment of this data set are calculated to obtain the load shedding value C of each node i,t in the random number characteristics within each time period. i,t

[0026] 2) Construct a simulation model for quantity factors and a method for extracting their random number characteristics;

[0027] The quantity of load shedding depends on the value of the lowest point of the frequency after the system is disturbed. Since the dynamic process of the system frequency after the system is disturbed has spatio-temporal distribution characteristics, the quantity of load shedding at each node is not consistent. The spatio-temporal distribution characteristics of the frequency depend on factors such as the operating conditions and large disturbances. The inherent random characteristics of these factors will cause the quantity of load shedding at each node to have different spatio-temporal distribution random characteristics. Since these influencing factors determine the quantity of load shedding, they can be simply referred to as quantity factors. The dynamic process of the frequency at each node after the system is disturbed is not the same, that is, the frequency change has spatio-temporal distribution characteristics. Both the disturbance (including the disturbance time and the amount of power imbalance caused by the disturbance) and the system operating conditions (including factors such as the frequency response regulation ability and the system frequency at the moment of disturbance) have strong random characteristics, resulting in the quantity of load shedding at each node in the power system showing random characteristics, which is also reflected in both time and space distributions.

[0028] In terms of time, the random characteristics of the disturbance and the system operating conditions will result in that even for the same disturbance, the quantity of load shedding at each node at different times may be different. This indicates that the quantity of load shedding at the node has random characteristics in time distribution.

[0029] In terms of space, after the system is disturbed, the dynamic processes of the frequencies at each node are not the same. This characteristic determines that the location (node) and quantity of the system's lost load are not the same. This indicates that the quantity of load shedding at the node has random characteristics in space distribution.

[0030] Furthermore, the time distribution characteristic and the space distribution characteristic interact with each other, resulting in the distribution pattern of the quantity of load shedding at the nodes in different sections, that is, the load shedding mode at the nodes is also not the same and shows random characteristics. Therefore, through theoretical analysis, a simulation model is constructed to obtain the random number characteristics of the quantity factors. The specific steps are as follows:

[0031] First, a simulation model of the power system is built by using a structure-preserving quasi-concentrated model. Each node of the power system is partitioned, and a multi-machine equivalent frequency response model considering the spatial distribution of the frequency is constructed. The mathematical model of its N-machine equivalent frequency response model is as follows: ​

[0032]

[0033] Among them, Δf represents the frequency deviation; P e represents the disturbance quantity; N represents the total number of units; F Hn represents the high-pressure cylinder coefficient; R n represents the frequency regulation coefficient of unit n; T Rn represents the reheating time constant of unit n; K mn represents the mechanical power gain factor; H represents the inertia time constant; D represents the load damping coefficient.

[0034] The frequency deviation Δf of each area can be obtained from Equation (3).

[0035] In addition, appropriate reheating time constant T R , inertia time constant H, load damping coefficient D, frequency regulation coefficient R and other parameters should be set when establishing the simulation model.

[0036] Subsequently, by adding disturbances with different magnitudes and probabilities in any area, the frequency change diagrams of each area can be obtained from Equation (3), and at the same time, the lowest frequency point and the probability of its occurrence are recorded.

[0037] Finally, according to the setting principles of relay protection for each node and the method of dividing the low-frequency load shedding rounds, the random digital characteristics of the load loss quantity in each area are obtained, and then the random digital characteristics of the quantity factor can be obtained. In this step, the random digital characteristics are the first moment and the second moment of the load cut-off quantity due to low-frequency load shedding in each area.

[0038] 3) Construct a mathematical model of the value of load loss and a method for obtaining its random digital characteristics

[0039] The random characteristics of the spatio-temporal distribution of the value of load loss, superimposed on the random characteristics of the load loss mode of the nodes, result in an extremely complex random distribution characteristic of the value of load loss. The present invention provides a method for calculating the value of load loss, and its mathematical model is:

[0040]

[0041] In the formula: C t is the value of load loss of the power system at time t, Q i,t is the load loss quantity of node i at time t, and n represents the total number of nodes.

[0042] The random digital characteristics of the system load loss value can be obtained from Equation (4). In this step, the random digital characteristics are the first moment and the second moment of the load loss value during daily operation.

[0043] The beneficial effects of the present invention are:

[0044] (1) The method for analyzing the stochastic characteristics of the value of lost load proposed by the present invention is a method for analyzing the stochastic characteristics of the node distribution pattern of the value of lost load considering the dynamic randomness of nature factors. The beneficial effects are as follows: a method for depicting the stochastic process of the value of lost load for different electricity consumption industries considering the stochasticity of multiple parameters in the input-output model is proposed; a method for describing the stochastic characteristics of the node load components considering the load industry label is proposed; and a method for calculating the stochastic characteristics of the node value of lost load by revealing the analytical mapping relationship between multiple stochastic variables is proposed. It can solve the problem that the static parameter theory of the value of lost load cannot consider the stochastic fluctuations of different types of loads, thereby realizing the analysis of the stochastic characteristics of the value of lost load caused by stochastic factors related to the load.

[0045] (2) The method for analyzing the stochastic characteristics of the value of lost load proposed by the present invention is also a method for analyzing the stochastic characteristics of the lost load pattern considering the dynamic randomness of quantity factors. The beneficial effects are as follows: a method for constructing a stochastic process model of the system lost load pattern considering the stochasticity of disturbances and operating conditions and a method for analyzing the stochastic characteristics of the quantity of lost load at nodes are proposed. It can solve the problem that the static parameter theory of the value of lost load ignores the stochastic characteristics of the spatio-temporal distribution of the quantity of system lost load, thereby realizing the analysis of the stochastic characteristics of the value of lost load caused by stochastic factors other than the load; and a method for analytically analyzing the stochastic characteristics of the value of system lost load under the superposition of multiple stochastic factors of nature and quantity is proposed. Compared with the static parameter theory of the value of lost load, it can more accurately determine the quantity of the value of lost load, providing a basic analysis tool for the analysis and control of the safe and stable operation of the power system.

[0046] In summary, the method proposed by the present invention provides a more accurate basic analysis tool for the analysis and control of the safe and stable operation of the power system, and can better realize the maximization of the overall benefits of the whole society. Brief Description of the Drawings

[0047] Figure 1 It is a schematic diagram of the action logic of the influencing factors of the dynamic stochastic characteristics of the value of lost load;

[0048] Figure 2 It is a schematic diagram of the system partition structure;

[0049] Figure 3 It is a schematic diagram of the spatio-temporal distribution stochastic characteristics of the value of lost load;

[0050] Figure 4 It is the load curve of the N-#1-#1 load on a certain day;

[0051] Figure 5 It is the load curve of the N-#1-#2 load on a certain day;

[0052] Figure 6 It is the four-region frequency response curve with a 12% disturbance in the N-#1 area;

[0053] Figure 7 The frequency response curve of four regions with a 15% disturbance in the N-#1 region;

[0054] Figure 8 Schematic diagram of the parameters related to the value of lost load in the power system under section t

[0055] Figure 9 The expectation and variance of the value of lost load in the power system; Figure 9 (a) The expectation of the value of lost load in the power system; Figure 9 (b) The variance of the value of lost load in the power system;

[0056] Figure 10 The probability density function of the value of lost load in the power system. Specific implementation manners

[0057] The present invention proposes a theory for analyzing the stochastic characteristics of the value of lost load. This theory is applied to all work related to the safe and optimal operation of the power system, especially in the analysis of formulating power system planning schemes, preparing operation plans, and designing control strategies. The present invention proposes that the dynamic characteristics of the value of lost load have stochastic characteristics, which are extremely complex under the superposition of two influencing factors, namely nature and quantity. The action logic of the influencing factors of the dynamic stochastic characteristics of the value of lost load is schematically shown in the appendix Figure 1 as shown.

[0058] The following further illustrates the present invention in conjunction with the accompanying drawings and specific examples, but it is not a limitation of the present invention.

[0059] It is worth noting that the present invention divides the target regional power grid into four regions, namely N-#1, N-#2, N-#3, and N-#4, as shown in the appendix Figure 2 as shown.

[0060] 1) Extraction of the stochastic digital characteristics of property factors

[0061] The nature of load loss is determined by the types of relevant loads. It is closely related to factors such as the production and living patterns of users. The inherent random characteristics of these factors will result in the random characteristics of the spatial and temporal distribution of the load loss value. Since these influencing factors determine the nature of load loss, they can be simply referred to as nature factors. In terms of time, for a single type of load, the random characteristics of the load quantity fluctuations caused by factors such as production and living patterns will lead to the load loss value of a single type of load may change dynamically over time and have random characteristics; for a load node composed of different types of electrical loads, its load components may also change dynamically over time and have random characteristics. The combined effect of this factor and the random characteristics of a single type of load will result in the random characteristics of the quantity of the load loss value of the node; In terms of space, generally speaking, the load components of each node in the power system are different, which leads to different load loss values for each node, which is manifested in the appendix Figure 3 That is, in section t, the load loss values of nodes i, j, and k may be different.

[0062] In this embodiment, the combination of the time distribution characteristics and the spatial distribution characteristics results in the distribution of the load loss value of the nodes in each section, that is, the load loss value node distribution pattern, which is also different and shows random characteristics. Therefore, through analysis, a mathematical model for calculating nature factors is constructed.

[0063] Since the proposed calculation method is still applicable to other regions, without loss of generality, the N-#1 region is taken as an example to illustrate the proposed calculation method. The specific steps are as follows:

[0064] First, through the power grid data center, the power consumption of all loads in the N-#1 region power grid during all daily operations in the past 5 years can be queried. The load ratio of each associated load in each region of the target region power grid during daily operation can be calculated through Equation (5):

[0065]

[0066] Among them, the load curves of two loads in this region on a certain day are shown in the appendix Figure 4 、 5 as shown.

[0067] Then, the load loss value of this region can be obtained from Equation (6). The PDF is calculated through the "gaussian_kde" function in the "scipy.stats" module in the Python library, and the random digital characteristics of the load loss value of this region are obtained:

[0068]

[0069] The calculated results show that the first moment and the second moment of the load shedding value in the N-#1 area during daily operation are at most 23.61 yuan / kWh and 19.27 yuan 2 / kWh 2 respectively, indicating that the load shedding value in the N-#1 area has strong volatility, uncertainty and is difficult to predict, and the direct economic loss caused by cutting off the load at the peak is relatively large.

[0070] 2) Extraction of the stochastic digital characteristics of the quantity factor

[0071] The quantity of load shedding depends on the value of the lowest point of the system frequency after the system is disturbed. Since the dynamic process of the system frequency after the system is disturbed has spatio-temporal distribution characteristics, the quantity of load shedding at each node is not the same. The spatio-temporal distribution characteristics of the frequency depend on factors such as the operating conditions and large disturbances of the system. The inherent stochastic characteristics of these factors will cause the quantity of load shedding at each node to have different spatio-temporal distribution stochastic characteristics. Since these influencing factors determine the quantity of load shedding, they can be simply referred to as quantity factors. The dynamic process of the frequency at each node after the system is disturbed is not the same, that is, the frequency change has spatio-temporal distribution characteristics. Both the disturbance (including the disturbance moment and the quantity of power imbalance caused by the disturbance) and the system operating conditions (including factors such as the frequency response regulation ability and the system frequency at the moment of disturbance) have strong stochastic characteristics, resulting in the quantity of load shedding at each node in the power system showing stochastic characteristics, which is also reflected in both time and space distributions.

[0072] In terms of time, the stochastic characteristics of the disturbance and the system operating conditions will cause the quantity of load shedding at each node at different moments to be possibly different even if the same disturbance occurs. This shows that the quantity of load shedding at the node has stochastic characteristics of time distribution.

[0073] In terms of space, the dynamic processes of the frequencies at each node after the system is disturbed are not the same, and this characteristic determines the location (node) and quantity of the system's lost load to be different. This shows that the quantity of load shedding at the node has stochastic characteristics of space distribution.

[0074] In this embodiment, the time distribution characteristics and the space distribution characteristics interact with each other, resulting in the distribution pattern of the quantity of load shedding at the nodes in different sections, that is, the load shedding mode at the nodes is also different and shows stochastic characteristics. Therefore, through analysis, a simulation model is constructed to obtain the stochastic digital characteristics of the quantity factor.

[0075] In this embodiment, by building a simulation model of the target area and applying different magnitudes and probabilities of disturbances to the power grid in the target area, frequency change curves of four different areas are obtained. Meanwhile, the values of the lowest frequencies are recorded, and according to the setting principles of relay protection for each node and the method of dividing the low-frequency load shedding rounds, the stochastic digital characteristics of the load loss quantity in each area are obtained. When the disturbance is greater than 12%, the lowest frequency in some areas is lower than the frequency at which the UFLS acts step by step. That is, when the disturbance occurs in the N-#1 area and the disturbance magnitudes are 12% and 15% respectively, the lowest frequencies of the system are 49.4595 Hz and 49.3244 Hz respectively, both of which will cause low-frequency load shedding events in the N-#1 area. According to the settings of the simulation data, 7% of the load in this area needs to be cut off. Among them, when the disturbance magnitudes are 12% and 15%, the frequency changes in the four areas of the system are as shown in the appendix Figure 6 、 7 shown.

[0076] Since only the N-#1 area needs to cut off the load, the load cut-off quantity in the N-#1 area accounts for 100% of the total load loss quantity in the power system. In addition, the frequency response ability of the system is related to parameters such as the reheat time constant T R 、the inertia time constant H, the load damping coefficient D, and the frequency regulation differential coefficient R. Among them, with the change of the inertia time constant H, it directly affects the initial slope frequency, the lowest point, and the peak time of the frequency curve. As the load damping coefficient D increases, the lowest point of the corresponding system frequency curve is higher, and the ability of the system to resist disturbances is stronger. With the changes of the reheat time constant T R and the frequency regulation differential coefficient R, the corresponding lowest frequency point and peak time also change.

[0077] 3) Extraction of the stochastic digital characteristics of the load loss value;

[0078] The stochastic characteristics of the spatio-temporal distribution of the load loss value, superimposed on the stochastic characteristics of the load loss pattern at the node, result in an extremely complex stochastic distribution characteristic of the load loss value, as shown in the appendix Figure 8 shown.

[0079] The stochastic digital characteristics of the system load loss value can be obtained from the following formula and the foregoing results:

[0080]

[0081] Multiply the load loss value in each area obtained above by the quantity of the load cut off due to the disturbance triggering the low-frequency load shedding, and then divide by the total load loss quantity, and the load loss value of the system at the time section t can be obtained. The calculation results are: the first moment and the second moment during its daily operation are as shown in the appendix Figure 9 (a), (b), and the probability density function is as shown in the appendix Figure 10The calculation results show that the first-order moment and second-order moment in daily operation can reach up to 23.61 yuan / kWh and 19.27 yuan respectively. 2 / kWh 2 .

[0082] Through case simulation, it can be seen that under the theory proposed by the present invention, the high expected peak period usually corresponds to the load peak of the node, indicating that the direct economic loss caused by cutting off the node is more significant at this time; the high variance period indicates that the volatility of the loss load value in this period is large and the uncertainty is strong, which is difficult to predict. The low variance period indicates that the loss load value is relatively stable.

[0083] The lowest frequency point in the disturbance area is lower than the lowest frequency point in other areas. When a disturbance occurs, the closer to the disturbance area, the more violent the frequency fluctuation; the farther from the disturbance area, the smoother the frequency fluctuation. The specific relationship is: the more violent the disturbance, the lower the corresponding lowest frequency point of the system, and the more serious the load loss in the area. And this change is nonlinear.

[0084] Aiming at the dynamic random characteristics inherent in the load loss value, the present invention focuses on the proposed concept of random characteristics of the load loss value, reveals and quantifies the influencing factors of its random characteristics, and proposes a method for obtaining its random digital characteristics, providing a basic analysis tool for the analysis and control of safe and stable operation of the power system.

[0085] The above-described embodiments merely express the implementation methods of the present invention, but they should not be understood as limiting the scope of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A method for analyzing the random characteristics of load loss value, characterized in that: The dynamic characteristics of the load loss value in the analysis method have random characteristics, and the random digital characteristics of the load loss value are obtained by quantifying the influence of the qualitative factors and the quantitative factors; the method includes the following steps: The first step is to construct a mathematical model of property factors and a method for extracting random digital features; The property factors have time distribution characteristics and space distribution characteristics. Combining the two results in the distribution of the node loss load value of each section, that is, the node distribution pattern of the loss load value presents a random characteristic; a mathematical model for calculating the property factors is constructed; the specific steps are: Step 1.1: Calculate the load proportion α connected to the load node through historical load data ik,t ; Step 1.2, calculate the load loss value C of each node i,t , which is composed of the load loss value of each load connected to the node and the proportion of each load at the node; By C i,t The expression can obtain the load proportion of each load component in each node at each time, and the random number characteristics of the load proportion of the node load component in different time periods can be obtained through kernel density estimation; Step 1.3, through C i,t The expression for obtaining the load loss value of node i at time t requires not only the calculation of the load proportion, but also the data of the load loss value of each load connected to the node. i,t The expression of C i,t Then, the first-order moment and second-order moment calculations are performed on this data set, and finally the load loss value C of each node is obtained. i,t Random number characteristics in each time period; The second step is to build a simulation model of the quantitative factor and a method to extract its random digital features; The quantity factor, i.e., the number of node load loss, has a random time distribution characteristic and a random spatial distribution characteristic. The two interact with each other, resulting in a distribution pattern of the number of node load loss in different sections, i.e., the node load loss pattern presents a random characteristic. A simulation model is constructed to obtain the random digital characteristics of the quantity factor. The specific steps are: Step 2.1, first build a simulation model of the power system, partition each node of the power system, and construct an N-machine equivalent frequency response model taking into account the frequency space distribution, and then build a mathematical model of the N-machine equivalent frequency response model; Step 2.2, by adding disturbances of different sizes and probabilities in any area, the frequency change diagram of each area is obtained by the mathematical model in step 2.1, and the lowest frequency point and the probability of occurrence of the lowest point are recorded; Step 2.3, according to the relay protection setting principle of each node and the low-frequency load shedding round division method, the random digital characteristics of the number of load loss in each area are obtained, and the random digital characteristics of the quantity factor can be obtained; The third step is to construct a mathematical model of the load loss value, and then obtain the final random digital characteristics; The mathematical model of the load loss value is: Where: C t is the load loss value of the power system at time t, Q i,t is the number of load losses of node i at time t, and n represents the total number of nodes; The random digital characteristics of the system load loss value are obtained by formula (4), and the random digital characteristics are the first-order moment and the second-order moment of the load loss value in daily operation.

2. A method for analyzing random characteristics of load loss value according to claim 1, characterized in that: In step 1.1 of the first step, The expression is: In formula (1), is the load proportion of node i occupied by load k at time t, is the total load of node i at time t, is the load value of load type k in node i at time t, and K is the total number of loads connected to the node.

3. A method for analyzing random characteristics of load loss value according to claim 1, characterized in that: Step 1.2 of the first step C i,t The expression is: In formula (2), C i,t is the load loss value of node i at time t, V k is the unloaded value of load k.

4. A method for analyzing random characteristics of load loss value according to claim 1, characterized in that: In the step 2.1, a simulation model of the power system is constructed by adopting a quasi-centralized model with structure preservation.

5. The method for analyzing random characteristics of load loss value according to claim 1, characterized in that: In step 2.1, the mathematical model of the N-machine equivalent frequency response model is as follows: Wherein, Δf represents the frequency deviation; P e represents the disturbance amount; N represents the total number of units; F Hn Indicates the high pressure cylinder coefficient; R n Indicates the frequency adjustment coefficient of unit n; T Rn Represents the reheat time constant of unit n; K mn represents the mechanical power gain factor; Hs represents the inertia time constant; D represents the load damping coefficient; The frequency deviation Δf of each area is calculated using formula (3).

Images