Feeder load removal method and system considering distributed power supply

Abstract

The invention provides a feeder load removal method and system considering a distributed power supply. The method comprises the following steps: performing real-time iterative sorting on each feeder according to load removal cost and load removal effect; calculating the total low-frequency load shedding amount according to the steady-state frequency set value; when it is detected that the frequency of the power system is decreased to a flexible load removal set value, the flexible load is removed preferentially; if the frequency rises again after the flexible load is cut off, the feeder line cutting-off process is stopped; if it is detected that the frequency rise is not obvious or does not rise, when it is detected that the frequency of the power system drops to the first-round action frequency of low-frequency load shedding, a cutting instruction is issued; determining the load shedding position and capacity according to the frequency change rate, implementing load shedding, cutting off the feeder line with the lowest total score of the feeder line, detecting the rising condition of the frequency after the load cutting is completed, ending the low-frequency load shedding if the frequency rises, and switching to the next level of load cutting if the frequency further drops; and detecting the frequency change condition, and when the frequency rises again, stopping the low-frequency load shedding process, switching to a secondary frequency modulation stage, and adjusting the frequency back to the range required by the working frequency.

Description

Technical Field

[0001] The present invention belongs to the technical field of feeder load shedding, and in particular relates to a feeder load shedding method and system considering distributed power sources. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] As the third line of defense of the power system, low-frequency load shedding is an effective way to suppress frequency drop and maintain system frequency stability. The traditional low-frequency load shedding scheme adopts a successive approximation calculation method to cut off the pre-set load according to the action frequency of each round.

[0004] The main round of low-frequency load shedding relies on frequency relays and time relays. When the power system frequency drops to a certain corresponding frequency level, the system will cut off part of the load; if the system frequency rises back up, the load shedding step will be stopped. If the system frequency continues to drop, it will reach the corresponding frequency of the n+1 level low-frequency load shedding, and the system will cut off part of the load until the system frequency returns to normal.

[0005] like Figure 1 As shown in Figure 2, before the fault occurred, the system frequency was stable at the rated value f e , assuming that a large amount of active power shortage occurs in the system at t1, and the system frequency drops sharply. When the frequency drops to f2, the first round of frequency relays starts and disconnects part of the user load. Since part of the load is cut off, the frequency will continue to drop according to the 2-3 curve instead of the 2-4 curve. When the frequency drops to f3, the second round of frequency relays for low-frequency load shedding starts and disconnects the second round of set loads. Figure 1 In the case of disconnecting the load set in the second round, the frequency starts to rise along the 3-6 curve and finally stabilizes at the recovery frequency f ss , which indicates that the total amount of load shedding in the first two times is roughly equal to the power shortage. If the total amount of disconnected load is less than the power shortage after the second round of action, the frequency will continue to drop and gradually approach the value of the system power shortage through low-frequency load shedding until the system frequency stabilizes or rises, and the whole process of low-frequency load shedding ends.

[0006] However, with the implementation of new energy policies such as "whole-county photovoltaic", massive distributed power sources are connected to the distribution network, which brings a series of problems to the traditional low-frequency load reduction strategy:

[0007] 1) With the access of distributed power sources, the output proportion of traditional power sources decreases, the inertia decreases, and the frequency changes are more obvious when responding to disturbances;

[0008] 2) After the distributed power source is connected to the feeder, the net load is reduced. If the load is reduced in turns according to the load ratio, the distributed power source may be miscut and the number of turns may be increased;

[0009] 3) After the distributed power generation is connected, the power flow of some feeders may be reversed. The traditional low-frequency load reduction scheme cannot identify this situation, resulting in the cutting off of these feeders, which further deteriorates the frequency.

[0010] A comparative simulation was conducted before and after the distributed power supply was connected. The results are as follows: Figure 2 After distributed photovoltaic access, the traditional low-frequency load reduction scheme needs to remove more rounds of loads to keep the system frequency relatively stable, and the lowest frequency point is lower than before access.

[0011] In addition, although there are removal schemes based on intelligent algorithms in the prior art, for example: CN113312839B_A method and device for emergency auxiliary load shedding decision-making based on reinforcement learning, CN107749620B_A method for power supply restoration of a distribution network containing distributed power sources, and the document "A method for removing distributed power sources based on communication protection fusion", the problems existing in the above technical schemes are:

[0012] The existing technology does not closely integrate distributed power sources, flexible loads and low-frequency load shedding processes. At the same time, flexible loads are not used as low-frequency load shedding frequency modulation resources. The cost of flexible load shedding is lower than that of other loads. Considering flexible loads as a frequency modulation resource can reduce the load shedding cost of low-frequency load shedding. In addition, the output of distributed power sources cannot be measured in real time, and the existing technology cannot predict it. Summary of the invention

[0013] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a feeder load shedding method considering distributed power sources. The feeder load shedding method considering distributed power sources is based on a new feeder sorting method, retains feeders with more distributed power source output, realizes accurate shedding of feeder loads, and also considers giving priority to shedding a part of flexible loads to achieve better economic benefits.

[0014] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0015] In a first aspect, a feeder load shedding method considering distributed power sources is disclosed, comprising:

[0016] Iterate and sort each feeder in real time according to the load shedding cost and load shedding effect;

[0017] Calculate the total amount of low-frequency load reduction based on the steady-state frequency setting value;

[0018] When it is detected that the power system frequency drops to the flexible load removal setting value, the flexible load is removed first;

[0019] If the frequency rises again after the flexible load is cut off, the feeder cutting process is stopped;

[0020] If it is detected that the frequency recovery is not obvious or does not recover, when it is detected that the power system frequency drops to the first round of low-frequency load shedding action frequency, a shedding command is issued;

[0021] Determine the load shedding position and capacity based on the frequency change rate, implement load shedding, and cut off the feeder with the lowest total score. After the load shedding is completed, detect the frequency recovery. If the frequency recovers, end the low-frequency load shedding. If the frequency drops further, proceed to the next level of load shedding.

[0022] Detect frequency changes. When the frequency rises again, stop the low-frequency load reduction process and enter the secondary frequency regulation stage to adjust the frequency back to the required working frequency range.

[0023] As a further technical solution, the feeder priorities are sorted, and the specific steps are as follows:

[0024] The set indicators include: load category, load frequency characteristics and the impact of distributed generation;

[0025] Divide each indicator into levels according to its importance, and obtain the judgment matrix according to the level division;

[0026] The weight of each indicator is calculated based on the judgment matrix;

[0027] The score of each indicator is multiplied by the corresponding weight to finally obtain the total score of each feeder. All feeders are ranked according to the total score of each feeder.

[0028] As a further technical solution, each indicator is divided into levels according to its importance, including: load category as the first-level target; distributed generation as the second-level target; load characteristic frequency as the third-level indicator.

[0029] As a further technical solution, the score of each indicator includes the load category score, load frequency characteristic score and distributed generation impact score;

[0030] The load category score is: the ratio of the unit outage cost to the unit outage cost benchmark value;

[0031] The load frequency characteristic score is: the ratio of the product of the power of the load on the feeder and the frequency characteristic coefficient of the load on the feeder to the reference power;

[0032] Distributed generation impact score: the ratio of the sum of actual load power and distributed generation power to net load power.

[0033] As a further technical solution, the load shedding position and capacity are determined according to the frequency change rate, and load shedding is implemented. When load shedding, the net load power of the feeder is used as the basis for load shedding, and the output of the distributed generation is separated from the net load power of the feeder.

[0034] As a further technical solution, the output of the distributed power source is predicted by historical data, feeder net load power and environmental factors.

[0035] As a further technical solution, when predicting the output of the distributed power source:

[0036] Obtain the measurement sequence corresponding to the light intensity, temperature and wind intensity in the area over a period of time;

[0037] Set the total power sequence of distributed photovoltaic power generation and the total power sequence of distributed wind power generation;

[0038] Assume that there is a nonlinear mapping relationship between the total power of distributed photovoltaic power generation and the light intensity and temperature; there is a nonlinear mapping relationship between the total power of distributed wind power generation and the wind intensity;

[0039] Use BP neural network to establish a separate prediction model for distributed power output, including: establishing a prediction model for distributed photovoltaic real-time output; establishing a prediction model for distributed wind power generation;

[0040] The real-time output of distributed power sources is predicted based on the established model.

[0041] In a second aspect, a feeder load shedding system considering distributed power sources is disclosed, comprising:

[0042] The sorting module is configured to: iteratively sort each feeder in real time according to the load shedding cost and the load shedding effect;

[0043] The load shedding module is configured to: calculate the total amount of low-frequency load shedding according to the steady-state frequency setting value, and when it is detected that the power system frequency drops to the flexible load shedding setting value, preferentially shedding the flexible load;

[0044] If the frequency rises again after the flexible load is cut off, the feeder cutting process is stopped;

[0045] If it is detected that the frequency recovery is not obvious or does not recover, when it is detected that the power system frequency drops to the first round of low-frequency load shedding action frequency, a shedding command is issued;

[0046] Determine the load shedding position and capacity based on the frequency change rate, implement load shedding, and cut off the feeder with the lowest total score. After the load shedding is completed, detect the frequency recovery. If the frequency recovers, end the low-frequency load shedding. If the frequency drops further, proceed to the next level of load shedding.

[0047] The frequency adjustment module is configured to: detect frequency changes, and when the frequency rises, stop the low-frequency load reduction process, enter the secondary frequency adjustment stage, and adjust the frequency back to the required working frequency range.

[0048] One or more of the above technical solutions have the following beneficial effects:

[0049] Considering the feeder load shedding method of distributed power generation, based on the new feeder sorting method, the feeders with more distributed power generation output are retained, the feeder load is accurately sheared, and the priority shedding of some flexible loads is considered to achieve better economic benefits. In the period when the distributed power generation output is large, this scheme can effectively reduce the number of load shedding rounds and the amount of load shedding, with a higher minimum frequency and better economy; in the period when the distributed power generation output is small, it can achieve slightly better results than the traditional scheme.

[0050] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0052] Figure 1 This is the schematic diagram of low frequency load reduction;

[0053] Figure 2 This is a schematic diagram showing the comparison of simulation results before and after the distributed power supply is connected;

[0054] Figure 3 A schematic diagram of a feeder load shedding process considering a distributed power source according to an embodiment of the present invention;

[0055] Figure 4 It is the network diagram of the simulation system;

[0056] Figure 5 It is the simulation frequency variation diagram;

[0057] Figure 6 This is a load variation diagram. DETAILED DESCRIPTION

[0058] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0059] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0060] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0061] Embodiment 1

[0062] See attached Figure 3 As shown, this embodiment discloses a feeder load shedding method considering distributed power sources. When a disturbance occurs, the load shedding master station sends instructions according to the feeder sequencing of the previous cycle, and the load shedding substation performs the shedding.

[0063] First, analyze the impact of distributed power output on feeder power flow, and use the following formula to express the relationship between them:

[0064] p(t)=p L (t)-p DG (t) (1)

[0065] Where: p(t) is the net load power of the feeder, p L (t) is the actual load power of the feeder, p DG (t) is the power of distributed generation. The traditional low-frequency load reduction decision is based on p L (t) is used as the basis, and the sum of the actual load powers of the feeders that need to be cut off satisfies the system power shortage. However, from formula (1), it can be seen that for different feeders with the same actual load power p(t), the distributed power power p connected to the grid DG The feeder with higher net load power p(t) has smaller net load power p(t), and the cut-off of feeders with smaller net load power p(t) has less impact on the stability of the power system. Therefore, considering the impact of distributed generation output, the net load power of feeders should be used as the basis for load shedding.

[0066] A high proportion of distributed power generation on the user side is connected to the grid to supply some local load electricity, but the output of distributed power generation affects the measurement of actual load. In the subsequent feeder priority sorting scheme, it is necessary to separate the output of distributed power generation from the net load power of the feeder.

[0067] The specific steps of the feeder load shedding method considering distributed power sources include:

[0068] Step 1): After the load shedding master station collects data such as feeder net load power and load frequency characteristics, the load shedding master station performs real-time iterative sorting of each feeder according to the load shedding cost and load shedding effect;

[0069] Step 2): When the power system frequency is detected to drop to the flexible load removal setting value, the flexible load is automatically removed first;

[0070] Step 3): If the frequency rises after the flexible load is cut off, the feeder cut-off process is stopped. If the frequency rises slightly or not at all, it means that the cut-off of the flexible load has no effect. When the power system frequency drops to the first round of low-frequency load shedding action frequency, a cut-off command is issued.

[0071] Step 4): After the j-th level load is removed, the frequency recovery is detected. If the frequency recovers, the current low-frequency load shedding is terminated. If the frequency drops further, the j+1-th level load shedding is switched to. Note that when the j+1-th level low-frequency load shedding is not triggered at the low frequency point, the secondary round is started for load shedding, which is still counted as the j+1-th low-frequency load shedding.

[0072] Step 5): Detect the frequency change. When the frequency rises again, stop the low-frequency load reduction process and enter the secondary frequency regulation stage to adjust the frequency back to the required working frequency range.

[0073] In step 1), the specific steps of feeder priority sorting are:

[0074] The feeder priority sorting needs to consider the following indicators: load category, load frequency characteristics, distributed generation, and flow direction.

[0075] The analytic hierarchy process is used to quantitatively describe the load shedding priority of the feeder, with the following three indicators:

[0076] (1) Load category (w1);

[0077] (2) Load frequency characteristics (w2);

[0078] (3) Impact of distributed generation (w3);

[0079] Construct a judgment matrix and calculate the weight of each indicator according to the importance of comparison. The criteria for forming the judgment matrix are shown in Table 1. Divide each indicator into levels according to its importance: the load category directly determines the order of low-frequency load shedding, which is taken as the first-level target; the new scheme attaches importance to the influence of the flow brought by distributed generation, and takes distributed generation as the second-level target; the load characteristic frequency is a third-level indicator.

[0080] Table 1 Judgment matrix formation criteria

[0081]

[0082]

[0083] Note: ① Scales 2, 4, 6, and 8 represent the median of the two adjacent judgments above; ② Reciprocal: Judgment b is obtained by comparing factors i and j. ij , then factor j is compared with factor i to determine b ji =1 / bij .

[0084] According to the above level division, the judgment matrix is ​​as follows:

[0085]

[0086] Assume the weights are:

[0087] w1+w2+w3=1 (3)

[0088] Where: w1, w2, w3 represent the weights of load category, load frequency characteristics, and distributed generation impact, respectively. After matrix calculation, we can get:

[0089] w=[w1,w2,w3] T =[0.5813 0.1096 0.3091] T (4)

[0090] Multiply the score of each indicator by the corresponding weight to get the total score T of each feeder. i :

[0091] T i =(w1·S i1 +w2·S i2 +w3·S i3 )*δ (5)

[0092] Based on the total score T of each feeder i , sort all feeders, exclude feeders with a score of zero, and prioritize feeders with high scores. Feeders with low scores are cut off first. i1 S is the load category score, i2 is the load frequency characteristic score, S i3 is the distributed generation impact score, and δ is the direction detection variable.

[0093] The above feeder sequencing method takes into account the impact of distributed generation output and quantifies it as an indicator to participate in feeder sequencing and load shedding.

[0094] The scoring calculation steps for the feeder priority ranking index are as follows:

[0095] For each feeder i, each indicator is scored:

[0096] 1) Load category: The cost of load shedding for different categories is different. Generally speaking, industrial load > commercial and residential load > flexible load.

[0097] Define load category score S i1 :

[0098]

[0099] Where: VoLL i It is the unit outage cost, which is a common indicator to measure the cost of different load shedding. The unit is usually US dollars / kWh. base It is the unit power outage cost benchmark value. For example, VoLL base When the VoLL of industrial load is 100$ / kWh, g is 20$ / kWh, and its S g1 =0.2; VoLL for commercial load s is 10$ / kWh, and its S s1 =0.1, comparing the two values, it can be concluded that commercial load shedding has a higher priority. Flexible loads are different from other types of loads and usually participate in demand response and economic dispatch in the power market through incentive price mechanisms.

[0100] Flexible loads are loads with flexible characteristics that can actively participate in the operation and control of the power grid and interact with the power grid for energy. Flexible loads are divided into shiftable loads, transferable loads and curtailable loads, and the three types of loads have different incentive prices in different time periods. Power companies provide different incentive prices for flexible loads in different time periods to guide users to reduce electricity consumption during peak demand or when the power grid is under pressure, optimize system costs, and reduce losses caused by unnecessary load shedding. The incentive price of flexible loads is much lower than the shedding price of residential loads, and the shedding cost is relatively small, so adding a round of flexible load shedding before feeder shedding improves frequency dynamics.

[0101] 2) Load frequency characteristics: For traditional loads, the cut-off frequency standard is related to the frequency characteristics. Frequency characteristic coefficient K L Smaller loads are less sensitive to frequency changes and will not experience a significant reduction in power consumption when the system frequency drops, so they cannot help restore the system to balance through their own frequency response.

[0102] Define the load frequency characteristic score S i2 :

[0103]

[0104] Where: K Li is the frequency characteristic coefficient of the load on the feeder, P i is the power of the load on the feeder, P b is the reference power, used to normalize the power index.

[0105] The power system can measure the load frequency characteristics by installing frequency sensors and power meters on the load side of the feeder to record frequency changes and load response data in real time.

[0106] 3) Distributed generation: Feeders with distributed power sources that have a positive effect on frequency characteristics should be considered for retention, that is, the net load power of the feeder should be used instead of the actual load power of the feeder as the standard for low-frequency load shedding.

[0107] Defining Distributed Generation Impact Score i3 :

[0108]

[0109] Feeders with small indicators are cut off first. The output of distributed generation is predicted by the neural network model.

[0110] Finally, since distributed power sources will bring about bidirectional power flow problems, a special indicator for judging the power flow direction should be set: if the feeder power flow flows from the distribution network to the main network during detection, it will not be cut off.

[0111] Assume that the detection variable δ is δ=1 when the power flow direction is positive; δ=0 when the power flow direction is negative. When the scheme is executed, feeders with a score of zero are excluded first. After the distributed generation is connected, the power flow of some feeders may be reversed. The traditional low-frequency load reduction scheme cannot distinguish this situation, resulting in the removal of these feeders, which further deteriorates the frequency. The purpose of setting the detection variable is to identify the feeders with reverse power flow so that the new scheme can retain these feeders that are beneficial to system stability.

[0112] The net load power in the above formula (12) can be obtained by measurement, the distributed generation power is predicted by the following function, and the actual load power is calculated by formula (1) based on these two data.

[0113] The real-time output of distributed power sources cannot be measured in real time, so the system uses historical data, feeder net load power and environmental factors to make real-time output predictions. The above historical data refers to the weather, climate and other historical data recorded by the system in the area where the distributed power sources are located.

[0114] Assume that the net load power measurement value of the kth feeder over a period of time is:

[0115] p=[p1,p2,…,p N ] T (9)

[0116] Where N is the time length of the sequence.

[0117] Assume that the light intensity in the area during this period is s, the temperature is T, and the wind intensity is w, and the corresponding measurement sequence is:

[0118] s=[s1,s2,…,s N ] T (10)

[0119] T=[T1,T2,…,TN ] T (11)

[0120] w=[w1,w2,…,w N ] T (12)

[0121] Assume that the total power sequence of distributed photovoltaic power generation is p pv , the total power sequence of distributed wind power generation p w , assuming that the total distributed photovoltaic power generation power p pv There is a nonlinear mapping relationship between light intensity s and temperature T. pv ();Total power of distributed wind power generation p w There is a nonlinear mapping relationship between F and wind intensity w w ( ), using BP neural network to establish a separation prediction model for distributed power output:

[0122]

[0123] Where: V1,…,V n is the weight coefficient matrix of the neural network, and n is the number of neural network layers. The loss function is defined as:

[0124]

[0125] Where: E pv is the loss function used in distributed photovoltaic prediction, p PVi For the historical contribution of distributed photovoltaic, ^p PVi The predicted value of distributed photovoltaic is obtained by using the gradient descent method to minimize the loss function, and the weights and biases are updated until the loss function is less than a certain threshold. At this time, the prediction model F of distributed photovoltaic real-time output can be obtained. pv ().

[0126] To explain, in a neural network, the weight coefficient is the parameter that connects each neuron in the network. When the neural network starts training, the weight coefficient will be randomly initialized. With each iteration, the neural network gradually updates the weight according to the error and gradient, and finally fits the model. The weight parameter is an intermediate variable in the neural network algorithm and is only used to fit the model.

[0127] After obtaining the above-mentioned distributed photovoltaic real-time output prediction model, real-time distributed power output prediction can be performed based on the prediction model and the weather and environmental conditions of the day.

[0128] Prediction model for distributed wind power generation F w () can be obtained through the same training method:

[0129]

[0130] Where: E w is the loss function used in distributed wind power prediction, p wi For the historical output of distributed wind turbines, is the predicted value of distributed photovoltaic. The gradient descent method is used to minimize the loss function, and the weights and biases are updated until the loss function is less than a certain threshold. At this time, the prediction model F of distributed wind power generation can be obtained. w ( ).

[0131] Simulation comparison:

[0132] The distribution network shown in the figure is built by DIgSILENT\PowerFactory software. The distribution network is a feeder combination including load, photovoltaic power generation and wind power generation, such as Figure 4 As shown in the figure. The distribution system is a 110kV substation, where the load is represented as an aggregate load connected to 11 20kV feeders. Among the 11 feeders, Load 1-Load 7 are set as residential loads, Load 8-Load 10 are set as commercial loads, and Load 11 is set as industrial loads. Generators, transformers, loads, wind power generation, and photovoltaic power generation all use the standard model provided by DIgSILENT\PowerFactory. The distributed output is predicted, and three simulation scenarios are constructed at three times: 5:00, 13:00, and 17:00.

[0133] The frequency characteristics of active power and reactive power of different load models are as follows:

[0134]

[0135] Frequency parameter K of different types of loads pf and K qf As shown in Table 2; the output forecast of distributed generation in different time periods is shown in Table 3; the scores of each feeder at different times are shown in Table 4.

[0136] Table 2 Load frequency parameters

[0137]

[0138]

[0139] Table 3 Distributed power output forecast

[0140]

[0141] Table 4 Scores of different feeders at different times

[0142]

[0143] The cut-off frequency starts from 49.5 Hz, and a round of load is cut off every 0.15 Hz. A sudden load increase disturbance is applied to the system at 0.5 s. When the low-frequency load reduction does not respond, the disturbance will cause the system frequency to drop below 49 Hz, making the system unstable. The traditional low-frequency load reduction strategy and the new low-frequency load reduction strategy are used to observe the changes in system frequency and distribution network load after using different strategies.

[0144] like Figure 5 As shown in (a) in the figure, the simulated frequency change diagram at five points in time shows that the photovoltaic power generation effect is extremely weak in scenario one. The simulation results of the two schemes are very close. After cutting off the flexible load that accounts for a small proportion, three low-frequency load reductions were carried out. The lowest frequency of the system was around 49.2 Hz. After the third round of load shedding, it returned to the range allowed by the recovery frequency.

[0145] like Figure 5 As shown in (b) in the figure, the simulated frequency change diagram at 13 o'clock, the photovoltaic power generation effect is strong in scenario 2, and the two schemes are obviously contrasted. The traditional scheme operates for 4 rounds, cuts off 5 feeders, and the lowest frequency is about 49hz; the improved scheme cuts off 3 rounds, cuts off three loads, and the lowest frequency is about 49.2hz. Compared with the traditional scheme, 86.4MW is cut off.

[0146] As shown in (c) of the figure, the simulated frequency change diagram at 17 o'clock, in scenario 3, the lowest frequency of the traditional solution is 49.26Hz, and the lowest frequency of the improved solution is 49.22Hz. This is because the system frequency changes slowly and the overall action time is longer than the traditional method considering the load shedding method of distributed power sources, resulting in a lower frequency. In view of this, it is possible to consider appropriately reducing the load shedding frequency interval of the improved solution compared to the traditional solution during the period when photovoltaic power generation is weak.

[0147] The amount of load removed by the two schemes is as follows Figure 6 As shown, (a) is the load change at 5 o'clock, (b) is the load change at 13 o'clock, (c) is the load change at 17 o'clock, and the optimization of load shedding cost is shown in Table 5. In scenario 1, the load shedding of the two schemes is basically the same; in scenario 2, there is a large difference in the load shedding of the two schemes. The traditional scheme shedding 198.1MW load, the improved scheme shedding 111.7MW load, the improved scheme shedding 86.4MW less, and the optimization is about 43.61%. In scenario 3, the traditional scheme shedding 107.1MW load, the improved scheme shedding 91.9MW load, and the optimization is about 14.19%, but the load shedding level is not reduced. The simulation results show that in the time period of distributed power output, the new scheme can effectively improve the economic benefits of load shedding.

[0148] Table 5 Optimization of load shedding cost at different times

[0149]

[0150] Embodiment 2

[0151] The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the program.

[0152] Embodiment 3

[0153] The purpose of this embodiment is to provide a computer-readable storage medium.

[0154] A computer-readable storage medium stores a computer program, which executes the steps of the above method when executed by a processor.

[0155] Embodiment 4

[0156] The purpose of this embodiment is to provide a feeder load shedding system considering distributed power sources, including:

[0157] The sorting module is configured to: iteratively sort each feeder in real time according to the load shedding cost and the load shedding effect;

[0158] The load shedding module is configured to: calculate the total amount of low-frequency load shedding according to the steady-state frequency setting value, and when it is detected that the power system frequency drops to the flexible load shedding setting value, preferentially shedding the flexible load;

[0159] If the frequency rises again after the flexible load is cut off, the feeder cutting process is stopped;

[0160] If it is detected that the frequency recovery is not obvious or does not recover, when it is detected that the power system frequency drops to the first round of low-frequency load shedding action frequency, a shedding command is issued;

[0161] Determine the load shedding position and capacity based on the frequency change rate, implement load shedding, and cut off the feeder with the lowest total score. After the load shedding is completed, detect the frequency recovery. If the frequency recovers, end the low-frequency load shedding. If the frequency drops further, proceed to the next level of load shedding.

[0162] The frequency adjustment module is configured to: detect frequency changes, and when the frequency rises, stop the low-frequency load reduction process, enter the secondary frequency adjustment stage, and adjust the frequency back to the required working frequency range.

[0163] Embodiment 5

[0164] The purpose of this embodiment is to provide a computer program product containing instructions, which, when running on a computer, enables the computer to execute the methods and functions involved in any of the above embodiments.

[0165] The steps involved in the apparatus of the above embodiment correspond to the method embodiment 1, and the specific implementation method can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0166] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0167] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A feeder load shedding method considering distributed power sources, characterized in that: include: Iterate and sort each feeder in real time according to the load shedding cost and load shedding effect; Calculate the total amount of low-frequency load reduction based on the steady-state frequency setting value; When it is detected that the power system frequency drops to the flexible load removal setting value, the flexible load is removed first; If the frequency rises again after the flexible load is cut off, the feeder cutting process is stopped; If it is detected that the frequency recovery is not obvious or does not recover, when it is detected that the power system frequency drops to the first round of low-frequency load shedding action frequency, a shedding command is issued; Determine the load shedding position and capacity based on the frequency change rate, implement load shedding, and cut off the feeder with the lowest total score. After the load shedding is completed, detect the frequency recovery. If the frequency recovers, end the low-frequency load shedding. If the frequency drops further, proceed to the next level of load shedding. Detect frequency changes. When the frequency rises again, stop the low-frequency load reduction process and enter the secondary frequency regulation stage to adjust the frequency back to the required working frequency range.

2. A feeder load shedding method considering distributed power sources as claimed in claim 1, characterized in that: Sort the feeder priorities. The specific steps are as follows: The set indicators include: load category, load frequency characteristics and the impact of distributed generation; Divide each indicator into levels according to its importance, and obtain the judgment matrix according to the level division; The weight of each indicator is calculated based on the judgment matrix; The score of each indicator is multiplied by the corresponding weight to finally obtain the total score of each feeder. All feeders are ranked according to the total score of each feeder.

3. A feeder load shedding method considering distributed power sources as claimed in claim 1, characterized in that: Each indicator is divided into levels according to its importance, including: load category as the first-level target; distributed generation as the second-level target; and load characteristic frequency as the third-level indicator.

4. A feeder load shedding method considering distributed power sources as claimed in claim 1, characterized in that: The score for each indicator includes the load category score, load frequency characteristic score and distributed generation impact score; The load category score is: the ratio of the unit outage cost to the unit outage cost benchmark value; The load frequency characteristic score is: the ratio of the product of the power of the load on the feeder and the frequency characteristic coefficient of the load on the feeder to the reference power; Distributed generation impact score: the ratio of the sum of actual load power and distributed generation power to net load power.

5. A feeder load shedding method considering distributed power sources as claimed in claim 1, characterized in that: The load shedding position and capacity are determined according to the frequency change rate, and load shedding is implemented. When shedding load, the net load power of the feeder is used as the basis for load shedding, and the output of the distributed generation is separated from the net load power of the feeder.

6. A feeder load shedding method considering distributed power sources as claimed in claim 1, characterized in that: The output of the distributed power source is predicted by historical data, feeder net load power and environmental factors; Preferably, when predicting the output of the distributed power source: Obtain the measurement sequence corresponding to the light intensity, temperature and wind intensity in the area over a period of time; Set the total power sequence of distributed photovoltaic power generation and the total power sequence of distributed wind power generation; Assume that there is a nonlinear mapping relationship between the total power of distributed photovoltaic power generation and the light intensity and temperature; there is a nonlinear mapping relationship between the total power of distributed wind power generation and the wind intensity; Use BP neural network to establish a separate prediction model for distributed power output, including: establishing a prediction model for distributed photovoltaic real-time output; establishing a prediction model for distributed wind power generation; The real-time output of distributed power sources is predicted based on the established model.

7. A feeder load shedding system considering distributed power sources, characterized in that: include: The sorting module is configured to: iteratively sort each feeder in real time according to the load shedding cost and the load shedding effect; The load shedding module is configured to: calculate the total amount of low-frequency load shedding according to the steady-state frequency setting value, and when it is detected that the power system frequency drops to the flexible load shedding setting value, preferentially shedding the flexible load; If the frequency rises again after the flexible load is cut off, the feeder cutting process is stopped; If it is detected that the frequency recovery is not obvious or does not recover, when it is detected that the power system frequency drops to the first round of low-frequency load shedding action frequency, a shedding command is issued; Determine the load shedding position and capacity based on the frequency change rate, implement load shedding, and cut off the feeder with the lowest total score. After the load shedding is completed, detect the frequency recovery. If the frequency recovers, end the low-frequency load shedding. If the frequency drops further, proceed to the next level of load shedding. The frequency adjustment module is configured to: detect frequency changes, and when the frequency rises, stop the low-frequency load reduction process, enter the secondary frequency adjustment stage, and adjust the frequency back to the required working frequency range.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method described in any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are performed.

Images