Distribution network operation mode plan making method based on expert feedback knowledge updating

Abstract

The invention discloses a distribution network operation mode plan making method based on expert feedback knowledge update, which comprises the following steps: S1, aiming at the carding of a distribution network system operation plan generation process, stepwise processing the whole plan generation process, and by checking the advantages and disadvantages of expert experience intervention in each step, carding historical data to obtain a distribution network system operation plan generation process; evaluating whether expert experience in the steps needs to intervene or not, and completing updating of expert knowledge feedback in the plan by historical data; s2, after the requirement of an operation mode plan is put forward, data of boundary conditions are coded, the coded data are matched with historical data, whether expert knowledge needs to be involved in each step or not is checked, feasible knowledge is optimized according to the expert knowledge, and finally checking is conducted through mechanism analysis. According to the method, historical data formulated by a distribution network operation mode plan is analyzed before and after manual intervention, the advantages and disadvantages of manual expert knowledge intervention in each process are checked, and an evaluation strategy of the influence degree is given.

Description

Technical Field

[0001] The invention relates to the technical field of power system operation, and in particular to a method for formulating distribution network operation mode plans based on expert feedback knowledge update. Background Art

[0002] Expert feedback knowledge updating technology studies how to use human flexibility and machine efficiency to enable machines and humans to collaborate to complete complex tasks efficiently and accurately. With the development of artificial intelligence-related technologies and industries, more and more artificial intelligence systems and products are deeply integrated into various application fields. How to better interact and collaborate with people has become the focus of attention. By combining the advantages of human knowledge and experience and artificial intelligence technology, complex human-machine collaborative tasks can be better completed. The ultimate goal is to achieve the natural integration of humans and artificial intelligence in application scenarios. How to efficiently use the prior knowledge and experience of human experts to achieve effective collaboration between humans and machines and engineering applications of data flow interaction is still insufficient. There are still many challenges and problems in the technical aspects of human-machine collaboration such as knowledge representation, data sharing, expert evaluation feedback, and collaborative decision optimization that need further research.

[0003] In the context of new power systems, the topology of distribution networks is becoming more complex as the scale of the system grows, and the operation modes are becoming more diverse. At present, there is still a lack of effective human-machine collaboration means that can continuously improve the safety, rationality and effectiveness of dispatching agents in distribution network operation mode plan tasks through expert evaluation and feedback, and save and learn this knowledge. Therefore, by constructing a dispatcher evaluation feedback evaluation mechanism for agent behavior, establishing a "dispatcher-agent" human-machine collaborative two-way interactive architecture, and realizing the effective transformation and dynamic optimization of dispatching agent understanding learning and dispatcher feedback evaluation, it has become one of the feasible methods for generating distribution network operation mode plans. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and propose a distribution network operation mode plan formulation method based on expert feedback knowledge update. The method refines the process of distribution network operation mode plan formulation on the basis of combing the process of distribution network operation mode plan formulation in combination with mechanism analysis, and analyzes the historical data of distribution network operation mode plan formulation before and after manual intervention, checks the pros and cons of manual expert knowledge intervention in each process, and gives an evaluation strategy for the degree of impact. When a new operation mode plan demand is proposed, the plan is formulated by matching historical data and integrating the pros and cons of expert experience in the process.

[0005] A distribution network operation mode plan formulation method based on expert feedback knowledge update of the present invention comprises the following steps:

[0006] S1. Sorting out the distribution network system operation plan generation process, the entire plan generation process is step-by-step, and by checking the pros and cons of expert experience intervention in each step, historical data is sorted out to evaluate whether expert experience needs to be involved in the step, and the historical data is used to update the expert knowledge feedback in the plan, so as to prepare for the operation mode plan;

[0007] S2. When the need for an operation mode plan is raised, the boundary condition data is encoded and matched with the historical data to check whether expert knowledge is needed in each step. Based on the expert knowledge, feasible knowledge is selected and finally verified through mechanism analysis to support the preparation of the plan.

[0008] Preferably, S1 comprises the following sub-steps:

[0009] S1.1 According to the requirements of the dispatching department, sort out the data requirements for the formulation of the operation mode plan, including topological parameters, measurement data and prediction data;

[0010] S1.2 Sort out the requirements for the formulation of distribution network operation mode plans under different working conditions, including normal operation, maintenance, and accident recovery;

[0011] S1.3 Classify the historical operation mode plan data according to normal mode, maintenance mode, and accident recovery mode;

[0012] S1.4 In normal operation mode, the operation mode plan data is divided into steady-state, transient, dynamic and other steps; similarly, the maintenance mode adds the transfer plan formulation, and the accident recovery mode adds the accident recovery plan step;

[0013] S1.5 In each step, split the data into data before expert knowledge intervention and data after intervention;

[0014] S1.6 For each piece of historical data, through mechanism analysis, the effect of each step before and after the expert intervention is scored according to very poor -2, poor -1, no 0, good 1, very good 2;

[0015] S1.7 The boundary condition data before the operation mode plan is formulated shall be encoded according to the principle of topology first and measurement later;

[0016] S1.8 Classify historical data into three categories: normal mode, maintenance mode, and accident recovery mode;

[0017] S1.9 performs secondary classification based on the encoded data, first topology and then measurement;

[0018] S1.10 After the boundary data is classified under the three major categories, the expert knowledge intervention evaluation of each step in the plan is added;

[0019] S1.11 After the first-level and second-level boundary data are classified, the expert knowledge intervention evaluations for each step are merged. If it is positive, the surface expert intervention is active; if it is negative, the surface intervention is negative; and 0 means useless. This completes whether each step requires expert intervention and the role of the intervention under different requirements and different boundary conditions.

[0020] Preferably, the three processing methods in S1.2 are:

[0021] S1.2A In normal mode, under given boundary conditions, the plan formulation process is divided into: voltage and current over-limit verification under steady state, stability analysis under transient state, and analysis under dynamic stability;

[0022] S1.2B In the maintenance mode, it is necessary to first develop a special supply plan for important loads, update boundary conditions, and then perform analysis in the normal way;

[0023] S1.2C Under accident recovery, first clarify the accident recovery plan, update the boundary conditions, and then perform the analysis in the normal way.

[0024] Preferably, the encoding method of S1.7 is:

[0025] S1.7A topological data, generator is 0, load is 1, line is 2, followed by nodes, line is the first and last node, load and generator nodes are filled in twice;

[0026] S1.7B measurement data,

[0027] According to the node voltage, the upper and lower limits are coded from 1 to 7 according to -3% to +7% in 2% intervals;

[0028] Line load rate, from -100% to +100%, 50% interval, upper and lower limits are coded from 1-7;

[0029] Generators and loads, coded from 0-100%, 20% intervals, 1-5.

[0030] Preferably, S2 comprises the following sub-steps:

[0031] S2.1 Formulate a plan for the operation of the distribution network in accordance with the dispatching requirements and the requirements of the distribution network operation.

[0032] S2.2 Collect and predict the data of the operation mode plan, and complete the data encoding in the order of topology first and measurement later

[0033] S2.3 uses the operation requirements as indexes, selects historical data determined by the primary and secondary classifications, and matches them with the collected data accordingly. If they do not match, repeat step S2.3. If they match, proceed to the next step;

[0034] S2.4 After matching the data, check the pros and cons of the expert knowledge intervention in each step, select the intervention as positive, merge the same measures that are rated as good (i.e. 2) in the historical data of this type, sort them, and select the measures with the most

[0035] S2.5 If the analysis result is consistent with the matched data, then it is completed, otherwise go to step S2.3.

[0036] The advantages and technical effects of the present invention are:

[0037] According to the needs of distribution network operation mode plan formulation, the present invention proposes a distribution network operation mode plan formulation method based on expert feedback knowledge update, by splitting the steps of operation mode formulation and using historical data to verify the role of expert knowledge in each step to evaluate, thereby providing a reference for mode plan formulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method S1 of the present invention;

[0039] Figure 2 This is a flow chart of the method S2 of the present invention. DETAILED DESCRIPTION

[0040] The embodiments of the invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the invention, and cannot be understood as limiting the invention.

[0041] The present invention proposes a distribution network operation mode plan formulation method based on expert feedback knowledge update. On the basis of combing the process of distribution network operation mode plan formulation in combination with mechanism analysis, the process is refined, and based on this, the historical data of distribution network operation mode plan formulation before and after human intervention is analyzed, the pros and cons of human expert knowledge intervention in each process are checked, and an evaluation strategy for the degree of impact is given. Then, when a new operation mode plan demand is proposed, the plan is formulated by matching historical data and integrating the pros and cons of expert experience in the process.

[0042] The present invention mainly includes two parts. The first part is to sort out the distribution network system operation plan generation process, and to step the entire plan generation process. By checking the pros and cons of expert experience intervention in each step, historical data is sorted out to evaluate whether expert experience needs to be intervened in the step. If the pros and cons of expert experience are involved, the expert knowledge feedback in the plan is updated through historical data; the second part is to clarify the steps where experts need to intervene and the suggestions of expert feedback knowledge in given steps through matching current data and boundary conditions with historical data after new operation mode requirements are met, so as to realize the requirement of formulating operation mode plans quickly through updating expert feedback knowledge.

[0043] In the first part, the key focus is on how to evaluate whether the expert knowledge feedback in the historical data is positive or negative based on the steps obtained by mechanism analysis. The corresponding block diagram is as follows Figure 1 As shown, the specific steps include:

[0044] (1) According to the requirements of the dispatching department, sort out the data requirements for the formulation of the operation mode plan, including topological parameters, measurement data, and prediction data.

[0045] (2) Sort out the requirements for the formulation of distribution network operation mode plans under different working conditions, including normal operation, maintenance, and accident recovery.

[0046] a) In normal mode, under given boundary conditions, the plan formulation process is divided into: voltage and current over-limit verification under steady state, stability analysis under transient state, and analysis under dynamic stability.

[0047] b) In the maintenance mode, it is first necessary to develop a special supply plan for the important loads, update the boundary conditions, and then conduct the analysis in the normal way.

[0048] c) Under accident recovery, first clarify the accident recovery plan, update the boundary conditions, and then analyze according to the normal mode. (3) Classify the historical operation mode plan data according to the normal mode, maintenance mode, and accident recovery mode.

[0049] (4) Under normal operation, the operation mode plan data is divided into steady-state, transient, dynamic and other steps; similarly, the maintenance mode adds the power transfer plan formulation, and the accident recovery adds the accident recovery plan step.

[0050] (5) In each step, the data is split into data before expert knowledge intervention and data after expert knowledge intervention.

[0051] (6) For each piece of historical data, through mechanism analysis, the effect of each step before and after the expert intervention is scored according to very poor (-2), poor (-1), no (0), good (1), and very good (2).

[0052] (7) The boundary condition data before the operation mode plan is formulated shall be encoded according to the principle of topology first and measurement later.

[0053] a) The topological data is as follows: generator is 0, load is 1, line is 2, followed by nodes, the line is the first and last nodes, and the nodes of load and generator are filled in twice.

[0054] b) Measurement data: node voltage (from -3% to +7%, 2% interval, upper and lower limits coded from 1-7), line load rate (from -100% to +100%, 50% interval, upper and lower limits coded from 1-7), generator and load (0-100%, 20% interval, coded from 1-5).

[0055] (8) Classify historical data into three categories: normal mode, maintenance mode, and accident recovery mode.

[0056] (9) According to the encoded data, first topology and then measurement are performed to perform secondary classification.

[0057] (10) After the boundary data are classified into three categories, the expert knowledge intervention evaluation is added to each step of the emergency plan.

[0058] (11) After the first-level and second-level boundary data are classified, the expert knowledge intervention evaluation for each step is merged. If it is positive, the surface expert intervention is active; if it is negative, the surface intervention is negative; and 0 means useless. This completes whether each step needs expert intervention under different requirements and different boundary conditions, as well as the role of the intervention.

[0059] In this way, based on the results of the mechanism analysis, the data classification is completed for the historical basic data, which serves as the basis for the next step of matching.

[0060] In the second part, when the demand for the operation mode plan is put forward, the boundary condition data is encoded, and the encoded data is matched with the historical data to check whether expert knowledge is needed in each step, and feasible knowledge is selected based on expert knowledge. Finally, the mechanism analysis is used to verify and support the preparation of the plan. The corresponding block diagram is as follows: Figure 2 As shown, the specific steps include:

[0061] (1) Formulate a plan for the operation of the distribution network in accordance with the dispatching requirements and the requirements of the distribution network operation.

[0062] (2) Collect and predict the data for the operation mode plan, and complete the data encoding in the order of topology first and measurement later

[0063] (3) Using the operation requirements as an index, select the historical data determined by the primary and secondary classifications, and match them with the collected data. If there is a mismatch, repeat step (3). If there is a match, proceed to the next step.

[0064] (4) After matching the data, check the pros and cons of the expert knowledge intervention in each step, select the intervention as positive, merge the same measures that are rated as good (i.e., 2) in the historical data of this type, sort them, and select the measures with the most

[0065] If the analysis result is consistent with the matched data, then it is complete, otherwise go to step (3).

[0066] Finally, all the parts not described in the present invention adopt mature products and mature technical means in the prior art.

[0067] In the description of this specification, reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in the embodiment or example of the invention.

[0068] Although embodiments of the invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A distribution network operation mode plan formulation method based on expert feedback knowledge update, characterized in that: The following steps are involved: S1. Sorting out the distribution network system operation plan generation process, the entire plan generation process is step-by-step, and by checking the pros and cons of expert experience intervention in each step, historical data is sorted out to evaluate whether expert experience needs to be involved in the step, and the historical data is used to update the expert knowledge feedback in the plan, so as to prepare for the operation mode plan; S2. When the need for an operation mode plan is raised, the boundary condition data is encoded and matched with the historical data to check whether expert knowledge is needed in each step. Based on the expert knowledge, feasible knowledge is selected and finally verified through mechanism analysis to support the preparation of the plan.

2. According to claim 1, a distribution network operation mode plan formulation method based on expert feedback knowledge update is characterized by: The S1 comprises the following sub-steps: S1.1 According to the requirements of the dispatching department, sort out the data requirements for the formulation of the operation mode plan, including topological parameters, measurement data and prediction data; S1.2 Sort out the requirements for the formulation of distribution network operation mode plans under different working conditions, including normal operation, maintenance, and accident recovery; S1.3 Classify the historical operation mode plan data according to normal mode, maintenance mode, and accident recovery mode; S1.4 In normal operation mode, the operation mode plan data is divided into steady-state, transient, dynamic and other steps; similarly, the maintenance mode adds the transfer plan formulation, and the accident recovery mode adds the accident recovery plan step; S1.5 In each step, split the data into data before expert knowledge intervention and data after intervention; S1.6 For each piece of historical data, through mechanism analysis, the effect of each step before and after the expert intervention is scored according to very poor -2, poor -1, no 0, good 1, very good 2; S1.7 The boundary condition data before the operation mode plan is formulated shall be encoded according to the principle of topology first and measurement later; S1.8 Classify historical data into three categories: normal mode, maintenance mode, and accident recovery mode; S1.9 performs secondary classification based on the encoded data, first topology and then measurement; S1.10 After the boundary data is classified under the three major categories, the expert knowledge intervention evaluation of each step in the plan is added; S1.11 After the first-level and second-level boundary data are classified, the expert knowledge intervention evaluations for each step are merged. If it is positive, the surface expert intervention is active; if it is negative, the surface intervention is negative; and 0 means useless. This completes whether each step requires expert intervention and the role of the intervention under different requirements and different boundary conditions.

3. The method for formulating a distribution network operation mode plan based on expert feedback knowledge update according to claim 2, characterized in that: The three processing methods in S1.2 are: S1.2A In normal mode, under given boundary conditions, the plan formulation process is divided into: voltage and current over-limit verification under steady state, stability analysis under transient state, and analysis under dynamic stability; S1.2B In the maintenance mode, it is necessary to first develop a special supply plan for important loads, update boundary conditions, and then perform analysis in the normal way; S1.2C Under accident recovery, first clarify the accident recovery plan, update the boundary conditions, and then perform the analysis in the normal way.

4. The method for formulating a distribution network operation mode plan based on expert feedback knowledge update according to claim 2, characterized in that: The encoding method of S1.7 is: S1.7A topological data, generator is 0, load is 1, line is 2, followed by nodes, line is the first and last node, load and generator nodes are filled in twice; S1.7B measurement data, According to the node voltage, the upper and lower limits are coded from 1 to 7 according to -3% to +7% in 2% intervals; Line load rate, from -100% to +100%, 50% interval, upper and lower limits are coded from 1-7; Generators and loads, coded from 0-100%, 20% intervals, 1-5.

5. The method for formulating distribution network operation mode plan based on expert feedback knowledge update according to claim 1 is characterized in that: The S2 comprises the following sub-steps: S2.1 Formulate a plan for the operation of the distribution network in accordance with the dispatching requirements and the requirements of the distribution network operation. S2.2 Collect and predict the data of the operation mode plan, and complete the data encoding in the order of topology first and measurement later S2.3 uses the operation requirements as indexes, selects historical data determined by the primary and secondary classifications, and matches them with the collected data accordingly. If they do not match, repeat step S2.

3. If they match, proceed to the next step; S2.4 After matching the data, check the pros and cons of the expert knowledge intervention in each step, select the intervention as positive, merge the same measures that are rated as good (i.e. 2) in the historical data of this type, sort them, and select the measures with the most S2.5 If the analysis result is consistent with the matched data, then it is completed, otherwise go to step S2.3.

Images