CN103544545B - Electric system risk evaluation accelerating sampling method based on fault set matching - Google Patents

Abstract

An accelerated sampling method for power system risk assessment based on fault set matching, the method includes: given a non-load shedding fault set of a power system, any fault state in the fault set will not cause load shedding of the power system. When the fault state of a certain power system is obtained by sampling, it is matched with the fault state in the fault set. If the two match, the optimization calculation will not be performed; if the two do not match, the optimization calculation will be performed. If the calculation result does not shed the load, the sampling state will be added to the non-load shedding fault set for matching in future calculations. . Since the failure rate of components in the power system is low, and the failure of individual components generally does not lead to load shedding, the optimization calculation for calculating the load shedding after sampling consumes a lot of calculation time, so this method avoids many unnecessary optimization calculations. A large amount of computing time is saved, thereby improving the efficiency of sampling calculation procedures in power system risk assessment.

Description

An Accelerated Sampling Method for Power System Risk Assessment Based on Fault Set Matching

technical field

The invention relates to an accelerated sampling method for power system risk assessment based on fault set matching, and belongs to the technical field of power system risk assessment.

Background technique

In power system risk assessment, the Monte Carlo method is commonly used for sampling, and the results of several sampling are averaged to obtain the result of risk assessment. In each sampling process, the deterministic state of each component is obtained according to the failure rate of each component, and then the optimization calculation is carried out, and a series of indicators such as the load shedding of the power system under the sampling state are obtained by an optimized method. The process of optimization calculation is the most computationally expensive part of each sample. However, in the calculation process of power system risk assessment, since the failure rate of power system components is generally small, and the failure of a single component generally does not cause power system load loss, the results of load shedding obtained by optimal calculation are mostly zero. That is, there is no load shedding fault in the power system. Therefore, in power system risk assessment, if it is clearly known that load shedding will not occur in a certain sampling state obtained, the optimization calculation is skipped, and the load shedding amount is directly given as zero, which can save the time of this optimization calculation. Improve computational efficiency. Therefore, based on the above ideas, a set of effective methods is needed to realize accelerated sampling of risk assessment and save computational overhead.

Contents of the invention

The purpose of the present invention is to solve the problem that the optimization calculation of each sampling process in the power system risk assessment has a large calculation cost and poor economic efficiency, and proposes an accelerated sampling method for power system risk assessment based on fault set matching. The matching of the state and the fault state in the non-load shedding fault set is used to judge whether optimization calculation is needed, thereby saving the time of each sampling and reducing the calculation cost.

The technical solution for realizing the present invention is that the method of the present invention judges whether optimization calculation needs to be performed by matching the sampling state and the fault state of the non-load-cutting fault set under the condition of a given system non-load-cutting fault set, and continuously updates the un-cutting fault set. Load fault sets, thereby reducing unnecessary optimization calculations, saving time for each sampling, and improving program calculation efficiency.

The inventive method comprises the following steps:

(1) Set the maximum number of fault states n _{FaultSetMatch} considered in the fault set matching algorithm (n _{FaultSetMatch} ≥ 0);

(2) Obtain a sampling state according to the failure rate of each component in the power system, record the total number of line faults n _LineFaults at this time, the total maximum output of the power unit P _{TotalUnitPowerMax} , the total load of non-faulty nodes P _{TotalNodeLoad} , and the total number of faulty nodes n _NodeFaults ;

(3) If the total number of faulty nodes _nNodeFaults is zero, and the total number of line faults _nLineFaults is not greater than the maximum number of fault states _{nFaultSetMatch} considered in the fault set matching algorithm, and the total maximum output of the power unit P _{TotalUnitPowerMax} is greater than the load of non-faulty nodes 1.05 times of the sum P _{TotalNodeLoad} , then carry out the matching of the fault set, if the condition is not satisfied, it is considered that the fault set matching is unsuccessful, and the optimization calculation of step (6) is directly carried out; that is, if n _NodeFaults = 0, and n _LineFaults ≤ n _{FaultSetMatch} , and P _{TotalUnitPowerMax} >1.05P _{TotalNodeLoad} , then carry out fault set matching, otherwise it is considered that the fault set matching is unsuccessful, and directly proceed to step (6) optimization calculation;

(4) If fault set matching is required, proceed to steps (4.1) to (4.3);

(4.1) Traverse the non-load shedding fault set to obtain the line state of a certain fault state in the non-load shedding fault set;

(4.2) If all the line states in this fault state in the non-load shedding fault set are completely consistent with the line state obtained by sampling, then it is considered that the fault set matching is successful, and the matching is no longer performed, and step (5) is directly carried out;

(4.3) If after the traversal of the non-load shedding fault set is completed, there is no match with the line state obtained by sampling, then it is considered that the fault set matching is unsuccessful, and step (5) is performed;

(5) If the fault set matching is successful, the optimization calculation in step (6) is not performed, and the load shedding of all nodes is directly assigned to zero, and step (8) is performed; if the fault set matching is unsuccessful, the step ( 6) Optimal calculation;

(6) optimize the calculation to obtain the load shedding amount of each node, and calculate the total load shedding P _{TotalCutdownLoad} of all nodes;

(7) Update the non-load shedding fault set:

(7.1) If the total number of faulty nodes _nNodeFaults is zero, and the total number of line faults _nLineFaults is not greater than the maximum number of fault states _{nFaultSetMatch} considered in the fault set matching algorithm, and the total maximum output of the power unit P _{TotalUnitPowerMax} is greater than the load of non-faulty nodes 1.05 times of the sum P _{TotalNodeLoad} , and the total load shedding P _{TotalCutdownLoad} of all nodes is zero, then proceed to step (7.2) to update the fault set, otherwise proceed to step 8); that is, if n _NodeFaults = 0, and n _LineFaults ≤ n _{FaultSetMatch} , and P _{TotalUnitPowerMax} >1.05P _{TotalNodeLoad} , and P _{TotalCutdownLoad} =0, then proceed to step (7.2) to update the failure set, otherwise proceed to step (8);

(7.2) Record the fault state of each line in the sampling state of the power system obtained in this sampling, and add it to the non-load shedding fault set as a non-load shedding fault;

(8) Repeat step (2) until the sampling meets the requirements, and the risk assessment procedure ends.

The beneficial effect of the present invention is that the accelerated sampling method proposed by the present invention judges whether optimization calculation is needed by matching the sampling state and the fault state in the non-load shedding fault set on the basis of the matching of the non-load shedding fault set in the power system, And continuously update the non-load shedding fault set, thereby reducing unnecessary optimization calculations, saving the time of each sampling, and improving the calculation efficiency of the program. The invention makes full use of the information obtained in the sampling process to update the system non-load shedding fault set, effectively reduces the number of optimization calculations, saves sampling time, and improves the efficiency of the risk assessment calculation program.

Description of drawings

Fig. 1 is a flowchart of an accelerated sampling method for power system risk assessment based on fault set matching in the present invention.

detailed description

The specific implementation of the accelerated sampling method for power system risk assessment of the present invention is described in detail as follows:

(1) Set the maximum number of fault states n _{FaultSetMatch} =2 considered in the fault set matching algorithm;

(2) Obtain a sampling state according to the failure rate of each component in the power system, record the total number of line faults at this time n _LineFaults = 1, the total maximum output of the starting unit P _{TotalUnitPowerMax} = 3385MW, and the total load of non-faulty nodes P _{TotalNodeLoad} = 2850MW , the total number of faulty nodes n _NodeFaults = 0;

(3) The above values satisfy that the total number of faulty nodes n _NodeFaults is zero, and the total number of line faults n _LineFaults is not greater than the maximum number of fault states n _{FaultSetMatch} considered in the fault set matching algorithm, and the total maximum output of the power unit P _{TotalUnitPowerMax} is greater than the non-faulted The condition of 1.05 times of the load sum of nodes P _{TotalNodeLoad} needs to match the failure set;

(4) Carry out fault set matching, traverse the non-load shedding fault set, and obtain the line state of a certain fault state in the non-load shedding fault set; Failure set matching is unsuccessful, proceed to step (5);

(5) failure set matching is unsuccessful, then carry out step (6) optimization calculation;

(6) Optimal calculation obtains the load shedding amount of each node, and calculates the total load shedding P _{TotalCutdownLoad} =0 of all nodes;

(7) Update the non-load shedding fault set:

(7.1) above-mentioned conditions satisfy condition n _NodeFaults =0, and n _{LineFaults≤n FaultSetMatch} , and P _{TotalUnitPowerMax} >1.05P _{TotalNodeLoad} , and P _{TotalCutdownLoad} =0, carry out step (7.2) update fault set;

(7.2) Record the fault state of each line in the sampling state of the power system obtained in this sampling, and add it to the non-load shedding fault set as a non-load shedding fault;

(8) Repeat step 2) until the sampling meets the requirements, a total of 260,000 samples are taken, and the risk assessment procedure ends, taking a total of 11.984 seconds.

Compared with the above accelerated sampling method, without the accelerated sampling sampling method, it takes 331.015 seconds to sample 260,000 times.

Claims (1)

1. An accelerated sampling method for power system risk assessment based on fault set matching, characterized in that, under the condition of a given system non-load shedding fault set, by matching the sampling state and the fault state in the non-load shedding fault set To judge whether it is necessary to perform optimization calculation, and constantly update the non-load shedding fault set, thereby reducing unnecessary optimization calculation, saving the time of each sampling, and improving the calculation efficiency of the program; the steps of the method are: (1) Set the maximum number of fault states n _{FaultSetMatch} considered in the fault set matching algorithm (n _{FaultSetMatch} ≥ 0); (2) Obtain a sampling state according to the failure rate of each component in the power system, record the total number of line faults n _LineFaults at this time, the total maximum output of the power unit P _{TotalUnitPowerMax} , the total load of non-faulty nodes P _{TotalNodeLoad} , and the total number of faulty nodes n _NodeFaults ; (3) If the total number of faulty nodes _nNodeFaults is zero, and the total number of line faults _nLineFaults is not greater than the maximum number of fault states _{nFaultSetMatch} considered in the fault set matching algorithm, and the total maximum output of the power unit P _{TotalUnitPowerMax} is greater than the load of non-faulty nodes 1.05 times of the sum P _{TotalNodeLoad} , then carry out the matching of the fault set, if the condition is not satisfied, then it is considered that the fault set matching is unsuccessful, and the optimization calculation of step (6) is directly carried out; that is, if n _NodeFaults = 0, and n _LineFaults ≤ n _{FaultSetMatch} , and P _{TotalUnitPowerMax} > 1.05P _{TotalNodeLoad} , then the fault set matching is performed, otherwise it is considered that the fault set matching is unsuccessful, and the optimization calculation of step (6) is directly carried out; (4) If fault set matching is required, proceed to steps (4.1) to (4.3): (4.1) Traverse the non-load shedding fault set to obtain the line state of a certain fault state in the non-load shedding fault set; (4.2) If all the line states in this fault state in the non-load shedding fault set are completely consistent with the line state obtained by sampling, then it is considered that the fault set matching is successful, and the matching is no longer performed, and step (5) is directly carried out; (4.3) If after the traversal of the non-load shedding fault set is completed, there is no match with the line state obtained by sampling, then it is considered that the fault set matching is unsuccessful, and step (5) is performed; (5) If the fault set matching is successful, the optimization calculation in step (6) is not performed, and the load shedding of all nodes is directly assigned to zero, and step (8) is performed; if the fault set matching is unsuccessful, the step ( 6) Optimal calculation; (6) optimize the calculation to obtain the load shedding amount of each node, and calculate the total load shedding P _{TotalCutdownLoad} of all nodes; (7) Update the non-load shedding fault set: (7.1) If the total number of faulty nodes _nNodeFaults is zero, and the total number of line faults _nLineFaults is not greater than the maximum number of fault states _{nFaultSetMatch} considered in the fault set matching algorithm, and the total maximum output of the power unit P _{TotalUnitPowerMax} is greater than the load of non-faulty nodes 1.05 times of the sum P _{TotalNodeLoad} , and the total load shedding P _{TotalCutdownLoad} of all nodes is zero, then proceed to step (7.2) to update the fault set, otherwise proceed to step (8); that is, if n _NodeFaults = 0, and n _{LineFaults ≤} n _{FaultSetMatch} , and P _{TotalUnitPowerMax} > 1.05P _{TotalNodeLoad} , and P _{TotalCutdownLoad} = 0, then proceed to step (7.2) to update the fault set, otherwise proceed to step (8); (7.2) Record the fault state of each line in the sampling state of the power system obtained in this sampling, and add it to the non-load shedding fault set as a non-load shedding fault; (8) Repeat step (2) until the sampling meets the requirements, and the risk assessment procedure ends.