CN107516170A - A kind of difference self-healing control method based on probability of equipment failure and power networks risk - Google Patents

A kind of difference self-healing control method based on probability of equipment failure and power networks risk Download PDF

Info

Publication number
CN107516170A
CN107516170A CN201710765450.3A CN201710765450A CN107516170A CN 107516170 A CN107516170 A CN 107516170A CN 201710765450 A CN201710765450 A CN 201710765450A CN 107516170 A CN107516170 A CN 107516170A
Authority
CN
China
Prior art keywords
load
equipment
power
probability
failure
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201710765450.3A
Other languages
Chinese (zh)
Other versions
CN107516170B (en
Inventor
张化光
刘鑫蕊
孙秋野
张孝波
王智良
杨珺
赵鑫
郑瑶瑶
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northeastern University China
Original Assignee
Northeastern University China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Northeastern University China filed Critical Northeastern University China
Priority to CN201710765450.3A priority Critical patent/CN107516170B/en
Publication of CN107516170A publication Critical patent/CN107516170A/en
Application granted granted Critical
Publication of CN107516170B publication Critical patent/CN107516170B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Primary Health Care (AREA)
  • Water Supply & Treatment (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • General Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a kind of difference self-healing control method based on probability of equipment failure and power networks risk, it includes：1st, service data and weather data are gathered and calculates each self-corresponding resultant fault probability of each Distribution Network Equipment；2nd, risk threshold value is determined；3rd, judge whether resultant fault probability exceedes risk threshold value corresponding to the Distribution Network Equipment, if being then based on the first multiple objective function, optimize power system operating mode；Otherwise the second multiple objective function is based on, optimizes power system operating mode；Power system operating mode priority is set simultaneously；4th, predict the payload of power network subsequent time based on more post-class processing algorithms and calculate the synthesis stoppage in transit probability and resultant fault risk of subsequent time equipment；5th, determine whether to perform Optimized Measures.The present invention has taken into full account the operation risk of probability of equipment failure and power network, while uses difference self-healing control, realizes the economic and reliable operation that power network is ensure that in the case of not removal of load as far as possible.

Description

Difference self-healing control method based on equipment failure probability and power grid operation risk
Technical Field
The invention relates to the field of self-healing control of a power distribution network, in particular to a difference self-healing control method based on equipment fault probability and power grid operation risks.
Background
With the rapid development of the smart power grid, a large number of uncertain accesses of distributed power sources, the structure of the power distribution network is more and more complex, and the assessment and prediction of the operation risk of the power distribution network and the rapid diagnosis and recovery of faults are more and more difficult. In recent years, domestic and foreign scholars have proposed a series of self-healing control methods from different perspectives, and although some progress and results have been made in the research work, the shortcomings still exist.
The traditional power grid operation risk assessment adopts the product of probability and fault consequences, although the fault with small occurrence probability and serious consequences is considered, the fault with large occurrence probability and not serious consequences is ignored, and compared with the fault with the small occurrence probability and not serious consequences, the fault has higher probability, so the power grid operation risk assessment obtained by the method is not satisfactory. At present, the self-healing goal of the power grid is mainly to improve the operation reliability of the power grid, the relation between the economic operation and the reliability of the power grid is not fully considered, and the economic efficiency and the reliability cannot be considered in the obtained optimization scheme.
Disclosure of Invention
In view of the defects in the prior art, the invention aims to provide a difference self-healing control method based on equipment failure probability and power grid operation risk.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a difference self-healing control method based on equipment failure probability and power grid operation risk is characterized by comprising the following steps:
step 1, collecting operation data of each power distribution network device in a power grid and weather data corresponding to an area where the power grid is located, and calculating a comprehensive fault probability corresponding to each power distribution network device based on a set device outage probability model;
step 2, determining a risk threshold value required by evaluation of each power distribution network device;
step 3, judging whether the calculated comprehensive fault probability exceeds a risk threshold corresponding to the power distribution network equipment, and optimizing a power grid operation mode based on a first multi-objective function if the calculated comprehensive fault probability exceeds the risk threshold corresponding to the power distribution network equipment; otherwise, optimizing the operation mode of the power grid based on the second multi-objective function; meanwhile, the priority of the power grid operation mode is set, so that when the power grid is in a set sudden large fault, the power grid operation mode based on a multi-time optimization method can be preferentially adopted for rapid network reconstruction;
step 4, predicting the load size of the power grid at the next moment based on a multi-classification regression tree algorithm, and calculating the comprehensive outage probability and the comprehensive fault risk of equipment at the next moment, wherein the comprehensive fault risk specifically refers to the product of the comprehensive outage probability and the fault consequence;
and 5: and determining whether to execute optimization measures according to the power grid operation mode determined in the step 3 and the comprehensive operation probability and comprehensive fault risk predicted in the step 4 at the next moment.
Further preferably, the equipment outage probability model refers to distribution network equipment X at time t i Risk value of power failureIn particular to a certain time t, the power distribution network equipment X i The model determined by multiplying the probability of occurrence of the fault by the power failure influence degree of the fault is expressed by the following formula
R t (X i )＝P(X i |t)L(X i )
Wherein, P (X) i I t) is power distribution network equipment X at the time t i Probability of failure; l (X) i ) Is a distribution network device X i The degree of influence of the fault power failure.
Further preferably, the distribution network equipment X at time t i Probability of failure P (X) i T) is equal to the sum of the failure probabilities of the equipment corresponding to different risk sources, as shown in the following formula
Wherein, P k (X i I t) is that at time t, risk source k has device X i Probability of failure; m is the number of risk sources;
the risk sources include at least some or all of the following fault factors:
(1) The failure rate λ of the distribution equipment itself is represented by the following formula, where the corresponding function formula is λ (Y)
Wherein Y represents the service life, and beta is more than or equal to 0 and less than or equal to 1; p represents the fixed failure rate, t 1 Indicates the age, t, corresponding to the area of equipment descent 2 Representing the corresponding age of the equipment belonging to the ascending area;
(2) The fault rate P of the equipment overload, P is the probability of the equipment fault caused by the overload, and specifically is based on a utility function, and an overload value L of the equipment is introduced on the basis of the fault rate lambda of the distribution equipment OD The corresponding function formula is shown as the following formula
Wherein the content of the first and second substances,l is the proportion of the current flowing through the distribution equipment to the rated current thereof, and a is the rated current proportion threshold value freely set by a user;
(3) Equipment failure rate caused by external factors, wherein the external factors at least comprise a strong wind and strong rain factor, a lightning stroke factor and a construction damage factor;
the power failure probability P of the equipment fault corresponding to the factors of strong wind and strong rain WR The corresponding function formula is shown in the following formula
P WR ＝N GWR /(N D N WR )
Wherein N is GWR The number of power failure times of the equipment caused by strong wind and strong rain, N D Is the total number of certain types of equipment in the power distribution network, N WR The number of times of strong wind and strong rain of the power distribution network in the set statistical period is set;
the power failure probability P of equipment failure corresponding to the lightning strike factor TH The corresponding function formula is shown as the following formula
P TH ＝N GTH /(N D N TH )
Wherein N is GTH Number of power failures in such equipment due to lightning strike, N D Is the total number of certain types of equipment in the power distribution network, N TH The number of lightning strikes of the power distribution network in the set statistical period is set;
the power failure probability P of equipment failure corresponding to the lightning strike factor S The corresponding function formula is shown in the following formula
P S ＝N GS /N S
Wherein N is GS Number of times of failure of such equipment due to nearby construction, N S And finding the total construction times near certain equipment of the power distribution network in the set statistical period.
More preferably, the fault power failure influence degree L (X) i ) The index of the degree of influence of power failure caused by corresponding equipment failure comprises loss load L L Loss of electric quantity E L Hours of power failure user h POU Weighted number of users W in power failure user level POU Indexes;
loss load L L The loss load is the loss load of each load in a load set S in power failure after any equipment of the distribution network fails, the loss load is expressed by the annual average load of the load distribution and transformation, and a calculation formula is shown as the following formula
Wherein A is i Is the annual reading electric quantity T of the distribution transformer corresponding to the load i in the load set S A Is the annual average running time of distribution in the distribution network,
loss electric quantity E L Means the average repair time T of the device of the class to which the device belongs R During the period, the calculation formula of the electric quantity data corresponding to the continuous power failure of the load in the load set S is shown as the following formula
E L ＝L L T R
Hours h of power failure user POU Is shown as the following formula
Wherein, W LUi The number of users corresponding to the load i in the load set S;
weighted number of users W in power outage user level POU Is shown in the following formula
Wherein alpha is m Represents the mth one of the load nodes i in the load set SThe corresponding value of the user under a certain load level;
determining the weight of each evaluation index by adopting an Analytic Hierarchy Process (AHP), and finally determining the power failure influence degree of each equipment fault;
further preferably, the step 2 comprises the following steps:
step 21, selecting n groups of historical data d i Calculating comprehensive fault probability sample data corresponding to each equipment in the power distribution network based on the set equipment outage probability model;
step 22, setting a weight range and a weight distribution function for each comprehensive failure probability sample data, determining a corresponding weight based on Monte Carlo simulation, and normalizing each determined weight to obtain a weight of each year, wherein the corresponding formula isAnd calculating the corresponding simulation threshold value:
and step 23, determining the average value in the simulation threshold value as the required risk threshold value.
Further preferably, in step 3, the fast network reconfiguration based on the power grid operation mode of the multiple optimization method specifically includes the following steps:
step 31, determining the load importance degree in a power grid, measuring the load importance degree by using the power loss load value, namely the economic loss caused by power loss and the influence degree on the society, and defining the socioeconomic operation load seriously influenced by power failure as a non-power-loss load;
step 32, performing first optimization: if the serious fault occurs and all the loads can not be recovered, selecting the switches directly connected with the loads and numbering the switches; simultaneously, searching all the non-loss load switches according to an enumeration method, calculating all switch combination states meeting the generated energy, and performing descending arrangement on the calculated load importance degree total values;
step 33 performs a second optimization: sequentially carrying out load flow calculation according to the obtained total value of the importance degrees of the loads in descending order, judging whether constraint conditions are met, and preferentially recovering power supply to the loads which cannot be powered off; the constraint conditions comprise equality constraint conditions and inequality constraint conditions expressed by a nonlinear power flow equation;
the equation constraint is the formula
In the formula: p Gi 、Q Di The active output and the reactive output of the generator of the node i are respectively; q Gi 、Q Di Respectively an active load and a reactive load of a node i; theta.theta. ij Is the voltage phase angle difference between nodes i and j; g ij 、B ij Conductance and susceptance for branches i-j, respectively; the inequality constraints include control variable constraints and state variable constraints, i.e., the formula
In the formula: q i Is the reactive power of the generator i; p imin 、P imax Lower and upper limits, Q, respectively, of the active power output of the generator imin 、Q imax Lower limit and upper limit of reactive power output of the generator, T i 、T imax The power transmitted for branch i and its capacity limit, S, respectively D 、S L Respectively a load bus set and a branch set;
step 34, carrying out third optimization: locking the determined switches directly connected with the non-loss-of-supply load, and re-optimizing the rest switches; calculating all switch combination states meeting the generated energy in the optimization process, and calculating the total value of the load importance degree in descending order;
step 35, performing fourth optimization: and carrying out load flow calculation according to the obtained total value of the load importance degrees in descending order, judging whether constraint conditions are met, and determining the maximum value which can meet recovery conditions as an optimal solution when the constraint conditions are met.
Further preferably, the step 4 comprises:
step 41, collecting historical sample data, wherein the historical sample data selects data containing factors influencing the power load, namely the data at least comprises meteorological factor information, day type, load information and large user production plan data, and the characteristic vector of the sample on the ith day is set as:
Y i ＝[y i1 y i2 …y iM ]i＝1,2,…,n
in the formula: n is the total number of the historical samples; y is iM The Mth influence factor value of the ith sample;
step 42, constructing a gray correlation judgment matrix, taking the day vector to be predicted as a mother sequence of the matrix, and dividing each row by the data of the first row to initialize the gray correlation judgment matrix;
step 43, determining the weight of each influencing factor by adopting an entropy weight method, wherein the weight vector is W = [ W ] 1 w 2 …w m ]To obtain a weighted grey correlation decision matrix
Step 44, regarding the first row in the weighted gray correlation decision matrix F' as a row vector, and then marking the row 1 as a row vector to be predicted, which is marked as a 0 And the row vector of each other historical sample is marked as A i Each sample A i And A 0 The included angle between the vectors is the gray projection angle of the sample; the grey correlation projection value of each historical daily row vector and the daily row vector to be predicted is：
In the formula: d i Representing the projection value of the ith sample vector on the day vector to be predicted; i =1,2, …, n, w j Represents the weight corresponding to the jth influencing factor, j =1,2, …, m, F ij Representing the correlation degree of the jth influence factor in the ith sample and the mother sequence;
step 45, sorting the gray projection values of the historical day vectors from large to small, setting a projection value threshold, and sequentially selecting partial samples from large to small to form a similar day sample set;
step 46, taking the similar day sample sets as original training data sets, assuming that the total data amount is X, sampling from the original training data sets to generate k training sample sets, and assuming that the data in each training sample set is X, each sample set is all training data of each classification tree in the multi-classification regression tree algorithm; meanwhile, judging whether the total amount of data in the extracted k training sample sets does not exceed the total amount of data in the original training data set, if so, not processing the extracted training sample sets; otherwise, detecting whether all data can be extracted, if the data which is not extracted exists, executing a covering operation, namely randomly covering the extracted data by the data which is not extracted, wherein the covering quantity is (k.x)/X, and detecting whether all the data can be extracted again, otherwise, repeating the covering operation;
step 47, growing each training sample set into a classification tree without pruning leaves, randomly selecting M features from the M features at nodes of the tree, and selecting the optimal feature from the M features on each node according to the kini index for branch growth; detecting whether the features which are not selected and have the weight exceeding a threshold value exist or not, randomly covering the selected features with the unselected features at the moment, wherein the covering quantity is (M multiplied by k)/M, detecting whether all important features are selected or not, and repeating the covering operation if not, wherein the important features are the features which are determined by a user and correspond to certain influence factors on the predicted load in all the features;
the Gini index (Gini index) used in the present algorithm means that if the data set D contains m classes, the Gini index G is assumed to be D The calculation formula of (2) is as follows:
in the formula: p is a radical of j Is the frequency of occurrence of the class j element; while the Keyni index requires consideration of a binary partition of each attribute, assuming a binary partition of attribute A divides data set D into D 1 And D 2 Then, the kini index of the sample set D divided by some attribute a at the child node this time is represented as:
wherein, the first and the second end of the pipe are connected with each other,represented in data set D 1 The index of the above kini number,is represented in a data set D 2 The step above, which is actually to obtain a weighted average of the two kini indexes;
for each attribute, considering each possible binary partition, finally selecting a subset of the smallest kini index of the attribute Chen Sheng as its split subset; under the rule, continuously splitting from top to bottom until the growth of the whole decision tree is finished;
and 48, obtaining a prediction result according to the generated average value of all the tree prediction values.
Further preferably, the step 5: and determining whether to execute optimization measures according to the power grid operation mode determined in the step 3 and the comprehensive operation probability and comprehensive fault risk predicted in the step 4 at the next moment.
Further preferably, the following criteria are used for judging whether to execute the current optimization measure:
(51) When the given optimization measure does not shed load, only power output and network reconstruction need to be increased, and the current optimization measure is directly adopted;
(52) When the optimization measures need load shedding, whether the comprehensive operation risk and the comprehensive equipment fault probability at the next moment exceed the threshold value needs to be compared:
(521) When the predicted equipment outage probability exceeds a threshold value, executing load shedding operation;
(522) When the predicted equipment outage probability does not exceed the threshold value and when the real-time comprehensive fault risk f1< the comprehensive fault risk predicted at the next moment, performing load shedding and other operations;
(523) And when the predicted equipment outage probability does not exceed the threshold value and the real-time comprehensive fault risk f1> predicts the comprehensive operation risk at the next moment, the load shedding measure is not executed, so that the power output is increased and the network reconstruction is taken as a main measure to reduce the risk of the power grid.
Compared with the prior art, the invention has the following beneficial effects:
the invention fully considers the equipment failure probability and the operation risk of the power grid, not only considers the failures with low occurrence probability and serious consequences, but also fully considers the failures with high occurrence probability and non-serious consequences; meanwhile, the difference self-healing control is adopted, so that different objective functions are adopted under different operation conditions of the power grid, and the economic and reliable operation of the power grid is ensured under the condition of not shedding load as much as possible; therefore, the difference self-healing control method based on the equipment failure probability and the power grid operation risk has certain practical significance.
Drawings
FIG. 1 is a flow chart of the main steps of the method of the present invention;
FIG. 2 is a diagram of core essential steps corresponding to the steps of the method of the present invention;
FIG. 3 is a flowchart of a multi-class regression tree algorithm for highlighting attribute features according to the present invention;
figure 4 is a wiring diagram of a power distribution system according to the described example of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1 to 4, the differential self-healing control method based on the equipment failure probability and the grid operation risk according to the present invention is characterized by comprising the following steps:
step 1, collecting operation data of each power distribution network device in a power grid and weather data corresponding to an area where the power grid is located, and calculating a comprehensive fault probability corresponding to each power distribution network device based on a set device outage probability model; the method mainly comprises the steps of performing real-time operation risk assessment, wherein the method comprises the steps of collecting initial data of a power grid, determining all equipment sets influenced by the initial data according to actual weather data of a meteorological department, calculating fault probability according to an established equipment outage probability model, and further determining an expected fault probability set, wherein the expected fault set considers two dimensions of equipment risk and equipment outage probability, so that faults with low occurrence probability and serious consequences are considered, faults with high occurrence probability and serious consequences are also fully considered;
step 2, determining a risk threshold value required by evaluation of each power distribution network device; firstly, calculating the fault risk probability of each power device before the fault by using the model, setting each weight range and a weight distribution function, then determining a threshold value through Monte Carlo simulation, and taking the mean value of the simulation result as the most probable value of the threshold value, namely the risk threshold value to be selected;
step 3, judging whether the calculated comprehensive fault probability exceeds a risk threshold corresponding to the power distribution network equipment, and optimizing a power grid operation mode based on a first multi-objective function if the calculated comprehensive fault probability exceeds the risk threshold corresponding to the power distribution network equipment; otherwise, optimizing the operation mode of the power grid based on the second multi-objective function; meanwhile, the priority of the power grid operation mode is set, so that when the power grid is in a set sudden large fault, the power grid operation mode based on a multi-time optimization method can be preferentially adopted for rapid network reconstruction;
and 4, step 4: predicting the load size at the next moment by using a multi-classification regression tree algorithm with outstanding attribute characteristics, and then calculating the comprehensive outage probability and the comprehensive fault risk of the equipment at the next moment according to the established equipment outage probability model;
and 5: determining the execution of the optimization measures according to the optimization measures determined in the step 3 and the comprehensive operation probability and comprehensive fault risk at the next moment predicted in the step 4;
further preferably, the equipment outage probability model refers to distribution network equipment X at time t i The risk value of power failure caused by fault refers to the potential possibility and loss of fault, specifically to a certain time t, the distribution network equipment X i The model determined by multiplying the probability of occurrence of the fault by the power failure influence degree of the fault corresponds to the formula shown in the following formula
R t (X i )＝P(X i |t)L(X i )
Wherein, P (X) i I t) is power distribution network equipment X at time t i Probability of failure; l (X) i ) Is a distribution network device X i The degree of influence of the fault power failure.
Further preferably, the distribution network equipment X at time t i Probability of failure P (X) i T) is equal to the sum of the failure probabilities of the equipment corresponding to different risk sources, as shown in the following formula
Wherein，P k (X i I t) is that at time t, risk source k has device X i Probability of failure; m is the number of risk sources; for device X i If no risk source k exists, then P k (X i |t)＝0。
The occurrence of equipment failure is the result of the combined action of internal factors and external factors, the main internal factors comprise the factors of the equipment, overload of equipment operation and the like, the main external factors comprise weather factors such as strong wind, strong rain, lightning stroke and the like, and factors such as construction damage and the like, and the factors form a risk source of the equipment failure of the power distribution network; specifically, the risk sources include at least some or all of the following fault factors:
(1) The failure rate lambda of the equipment per se tends to exist, and the failure rate lambda (Y) of most elements is generally considered to be in a bathtub curve shape, wherein Y is the service life; and the change of the fault rate of each device on the feeder line has the following rule: the failure rate can be generally divided into 3 areas, namely an area I (descending area), an area II (invariable area) and an area III (ascending area); the equipment in the area I is in a running-in stage, and the fault rate of the equipment is in a continuously-reduced but higher than stable fault rate change trend; for the area III, the failure rate of the equipment is in a continuously rising change trend, which is mainly caused by long service life of the equipment and obvious aging problem of the equipment; based on the above rules, the operation statistics and equipment state evaluation data of various equipments in the power distribution network under different service lives, it can be determined that the fault rate corresponding to various equipment under different service lives, i.e. the corresponding relationship between the fault rate lambda (Y) and the service life Y, theoretically, the fault rate of various equipments changes with the service life to satisfy the bathtub curve rule, i.e. the fault rate is generally represented by Weibull distribution from the basic stability to the continuous rising trend,
the corresponding function formula is λ (Y), which is shown in the following formula
Wherein Y represents the service life, beta is more than or equal to 0 and less than or equal to 1;p represents the fixed failure rate, t 1 Indicates the age, t, corresponding to the area of equipment descent 2 Representing the corresponding age of the equipment belonging to the ascending region;
(2) The overload operation of the distribution equipment does not always have a fault, but the overload degree of the distribution equipment has a positive relation with the equipment fault, namely the larger the overload degree is, the more easily the distribution equipment has the fault; when the current of the distribution equipment is less than or equal to a certain proportion a of the rated current (a can be freely set by a user according to the purpose of system evaluation), the probability of causing equipment failure is 0; but as the current flowing through the device increases, the probability of device failure increases, and the rate of increase will also be faster; therefore, the fault rate P of the equipment overload or the probability of equipment fault caused by overload can be introduced into the overload value L of the equipment on the basis of the fault rate lambda of the distribution equipment on the basis of the utility function OD The corresponding function formula is shown as the following formula
Wherein, the first and the second end of the pipe are connected with each other,l is the proportion of the current flowing through the distribution equipment to the rated current thereof, and a is the rated current proportion threshold value freely set by a user;
(3) The equipment failure rate caused by external factors is that the equipment of the power distribution network is damaged by strong wind and strong rain, lightning and construction factors, so that a certain failure and power failure probability is brought, so that the external factors are considered, and the external factors at least specifically comprise the strong wind and strong rain, the lightning factor and the construction failure factor;
the power failure probability P of equipment failure corresponding to the factors of strong wind and strong rain WR The corresponding function formula is shown as the following formula
P WR ＝N GWR /(N D N WR )
Wherein N is GWR The number of power failure times of the equipment caused by the factors of strong wind and strong rain,N D is the total number of certain types of equipment in the power distribution network, N WR The times of strong wind and heavy rain of the power distribution network in the set statistical period are set;
the power failure probability P of equipment failure corresponding to the lightning strike factor TH The corresponding function formula is shown as the following formula
P TH ＝N GTH /(N D N TH )
Wherein, N GTH Number of power failures in such equipment due to lightning strike, N D Is the total number of certain types of equipment in the power distribution network, N TH The number of lightning strikes of the power distribution network in the set statistical period is set;
the power failure probability P of equipment failure corresponding to the lightning strike factor S The corresponding function formula is shown in the following formula
P S ＝N GS /N S
Wherein N is GS Number of times of failure of such equipment due to nearby construction, N S And finding the total construction times near certain equipment of the power distribution network in the set statistical period.
In summary, the failure probability of a certain type of equipment in the power distribution network should comprehensively consider the self-factor, the overload condition and the external factor of the equipment, and the sum of the failure probabilities of various risk sources of the equipment is taken as the failure rate of the equipment.
Further preferably, since the degree of influence of the power failure due to the failure of each device in the power distribution network is related to the consequences of the failure, it is necessary to determine an expected set of failures at the device level, that is, when a device on the feeder line fails and has a power failure, each load in the corresponding load set S has a continuous power failure. Analyzing the expected failure power failure consequences of each device on the feeder line according to different feeder lines in an expected accident set by using a failure traversal method, and constructing a failure power failure load set S corresponding to the failure of each device on different feeder lines after comprehensively considering the transfer capacity of the connecting line; specifically, the degree of influence L (X) on the fault power failure i ) The consequences of equipment failure are mainly for the power supply enterprise, load or power consumer, and are strongly related to power supply reliability. Different equipment faults are caused by different power failure rangesThere is also a difference in the magnitude of the effect of (c). The factors of lost load, lost electric quantity, power failure user number, user important level and power failure duration are main factors reflecting the influence degree of the failure and the power failure. However, as part of the loads are cut off to supply power, the load flow of the feeder line is reduced, even the length of the operating feeder line is shortened, and in addition, automatic switching capacitor banks meeting the requirements are installed in a centralized and dispersed mode in the power distribution network for reactive compensation, the possibility that the loads continuously supplying power have low voltage is low and is not considered. Therefore, the following indexes reflecting the influence degree of the equipment failure and power failure are constructed to include the loss load L L Loss of electric power E L Hours of power failure user h POU Weighted number of users W in power failure user level POU Indexes;
loss load L L The loss load of each load power failure in a failure power failure load set S after any equipment of the power distribution network fails is expressed by the average load of the load distribution year, and a calculation formula is shown as the following formula
Wherein, A i Is the annual reading electric quantity T of the distribution transformer corresponding to the load i in the load set S A Is the annual average running time of distribution in the distribution network,
loss of electric quantity E L Means the average repair time T of the device of the class to which the device belongs R During the period, the calculation formula of the electric quantity data corresponding to the continuous power failure of the load in the load set S is shown as the following formula
E L ＝L L T R
Hours h of power failure user POU Is shown as the following formula
Wherein, W LUi The number of users corresponding to the load i in the load set S;
according to load classification, the primary load has great influence on power failure, the secondary load has the lowest secondary load and the tertiary load is the smallest, and the power failure user level weights the number of households W POU Is shown as the following formula
Wherein alpha is m Representing the corresponding value of the mth user in the load node i in the load set S under a certain load level, and if the load level is three levels, respectively taking K for the first-level load U1 Second order load of K U2 Taking 1 as the third-level load, and weighting according to the expert evaluation result; different indexes reflect the degree of influence of the power failure caused by equipment failure from different sides, so the level of the influence needs to be comprehensively reflected by a comprehensive evaluation method. The contribution of the 4 evaluation indexes to the equipment fault power failure influence degree is different, the weight of each evaluation index is determined by an Analytic Hierarchy Process (AHP), and finally, the equipment fault power failure influence degree is determined.
Further preferably, the step 2 comprises the following steps:
step 21, selecting n groups of historical data d i Calculating comprehensive fault probability sample data corresponding to each equipment in the power distribution network based on the set equipment outage probability model;
step 22, setting a weight range and a weight distribution function for each comprehensive failure probability sample data, determining a corresponding weight based on Monte Carlo simulation, and normalizing each determined weight to obtain a weight of each year, wherein the corresponding formula isAnd calculating the corresponding simulation threshold value:
the weight selection range is selected according to the principle that the longer the current time is, the smaller the weight range is;
and step 23, determining the mean value in the simulation result as the required risk threshold value by taking the minimum value of the simulation result as the lower limit of the threshold value and the maximum value of the simulation result as the upper limit of the threshold value.
Further preferably, the step 3: judging whether the calculated comprehensive fault probability exceeds a risk threshold value corresponding to the power distribution network equipment, and optimizing a power grid operation mode on the basis of a first multi-objective function A (namely, eliminating faults by taking the comprehensive fault risk of the power equipment and other equipment exceeding the threshold value as a target) if the calculated comprehensive fault probability exceeds the risk threshold value corresponding to the power distribution network equipment; otherwise, optimizing the operation mode of the power grid based on a second multi-objective function B (namely, removing the fault by taking the comprehensive fault risk as a target); meanwhile, the priority of the power grid operation mode is set, so that when the power grid is in the set sudden large fault, the power grid operation mode based on a multi-time optimization method can be preferentially adopted for rapid network reconstruction, and rapid recovery of the power grid is realized; furthermore, the first multi-objective function is to perform descending order arrangement on the obtained comprehensive fault risk, namely the product of the outage probability of the equipment and the corresponding outage consequence, when the evaluated equipment fault probabilities are not greater than the risk threshold, so as to reduce the fault risk and the economic cost, and determine an objective function set for the multi-objective function:
minY＝[R 1 ,F C ]
in the formula: y is an objective function vector; r 1 For the risk of complex faults of the electrical apparatus, F C For the economic cost of the power grid, a quadratic function can be generally adopted for representation;
in the formula: s G Is a generator set; p i Active power output of the generator i; alpha is alpha i 、β i 、γ i Is the fuel cost coefficient of generator i; the second multi-objective function B is used for reducing the risk probability of the evaluated equipment fault when the risk probability is larger than the risk threshold valueDetermining an objective function set for the multi-objective function according to the low risk probability of the equipment, the comprehensive fault risk except the equipment and the economic cost:
minY＝[R 1 ,R 2 F C ]
in the formula: y is an objective function vector; r 1 For comprehensive risks of power plants, R 2 Probability of failure of power plant, F C Economic cost for the power grid;
further preferably, in step 3, the fast network reconfiguration based on the power grid operation mode of the multiple optimization method specifically includes the following steps:
step 31, determining the load importance degree in a power grid, measuring the load importance degree by using the power loss load value, namely the economic loss caused by power loss and the influence degree on the society, and defining the socioeconomic operation load seriously influenced by power failure as a non-power-loss load;
step 32, performing first optimization: if the serious fault occurs and all the loads cannot be recovered, selecting a switch directly connected with the load, and numbering the switches; simultaneously, searching all the non-loss load switches according to an enumeration method, calculating all switch combination states meeting the generated energy, and performing descending arrangement on the calculated load importance degree total values;
step 33 performs a second optimization: sequentially carrying out load flow calculation according to the obtained total value of the importance degrees of the loads in descending order, judging whether constraint conditions are met, and preferentially recovering power supply to the loads which cannot be powered off; the constraint conditions comprise equality constraint and inequality constraint, wherein the equality constraint represented by the nonlinear power flow equation is adopted
In the formula: p is Gi 、Q Di The active output and the reactive output of the generator of the node i are respectively; q Gi 、Q Di Respectively an active load and a reactive load of a node i; theta ij Is the phase angle of the voltage between nodes i and jA difference; g ij 、B ij Conductance and susceptance for branches i-j, respectively; the inequality constraints include control variable constraints and state variable constraints, which may be expressed as
In the formula: q i Is the reactive power of the generator i; p imin 、P imax The lower limit and the upper limit of the active output of the generator are respectively; q imin 、Q imax Respectively lower limit and upper limit of reactive power output of the generator; t is i 、T imax The power transmitted by the branch i and the capacity limit thereof are respectively; s D 、S L Respectively a load bus set and a branch set;
step 34, carrying out third optimization: locking the determined switches directly connected with the non-loss-of-supply load, and re-optimizing the rest switches; calculating all switch combination states meeting the generated energy in the optimization process, and calculating the total value of the load importance degree in descending order;
step 35, performing fourth optimization: and carrying out load flow calculation according to the obtained total value of the load importance degrees in descending order, judging whether constraint conditions are met, and determining the maximum value which can meet recovery conditions as an optimal solution when the constraint conditions are met.
Further preferably, the step 4 comprises:
step 41, collecting historical sample data, wherein the historical sample data selects data containing factors influencing the power load, namely the data at least comprises meteorological factor information, day type, load information, a large user production plan and the like, the meteorological factor information, the day type, the load information and the large user production plan data are real-time data, the load information selects data collected at the last moment, and the characteristic vector of the sample at the ith day is set as:
Y i ＝[y i1 y i2 …y iM ]i＝1,2,…,n
in the formula: n is the total number of the historical samples; y is iM The Mth influence factor value of the ith sample;
step 42, constructing a gray correlation judgment matrix, taking the day vector to be predicted as a mother sequence of the matrix, and dividing each row by the data of the first row to initialize the gray correlation judgment matrix;
step 43, determining the weight of each influence factor by adopting an entropy weight method, wherein the weight vector is W = [ W = 1 w 2 …w m ]To obtain a weighted grey correlation decision matrix
Step 44, regarding the first row in the weighted gray correlation decision matrix F' as a row vector, and then the row vector of the row 1 to be predicted is marked as A 0 And the row vector of each other historical sample is marked as A i Each sample A i And A 0 The included angle between the vectors is the gray projection angle of the sample; the gray associated projection value of each historical daily vector and the daily vector to be predicted is as follows:
in the formula: d i Representing the projection value of the ith sample vector on the day vector to be predicted; i =1,2, …, n, w j Represents the weight corresponding to the jth influencing factor, j =1,2, …, m, F ij Representing the correlation degree of the jth influence factor in the ith sample and the mother sequence;
step 45, sorting the gray projection values of the historical day vectors from large to small, setting a projection value threshold, and sequentially selecting partial samples from large to small to form a similar day sample set;
step 46, taking the similar day sample sets as original training data sets, assuming that the total data amount is X, sampling from the original training data sets to generate k training sample sets, and assuming that the data in each training sample set is X, each sample set is all training data of each classification tree in the multi-classification regression tree algorithm; meanwhile, judging whether the total amount of data in the extracted k training sample sets does not exceed the total amount of data in the original training data set, if so, not processing the extracted training sample sets; otherwise, detecting whether all data can be extracted, if the data which is not extracted exists, executing a covering operation, namely, randomly covering the extracted data by the data which is not extracted, wherein the covering quantity is (k.x)/X, and detecting whether all the data can be extracted again, otherwise, repeating the covering operation;
step 47, growing each training sample set into a classification tree without pruning leaves, randomly selecting M features from the M features at nodes of the tree, and selecting the optimal feature from the M features on each node according to the kini index for branch growth; simultaneously detecting whether the features which are not selected and have the weight exceeding a threshold value exist, randomly covering the unselected features with the selected features, wherein the covering quantity is (M multiplied by k)/M, detecting whether all important features set by a user are selected, and otherwise, repeating the covering operation;
the Gini index (Gini index) used in the present algorithm means that if the data set D contains m classes, the Gini index G is assumed to be D The calculation formula of (c) is:
in the formula: p is a radical of j Is the frequency of occurrence of the class j element; while the Gini index requires consideration of the binary partition of each attribute, assuming the binary partition of attribute A divides dataset D into D 1 And D 2 Then, the kini index of the sample set D divided by some attribute a at the child node this time is:
for each attribute, considering each possible binary partition, finally selecting a subset of the smallest kini index of the attribute Chen Sheng as its split subset; under the rule, continuously splitting from top to bottom until the growth of the whole decision tree is finished;
and 48, obtaining a prediction result according to the generated average value of all the tree prediction values.
Step 5 mainly corresponds to the execution situation of the optimization measures; the self-healing control method aims at ensuring the reliability of a power grid and the continuous power supply of loads, when the obtained optimization measures need load shedding to reduce the risk, the risk of the power grid at the next stage needs to be predicted to judge whether to cut the loads, and the specific conditions are as follows:
the following criteria exist for judging whether to execute the current optimization measures:
(51) When the given optimization measure does not shed load, only power output and network reconfiguration need to be increased, and the current optimization measure is directly adopted;
(52) When the optimization measure needs load shedding, whether the comprehensive operation risk and the comprehensive equipment fault probability at the next moment exceed the threshold value needs to be compared:
(521) When the predicted equipment outage probability exceeds a threshold value, executing load shedding operation;
(522) When the predicted equipment outage probability does not exceed the threshold value and when the real-time comprehensive fault risk f1 is less than the comprehensive fault risk predicted at the next moment, performing load shedding and other operations;
(523) And when the predicted equipment outage probability does not exceed the threshold value and the real-time comprehensive fault risk f1> predicts the comprehensive operation risk at the next moment, the load shedding measure is not executed, so that the power output is increased and the network reconfiguration is taken as a main measure to reduce the risk of the power grid.
The detailed implementation is described below, as shown in fig. 4, wherein the node 7 is a distributed power supply, the capacity of each node is 200KW, and each node is equipped with an energy storage power supply with a capacity of 100kw.h. The rated voltage of the system is 12.66KV, wherein the available capacities of the interconnection switches 6-14, 11-20 and 17-19 are 200KW, 150KW and 200KW respectively.
The fault probability and the power failure influence degree of each device are calculated, the beta value of the bathtub curve of each device is obtained through historical fault data fitting, and the fault rate and the average repair time of each device in the power distribution network are shown in table 1.
TABLE 1 failure Rate and mean repair time for various types of devices
The distribution network weather condition is in thunderstorm weather, strong wind and strong rain and accumulation, the line 5-6 runs in overload, and the fault rate of each device is calculated according to the historical record and the formulas (2) - (8):
according to the historical data and the formulas (9) - (12), the influence degree of the fault power failure is obtained, as shown in the table 2
Name of the device Influence value Name of the device Influence value Name of the device Influence value
Circuit breaker B1 0.82 Line 0-1 0.49 Line 1-2 0.47
Transformer T 0.79 Lines 2-3 0.35 Lines 3-4 0.29
Lines 4-5 0.29 Lines 5-6 0.29 Lines 6-7 0.26
Lines 7-8 0.22 Lines 8-9 0.21 Lines 9-10 0.21
Lines 10-11 0.22 Lines 2-12 0.29 Lines 12-13 0.27
Lines 13-14 0.26 Lines 3-15 0.29 Lines 15-16 0.25
Lines 16-17 0.21 Lines 6-14 Interconnection switch Lines 17-19 Interconnection switch
Lines 11-20 Interconnection switch
TABLE 2 Fault Power failure influence degree
Calculating the power failure risk value of each equipment fault according to the formula (1) shown in table 3:
name of the device Value of risk Name of the device Value of risk Name of the device Value of risk
Circuit breaker B1 0.082 Line 0-1 0.102 Line 1-2 0.094
Transformer T 0.158 Lines 2-3 0.091 Lines 3-4 0.069
Lines 4-5 0.067 Lines 5-6 0.121 Lines 6-7 0.049
Lines 7-8 0.048 Lines 8-9 0.058 Lines 9-10 0.044
Lines 10-11 0.055 Lines 2-12 0.060 Lines 12-13 0.062
Lines 13-14 0.067 Lines 3-15 0.063 Lines 15-16 0.062
Lines 16-17 0.044 Lines 6-14 Interconnection switch Lines 17-19 Interconnection switch
Lines 11-20 Interconnection switch
TABLE 3 Power failure risk value of each equipment failure
Determining the future threshold of each device by using historical data, taking the mean value of Monte Carlo simulation results as the threshold, and selecting the threshold of each device as shown in Table 4:
name of the device Threshold value Name of the device Threshold value Name of the device Threshold value
Breaker B1 0.10 Overhead line 0.31 Transformer T 0.11
Interconnection switch
TABLE 4 failure probability threshold for devices
Determining an optimization function according to the obtained equipment fault power failure risk values and equipment fault rates, wherein the equipment fault rate of the lines 5-6 exceeds a threshold value of 0.3, so that the reduced lines 5-6 are taken as a risk objective function R 1 Then, determining other objective functions according to the fault power failure risk values, wherein the risk values are 0-1 and 1-2 of the line, and the transformer T is used as an objective function R 2 、R 3 、R 4 And economic cost F of the power grid C 。
Objective function minY = [ R 1 ,R 2 ，R 3 ，R 4 ，F C ]
And predicting the load size at the next moment by using a multi-classification regression tree algorithm with outstanding attribute characteristics, and then calculating the comprehensive outage probability and the comprehensive fault risk of the equipment at the next moment according to the established equipment outage probability model. Part of the raw data set is shown in table 5:
properties Value of
Type of day Sunday
Maximum air temperature 23.0
Minimum air temperature 8.0
Amount of rainfall 45mm
Whether there is lightning strike Is provided with
Wind speed 9.8m/s
Load of the same period of the last week 1529.25w
Load in the same period of the previous month 1492.26w
Load(s) 1529.87w
Table 5 partial raw data set data
Predicting the magnitude of the load, calculating the fault probability and the risk value at the next moment according to the fault probability and the risk value calculation method of each device, and predicting the fault rate of each device at the next moment as shown in table 6:
TABLE 6 predicted device failure rates
Predicting the risk at the next moment as:
name of the device Value of risk Name of the device Value of risk Name of the device Value of risk
Circuit breaker B1 0.082 Line 0-1 0.102 Line 1-2 0.096
Transformer T 0.158 Lines 2-3 0.092 Lines 3-4 0.069
Lines 4-5 0.067 Lines 5-6 0.119 Lines 6-7 0.049
Lines 7-8 0.048 Lines 8-9 0.058 Lines 9-10 0.044
Lines 10-11 0.055 Lines 2-12 0.061 Lines 12-13 0.062
Lines 13-14 0.067 Lines 3-15 0.063 Lines 15-16 0.062
Lines 16-17 0.044 Lines 6-14 Interconnection switch Lines 17-19 Interconnection switch
Lines 11-20 Interconnection switch
TABLE 7 prediction of individual equipment risk values
And finally, executing the optimized recovery operation according to the following principles:
the following criteria exist for judging whether to execute the current optimization measures:
1) When the given optimization measure does not shed load, only power output and network reconstruction need to be increased, and the current optimization measure is directly and directly adopted
2) When the optimization measures need load shedding, whether the comprehensive operation risk and the equipment outage probability at the next moment exceed the threshold value needs to be compared:
(1) When the predicted equipment outage probability has a threshold value, executing load shedding and other operations;
(2) When the predicted equipment outage probability does not exceed the threshold value and when the real-time comprehensive fault risk f1 is less than the comprehensive fault risk predicted at the next moment, performing load shedding and other operations;
(3) When the predicted equipment outage probability does not exceed the threshold value and the real-time comprehensive fault risk f1> predicts the comprehensive operation risk at the next moment, the load shedding measure is not executed, so that the power output is increased and the network reconstruction is taken as a main measure to reduce the risk of the power grid;
according to the criteria and the obtained real-time equipment fault rate and risk value, the obtained optimization scheme is as follows:
the above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art should be considered to be within the technical scope of the present invention, and the technical solutions and the inventive concepts thereof according to the present invention should be equivalent or changed within the scope of the present invention.

Claims (7)

1. A difference self-healing control method based on equipment failure probability and power grid operation risk is characterized by comprising the following steps:
step 1, collecting operation data of each power distribution network device in a power grid and weather data corresponding to an area where the power grid is located, and calculating a comprehensive fault probability corresponding to each power distribution network device based on a set device outage probability model;
step 2, determining a risk threshold value required by evaluation of each power distribution network device;
step 3, judging whether the calculated comprehensive fault probability exceeds a risk threshold value corresponding to the power distribution network equipment, and optimizing a power distribution network operation mode based on a first multi-objective function if the calculated comprehensive fault probability exceeds the risk threshold value; otherwise, optimizing the operation mode of the power grid based on the second multi-objective function; meanwhile, the priority of the power grid operation mode is set, so that when the power grid is in a set sudden large fault, the power grid operation mode based on a multi-time optimization method can be preferentially adopted for rapid network reconstruction;
step 4, predicting the load size of the power grid at the next moment based on a multi-classification regression tree algorithm, and calculating the comprehensive outage probability and the comprehensive fault risk of equipment at the next moment, wherein the comprehensive fault risk specifically refers to the product of the comprehensive outage probability and the fault consequence;
and 5: and (4) determining whether to execute optimization measures according to the power grid operation mode determined in the step (3) and the comprehensive operation probability and comprehensive fault risk at the next moment predicted in the step (4).
2. The method of claim 1, wherein:
the equipment outage probability model refers to power distribution network equipment X at the moment t i The risk value of power failure in case of failure specifically refers to a certain time t, and the power distribution network equipment X i The model determined by multiplying the probability of occurrence of the fault by the power failure influence degree of the fault corresponds to the formula shown in the following formula
R t (X i )＝P(X i |t)L(X i )
Wherein, P (X) i I t) is power distribution network equipment X at time t i Probability of failure; l (X) i ) Is a distribution network device X i The degree of influence of the fault power failure.
3. The method of claim 2, wherein: distribution network equipment X at time t i Probability of failure P (X) i T) is equal to the sum of the failure probabilities of the equipment corresponding to different risk sources, as shown in the following formula
Wherein, P k (X i I t) is that at time t, risk source k has device X i Probability of failure; m is the number of risk sources;
the risk sources include at least some or all of the following fault factors:
(1) The failure rate λ of the distribution equipment itself is represented by the following formula, where the corresponding function formula is λ (Y)
Wherein Y represents the service life, beta is more than or equal to 0 and less than or equal to 1; p represents the fixed failure rate, t 1 Indicates the age, t, corresponding to the area of equipment descent 2 Representing the corresponding age of the equipment belonging to the ascending area;
(2) Over-loading of equipmentThe load fault rate P, P is the probability of equipment fault caused by overload, and specifically is based on utility function, and introduces the overload value L of the equipment on the basis of the fault rate lambda of the distribution equipment OD The corresponding function formula is shown as the following formula
Wherein the content of the first and second substances,l is the proportion of the current flowing through the distribution equipment to the rated current thereof, and a is the rated current proportion threshold value freely set by a user;
(3) Equipment failure rate caused by external factors, wherein the external factors at least comprise a strong wind and strong rain factor, a lightning stroke factor and a construction damage factor;
the power failure probability P of equipment failure corresponding to the factors of strong wind and strong rain WR The corresponding function formula is shown as the following formula
P WR ＝N GWR /(N D N WR )
Wherein N is GWR The number of power failure times of the equipment caused by strong wind and strong rain, N D Is the total number of certain types of equipment in the power distribution network, N WR The number of times of strong wind and strong rain of the power distribution network in the set statistical period is set;
the power failure probability P of equipment failure corresponding to the lightning strike factor TH The corresponding function formula is shown as the following formula
P TH ＝N GTH /(N D N TH )
Wherein N is GTH Number of power failures in such installations due to lightning strike factors, N D Is the total number of certain types of equipment in the power distribution network, N TH The number of lightning strikes of the power distribution network in the set statistical period is set;
the power failure probability P of the equipment fault corresponding to the lightning strike factor S The corresponding function formula is shown as the following formula
P S ＝N GS /N S
Wherein N is GS Number of times of failure of such equipment due to nearby construction, N S And finding the total construction times near certain equipment of the power distribution network in the set statistical period.
4. The method of claim 2, wherein: the fault power failure influence degree L (X) i ) The index of the degree of influence of power failure caused by corresponding equipment failure comprises loss load L L Loss of electric power E L Hours of power failure user h POU Weighted number of users W in power failure user level POU Indexes;
loss load L L The loss load is the loss load of each load in a load set S in power failure after any equipment of the distribution network fails, the loss load is expressed by the annual average load of the load distribution and transformation, and a calculation formula is shown as the following formula
Wherein A is i Is the annual reading electric quantity T of the distribution transformer corresponding to the load i in the load set S A Is the annual average running time of distribution in the distribution network,
loss of electric quantity E L Means the average repair time T of the device of the class to which the device belongs R During the period, the calculation formula of the electric quantity data corresponding to the continuous power failure of the load in the load set S is shown as the following formula
E L ＝L L T R
Hours h of power failure user POU Is shown as the following formula
Wherein, W LUi The number of users corresponding to the load i in the load set S;
weighted number of users W in power failure user level POU Is shown as the following formula
Wherein alpha is m Representing a value corresponding to the mth user in a load node i in the load set S under a certain load level;
and determining the weight of each evaluation index by adopting an Analytic Hierarchy Process (AHP), and finally determining the power failure influence degree of each equipment fault.
5. The method of claim 1, wherein: the step 2 comprises the following steps:
the step 2 comprises the following steps:
step 21, selecting n groups of historical data d i Calculating comprehensive fault probability sample data corresponding to each equipment in the power distribution network based on the set equipment outage probability model;
step 22, respectively setting a weight range and a weight distribution function for each comprehensive fault probability sample data, determining corresponding weights based on Monte Carlo simulation, and normalizing each determined weight to obtain the weight of each year, wherein the corresponding formula isAnd calculating the corresponding simulation threshold value:
and step 23, determining the average value in the simulation threshold value as the required risk threshold value.
6. The method of claim 1, wherein: the step 3 of performing rapid network reconfiguration based on the power grid operation mode of the multi-time optimization method specifically comprises the following steps:
step 31, determining the load importance degree in a power grid, measuring the load importance degree by using the power loss load value, namely the economic loss caused by power loss and the influence degree on the society, and defining the social economic operation load seriously influenced by power failure as an uninterruptible load;
step 32, carrying out first optimization: if the serious fault occurs and all the loads can not be recovered, selecting the switches directly connected with the loads and numbering the switches; simultaneously, searching all the non-loss load switches according to an enumeration method, calculating all switch combination states meeting the generated energy, and performing descending arrangement on the calculated load importance degree total values;
step 33 performs a second optimization: sequentially carrying out load flow calculation according to the obtained total value of the importance degrees of the loads in descending order, judging whether constraint conditions are met, and preferentially recovering power supply to the loads which cannot be powered off; the constraint conditions comprise equality constraint conditions and inequality constraint conditions expressed by a nonlinear power flow equation;
the equation constraint, i.e. the formula
In the formula: p Gi 、Q Di The active output and the reactive output of the generator of the node i are respectively; q Gi 、Q Di Respectively an active load and a reactive load of a node i; theta.theta. ij Is the voltage phase angle difference between nodes i and j; g ij 、B ij Conductance and susceptance for branches i-j, respectively; the inequality constraints include control variable constraints and state variable constraints, i.e., the formula
In the formula: q i Is the reactive power of the generator i; p imin 、P imax Lower and upper limits, Q, of the active power output of the generator, respectively imin 、Q imax Lower limit and upper limit of reactive power output of the generator, T i 、T imax Are respectively a branchi power transmitted and its capacity limit, S D 、S L Respectively a load bus set and a branch set;
step 34, carrying out third optimization: locking the determined switches directly connected with the non-loss-of-supply load, and re-optimizing the rest switches; calculating all switch combination states meeting the generated energy in the optimization process, and calculating the total value of the load importance degree in descending order;
step 35, performing fourth optimization: and carrying out load flow calculation according to the obtained total value of the load importance degrees in descending order, judging whether constraint conditions are met, and determining the maximum value which can meet recovery conditions as an optimal solution when the constraint conditions are met.
7. The method of claim 1, wherein: the step 4 comprises the following steps:
step 41, collecting historical sample data, wherein the historical sample data selects data containing factors influencing the power load, namely the data at least comprises meteorological factor information, day type, load information and large user production plan data, and the characteristic vector of the sample on the ith day is set as:
Y i ＝[y i1 y i2 …y iM ]i＝1,2,…,n
in the formula: n is the total number of the historical samples; y is iM The Mth influence factor value of the ith sample;
step 42, constructing a gray correlation judgment matrix, taking the day vector to be predicted as a mother sequence of the matrix, and dividing each row by the data of the first row to initialize the gray correlation judgment matrix;
step 43, determining the weight of each influencing factor by adopting an entropy weight method, wherein the weight vector is W = [ W ] 1 w 2 …w m ]To obtain a weighted grey correlation decision matrix
Step 44, regarding the first row in the weighted gray correlation decision matrix F' as a row vector, and then the row vector of the row 1 to be predicted is marked as A 0 And the row vector of each other historical sample is marked as A i Each sample A i And A 0 The included angle between the vectors is the gray projection angle of the sample; the gray associated projection value of each historical daily vector and the daily vector to be predicted is as follows:
in the formula: d i Representing the projection value of the ith sample vector on the day vector to be predicted; i =1,2, …, n, w j Represents the weight corresponding to the jth influencing factor, j =1,2, …, m, F ij Representing the correlation degree of the jth influence factor in the ith sample and the mother sequence;
step 45, sorting the gray projection values of the historical day vectors from large to small, setting a projection value threshold, and sequentially selecting partial samples from large to small to form a similar day sample set;
step 46, taking the similar day sample sets as original training data sets, assuming that the total data amount is X, sampling from the original training data sets to generate k training sample sets, and assuming that the data in each training sample set is X, each sample set is all training data of each classification tree in the multi-classification regression tree algorithm; meanwhile, judging whether the total amount of data in the extracted k training sample sets does not exceed the total amount of data in the original training data set, if so, not processing the extracted training sample sets; otherwise, detecting whether all data can be extracted, if the data which is not extracted exists, executing a covering operation, namely, randomly covering the extracted data by the data which is not extracted, wherein the covering quantity is (k.x)/X, and detecting whether all the data can be extracted again, otherwise, repeating the covering operation;
step 47, growing each training sample set into a classification tree without pruning leaves, randomly selecting M features from the M features at nodes of the tree, and selecting the optimal feature from the M features on each node according to the kini index for branch growth; detecting whether the features which are not selected and have the weight exceeding a threshold value exist or not, randomly covering the selected features with the unselected features at the moment, wherein the covering quantity is (M multiplied by k)/M, detecting whether all important features are selected or not, and repeating the covering operation if not, wherein the important features are the features which are determined by a user and correspond to certain influence factors on the predicted load in all the features;
the Gini index used in the algorithm is the Gini index G assuming that the data set D contains m classes D The calculation formula of (2) is as follows:
in the formula: p is a radical of formula j Is the frequency of occurrence of the class j element; while the Gini index requires consideration of the binary partition of each attribute, assuming the binary partition of attribute A divides dataset D into D 1 And D 2 Then, the kini index of the sample set D divided by some attribute a at the child node this time is represented as:
wherein the content of the first and second substances,is represented in a data set D 1 The index of the above kini number,is represented in a data set D 2 The kini index above;
for each attribute, considering each possible binary partition, finally selecting a subset of the smallest kini index of the attribute Chen Sheng as its split subset; under the rule, continuously splitting from top to bottom until the growth of the whole decision tree is finished;
and 48, obtaining a prediction result according to the generated average value of all the tree prediction values.
CN201710765450.3A 2017-08-30 2017-08-30 Difference self-healing control method based on equipment failure probability and power grid operation risk Active CN107516170B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710765450.3A CN107516170B (en) 2017-08-30 2017-08-30 Difference self-healing control method based on equipment failure probability and power grid operation risk

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710765450.3A CN107516170B (en) 2017-08-30 2017-08-30 Difference self-healing control method based on equipment failure probability and power grid operation risk

Publications (2)

Publication Number Publication Date
CN107516170A true CN107516170A (en) 2017-12-26
CN107516170B CN107516170B (en) 2020-07-28

Family

ID=60724692

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710765450.3A Active CN107516170B (en) 2017-08-30 2017-08-30 Difference self-healing control method based on equipment failure probability and power grid operation risk

Country Status (1)

Country Link
CN (1) CN107516170B (en)

Cited By (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108375715A (en) * 2018-03-08 2018-08-07 中国电力科学研究院有限公司 A kind of distribution network line fault risk day prediction technique and system
CN108416469A (en) * 2018-02-10 2018-08-17 杭州后博科技有限公司 A kind of extreme weather causes the prediction technique and system of steel tower security risk
CN108695846A (en) * 2018-05-29 2018-10-23 国电南瑞科技股份有限公司 A kind of unit style power distribution network operation risk assessment method
CN109145364A (en) * 2018-07-04 2019-01-04 西北工业大学 Sensitivity Analysis Method based on Geordie inequality
CN109190996A (en) * 2018-09-18 2019-01-11 中国电力科学研究院有限公司 A kind of distribution net platform region operation risk monitoring and pre-alarming method and device
CN109376877A (en) * 2018-10-11 2019-02-22 华自科技股份有限公司 Equipment O&M method for early warning, device, computer equipment and storage medium
CN109460004A (en) * 2018-10-26 2019-03-12 国网天津市电力公司 Distribution network failure prediction technique and system based on big data
CN109521672A (en) * 2018-10-22 2019-03-26 东北大学 A kind of intelligent selecting method of electric arc furnaces power supply curve
CN109784502A (en) * 2018-12-11 2019-05-21 深圳供电局有限公司 On-site maintenance processing method and system based on fault equipment
CN110071500A (en) * 2019-04-09 2019-07-30 国网山东省电力公司济南供电公司 Distribution line restores Sequence Decision method and system after a kind of bus-bar fault power loss
CN110210632A (en) * 2019-05-31 2019-09-06 国网河北省电力有限公司沧州供电分公司 Failure emergency processing method, device and terminal based on ubiquitous electric power Internet of Things
CN110705801A (en) * 2019-10-10 2020-01-17 国网山东省电力公司泰安供电公司 Power grid accident economic loss estimation method and system based on fault loss electric quantity
CN110739691A (en) * 2019-11-06 2020-01-31 南方电网科学研究院有限责任公司 power grid branch fault prediction method and device
CN111044850A (en) * 2020-01-14 2020-04-21 广东电网有限责任公司 Power grid fault diagnosis method based on equipment fault probability
CN111160603A (en) * 2019-11-21 2020-05-15 中国电力科学研究院有限公司 Method and system for guaranteeing reliability of end-to-end power communication service
CN111488896A (en) * 2019-01-28 2020-08-04 国网能源研究院有限公司 Distribution line time-varying fault probability calculation method based on multi-source data mining
CN112116276A (en) * 2020-09-29 2020-12-22 广东电网有限责任公司广州供电局 Transformer substation operation risk assessment method considering time-varying state of electrical main equipment
CN112668157A (en) * 2020-12-14 2021-04-16 河北电力交易中心有限公司 Node power failure risk invariant node reserve capacity rapid assessment method
CN113052473A (en) * 2021-03-31 2021-06-29 贵州电网有限责任公司 Power grid risk analysis method based on fault rate and static safety analysis
CN113570200A (en) * 2021-06-30 2021-10-29 国电南瑞科技股份有限公司 Power grid operation state monitoring method and system based on multidimensional information
CN113610270A (en) * 2021-07-01 2021-11-05 广西电网有限责任公司电力科学研究院 Distribution transformer operation risk prediction method and system considering branch slot influence
CN114566964A (en) * 2022-04-29 2022-05-31 国网天津市电力公司电力科学研究院 Power distribution network feeder automation control method, device, equipment and storage medium
CN115018100A (en) * 2022-06-23 2022-09-06 国网福建省电力有限公司电力科学研究院 Operation maintenance and overhaul decision method based on health state of power transformation equipment
CN116937561A (en) * 2023-07-21 2023-10-24 武汉星创源科技有限公司 Active power distribution network risk assessment and fault early warning method
CN117370872A (en) * 2023-12-07 2024-01-09 国网江苏省电力有限公司南通供电分公司 Optimizing control method and device for power equipment
CN117910639A (en) * 2024-01-19 2024-04-19 湖北省聚广鑫科技有限公司 Power distribution network reliability optimization method and system

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130218355A1 (en) * 2012-02-16 2013-08-22 Spyros James Lazaris Electricity grid data analytics as a moduled service for production, delivery, and distribution of power as a dynamic demand response within a renewable energy-based electricity grid infrastructure
CN106786546A (en) * 2017-01-11 2017-05-31 南京工业大学 Power distribution network fault recovery strategy optimization method based on risk assessment

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130218355A1 (en) * 2012-02-16 2013-08-22 Spyros James Lazaris Electricity grid data analytics as a moduled service for production, delivery, and distribution of power as a dynamic demand response within a renewable energy-based electricity grid infrastructure
CN106786546A (en) * 2017-01-11 2017-05-31 南京工业大学 Power distribution network fault recovery strategy optimization method based on risk assessment

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
DARMAWAN SUTANTO等: "Design of an Intelligent Self-Healing Smart Grid using a Hybrid Multi-Agent Framework", 《JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY》 *
RAN LI等: "Economic Optimization of Self-healing Control of Power Grid Based on Multi-agent system", 《PROCEEDINGS OF 2016 IEEE ADVANCED INFORMATION MANAGEMENT,COMMUNICATES,ELECTRONIC AND AUTOMATION CONTROL CONFERENCE(IMCEC 2016)》 *
姚玉海: "基于预想事故集的配电网自愈控制研究", 《中国博士学位论文全文数据库》 *

Cited By (38)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108416469A (en) * 2018-02-10 2018-08-17 杭州后博科技有限公司 A kind of extreme weather causes the prediction technique and system of steel tower security risk
CN108416469B (en) * 2018-02-10 2021-02-02 杭州后博科技有限公司 Method and system for predicting potential safety hazard of iron tower caused by extreme weather
CN108375715A (en) * 2018-03-08 2018-08-07 中国电力科学研究院有限公司 A kind of distribution network line fault risk day prediction technique and system
CN108375715B (en) * 2018-03-08 2023-11-03 中国电力科学研究院有限公司 Power distribution network line fault risk day prediction method and system
CN108695846A (en) * 2018-05-29 2018-10-23 国电南瑞科技股份有限公司 A kind of unit style power distribution network operation risk assessment method
CN108695846B (en) * 2018-05-29 2021-11-09 国电南瑞科技股份有限公司 Unit distribution power grid operation risk assessment method
CN109145364A (en) * 2018-07-04 2019-01-04 西北工业大学 Sensitivity Analysis Method based on Geordie inequality
CN109190996A (en) * 2018-09-18 2019-01-11 中国电力科学研究院有限公司 A kind of distribution net platform region operation risk monitoring and pre-alarming method and device
CN109376877A (en) * 2018-10-11 2019-02-22 华自科技股份有限公司 Equipment O&M method for early warning, device, computer equipment and storage medium
CN109376877B (en) * 2018-10-11 2021-01-26 华自科技股份有限公司 Equipment operation and maintenance early warning method and device, computer equipment and storage medium
CN109521672A (en) * 2018-10-22 2019-03-26 东北大学 A kind of intelligent selecting method of electric arc furnaces power supply curve
CN109521672B (en) * 2018-10-22 2021-08-13 东北大学 Intelligent selection method for power supply curve of electric arc furnace
CN109460004A (en) * 2018-10-26 2019-03-12 国网天津市电力公司 Distribution network failure prediction technique and system based on big data
CN109784502A (en) * 2018-12-11 2019-05-21 深圳供电局有限公司 On-site maintenance processing method and system based on fault equipment
CN111488896B (en) * 2019-01-28 2023-08-11 国网能源研究院有限公司 Distribution line time-varying fault probability calculation method based on multi-source data mining
CN111488896A (en) * 2019-01-28 2020-08-04 国网能源研究院有限公司 Distribution line time-varying fault probability calculation method based on multi-source data mining
CN110071500A (en) * 2019-04-09 2019-07-30 国网山东省电力公司济南供电公司 Distribution line restores Sequence Decision method and system after a kind of bus-bar fault power loss
CN110071500B (en) * 2019-04-09 2023-03-21 国网山东省电力公司济南供电公司 Method and system for decision-making of power distribution line recovery sequence after bus fault power loss
CN110210632A (en) * 2019-05-31 2019-09-06 国网河北省电力有限公司沧州供电分公司 Failure emergency processing method, device and terminal based on ubiquitous electric power Internet of Things
CN110705801A (en) * 2019-10-10 2020-01-17 国网山东省电力公司泰安供电公司 Power grid accident economic loss estimation method and system based on fault loss electric quantity
CN110739691A (en) * 2019-11-06 2020-01-31 南方电网科学研究院有限责任公司 power grid branch fault prediction method and device
CN110739691B (en) * 2019-11-06 2021-05-14 南方电网科学研究院有限责任公司 Power grid branch fault prediction method and device
CN111160603A (en) * 2019-11-21 2020-05-15 中国电力科学研究院有限公司 Method and system for guaranteeing reliability of end-to-end power communication service
CN111044850A (en) * 2020-01-14 2020-04-21 广东电网有限责任公司 Power grid fault diagnosis method based on equipment fault probability
CN112116276A (en) * 2020-09-29 2020-12-22 广东电网有限责任公司广州供电局 Transformer substation operation risk assessment method considering time-varying state of electrical main equipment
CN112668157A (en) * 2020-12-14 2021-04-16 河北电力交易中心有限公司 Node power failure risk invariant node reserve capacity rapid assessment method
CN113052473B (en) * 2021-03-31 2024-03-22 贵州电网有限责任公司 Power grid risk analysis method based on fault rate and static safety analysis
CN113052473A (en) * 2021-03-31 2021-06-29 贵州电网有限责任公司 Power grid risk analysis method based on fault rate and static safety analysis
CN113570200A (en) * 2021-06-30 2021-10-29 国电南瑞科技股份有限公司 Power grid operation state monitoring method and system based on multidimensional information
CN113570200B (en) * 2021-06-30 2023-10-03 国电南瑞科技股份有限公司 Power grid running state monitoring method and system based on multidimensional information
CN113610270A (en) * 2021-07-01 2021-11-05 广西电网有限责任公司电力科学研究院 Distribution transformer operation risk prediction method and system considering branch slot influence
CN113610270B (en) * 2021-07-01 2024-03-26 广西电网有限责任公司电力科学研究院 Distribution transformer operation risk prediction method and system considering branch groove influence
CN114566964A (en) * 2022-04-29 2022-05-31 国网天津市电力公司电力科学研究院 Power distribution network feeder automation control method, device, equipment and storage medium
CN115018100A (en) * 2022-06-23 2022-09-06 国网福建省电力有限公司电力科学研究院 Operation maintenance and overhaul decision method based on health state of power transformation equipment
CN116937561A (en) * 2023-07-21 2023-10-24 武汉星创源科技有限公司 Active power distribution network risk assessment and fault early warning method
CN117370872A (en) * 2023-12-07 2024-01-09 国网江苏省电力有限公司南通供电分公司 Optimizing control method and device for power equipment
CN117370872B (en) * 2023-12-07 2024-03-26 国网江苏省电力有限公司南通供电分公司 Optimizing control method and device for power equipment
CN117910639A (en) * 2024-01-19 2024-04-19 湖北省聚广鑫科技有限公司 Power distribution network reliability optimization method and system

Also Published As

Publication number Publication date
CN107516170B (en) 2020-07-28

Similar Documents

Publication Publication Date Title
CN107516170B (en) Difference self-healing control method based on equipment failure probability and power grid operation risk
CN112749904B (en) Power distribution network fault risk early warning method and system based on deep learning
CN112217202A (en) Distributed new energy, energy storage and power distribution network planning method considering flexibility investment
CN108876163B (en) Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning
CN107292478B (en) Method for acquiring influence situation of disaster on power distribution network
CN110570016B (en) Disaster-resistant backbone net rack optimization method considering multi-stage performance
CN104037776A (en) Reactive power grid capacity configuration method for random inertia factor particle swarm optimization algorithm
CN112149873A (en) Low-voltage transformer area line loss reasonable interval prediction method based on deep learning
CN109993665B (en) Online safety and stability assessment method, device and system for power system
CN110298494A (en) A kind of wind power forecasting method based on Segment Clustering and Combinatorial Optimization
CN114742283A (en) Medium-voltage line loop closing current estimation and risk assessment method based on deep learning network
CN114386753A (en) Equipment risk comprehensive analysis early warning method based on main transformer load condition
CN108075471B (en) Multi-objective constraint optimization power grid scheduling strategy based on stochastic power output prediction
CN110783913B (en) Group-based optimal power grid topology online optimization method considering expected accident set
CN115758151A (en) Combined diagnosis model establishing method and photovoltaic module fault diagnosis method
CN113673065B (en) Loss reduction method for automatic reconstruction of power distribution network
Meng et al. A new PV generation power prediction model based on GA-BP neural network with artificial classification of history day
CN113283492B (en) Photovoltaic output scene generation method and scene generation device based on Catboost algorithm
CN114142461A (en) New energy differentiated configuration method considering power grid form evolution and structure development
CN105207255B (en) A kind of power system peak regulation computational methods suitable for wind power output
Wang et al. Early warning of distribution transformer based on bp neural network considering the influence of extreme weather
CN111697607A (en) Multi-terminal flexible direct-current transmission receiving-end power grid access method and system
CN115118015A (en) Platform district power supply stability monitoring system based on fuse terminal
CN112241812B (en) Topology identification method for low-voltage distribution network based on single-side optimization and genetic algorithm cooperation
CN114707713A (en) Low-voltage distribution network load prediction method considering distributed power supply influence

Legal Events

Date Code Title Description
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant