CN115640969A - Power grid operation and maintenance cost distribution method based on equipment state and operation age - Google Patents
Power grid operation and maintenance cost distribution method based on equipment state and operation age Download PDFInfo
- Publication number
- CN115640969A CN115640969A CN202211281098.3A CN202211281098A CN115640969A CN 115640969 A CN115640969 A CN 115640969A CN 202211281098 A CN202211281098 A CN 202211281098A CN 115640969 A CN115640969 A CN 115640969A
- Authority
- CN
- China
- Prior art keywords
- power grid
- grid equipment
- equipment
- prediction model
- fault
- 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.)
- Pending
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000009826 distribution Methods 0.000 title claims abstract description 13
- 230000036541 health Effects 0.000 claims abstract description 26
- 241000282461 Canis lupus Species 0.000 claims description 34
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000010586 diagram Methods 0.000 claims description 5
- 238000012544 monitoring process Methods 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 230000005251 gamma ray Effects 0.000 claims description 3
- 241000160777 Hipparchia semele Species 0.000 claims 1
- 238000005259 measurement Methods 0.000 claims 1
- 238000007781 pre-processing Methods 0.000 abstract description 2
- 238000005457 optimization Methods 0.000 description 10
- 230000009466 transformation Effects 0.000 description 8
- 210000002569 neuron Anatomy 0.000 description 7
- 238000012545 processing Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 230000000007 visual effect Effects 0.000 description 6
- 230000006399 behavior Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 5
- 210000000988 bone and bone Anatomy 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 206010001488 Aggression Diseases 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000016571 aggressive behavior Effects 0.000 description 1
- 208000012761 aggressive behavior Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013278 delphi method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000009666 routine test Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012876 topography Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of data preprocessing and identification, and discloses a power grid operation and maintenance cost distribution method based on equipment state and operation age, S1, introducing an equipment health index concept, and taking the equipment health index as a first influence element of power grid equipment failure; s2, identifying a second influence element which influences the power grid equipment fault except the first influence element; s3, constructing a power grid equipment fault prediction model based on the deep belief network; s4, inputting the input variables into an optimal power grid equipment fault prediction model, and predicting to obtain the corresponding fault rates of all power grid equipment through the optimal power grid equipment fault prediction model; and S5, according to the comprehensive level ratio of the failure rates of the power grid equipment in different regions, the operation and maintenance of the power grid equipment in the next year is invested into total funds to be distributed. The invention fully considers the health index of the equipment and other influence factors influencing the failure rate of the equipment, and can reflect the problem of practical requirements of equipment operation and maintenance under the real operation condition.
Description
Technical Field
The invention relates to the technical field of data preprocessing and identification, in particular to a power grid operation and maintenance cost distribution method based on equipment states and operation years.
Background
The power energy supply is an important basis for guaranteeing the health and stable development of social economy, and the scientific and reasonable operation and maintenance investment is an important hand for guaranteeing the operation safety and the power supply reliability of the power grid equipment. Under the new situation, with the continuous promotion and deepening of the national power system reform, the reasonable configuration of enhancing the operation and maintenance cost of the power grid becomes one of the important means for improving the quality and the efficiency of the power grid enterprises.
At present, methods for distributing operation and maintenance costs of power grid enterprises mainly include two types:
firstly, the operation and maintenance of traditional power grid enterprise invests and distributes and adopts "the quotation system" of "top-down", promptly: firstly, the operation and maintenance input scale of a superior unit of a power grid enterprise in one year is determined, and the overall scale of the operation and maintenance input in the next year is determined according to a certain proportion (usually 10% -20%); secondly, comprehensively considering the assets, equipment scale, power grid structure and other elements of the lower level unit, and determining the distribution proportion of the lower level unit; and finally, dynamically adjusting the input proportion of the lower-level unit by referring to the benefit output level (usually the electric quantity output level) of the lower-level unit, thereby obtaining the final operation and maintenance input scale of each unit. The method firstly determines the overall scale of the dish, mainly according to historical experience and lacks certain scientificity; meanwhile, when the operation and maintenance invest capital allocation, although the scale level of lower-level unit assets and equipment is considered, the operation state and the operation age of the equipment are not considered enough, so that the problem of unbalanced capital allocation is generated. For example, the power grid equipment in a certain area has relatively large asset scale and large electricity output benefit, but the equipment in the area has good running state, takes new production equipment as a main part, has low probability of total equipment failure and power grid system failure occurrence, and has less operation and maintenance investment required by the unit actually; however, under the condition of 'split-disk' control, the region can obtain a larger amount of operation and maintenance investment, so that the operation and maintenance investment capital is wasted, and the lean management and control of enterprises are not facilitated.
And secondly, the superior unit establishes a comprehensive evaluation analysis index system by combing internal and external factors influencing the operation and maintenance investment of the power grid, and calculates the comprehensive evaluation score ratio of each inferior unit so as to determine the operation and maintenance investment scale of different regions. Although key indexes such as the operation state and the operation age of the power grid equipment are considered to a certain extent, the method also has the problem of reasonable operation and maintenance investment distribution due to the fact that the considered elements are too comprehensive. For example, in a certain area, due to the fact that the equipment operation age is long, the equipment operation state index is relatively poor, and the probability of equipment failure is high; however, the proportion of the two indexes in the comprehensive evaluation score is small, so that the operation and maintenance invested funds actually allocated to the region are relatively small, and the operation safety and stability of the power grid equipment in the region are not facilitated.
Disclosure of Invention
The invention provides a power grid operation and maintenance cost distribution method based on equipment states and operation years, which is used for solving the problem that the actual requirements of equipment operation and maintenance under the real operation condition cannot be reflected by a constructed model due to the fact that the existing method relies on historical experiences too much and the equipment operation states and the operation years are not considered fully.
The invention is realized by the following technical scheme:
a power grid operation and maintenance cost distribution method based on equipment states and operation years comprises the following steps:
s1, introducing an equipment health index concept, determining an equipment health index calculation method, and taking the equipment health index as a first influence element of a power grid equipment fault;
s2, identifying a second influence element which influences the power grid equipment fault except the first influence element, measuring and analyzing the influence degree of the second influence element on the power grid equipment fault to obtain a preferred second influence element, and taking the preferred second influence element and the first influence element as input variables of a power grid equipment fault prediction model;
s3, constructing a power grid equipment fault prediction model based on a deep belief network, and optimizing parameters of the power grid equipment fault prediction model to obtain an optimal power grid equipment fault prediction model;
s4, inputting input variables into an optimal power grid equipment fault prediction model, predicting through the optimal power grid equipment fault prediction model to obtain corresponding fault rates of all power grid equipment, and obtaining a comprehensive level of the fault rates of the power grid equipment in the region through weighted average;
and S5, according to the comprehensive level ratio of the failure rates of the power grid equipment in different regions, the operation and maintenance of the power grid equipment in the next year is invested into total funds to be distributed.
As optimization, the specific steps of S1 are:
s1.1, collecting historical statistical data of the power grid equipment;
s1.2, introducing an equipment health index concept, combining historical statistical data of the power grid equipment with technical standards and monitoring operation and maintenance information of the power grid equipment to obtain deduction values of all parts of the power grid equipment, and summarizing the deduction values of all the parts belonging to the same power grid equipment to form a health index of the power grid equipment.
As an optimization, in S2, the second influencing element is mainly identified by a fishbone diagram method system.
In the optimization step S2, the influence degree of the second influence element on the power grid equipment fault is measured and analyzed mainly by a grey correlation analysis method.
As optimization, the specific steps of measuring and analyzing the influence degree of the second influence element on the power grid equipment fault through the grey correlation degree are as follows:
s2.1, determining a reference sequence and a comparison sequence, wherein the reference sequence is a data sequence of the fault rate of the power grid equipment fault, and the comparison sequence is a data sequence of a second influence factor of the power grid equipment fault;
s2.2, carrying out non-dimensionalization processing on the reference sequence and the comparison sequence to obtain a reference number sequence X 0 ＝{x 0 (t) } and comparison series X i ＝{x i (t)}；
S2.3, calculating the reference sequence X 0 ＝{x 0 (t) } and comparison series X i ＝{x i (t) } comparison correlation coefficient δ oi (j)：
Δ 0i (j)＝|x 0 (j)-x i (j)|；
Δ min ＝min i min j Δ 0i (j),Δ max ＝max i max j Δ 0i (j)；
n is the total time, k is the number of second influencing elements, delta oi (j) To normalize the parameter, Δ min As a minimum normalization parameter, Δ min As maximum normalized parameter, x 0 (j) Is the size of the reference sequence at time t = j, x i (t) the magnitude ρ of the ith comparison series at time t = j represents the resolution coefficient;
s2.4, drawing a comparison correlation coefficient curve of a certain second influence element and the fault rate of the power grid equipment by taking the abscissa as time t and the ordinate as the comparison correlation coefficient, obtaining the correlation area of a comparison sequence according to the comparison correlation coefficient curve, and then calculating the correlation degree of the certain second influence element and the fault rate of the power grid equipment:
S 00 representing the area of association of the reference series, S 0i Representing the correlation area of the comparison sequence, wherein the correlation coefficient corresponding to the correlation area of the reference sequence is 1;
s2.5, sequencing the association degrees of all second influence elements and the fault rate of the power grid equipment, selecting the second influence elements sequenced at the first A position as preferred second influence elements, and combining the preferred second influence elements with the first influence elements to serve as input variables of the power grid equipment fault prediction model.
As optimization, in S3, the parameters of the power grid equipment fault prediction model are optimized specifically by a grey wolf algorithm.
As optimization, in S3, the specific steps of constructing a power grid device fault prediction model based on the deep belief network are as follows:
s3.1, inputting the input variable serving as an input layer vector to a first layer RBM to finish unsupervised training;
s3.2, obtaining feature data after completing feature learning, inputting the obtained feature data serving as an input variable of a new layer into an RBM of a next layer, and continuing to perform unsupervised training;
and S3.3, repeating the S3.2 until the RBM of each layer is trained and learned, and taking the characteristics obtained in the RBM of the last layer as output characteristics.
As optimization, in S3, the specific steps of optimizing the parameters of the power grid equipment fault prediction model through the grey wolf algorithm are as follows:
s3.4, initializing parameters of a gray wolf algorithm, wherein the parameters of the gray wolf algorithm comprise population scale, maximum iteration times and initialized population position, and the population is the output characteristic of the power grid equipment fault prediction model;
s3.5, performing opponent search, and judging by taking the error rate of the power grid equipment fault prediction model as a fitness value, if the fitness of the obtained opponent is superior to that of the original individual, generating an initial wolf pack by using the opponent, otherwise, generating the initial wolf pack by using the original individual, wherein the individual is the fault rate output by the power grid equipment fault prediction model;
and S3.6, carrying out grade division on the individuals according to the fitness, taking the first three individuals to guide the candidate solution of the fault rate output by the power grid equipment fault prediction model to carry out position updating, and obtaining the optimal power grid equipment fault prediction model after the updating is finished.
As optimization, the specific calculation mode of S4 is:
wherein, gamma represents the comprehensive level value of the fault rates of all the power grid equipment in a certain area, and lambda g And a fault rate value predicted by the g-th type of power grid equipment through the optimal power grid equipment fault prediction model is represented, d represents the type number of the power grid equipment, and f is the number of the power grid equipment.
As optimization, in S5, according to the comprehensive level proportion of the failure rates of the power grid devices in different areas, the calculation method for allocating the total capital invested in the operation and maintenance of the power grid devices in the next year is as follows:
wherein phi is r The method comprises the steps of representing the operation and maintenance investment fund allocation limit of the power grid equipment in the next year in the r-th area; a represents the total scale of capital invested in operation and maintenance of the power grid equipment in the next year in all regions; gamma ray r And (4) representing the comprehensive level value of the fault rate of the power grid equipment in the current year in the r-th area.
Compared with the prior art, the invention has the following advantages and beneficial effects:
the power grid equipment fault prediction model built by the invention does not depend on historical experience too much, fully considers the equipment health index and other influence factors influencing the equipment fault rate, and can reflect the problem of actual demand of equipment operation and maintenance investment under the real operation condition.
Drawings
In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and that for those skilled in the art, other related drawings can be obtained from these drawings without inventive effort. In the drawings:
fig. 1 is a flowchart of a power grid operation and maintenance cost distribution method based on a device state and an operation age according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and the accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not used as limiting the present invention.
Examples
As shown in fig. 1, a method for allocating operation and maintenance costs of a power grid based on a device status and an operation age according to the present invention includes:
s1, introducing an equipment health index concept, determining an equipment health index calculation method, and taking the equipment health index as a first influence element of a power grid equipment fault;
in this embodiment, the specific steps of S1 are:
s1.1, collecting historical statistical data of the power grid equipment; the historical statistical data comprises equipment operation time data, statistical data of fault rates of different operation years of equipment, online monitoring data of power grid enterprise equipment, routine test data, equipment inspection data and the like.
S1.2, introducing an equipment health index concept, combining historical statistical data of the power grid equipment with technical standards and monitoring operation and maintenance information of the power grid equipment to obtain deduction values of all parts of the power grid equipment, and summarizing the deduction values of all parts of the same power grid equipment to form a health index of the power grid equipment.
Specifically, an equipment health index concept is introduced and used for describing a numerical value of the health degree of the power grid equipment. According to technical standards such as 'inspection test regulations on the state of power transmission and transformation equipment' and 'inspection guide rules on the state of power transmission and transformation equipment' issued by national grid companies, and the like, as calculation bases, deduction values of all parts are obtained through state monitoring and operation and maintenance test information, and finally, a comprehensive deduction value of the power transmission and transformation equipment is obtained, namely, a health index of the equipment.
Taking a transformer as an example, the components such as the main body, the insulating sleeve, the tap switch, the cooling system, the relay protection device and the like are respectively analyzed. If the states of all the components are normal, the transformer state is normal, and the health index is the sum of the deduction values of all the components multiplied by corresponding weights; if some parts in the transformer are in abnormal states, the state of the transformer is represented by the state of the part with the most serious degradation, and the deduction value of the part is the health index of the transformer. The division of the transformer states and the weighting of the components are clearly defined in the above-mentioned standards, while the deduction values of other components of the grid installation are also clearly defined in the above-mentioned standards, and are not described in detail here.
S2, identifying a second influence element which influences the power grid equipment fault and is obtained by removing the first influence element by combining a fishbone diagram method system, measuring and analyzing the influence degree of the second influence element on the power grid equipment fault through a grey correlation degree analysis method to obtain a preferred second influence element, and taking the preferred second influence element and the first influence element as input variables of a power grid equipment fault prediction model.
The fishbone diagram method and system are favorable for identifying the influence factors of the power grid equipment faults from different dimensions such as society, economy, environment and the like and identifying the influence factors such as commissioning life, topography, meteorological environment, transformer substation types and load rate in combination with actual operation and maintenance current situations of equipment.
The fishbone diagram method system extracts factors from the aspects of the structure, the flow and the like of an analyzed object, and then draws the factors into a graph with clear layering and clear arrangement according to the relevance among the factors.
The application of the fishbone map comprises two steps, namely analyzing factors and drawing the fishbone map.
1) The procedure for analyzing privacy is as follows.
Step 10: for the study subjects, the classification mode, i.e. the major factor, was chosen.
Step 20: and (3) respectively finding out all possible factors in each class by using a brain storm method, a Delphi method and the like.
And step 30: and (5) sorting the factors to determine the attributes of the factors.
Step 40: the factors are briefly described.
2) The steps for drawing the fishbone map are as follows.
And special software is adopted to draw fishbone images, such as visio, XMind and the like. Simple fishbone maps can also be drawn in Word and Excel. Specifically, the problem to be researched is marked on the fish head, the large bone is drawn, the large factor is filled, the middle bone and the small bone are extended from the large bone, the middle factor and the small factor are respectively filled, and finally the second influence factor is obtained.
The grey correlation degree analysis theory is a multivariate (factor) statistical analysis method, which uses the sample data of each variable as the basis and uses the grey correlation degree to describe the strength, size and order of the relationship between them. The basic idea is that sample data of variables are used as sequences, the sequences form a curve, the closeness degree of a certain two variables is judged by the closeness degree of the two curves, the closer the curves are, the greater the relevance degree between the corresponding sequences is, otherwise, the smaller the relevance degree is, and correlation analysis is performed according to the following steps. Specifically, in combination with the sample data statistical analysis, relevance analysis is performed on the fault and the influence factors, and the factors with high relevance are identified as input nodes of the following prediction model, wherein the elements with the relevance greater than 0.6 are used as input variables of the following prediction model.
Specifically, in this embodiment, the specific steps of measuring and analyzing the influence degree of the second influence element on the power grid equipment fault through the gray correlation degree include:
s2.1, determining a reference sequence and a comparison sequence, wherein the reference sequence is a data sequence reflecting the behavior characteristics of the system, and the comparison sequence is a data sequence reflecting factors influencing the behavior of the system. In this embodiment, the reference sequence is a data sequence of a failure rate of the power grid device failure, and the comparison sequence is a data sequence of a second influencing element of the power grid device failure.
S2.2, carrying out non-dimensionalization processing on the reference sequence and the comparison sequence to obtain a reference number sequence X 0 ＝{x 0 (t) } and comparison series X i ＝{x i (t)}，x 0 (t) is the set of failure rates at time t, x i (t) is a second set of influencing elements at time t. Since the physical significance of the second influencing elements in the system is different, the influence of the dimension must be eliminated.
Let X i ＝(x i (1),x i (2),...,x i (n)) is X i A sequence of actions of (a). The common processing method comprises the following steps of carrying out initial value transformation and calculating:
in general, for a relatively stable economic and social system sequence, initialization transformation is mostly adopted. The number series mostly shows a stable growth trend, the growth trend can be more obvious through initial processing, the second processing mode is averaging transformation, and the calculation process is as follows:
the third processing mode is interval transformation, and the calculation process is as follows:
the above-mentioned conversion to prior art, the parameters therein are also well known in the art and are not described in detail here.
A normalized transformation is performed as follows:
where μ and σ are the mean and variance of the sequence, respectively.
Calculating correlation coefficient based on the above processing procedures, and performing non-dimensionalization processing to obtain reference number X 0 ＝{x 0 (t) the comparison number is X i ＝{x i (t)}。
S2.3, calculating the reference sequence X 0 ＝{x 0 (t) } and a comparison series X i ＝{x i (t) } comparison correlation coefficient δ oi (j)：
Δ 0i (j)＝|x 0 (j)-x i (j)|；
Δ min ＝min i min j Δ 0i (j),Δ max ＝max i max j Δ 0i (j)；
n is the total time, k is the number of second influencing elements, delta oi (j) To normalize the parameter, Δ min As minimum normalization parameter, Δ min As maximum normalized parameter, x 0 (j) Is the size of the reference sequence at time t = j, x i (t) is the magnitude of the ith comparison series at time t = j, and ρ represents the resolution factor in the sense that Δ is attenuated max Distortion caused by too large value, increasing the turn-offThe difference between the connection coefficients is significant, rho is more than 0 and less than 1, and the general situation rho is 0.5. The correlation coefficient reflects the reference sequence X 0 And comparing the series X i Tightness 0 < delta at time j 0i ≤1。；
S2.4, calculating the association degree based on the parameters, drawing a comparison association coefficient curve of the fault rate of the certain second influence element and the power grid equipment by taking the abscissa as time t and the ordinate as the comparison association coefficient, obtaining the association area of a comparison sequence according to the comparison association coefficient curve, and then calculating the association degree of the fault rate of the certain second influence element and the power grid equipment:
S 00 indicating the area of association of the reference number series, S 0i The correlation coefficient corresponding to the correlation area of the comparison number series is 1 everywhere.
S2.5, sequencing the relevance of all second influence elements and the fault rate of the power grid equipment, selecting the second influence elements positioned at the front A position in the sequence as preferred second influence elements, and combining the preferred second influence elements with the first influence elements as input variables of the power grid equipment fault prediction model.
And sequencing the relevance of the plurality of second influence elements to the fault, wherein the greater the relevance, the greater the influence degree of the second influence elements to the fault is. And selecting the operation age, the landform and landform, the health index, the meteorological environment, the transformer substation type and the load rate as preferred second influence elements by combining the correlation evaluation result, and then using the preferred second influence elements and the first influence elements together as the neuron nodes of the input layer of the power grid equipment fault prediction model.
And S3, constructing a power grid equipment fault prediction model based on the deep belief network, and optimizing parameters of the power grid equipment fault prediction model through a wolf algorithm to obtain an optimal power grid equipment fault prediction model.
Specifically, the method comprises the following steps:
s3.1, inputting the input variable serving as an input layer vector to a first layer RBM to finish unsupervised training;
s3.2, obtaining feature data after completing feature learning, inputting the obtained feature data serving as an input variable of a new layer into a RBM of a next layer, and continuing unsupervised training;
and S3.3, repeating the S3.2 until the RBM of each layer is trained and learned, and taking the characteristics obtained in the RBM of the last layer as output characteristics. The deep belief network uses a plurality of RBM units to form a basic network architecture, so that the deep belief network has unsupervised pre-training and supervised fine adjustment, and the combination of unsupervised and unsupervised conditions not only solves the problem of gradient dispersion existing in the traditional method, but also can solve the problem that the network is easy to fall into local optimum.
For the deep confidence network topology structure chart, the detailed training process is as follows:
the first step is as follows: determining an energy function E (v, h | theta) according to the neuron nodes of the input layer and the hidden layer of the deep confidence network;
wherein v is x ∈{v 1 ，v 2 ，...，v p Denotes the value of the xth neuron node of the visual layer, the visual layer is the input layer of the data, the neuron node of the visual layer is the preferred second influence element, p is the number of the neuron nodes of the visual layer, h zy Representing the yth neuron node of the z-th hidden layer, q being the number of neurons of the hidden layer, o being the number of hidden layers, θ = { w, a, b } being network parameters, w being weights between the visible layer and the hidden layer, a and b being offsets of the visible layer and the hidden layer, respectively;
secondly, obtaining a joint probability distribution function according to the energy function E (v, h | theta):
in the above formula, Z (theta) = ∑ Σ v，h e -E(v，h|θ) Expressed as normalization factors, representing the algebraic sum of all energy functions;
thirdly, after the state of the visual layer is determined, the activation probability of the hidden layer unit is as follows:
after the h state of the hidden layer is determined, the activation probability of the visual layer unit is as follows:
when the training sample number is K, the parameter θ can be determined by solving a log-likelihood function maximization problem, where an objective function of the log-likelihood function maximization problem is as follows:
in the above formula, maxL (θ) is obtained by a random gradient method.
By repeating Gibbs sampling, the update rule of the RBM parameter can be obtained as follows:
Δw xy ＝ε(<v x h y > data -<v x h y > recon )
Δa x ＝ε(<V x > data -<v x > recon )
Δb y ＝ε(<h y > data -<h y > recon )
wherein epsilon is the RBM learning rate,<·> data and<·> rec o n mathematical expectations for input data and reconstructed data, respectively。
Therefore, in this embodiment, in S3, the specific steps of optimizing the parameters of the power grid equipment fault prediction model through the grayish wolf algorithm are as follows:
s3.4, initializing parameters of a gray wolf algorithm, wherein the parameters of the gray wolf algorithm comprise population scale, maximum iteration times and initialized population position, and the population is the output characteristic of the power grid equipment fault prediction model;
s3.5, performing opponent search, and judging by taking the error rate of the power grid equipment fault prediction model as a fitness value, if the fitness of the obtained opponent is superior to that of the original individual, generating an initial wolf pack by using the opponent, otherwise, generating the initial wolf pack by using the original individual, wherein the individual is the fault rate output by the power grid equipment fault prediction model;
and S3.6, carrying out grade division on the individuals according to the fitness, taking the individuals in the first three ranks to guide the omega wolf to carry out position updating, and obtaining an optimal power grid equipment fault prediction model after the updating is finished (the maximum iteration times are reached).
The principle of the grey wolf algorithm is as follows:
(1) Enveloping behavior
The data model of the grayish wolf surrounding the prey can be expressed as.
D＝|C·X p (t)-X(t)|
X(t+1)＝X p (t)-A·D
Wherein D represents the distance between the wolf group and the prey, A =2 alpha r 1 -α，C＝2·r 2 T denotes the number of iterations, X p And X represents the positions of the prey and the wolf group, respectively, r 1 、r 2 Is a random quantity with a value range of [0,1%]And the value range of alpha is [0,2 ]]。
(2) Hunting behavior
Assuming that alpha, beta and delta represent the global optimal solution, the second solution and the third solution of the wolf body and are optimally positioned, the distances are respectively expressed as
D α ＝|C 1 ·X α -X|
D β ＝|C 2 ·X β -X|
D δ ＝|C 2 ·X δ -X|
In the formula, D α 、D β 、D δ Representing the approximate distances of the individuals alpha, beta, delta from the current position X, X α 、X β 、X δ Sequentially representing the positions of the global optimal solution, the second solution and the third solution; c 1 、C 2 、C 3 Represents a random vector having a value in the range of [0,1 ]]. X and X (t + 1) are represented by formula (18), formula (19), formula (20) and formula (21), respectively.
X 1 ＝X α -A 1 ·(D α )
X 2 ＝X β -A 2 ·(D β )
X 3 ＝X δ -A 3 ·(D δ )
Wherein X (t + 1) represents an update solution, A 1 、A 2 、A 3 Representing a random quantity.
(3) Aggressive behavior
The attack is the final stage of the wolf colony predation behavior and can be realized by adjusting the parameter alpha.
If A is less than or equal to 1, the wolf colony approaches the prey and attacks the prey intensively (X) * ,Y * ) (ii) a Conversely, the wolf pack gradually moves away from the prey.
Specifically, a transformer fault is taken as an example.
1) Acquiring transformer fault sample data, and analyzing the fault sample data through grey correlation;
2) Constructing a component and initializing a power grid equipment fault prediction model based on a deep belief network, and determining the number of network layers and the number of nodes of each layer of the power grid equipment fault prediction model so as to determine the dimensionality of a wolf vector;
3) Initializing grey wolf algorithm parameters, including population scale, maximum iteration times and initialized population position, performing opposite search, taking the error rate of the power grid equipment fault prediction model as a fitness value, adopting an opposite individual if the fitness of the opposite individual is superior to that of an original individual, and otherwise, adopting the original individual to generate an initial wolf group;
4) The initial wolf group is graded according to the fitness, the top three ranking wolfs are respectively alpha, beta and delta wolfs, and the alpha, beta and delta wolfs guide the omega wolfs to carry out position updating;
5) Updating the convergence factor a according to a formula, and updating the parameters A and C according to a hunting searching formula; and calculating the distance and the new individual position according to a formula position information formula.
6) Judging whether the iteration times reach the maximum iteration times, if so, terminating the IGWO optimization, otherwise, returning to the step 4)
7) And obtaining the optimal weight of the power grid equipment fault prediction model, and performing pre-training and reverse fine adjustment on the power grid equipment fault prediction model.
8) And calculating test data and outputting a fault classification result of the test set.
S4, inputting input variables into an optimal power grid equipment fault prediction model, predicting through the optimal power grid equipment fault prediction model to obtain corresponding fault rates of all power grid equipment, and obtaining a comprehensive level of the fault rates of the power grid equipment in the region through weighted average;
and S5, according to the comprehensive level proportion of the failure rates of the power grid equipment in different areas, the operation and maintenance of the power grid equipment in the next year is invested into total capital for distribution.
In this embodiment, the specific calculation manner of S4 is:
wherein, gamma represents the comprehensive level value of the failure rate of all the power grid equipment in a certain area, and lambda g And the failure rate value predicted by the category g power grid equipment through the optimal power grid equipment failure prediction model is represented, d represents the type number of the power grid equipment, and f is the number of the power grid equipment.
In this embodiment, in S5, according to the comprehensive level ratio of the failure rates of the power grid devices in different regions, the calculation method for allocating the total capital invested in the next year of the power grid devices is as follows:
wherein phi is r The method comprises the steps of representing the operation and maintenance investment fund allocation limit of the power grid equipment in the next year in the r-th area; a represents the total scale of capital invested in operation and maintenance of the power grid equipment in the next year in all regions; gamma ray r And (4) representing the comprehensive level value of the fault rate of the power grid equipment in the current year in the r-th area.
The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only examples of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A power grid operation and maintenance cost distribution method based on equipment states and operation years is characterized by comprising the following steps:
s1, introducing an equipment health index concept, determining an equipment health index calculation method, and taking the equipment health index as a first influence element of a power grid equipment fault;
s2, identifying a second influence element which influences the power grid equipment fault except the first influence element, measuring and analyzing the influence degree of the second influence element on the power grid equipment fault to obtain a preferred second influence element, and taking the preferred second influence element and the first influence element as input variables of a power grid equipment fault prediction model;
s3, constructing a power grid equipment fault prediction model based on a deep belief network, and optimizing parameters of the power grid equipment fault prediction model to obtain an optimal power grid equipment fault prediction model;
s4, inputting the input variables into an optimal power grid equipment fault prediction model, predicting the corresponding fault rates of all power grid equipment through the optimal power grid equipment fault prediction model, and obtaining the comprehensive level of the fault rates of the power grid equipment in the region through weighted average;
and S5, according to the comprehensive level proportion of the failure rates of the power grid equipment in different areas, the operation and maintenance of the power grid equipment in the next year is invested into total capital for distribution.
2. The method for allocating the operation and maintenance cost of the power grid based on the equipment state and the operation age according to claim 1, wherein the specific steps of S1 are as follows:
s1.1, collecting historical statistical data of the power grid equipment;
s1.2, introducing an equipment health index concept, combining historical statistical data of the power grid equipment with technical standards and monitoring operation and maintenance information of the power grid equipment to obtain deduction values of all parts of the power grid equipment, and summarizing the deduction values of all the parts belonging to the same power grid equipment to form a health index of the power grid equipment.
3. The method according to claim 1, wherein in S2, the identification of the second influencing element is mainly implemented by a fishbone diagram method system.
4. The method for allocating operation and maintenance costs of the power grid based on the equipment status and the operating life according to claim 1, wherein in S2, the measurement and analysis of the degree of influence of the second influencing element on the equipment fault of the power grid are mainly realized by a grey correlation degree analysis method.
5. The method for allocating the operation and maintenance cost of the power grid based on the equipment state and the operation life according to claim 4, wherein the specific steps of measuring and analyzing the influence degree of the second influence element on the fault of the power grid equipment through the grey correlation degree are as follows:
s2.1, determining a reference sequence and a comparison sequence, wherein the reference sequence is a data sequence of the fault rate of the power grid equipment fault, and the comparison sequence is a data sequence of a second influence element of the power grid equipment fault;
s2.2, carrying out dimensionless treatment on the reference sequence and the comparison sequence to obtain a reference number sequence X 0 ＝{x 0 (t) } and comparison series X i ＝{x i (t)}；
S2.3, calculating the reference sequence X 0 ＝{x 0 (t) } and a comparison series X i ＝{x i (t) } comparison correlation coefficient δ oi (j)：
Δ 0i (j)＝|x 0 (j)-x i (j)|；
Δ min ＝min i min j Δ 0i (j),Δ max ＝max i max j Δ 0i (j)；
n is the total time, k is the number of second influencing elements, delta oi (j) To normalize the parameter, Δ min As a minimum normalization parameter, Δ min As maximum normalized parameter, x 0 (j) Is the size of the reference series at time t = j, x i (t) the magnitude ρ of the ith comparison series at time t = j represents the resolution coefficient;
s2.4, drawing a comparison correlation coefficient curve of the fault rate of the certain second influence element and the power grid equipment by taking the abscissa as time t and the ordinate as the comparison correlation coefficient, obtaining the correlation area of a comparison sequence according to the comparison correlation coefficient curve, and then calculating the correlation degree of the fault rate of the certain second influence element and the power grid equipment:
S 00 indicating the area of association of the reference number series, S 0i Representing comparison seriesThe correlation coefficient corresponding to the correlation area of the reference number sequence is 1;
s2.5, sequencing the relevance of all second influence elements and the fault rate of the power grid equipment, selecting the second influence elements positioned at the front A position in the sequence as preferred second influence elements, and combining the preferred second influence elements with the first influence elements as input variables of the power grid equipment fault prediction model.
6. The method for allocating operation and maintenance costs of the power grid based on the equipment state and the operating life according to claim 1, wherein in S3, parameters of the power grid equipment fault prediction model are optimized specifically by a grayish wolf algorithm.
7. The power grid operation and maintenance cost distribution method based on the equipment state and the operation age according to claim 6, wherein in S3, the specific steps of constructing the power grid equipment fault prediction model based on the deep belief network are as follows:
s3.1, inputting the input variable serving as an input layer vector to a first layer RBM to finish unsupervised training;
s3.2, obtaining feature data after completing feature learning, inputting the obtained feature data serving as an input variable of a new layer into a RBM of a next layer, and continuing unsupervised training;
and S3.3, repeating the S3.2 until the RBM of each layer is trained and learned, and taking the features obtained in the last layer of RBM as output features.
8. The method for allocating operation and maintenance costs of the power grid based on the equipment status and the operating life according to claim 7, wherein in S3, the specific steps of optimizing the parameters of the power grid equipment fault prediction model through the grayling algorithm are as follows:
s3.4, initializing parameters of a grey wolf algorithm, wherein the parameters of the grey wolf algorithm comprise the size of a population, the maximum iteration times and the position of the initialized population, and the population is the output characteristic of the power grid equipment fault prediction model;
s3.5, performing opposite search, judging by taking the error rate of the power grid equipment fault prediction model as a fitness value, if the fitness of the obtained opposite individual is superior to that of the original individual, generating an initial wolf pack by using the opposite individual, and otherwise, generating the initial wolf pack by using the original individual, wherein the individual is the fault rate output by the power grid equipment fault prediction model;
and S3.6, carrying out grade division on the individuals according to the fitness, taking the first three ranked individuals to guide the candidate solution of the fault rate output by the power grid equipment fault prediction model to carry out position updating, and obtaining the optimal power grid equipment fault prediction model after the updating is finished.
9. The method for allocating the operation and maintenance cost of the power grid based on the equipment state and the operation age according to claim 1, wherein the specific calculation mode of S4 is as follows:
wherein, gamma represents the comprehensive level value of the fault rates of all the power grid equipment in a certain area, and lambda g And a fault rate value predicted by the g-th type of power grid equipment through the optimal power grid equipment fault prediction model is represented, d represents the type number of the power grid equipment, and f is the number of the power grid equipment.
10. The method according to claim 9, wherein in step S5, the calculation method for allocating the total capital invested for the next-year power grid equipment operation and maintenance based on the equipment status and the operating life is as follows according to the comprehensive level ratio of the failure rates of the power grid equipment in different regions:
wherein phi r Representing the operation and maintenance of the power grid equipment in the next year of the r-th areaInvesting capital allocation amount; a represents the total scale of capital investment of 'plate' of the operation and maintenance of the power grid equipment in the next year in all areas; gamma ray r And (4) representing the fault rate comprehensive level value of the current power grid equipment in the r-th area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211281098.3A CN115640969A (en) | 2022-10-19 | 2022-10-19 | Power grid operation and maintenance cost distribution method based on equipment state and operation age |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211281098.3A CN115640969A (en) | 2022-10-19 | 2022-10-19 | Power grid operation and maintenance cost distribution method based on equipment state and operation age |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115640969A true CN115640969A (en) | 2023-01-24 |
Family
ID=84944761
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211281098.3A Pending CN115640969A (en) | 2022-10-19 | 2022-10-19 | Power grid operation and maintenance cost distribution method based on equipment state and operation age |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115640969A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116523509A (en) * | 2023-07-04 | 2023-08-01 | 中能聚创(杭州)能源科技有限公司 | Power monitoring analysis method and monitoring analysis system |
| CN116684823A (en) * | 2023-04-26 | 2023-09-01 | 国网浙江省电力有限公司嘉兴供电公司 | Associated equipment positioning and fault determining method and system based on equipment health codes |
| CN117557113A (en) * | 2023-10-10 | 2024-02-13 | 国网湖北省电力有限公司经济技术研究院 | Power grid operation and maintenance scheme planning method and system considering equipment characteristics |
-
2022
- 2022-10-19 CN CN202211281098.3A patent/CN115640969A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116684823A (en) * | 2023-04-26 | 2023-09-01 | 国网浙江省电力有限公司嘉兴供电公司 | Associated equipment positioning and fault determining method and system based on equipment health codes |
| CN116684823B (en) * | 2023-04-26 | 2024-02-09 | 国网浙江省电力有限公司嘉兴供电公司 | Associated equipment positioning and fault determining method and system based on equipment health codes |
| CN116523509A (en) * | 2023-07-04 | 2023-08-01 | 中能聚创(杭州)能源科技有限公司 | Power monitoring analysis method and monitoring analysis system |
| CN116523509B (en) * | 2023-07-04 | 2023-09-19 | 中能聚创(杭州)能源科技有限公司 | Power monitoring analysis method and monitoring analysis system |
| CN117557113A (en) * | 2023-10-10 | 2024-02-13 | 国网湖北省电力有限公司经济技术研究院 | Power grid operation and maintenance scheme planning method and system considering equipment characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112488395B (en) | Method and system for predicting line loss of power distribution network | |
| CN115640969A (en) | Power grid operation and maintenance cost distribution method based on equipment state and operation age | |
| CN106055918B (en) | Method for identifying and correcting load data of power system | |
| CN110222897A (en) | A kind of distribution network reliability analysis method | |
| CN112418491B (en) | Dynamic configuration method for residual sediment storage capacity of reservoir | |
| CN112990627B (en) | Power quality evaluation method | |
| CN113344589B (en) | Intelligent identification method for collusion behavior of power generation enterprise based on VAEGMM model | |
| CN116579768B (en) | Power plant on-line instrument operation and maintenance management method and system | |
| CN112529683A (en) | Method and system for evaluating credit risk of customer based on CS-PNN | |
| CN115358437A (en) | Power supply load prediction method based on convolutional neural network | |
| CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
| CN116169670A (en) | Short-term non-resident load prediction method and system based on improved neural network | |
| Wang et al. | Sustainable investment forecasting of power grids based on the deep restricted Boltzmann machine optimized by the lion algorithm | |
| Ding et al. | Conflict management-based consensus reaching process considering conflict relationship clustering in large-scale group decision-making problems | |
| CN116776209A (en) | Method, system, equipment and medium for identifying operation state of gateway metering device | |
| Gao et al. | Establishment of Economic Forecasting Model of High‐tech Industry Based on Genetic Optimization Neural Network | |
| CN107644285A (en) | The screening of power sales profitability evaluation index and Weight Determination and system | |
| CN115081551A (en) | RVM line loss model building method and system based on K-Means clustering and optimization | |
| Liu et al. | Short-term power load forecasting via recurrent neural network with similar day selection | |
| Jiang et al. | Design of Line Loss Rate Calculation Method for Low-Voltage Desk Area Based on GA-LMBP Neural Network Model | |
| Mao et al. | QoS trust rate prediction for Web services using PSO-based neural network | |
| Giovanis | Application of feed-forward neural networks autoregressive models with genetic algorithm in gross domestic product prediction | |
| Fan | Synthetic Evaluation of Power Quality Based on Prospect Theory and Membership Grade | |
| Fan et al. | WIND SPEED PROBABILISTIC FORECASTING BASED ON THE CORRELATION BETWEEN WIND PATTERNS AND PREDICTION ERRORS | |
| Yao et al. | Application of Improved GRA-TOPSIS Model in Electric Power Engineering Construction Risk Evaluation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |