Method for preventive maintenance of distributed energy supply system equipment based on entropy theory
1. A preventive maintenance method for distributed energy supply system equipment based on entropy theory is characterized in that: according to actual operation and maintenance experience, according to the electric power market demand and the actual conditions and characteristics of equipment, equipment evaluation based on a Latin sampling method is established, and a model evaluates the importance and the hazard of each subsystem and related components of distributed energy supply equipment; creating ten operation and maintenance importance influence factors and four operation and maintenance hazard evaluation factors, describing influence degrees generated by the factors quantitatively by a hierarchical analysis method, and correcting influence weights of the factors in a closed loop by combining dynamic statistical analysis of fault data to obtain grading standards and weights of the corresponding influence factors of subsystems and components of the distributed energy supply equipment; combining the equivalent indexes of equipment maintenance cost, equipment importance and fault hazard degree, and providing a preventive maintenance method for the equipment of the distributed energy supply system based on the entropy theory to determine more scientific and reasonable maintenance opportunity and period; the method comprises the following steps:
step one, establishing an equipment importance model based on a Latin sampling method;
step two, establishing an equipment hazard degree model based on a Latin sampling method;
and step three, establishing a preventive maintenance method for the distributed energy supply system equipment based on the entropy theory.
2. The method for preventive maintenance of equipment of a distributed energy supply system based on entropy theory as claimed in claim 1, wherein: the ten important degree influence factors of the operation and maintenance of the equipment are as follows: the method comprises the following steps of influencing operation and maintenance personnel and environment safety after failure of distributed energy supply equipment, influencing system functions due to failure of the distributed energy supply equipment, influencing maintenance cost due to failure of the distributed energy supply equipment, influencing outage loss, influencing equipment state monitorability, influencing downtime, influencing the difficulty level of maintenance and repair, influencing failure frequency, influencing the timeliness of supply of spare parts, and influencing the failure level of the distributed energy supply equipment due to external environment change.
3. An entropy theory-based distributed energy supply system equipment preventive maintenance method as claimed in claim 2, characterized in that: dividing the ten influence factors into 5 grades and giving interval scores, thereby quantifying the scoring standard and weight of each subsystem and component of the distributed energy supply equipment corresponding to the influence factors;
firstly, constructing a judgment matrix; the relative importance among the ten operation and maintenance importance influencing factors is represented by a judgment matrix U as follows:
the element u in the matrix represents the relative importance of the evaluation factor;
recalculate the importance of the device andsorting; calculating the maximum characteristic root lambda of the judgment matrix UmaxSubstituting into homogeneous linear equation set formula (2):
the weight for solving each influence factor is w1,w2,w3,w4,w5,w6,w7,w8,w9,w0λ is the matrix eigenvalue, then the maximum eigenvalue λmaxThe corresponding feature vector W is as shown in equation (3):
W=(w1,w2,...,w0)T (3)
in the formula (3), T is a transposed symbol of the matrix; solving the weight of the ten influencing factors; calculating the priority sequence of each influence factor according to the weight;
finally, carrying out consistency check, as shown in formula (4):
CR=CI/RI (4)
in the formula, CR is the random consistency ratio of the judgment matrix; CI is a general consistency index of the judgment matrix, and the value CI is (lambda)max-n)/(n-1); RI is an average random consistency index of the judgment matrix;
when CR is less than 0.1, the calculation result of the current judgment matrix is considered to have satisfactory consistency, and the distribution of the weight numbers of different factors is reasonable; otherwise, the judgment matrix needs to be adjusted and recalculated;
during operation, the latin sampling dynamic closed loop is used for correcting the weight of each importance evaluation factor, so that the influence of the evaluation factors is more consistent with the actual working condition of the equipment, the robustness of the importance of each subsystem and component of the equipment is enhanced, and the influence of artificial interference on the importance evaluation is reduced;
dividing the input probability distribution into a plurality of independent segments by using Latin sampling, wherein the probability of each segment being extracted is the same, and randomly extracting a sample from each segment; calculating the importance weight of each subsystem and each component of the equipment through Latin sampling, and sequencing the importance weight of each subsystem and each component;
on the basis of determining the importance weight of the equipment, calculating the importance evaluation index of each subsystem and component of the equipment by adopting a linear weighted mathematical model, wherein the calculation formula is as follows:
wherein Index is an Index for evaluation of importance, aiIs the weight of the ith factor; m isiThe index i is the number of influencing factors.
4. The method for preventive maintenance of equipment in a distributed energy supply system based on the entropy theory as claimed in claim 1, wherein the method for establishing the equipment hazard degree model based on the Latin sampling method in the second step is as follows:
calculating the fault hazard degree of the distributed energy supply equipment by combining an analytic hierarchy process and Latin sampling; determining four harmfulness influence factors of equipment operation and maintenance: the method comprises the following steps of (1) distributing energy supply equipment fault degree, equipment historical fault influence degree, equipment fault occurrence probability and equipment fault influence severity degree;
dividing the four influence factors into 5 grades and giving interval scores, thereby quantifying the scoring standard and weight of each subsystem and component of the distributed energy supply equipment corresponding to the influence factors;
first, a decision matrix is constructed. The hazard degree analysis has four operation and maintenance evaluation factors, and the relative hazard degree of each factor can be represented by the following judgment matrix B:
the element b in the matrix represents the relative harmfulness of the evaluation factor;
calculating the equipment hazard degree and sequencing; calculating the maximum characteristic root lambda of the judgment matrix BmaxSubstitution intoHomogeneous linear equation set (7):
the weight for solving each influence factor is x1,x2,x3,x4Then the maximum characteristic root λmaxCorresponding feature vector X:
X=(x1,x2,x3,x4)T (8)
obtaining a weight matrix X corresponding to each influence factor; calculating and sequencing the priority of each influence factor according to the weight;
finally, consistency check is carried out, as shown in formula (9):
CR=CI/RI (9)
in the formula, CR is the random consistency ratio of the judgment matrix; CI is a general consistency index of the judgment matrix, and the value CI is (lambda)max-n)/(n-1); RI is an average random consistency index of the judgment matrix;
when CR is less than 0.1, the calculation result of the current judgment matrix is considered to have satisfactory consistency, and the distribution of the weight numbers of different factors is reasonable; otherwise, the judgment matrix needs to be adjusted and recalculated;
during operation, the weight of each hazard evaluation factor is corrected by applying dynamic closed-loop Latin sampling, so that the influence of the evaluation factors is more consistent with the actual working condition of equipment;
dividing the input probability distribution into a plurality of independent segments by using Latin sampling, wherein the probability of each segment being extracted is the same, and randomly extracting a sample from each segment; calculating the hazard weight of each subsystem and component of the equipment through Latin sampling, and sequencing the hazard weights of each subsystem and component;
on the basis of determining the harmfulness weight of the equipment, calculating the harmfulness evaluation index of each subsystem and component of the equipment by adopting a linear weighted mathematical model, wherein the calculation formula is as follows:
wherein Index2 is a risk evaluation Index, fjThe weight of the jth hazard evaluation factor; gjThe index j is the number of the influencing factor.
5. The method for preventive maintenance of the distributed energy supply system based on the entropy theory as claimed in claim 1, wherein the step three of establishing the method for preventive maintenance of the distributed energy supply system equipment based on the entropy theory comprises the following steps:
comprehensively considering equivalent indexes of component maintenance cost, importance and harmfulness, establishing a preventive maintenance method for distributed energy supply system equipment based on an entropy theory, and solving an optimal maintenance period S;
the preventive maintenance model of the distributed energy supply system equipment based on the entropy theory is shown as a formula (11):
minIn(S)=e1(S)zn1(S)+e2(S)zn2+e3(S)zn3 (11)
in the formula, Zn1(S) is the repair cost for the nth component at time S corresponding to the cost optimized model; zn2Is the corresponding importance, Z, of the nth component as determined by the Latin sampling methodn3Determining a corresponding criticality for the nth component according to a latin sampling method; e.g. of the type1(S)、e2(S)、e3(S) weights of the respective factors;
wherein the cost optimization model is as follows:
in the formula, c1For after-repair loss, c2For periodic maintenance loss, F(s) is a cumulative failure probability function, R(s) is a reliability function, Zn1(S) minimum preventive maintenance flowerThe cost is minimum, and the corresponding time S is the economic optimal maintenance period;
according to the principle of entropy theory, in the formula (11),
k=1/ln(n) (13)
in the formula, k is a calculation coefficient; pn1(S) the maintenance cost of the nth component at the time of S is in proportion to the sum of the importance, the hazard and the maintenance cost; pn2(S) is the ratio of the importance of the nth component to the sum of the importance, the hazard and the maintenance cost at time S, Pn3(S) the hazard degree of the nth component at the time of S is in proportion to the sum of the importance degree, the hazard degree and the maintenance cost;
e1(S)=(1-Q1(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (20)
e2(S)=(1-Q2(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (21)
e3(S)=(1-Q3(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (22)
in the formula, Q1(S)、Q2(S)、Q3(S) are each independently of the other e1(S)、e2(S)、e3(S) entropy of each risk factor;
therefore, the time S corresponding to the minimum value of the equipment maintenance decision index I (S) is obtained, and the optimal maintenance time interval S for ensuring economy, reliability and availability is obtained.
Background
With the continuous promotion of the development strategy of the national clean energy, the industry of the distributed energy supply system is continuously developed, the variety and the complexity of distributed energy supply equipment are increased, and how to carry out equipment maintenance work makes the equipment reliable and economic operation become a research hotspot. Distributed energy supply equipment can gradually degrade in the long-term operation process, the system parts are complex in process and expensive in price, the parts of the system are easy to generate linkage influence, and the possible consequences caused by damage or failure are very severe.
However, the maintenance of the distributed energy supply equipment at present excessively depends on the experience of operation and maintenance workers and data provided by original manufacturers, and the decision on specific problems occurring on the site is poor. On the other hand, the influence of different influence factors on the maintenance strategy is not considered in the traditional reliability-centered maintenance mode, and the change of corresponding maintenance decision making along with the change of the equipment running time is ignored. Therefore, the conventional reliability-centered maintenance method cannot fully meet the actual maintenance work requirement.
Preventative maintenance helps keep the equipment in operation through regular, routine maintenance, thereby avoiding costly losses from unplanned outages or malfunctions. Therefore, preventive maintenance is very important to ensure the normal operation of the system and avoid the loss caused by the production stop accident.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a distributed energy supply system preventive maintenance method based on an entropy theory and taking reliability as a center. The method provided by the invention optimizes a maintenance decision model of the distributed energy supply system with reliability as the center by applying an entropy theory, and makes more scientific and reasonable maintenance opportunity and period according to the actual engineering, thereby solving the over-maintenance and under-maintenance phenomena of the distributed energy supply equipment caused by unreasonable maintenance opportunity and period.
The equipment preventive maintenance method based on the entropy theory for the distributed energy supply system establishes an equipment evaluation model based on a Latin sampling method according to actual operation and maintenance experience and according to the electric power market demand and the actual conditions and characteristics of the equipment, and evaluates the importance and the hazard of each subsystem and related components of the distributed energy supply equipment; creating ten operation and maintenance importance influence factors and four operation and maintenance hazard evaluation factors, describing influence degrees generated by the factors quantitatively by a hierarchical analysis method, and correcting influence weights of the factors in a closed loop by combining dynamic statistical analysis of fault data to obtain grading standards and weights of the corresponding influence factors of subsystems and components of the distributed energy supply equipment; on the basis, equipment maintenance cost is introduced, a preventive maintenance cost model based on Weibull distribution is established, a preventive maintenance method for equipment of the distributed energy supply system is provided based on the entropy theory by combining equivalent indexes of the equipment maintenance cost, the equipment importance and the fault hazard degree, and more scientific and reasonable maintenance opportunities and periods are determined, and the specific steps are as follows:
the method comprises the following steps: establishing equipment importance model based on Latin sampling method
According to actual operation and maintenance experience, ten influencing factors of the equipment operation and maintenance are established according to the electric power market demand and the actual condition and characteristics of the equipment: the method comprises the following steps of influencing operation and maintenance personnel and environment safety after failure of distributed energy supply equipment, influencing system functions due to failure of the distributed energy supply equipment, influencing maintenance cost due to failure of the distributed energy supply equipment, influencing outage loss, influencing equipment state monitorability, influencing downtime, influencing the difficulty level of maintenance and repair, influencing failure frequency, influencing the timeliness of supply of spare parts, and influencing the failure level of the distributed energy supply equipment due to external environment change.
In each item of influence factor analysis, in order to reduce the complexity of the equipment importance degree analysis, an analytic hierarchy process is adopted, and the ten influence factors are divided into 5 grades and given interval scores, so that the scoring standards and weights of the corresponding influence factors of each subsystem and each component of the distributed energy supply equipment are quantized, and the method specifically comprises the following steps:
first, a decision matrix is constructed. Analyzing the importance of ten operation and maintenance evaluation factors, wherein the relative importance among the factors can be represented by a judgment matrix U as follows:
the element u in the matrix represents the relative importance of the evaluation factor.
The device importance is recalculated and ranked. Calculating the characteristic root of the matrix A, taking lambda as the characteristic value corresponding to the matrix, and calculating the maximum characteristic value lambdamaxSubstituting into homogeneous linear equation set formula (2):
the weight for solving each influence factor is w1,w2,w3,w4,w5,w6,w7,w8,w9,w0To obtain the maximum eigenvalue lambdamaxA corresponding feature vector W;
W=(w1,w2,...,w0)T (3)
in the formula (3), T is a transposed symbol of the matrix; weights of the ten influencing factors are obtained, so that the influence generated by each factor is quantized. And calculating the priority of each influencing factor according to the weight and sorting.
Finally, consistency check is performed, as shown in formula (4)
CR=CI/RI (4)
In the formula, CR is the random consistency ratio of the judgment matrix; CI is a general consistency index of the judgment matrix, and the value is as follows: CI ═ λmax-n)/(n-1); RI is the average random of the decision matrixThe RI value of the machine consistency index is shown in the following table, and n is the number of importance evaluation factors.
n
1
2
3
4
5
n
6
7
8
9
10
RI
0
0
0.52
0.89
1.12
RI
1.26
1.36
1.41
1.46
1.49
When CR is less than 0.1, the calculation result of the current judgment matrix is considered to have satisfactory consistency, and the distribution of the weight numbers of different factors is reasonable; otherwise, the judgment matrix needs to be adjusted, for example, the judgment of the inverse M-matrix and the parallel algorithm are applied, and the recalculation is carried out.
During operation, the latin sampling can be used to dynamically modify the weight of each importance evaluation factor in a closed loop, so that the evaluation factor influences more conform to the actual working conditions of the equipment, the robustness of the importance of each subsystem and component of the equipment is enhanced, and the influence of human interference on the importance evaluation is reduced.
The input probability distribution is divided into a plurality of independent segments by using Latin sampling, the probability of each segment being extracted is the same, and samples are randomly extracted from each segment. And analyzing the importance weight of each subsystem and each component of the equipment through Latin sampling statistics, and effectively sequencing the importance weight of each subsystem and each component.
On the basis of determining the importance weight of the equipment, calculating the importance evaluation index of each subsystem and component of the equipment by adopting a linear weighted mathematical model, wherein the calculation formula is as follows:
wherein Index is an Index for evaluation of importance, aiIs the weight of the ith factor; m isiThe index i is the number of influencing factors.
Step two: establishing equipment hazard degree model based on Latin sampling method
And calculating the fault hazard degree of the distributed energy supply equipment by combining an analytic hierarchy process and Latin sampling. Determining four harmfulness influence factors of equipment operation and maintenance: the method comprises the steps of distributed energy supply equipment fault degree, equipment historical fault influence degree, equipment fault occurrence probability and equipment fault influence severity degree.
And dividing the four influence factors into 5 grades and giving interval scores, thereby quantifying the scoring standard and weight of each subsystem and component of the distributed energy supply equipment corresponding to the influence factors.
First, a decision matrix is constructed. The relative criticality among the four criticality-affecting factors is represented by the following decision matrix B:
the element b in the matrix represents the relative criticality of the evaluation factor.
And calculating the equipment hazard degree and sequencing. Calculating the maximum characteristic root lambda of the judgment matrix BmaxSubstituting into homogeneous linear equation set formula (7):
the weight for solving each influence factor is x1,x2,x3,x4To obtain the maximum characteristic root lambdamaxCorresponding feature vector X:
X=(x1,x2,x3,x4)T (8)
the weight of each influencing factor can be obtained. Thereby quantifying the effect of each factor. And calculating the priority of each influencing factor through the weight and sorting.
Finally, carrying out consistency test, as formula (9):
CR=CI/RI (9)
in the formula, CR is referred to as a random consistency ratio of the judgment matrix; CI is a general consistency index of the decision matrix, and its value CI ═ λmax-n)/(n-1); the RI is called the average random consistency index of the judgment matrix and takes the values as the following table.
n
1
2
3
4
RI
0
0
0.52
0.89
When CR is less than 0.1, the calculation result of the current judgment matrix is considered to have satisfactory consistency, and the distribution of the weight numbers of different factors is reasonable; otherwise, the judgment matrix needs to be adjusted and recalculated.
During operation, the weight of each hazard degree evaluation factor can be dynamically corrected in a closed loop mode by applying Latin sampling, so that the influence of the evaluation factors is more consistent with the actual working condition of the equipment, the robustness of the importance of each subsystem and part of the equipment is enhanced, and the influence of artificial interference on hazard degree evaluation is reduced.
The input probability distribution is divided into a plurality of independent segments by using Latin sampling, the probability of each segment being extracted is the same, and samples are randomly extracted from each segment. And analyzing the hazard weight of each subsystem and component of the equipment through Latin sampling statistics, and effectively sequencing the hazard weight of each subsystem and component.
On the basis of determining the harmfulness weight of the equipment, calculating the harmfulness evaluation index of each subsystem and component of the equipment by adopting a linear weighted mathematical model, wherein the calculation formula is as follows:
wherein Index2 is a risk evaluation Index, fjThe weight of the jth hazard evaluation factor; gjThe index j is the number of the influencing factor.
Step three: method for establishing preventive maintenance of distributed energy supply system equipment based on entropy theory
The method comprehensively considers equivalent indexes of component maintenance cost, importance and harmfulness, establishes a preventive maintenance method for the distributed energy supply system equipment based on the entropy theory, obtains the optimal maintenance period S, and ensures the economy, reliability and availability of the distributed energy supply system equipment.
The preventive maintenance model of the distributed energy supply system equipment based on the entropy theory is shown as a formula (11):
minIn(S)=e1(S)zn1(S)+e2(S)zn2+e3(S)zn3 (11)
in the formula, Zn1(S) is the repair cost for the nth component at time S corresponding to the cost optimized model; zn2Is the corresponding importance, Z, of the nth component as determined by the Latin sampling methodn3The equipment hazard degree; e.g. of the type1(S)、e2(S)、e3And (S) weights of maintenance cost, importance and hazard degree when the cost optimal model is used.
Wherein the cost optimization model is as follows:
in the formula, c1For post repair losses; c. C2Loss for periodic maintenance; f(s) is a cumulative probability of failure function; r(s) is a reliability function. Is easy to know, Zn1(S) minimum preventive maintenance costsAnd at least, the corresponding time S is the economic optimal maintenance period.
According to the principle of entropy theory, in the formula (11),
k=1/ln(n) (13)
in the formula, k is a calculation coefficient; pn1(S) the maintenance cost of the nth component at the time of S is in proportion to the sum of the importance, the hazard and the maintenance cost; pn2(S) is the ratio of the importance of the nth component to the sum of the importance, the hazard and the maintenance cost at time S, Pn3(S) is the ratio of the criticality of the nth component at time S to the sum of the importance, the criticality, and the maintenance cost.
e1(S)=(1-Q1(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (20)
e2(S)=(1-Q2(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (21)
e3(S)=(1-Q3(S))/[(1-Q1(S))+(1-Q2(S))+(1-Q3(S))] (22)
In the formula, Q1(S)、Q2(S)、Q3(S) are each independently of the other e1(S)、e2(S)、e3Entropy of each risk factor of (S).
Therefore, the time S corresponding to the minimum value of the equipment maintenance decision index I (S) is obtained, and the optimal maintenance time interval S for ensuring economy, reliability and availability is obtained.
Drawings
FIG. 1 is a flow chart of a method for preventive maintenance of distributed energy supply equipment based on entropy theory according to the invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
As shown in FIG. 1, the embodiment of the method for preventive maintenance of distributed energy supply system equipment based on entropy theory of the invention comprises the following steps:
1. establishing equipment importance model based on Latin sampling method
According to actual operation and maintenance experience, ten important influence factors of the equipment operation and maintenance are established according to the electric power market demand and the actual condition and characteristics of the equipment: the method comprises the following steps of influencing operation and maintenance personnel and environment safety after failure of distributed energy supply equipment, influencing system functions due to failure of the distributed energy supply equipment, influencing maintenance cost due to failure of the distributed energy supply equipment, influencing outage loss, influencing equipment state monitorability, influencing downtime, influencing the difficulty level of maintenance and repair, influencing failure frequency, influencing the timeliness of supply of spare parts, and influencing the failure level of the distributed energy supply equipment due to external environment change. In each influence factor analysis, in order to reduce the complexity of the equipment importance degree analysis, a hierarchical analysis is adopted for analysis. And dividing the ten influence factors into 5 grades and giving interval scores, thereby quantifying the scoring standard and weight of each subsystem and component of the distributed energy supply equipment corresponding to the influence factors.
In the formula, A is an importance matrix, and an element u in the matrix represents the relative importance of the evaluation factor; lambda is a characteristic value corresponding to the matrix; w is the importance weight; CR is a judgment matrix consistency ratio; CI is a general consistency index of the judgment matrix; RI is an index for judging the average random consistency of the matrix; index is an importance evaluation Index, aiIs the weight of the ith factor; m isiThe index i is the number of influencing factors.
2. Establishing equipment hazard degree model based on Latin sampling method
And calculating the fault hazard degree of the distributed energy supply equipment by combining an analytic hierarchy process and Latin sampling. Determining four harmfulness influence factors of equipment operation and maintenance: the method comprises the steps of distributed energy supply equipment fault degree, equipment historical fault influence degree, equipment fault occurrence probability and equipment fault influence severity degree. And dividing the four influence factors into 5 grades and giving interval scores, thereby quantifying the scoring standard and weight of each subsystem and component of the distributed energy supply equipment corresponding to the influence factors.
Wherein B is a harmfulness matrix; the element b in the matrix represents the relative harmfulness of the evaluation factor; x is a hazard weight; CR is a judgment matrix consistency ratio; CI is a general consistency index of the judgment matrix; RI is an index for judging the average random consistency of the matrix; index2 is a hazard rating Index, fiIs the weight of the ith factor; giThe index j is the number of influencing factors.
3. Method for establishing preventive maintenance of distributed energy supply system equipment based on entropy theory
The preventive maintenance method for the distributed energy supply system equipment is provided based on the entropy theory, the equivalent indexes of the component maintenance cost, the importance degree and the fault hazard degree are comprehensively considered, the maintenance optimal period S is obtained, and the economy, the reliability and the availability of the distributed energy supply system equipment are guaranteed.
In the formula, Zn1(S) is the repair cost for the nth component at time S corresponding to the cost optimized model; zn2Is the corresponding importance determined by the nth component according to the latin sampling method; zn3The equipment hazard degree; e.g. of the type1(S)、e2(S)、e3(S) weights of the respective factors; c. C1For post repair losses; c. C2Loss for periodic maintenance; f(s) is a cumulative probability of failure function; r(s) is a reliability function; k is a calculation coefficient; pn1(S) the maintenance cost of the nth component at the time of S is in proportion to the sum of the importance, the hazard and the maintenance cost; pn2(S) the importance of the nth component at the time of S is in proportion to the sum of the importance, the hazard and the maintenance cost; pn3(S) the hazard degree of the nth component at the time of S is in proportion to the sum of the importance degree, the hazard degree and the maintenance cost; q1(S)、Q2(S)、Q3(S) are each independently of the other e1(S)、e2(S)、e3Entropy of each risk factor of (S).