Preventive maintenance decommissioning plan collaborative optimization method for aged transformer

Through the fuzzy entropy weight method and Weibull distribution model combined with particle swarm-simulated annealing algorithm, the preventive maintenance and decommissioning plan of the transformer was coordinated to optimize the problem of unreasonable maintenance and decommissioning plans in the existing technology, and the system reliability and economics were achieved balance of sexuality and reasonable management of aging transformers.

CN119089778APending Publication Date: 2024-12-06CHONGQING HECHUAN POWER GENERATION CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411176740.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-26
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

When formulating transformer maintenance and decommissioning plans, the existing technology ignores the decision-making relationship between maintenance and decommissioning plans. The aging process modeling is not accurate enough, and focuses on a single transformer, which ignores the connection between the transformer group and the system, resulting in maintenance. The plan and the retirement plan are unreasonable.

Method used

The importance of aging transformers is sorted by the fuzzy entropy weight method, and the aging failure model is established based on the correctable three-parameter Weibull distribution model, combined with the system reliability constraints to perform two-layer collaborative optimization of preventive maintenance and decommissioning plans, using particle swarm-simulated annealing algorithm Solve to minimize the total cost.

Benefits of technology

It realizes more accurate ranking of importance of aging transformer and aging failure reliability assessment, provides reasonable maintenance and decommissioning decisions, meets system reliability and economic requirements, and improves the scientificity and accuracy of equipment management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089778A_ABST
    Figure CN119089778A_ABST
Patent Text Reader

Abstract

The invention relates to the field of equipment maintenance methods, in particular to a preventive maintenance decommissioning plan collaborative optimization method for aged transformers, which comprises the following steps of: obtaining importance indexes of each aged transformer by adopting a fuzzy entropy weight method in combination with the structural vulnerability of the transformer and the contribution degree of aging failure to the risk of the transformer, and sequencing the importance indexes; establishing a transformer aging failure model based on the transformer aging failure of the correctable three-parameter distribution model, and dynamically correcting an aging threshold value and expected life in the model according to preventive overhaul and decommissioning time; according to an aged transformer preventive maintenance and decommissioning plan double-layer collaborative optimization model constrained by system reliability, the preventive maintenance time and decommissioning time of each aged transformer in a plan period are determined, the purpose of minimizing the total cost expenditure in the whole plan period is achieved, and the system reliability constraint is met at the same time; and solving by adopting a particle swarm-simulated annealing algorithm to obtain an optimal decision scheme. The method can cooperatively consider maintenance and decommissioning to make decisions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of equipment maintenance methods, and in particular to a collaborative optimization method for a preventive maintenance and decommissioning plan for an aging transformer. Background Art

[0002] In order to comprehensively improve the equipment and overall reliability level in the power system, it is necessary to manage the equipment throughout its life cycle, including maintenance and decommissioning. In the power system, the transformer is the core equipment for realizing power transmission and conversion, and its reliability directly affects the safety and stability of the entire power generation and transmission system. Therefore, it is very important to formulate a scientific and reasonable transformer maintenance and decommissioning plan that takes into account the reliability and economy of the system operation.

[0003] At present, relatively mature technical solutions have been formed at home and abroad on how to determine the transformer maintenance strategy and retirement strategy, but most of the technical solutions are to optimize the transformer maintenance plan and retirement plan separately. Such technical solutions have the following shortcomings:

[0004] (1) In terms of decision-making models: existing solutions usually ignore the decision-making dependency between maintenance and retirement plans (i.e., maintenance can improve the reliability of transformers and thus delay their retirement time), and may not be able to obtain a global optimal solution;

[0005] (2) In terms of decision analysis: the existing solutions do not accurately model the transformer aging process, and most of the qualitative analysis methods such as the service age regression factor are used to improve the reliability of aging transformers after maintenance and retirement, which cannot quantify the impact of maintenance and retirement measures from the degradation mechanism level;

[0006] (3) In terms of decision-making objects: The decision-making object is usually a single transformer, focusing on analyzing the impact of individual transformers on system reliability and economy, while ignoring the connection between the transformer group and the system. This ultimately leads to unreasonable maintenance and retirement plans for transformers. Summary of the invention

[0007] The present invention aims to provide a collaborative optimization method for preventive maintenance and decommissioning plans for aging transformers, so as to solve the problem of unreasonable results caused by separate processing of maintenance plans and decommissioning plans.

[0008] The collaborative optimization method for preventive maintenance and decommissioning planning of aging transformers in this solution includes:

[0009] Step 1: Combining the structural vulnerability of the transformer and the contribution of aging failure to the transformer risk, the fuzzy entropy weight method is used to obtain the importance index of each aging transformer and rank them;

[0010] Step 2, establishing a transformer aging failure model based on the calibrable three-parameter Weibull distribution model, and dynamically correcting the two parameters of the aging threshold and expected life in the model according to the preventive maintenance and retirement time;

[0011] Step 3: According to the two-level collaborative optimization model of preventive maintenance and retirement planning of aging transformers constrained by system reliability, the preventive maintenance time and retirement time of each aging transformer in the future planning period are determined, with the goal of minimizing the total cost expenditure in the entire planning period while satisfying the system reliability constraints;

[0012] Step 4: Use the particle swarm-simulated annealing algorithm to solve and obtain the optimal decision-making solution.

[0013] The beneficial effects of this program are:

[0014] Through the aging transformer importance ranking model based on the fuzzy entropy weight method; and based on the correctable three-parameter Weibull distribution model, a transformer aging failure reliability model taking into account the impact of preventive maintenance and decommissioning was established; a two-layer collaborative optimization model of preventive maintenance and decommissioning plans for aging transformers considering system reliability was constructed; finally, the particle swarm-simulated annealing algorithm was used to solve the proposed optimization model, combined with the consideration of preventive maintenance plans and decommissioning plans for target solving, which can provide system operation and maintenance personnel with a more reasonable decision-making basis for reliability-centered aging transformer group asset management, making the maintenance plan and decommissioning plan results more accurate and in line with the actual equipment needs.

[0015] Furthermore, in step 1, the electrical intermediate index is used to characterize the structural vulnerability of each aging transformer, and the contribution of each transformer aging failure to the system risk is used as a quantitative index to measure its impact on the system redundancy. The specific steps are:

[0016] Step 1.1, the formula for calculating the electrical dielectric index of each aging transformer in the system is expressed as:

[0017]

[0018] Among them, G Represents the set of unit nodes in the system; Ψ D Represents the set of load nodes in the system; Represents transformer T n The structural vulnerability index of n The first and last nodes of the branch; g and d represent the two node numbers in a single "unit-load" node pair (g, d); Represents the transformer T based on the node pair (g, d) n The flow betweenness index is calculated as follows:

[0019]

[0020] Among them, W g W represents the weight of the power generation node g, which is the rated capacity of the generator; d Represents the weight of load node d, which is the peak load of the node; I gd (i, j) indicates that after adding a unit injection current element between the "unit-load" node pair (g, d), the current in the transformer T n The current caused in the branch;

[0021] Step 1.2, the formula for calculating the contribution of aging failure of each aging transformer in the system to the system risk is expressed as:

[0022]

[0023] in, and They are respectively the transformer T before and after aging failure n load loss risk indicator;

[0024] Step 1.3, the weight formula for calculating the two indicators of structural vulnerability and contribution based on the fuzzy entropy weight method is expressed as:

[0025]

[0026] In the formula, ω l represents the fuzzy entropy weight of the lth indicator; Ь1 and Ь2 represent the preference weights of system operators for system structural vulnerability and adequacy respectively; IE1 and IE2 represent the structural vulnerability index H 1 and risk contribution index H 2 The information entropy of nl It represents the normalized value of the lth index of the nth transformer, and the calculation formula is expressed as:

[0027]

[0028] Where: N atl Indicates the number of aging transformers in the system;

[0029] Step 1.4, calculate and output the importance evaluation index of each transformer

[0030]

[0031] The beneficial effect is that by considering the structural vulnerability of the transformer and the degree of influence of aging failure on the risk contribution of the transformer, the parameters carrying the most characteristics can be accurately screened out for analysis.

[0032] Furthermore, in step 1, after obtaining the importance evaluation index of each transformer, they are sorted according to the index value, and the aging transformers with the highest sorting results are selected as the decision-making objects for subsequent preventive maintenance and decommissioning plans.

[0033] The beneficial effect is that the decision object is determined according to the order after parameter processing, thereby improving the accuracy of determining the decision object.

[0034] Further, the step 2 includes the following sub-steps:

[0035] Step 2.1, the calibrated three-parameter Weibull distribution model can be used to describe the quantitative relationship between the time-varying aging failure probability of the transformer and its aging threshold and expected life, which can be expressed as:

[0036]

[0037] In the formula, γ n With η n Transformer T n The aging threshold and expected life of the device are: c1, c2 and β are constant parameters obtained by fitting the failure statistics data;

[0038] Step 2.2, according to the execution time of preventive maintenance and decommissioning measures, determine the transformer aging threshold and life update value after preventive maintenance and decommissioning, substitute the updated aging threshold and expected life parameters into the aging failure model of step 2.1, and you can get the transformer aging failure probability function taking into account the impact of preventive maintenance and decommissioning.

[0039] The beneficial effect is that the quantitative calculation of various parameters can quantify the impact of maintenance and retirement at different times on aging transformers, so as to accurately evaluate the working condition of the equipment.

[0040] Furthermore, in step 2.2, the transformer T n The expected aging failure probability function after taking into account the impact of preventive maintenance and decommissioning in the future period is expressed as:

[0041]

[0042] In the formula, the aging transformer T n T has been running ser,n Year, and They represent the aging threshold and life expectancy evaluated at the current moment respectively; Indicates that in the T ser + D1 year estimated life expectancy after preventive maintenance; and Indicates that in the T ser+D2 years of ageing thresholds and life expectancy assessments after preventive decommissioning operations.

[0043] The beneficial effect is that through the quantitative definition of the formula, the transformer T n The expected aging failure probability function under the “maintenance before retirement” strategy. In engineering practice, it is also possible to perform maintenance on the transformer T in the future planning period. n Implement differentiated asset management strategies such as "neither maintenance nor retirement", "only maintenance" and "only retirement". Transformer T under any management strategy n The expected aging failure probability functions are obtained by modifying the above formula, which improves the accuracy and timeliness of strategy formulation and meets actual needs.

[0044] Further, the step 3 includes the following sub-steps:

[0045] Step 3.1, the calculation formula for the total cost in the future planning period is:

[0046]

[0047] Where K represents the total number of years of the planning period; N IM It represents the number of aging transformers to be decided within the planning period selected by importance sorting results; and They represent the transformer T in the kth year of the planning period. n The expected investment costs required to implement preventive maintenance and decommissioning measures; Represents transformer T n The disposal cost of random failures occurring in year k; Represents transformer T n Disposal costs incurred when aging failure occurs in the kth year; represents the reliability cost of the system in the kth year of the planning period;

[0048] Step 3.2, the constraints of the upper-level preventive maintenance and decommissioning plan decision model include maintenance time constraints, decommissioning time constraints, and maintenance-decommissioning coupling constraints, which are expressed as:

[0049] If the transformer T n There is a preventive maintenance plan in the future planning period, and its maintenance time d ohl,n Should satisfy: 0 <d ohl,n ≤K;

[0050] If the transformer T n There is a preventive decommissioning plan in the future planning period, and its decommissioning time is d ret i re,n Should satisfy: 0 <d retire,n ≤K;

[0051] If the transformer has both preventive maintenance and decommissioning plans, the maintenance time d ohl,n Retirement time retire,n The following should be satisfied: retire,n ≥d ohl,n +d set,val , d se t ,va l Indicates the acceptable preventive maintenance-retirement time interval;

[0052] The maintenance and decommissioning resource constraints are expressed as:

[0053]

[0054] Where: num ohl and num retire The upper limit of the number of transformers that can be subject to preventive maintenance and decommissioning measures each year respectively;

[0055] The expected load loss EENS index is used to construct the system reliability constraint, which is expressed as: EENS k ≤EENS pre ;

[0056] Step 3.3, the construction method of the lower layer reliability verification model is used to calculate the optimal average load reduction of the system in each year corresponding to the upper layer decision-making scheme, which is expressed as:

[0057]

[0058] Where: Q sam,k represents the total number of simulation cycles performed for reliability assessment in the kth year of the planning period; Θ q represents the set of all system states in the qth evaluation cycle; d q,s represents the state duration of the system state s in the qth evaluation cycle; and represents the optimal load reduction interval [LC q,s The upper and lower limits of ] are obtained by solving the optimal load reduction interval analysis model, and its compact form is expressed as:

[0059] min[LC]=[D][X];

[0060] st[A][X]-[B](≤)(=,≤)0;

[0061] [X]≥0;

[0062] Where: [X] = ([X ij ] n×1) represents the n-dimensional decision variable of the objective function, whose elements D represents the coefficient matrix of [X] in the objective function; A represents the coefficient matrix of the constraint condition; [B] = ([B ij ] m×1 ), whose elements Indicates the interval value of each node load; [LC] = [LC - ,LC + ], which represents the optimal value interval of load reduction under a given system state.

[0063] The beneficial effect is that by considering the double-layer collaborative optimization scheme of the maintenance and decommissioning plan of the aging transformer, the optimal maintenance and decommissioning time can be obtained, so that the system can simultaneously meet the reliability operation and the optimal economic expenditure.

[0064] Furthermore, in step 3.1, the expected investment cost calculation formula for the preventive maintenance is:

[0065]

[0066] Where: θ is the discount rate; Represents a 0-1 indicator variable. If T n There is a preventive maintenance plan in year k, The value is 1, otherwise it is 0; It indicates the probability that the preventive maintenance plan can be implemented; Indicates that the moisture content of transformer insulation paper is reduced by Δω aver the costs required;

[0067] The expected investment cost of preventive decommissioning is calculated as:

[0068]

[0069] Where: Represents a 0-1 indicator variable. If the transformer T n When there is a preventive decommissioning plan in year k, The value is 1, otherwise it is 0; It indicates the probability that the preventive decommissioning plan can be implemented; Represents transformer T n decommissioning costs;

[0070] The formula for calculating the disposal cost of random failure is:

[0071]

[0072] Where: C repair represents the cost required to repair a single random failure; and Respectively represent transformer T n The random failure rate and expected number of random failures in the kth year of the future planning period;

[0073] The calculation formula for the disposal cost of aging failure is:

[0074]

[0075] Where: Represents transformer T n The probability of aging failure occurring in the kth year;

[0076] Reliability cost is the load loss cost caused by load reduction under a given decision plan. and penalty costs for violating the expected reliability requirements of the system The calculation formula is:

[0077]

[0078] In the formula, EENS k represents the optimal average load reduction of the system in year k under a given decision-making scheme; EENS R represents the optimal average load reduction of the system in the kth year without considering the aging and failure of transformers; u co Represents the load loss factor, which is used to measure the loss cost per unit load loss; EENS pre represents the load reduction that the system can accept each year; u pu Represents the unit load loss penalty factor.

[0079] The beneficial effect is that the quantitative calculation of various costs can comprehensively and accurately judge the economic indicators.

[0080] Furthermore, in step 4, solving the particle swarm-simulated annealing algorithm for the collaborative optimization model of preventive maintenance and retirement planning of aging transformers considering system reliability includes the following sub-steps:

[0081] Step 4.1: Input the power system topology parameters, operation parameters, component reliability parameters, and load curve; set the PSO-SA optimization algorithm parameters, including: particle population size, maximum number of iterations R set , inertia weight coefficient, learning factor and temperature attenuation coefficient;

[0082] Step 4.2, initialize the number of iterations r = 1, and randomly generate a particle population that meets the constraints;

[0083] Step 4.3, calculate the system reliability index corresponding to each particle: For each particle in the population, calculate N based on the transformer unavailability rate calculation model. IMThe comprehensive unavailability rate of the aging transformers in each year during the planning period, and the reliability index of the system in each year;

[0084] Step 4.4, calculate the fitness value of each particle: for each particle in the population, calculate the reliability cost of each year in the planning period according to the result obtained in step 3, and then add it to the expected investment cost and fault handling cost to obtain the total cost to identify the optimal position of the individual in this round of iteration;

[0085] Step 4.5: Determine whether to use the individual optimal position obtained in this round of iteration as the new global optimal position according to the Metropolis criterion; if the new solution is accepted, update the annealing temperature;

[0086] Step 4.6, determine whether r reaches the maximum number of iterations R set : If not reached, set r = r + 1, and update the speed and position of each particle in the population, and then go to step 4.3; if r ≥ R set , then the final decision solution result pg and the corresponding fitness value f(pg) are output. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 A flowchart of an embodiment of a collaborative optimization method for preventive maintenance and decommissioning planning of aging transformers;

[0088] Figure 2 This is a flowchart for solving a double-layer collaborative optimization model based on the PSO-SA algorithm in the collaborative optimization method embodiment for preventive maintenance and decommissioning plans for aging transformers. DETAILED DESCRIPTION

[0089] The following is further explained in detail through specific implementation methods.

[0090] Example

[0091] A collaborative optimization approach for preventive maintenance decommissioning planning of aging transformers, such as Figure 1 As shown, including:

[0092] Step 1: Combine the structural vulnerability of the transformer and the contribution of aging failure to the transformer risk, and use the fuzzy entropy weight method to obtain the importance index of each aging transformer and rank them. The electrical betweenness index is used to characterize the structural vulnerability of each aging transformer, and the contribution of aging failure of each transformer to the system risk is used as a quantitative indicator to measure its impact on system redundancy. The specific steps are as follows:

[0093] Step 1.1, the formula for calculating the electrical dielectric index of each aging transformer in the system is expressed as:

[0094]

[0095] Among them, G Represents the set of unit nodes in the system; Ψ D Represents the set of load nodes in the system; Represents transformer T n The structural vulnerability index of n The first and last nodes of the branch; g and d represent the two node numbers in a single "unit-load" node pair (g, d); Represents the transformer T based on the node pair (g, d) n The flow betweenness index is calculated as follows:

[0096]

[0097] Among them, W g W represents the weight of the power generation node g, which is the rated capacity of the generator; d Represents the weight of load node d, which is the peak load of the node; I gd (i, j) indicates that after adding a unit injection current element between the "unit-load" node pair (g, d), the current in the transformer T n The current caused in the branch;

[0098] Step 1.2, the formula for calculating the contribution of aging failure of each aging transformer in the system to the system risk is expressed as:

[0099]

[0100] in, and They are respectively the transformer T before and after aging failure n load loss risk indicator;

[0101] Step 1.3, the weight formula for calculating the two indicators of structural vulnerability and contribution based on the fuzzy entropy weight method is expressed as:

[0102]

[0103] In the formula, ω l represents the fuzzy entropy weight of the lth indicator; Ь1 and Ь2 represent the preference weights of system operators for system structural vulnerability and adequacy respectively; IE1 and IE2 represent the structural vulnerability index H 1 and risk contribution index H 2 The information entropy of nl It represents the normalized value of the lth index of the nth transformer, and the calculation formula is expressed as:

[0104]

[0105] Where: N atl Indicates the number of aging transformers in the system;

[0106] Step 1.4, calculate and output the importance evaluation index of each transformer

[0107]

[0108] After obtaining the importance evaluation index of each transformer, they are sorted according to the index value, and the aging transformers with higher ranking results are selected as the decision-making objects for subsequent preventive maintenance and decommissioning plans.

[0109] Step 2: Establish a transformer aging failure model based on the correctable three-parameter Weibull distribution model. Dynamically correct the two parameters of the model, aging threshold and expected life, according to the preventive maintenance and retirement time, so as to quantitatively evaluate the impact of maintenance / retirement measures implemented at different times on the reliability of the aging transformer.

[0110] The specific process includes the following sub-steps:

[0111] Step 2.1, the calibrated three-parameter Weibull distribution model can be used to describe the quantitative relationship between the time-varying aging failure probability of the transformer and its aging threshold and expected life, which can be expressed as:

[0112]

[0113] In the formula, γ n With η n Transformer T n The aging threshold and expected life of the device are: c1, c2 and β are constant parameters obtained by fitting the failure statistics data;

[0114] Step 2.2, according to the execution time of preventive maintenance and decommissioning measures, determine the transformer aging threshold and life update value after preventive maintenance and decommissioning, substitute the updated aging threshold and expected life parameters into the aging failure model of step 2.1, and you can get the transformer aging failure probability function taking into account the impact of preventive maintenance and decommissioning.

[0115] Assuming the aging transformer T n T has been running ser,n years, and the aging threshold and life expectancy evaluated at the current moment (before the start of the planning period) are and Transformer n The expected aging failure probability function after taking into account the impact of preventive maintenance and decommissioning in the future period is expressed as:

[0116]

[0117] In the formula, the aging transformer T n T has been running ser,n Year, and They represent the aging threshold and life expectancy evaluated at the current moment respectively; Indicates that in the T ser + D1 year estimated life expectancy after preventive maintenance; and Indicates that in the T ser +D2 years of ageing thresholds and life expectancy assessments after preventive decommissioning operations.

[0118] Step 3, according to the two-layer collaborative optimization model of preventive maintenance and retirement plan of aging transformers constrained by system reliability, determine the preventive maintenance time and retirement time of each aging transformer in the future planning period, with the purpose of minimizing the total cost expenditure in the entire planning period, while satisfying the system reliability constraints, so as to determine the preventive maintenance time and retirement time of each aging transformer in the future planning period, with the purpose of minimizing the total cost expenditure in the entire planning period, while satisfying the system reliability constraints.

[0119] The specific process includes the following steps:

[0120] Step 3.1, the calculation formula for the total cost in the future planning period is:

[0121]

[0122] Where K represents the total number of years of the planning period; N IM It represents the number of aging transformers to be decided within the planning period selected by importance sorting results; and They represent the transformer T in the kth year of the planning period. n The expected investment costs required to implement preventive maintenance and decommissioning measures; Represents transformer T n The disposal cost of random failures occurring in year k; Represents transformer T n Disposal costs incurred when aging failure occurs in the kth year; represents the reliability cost of the system in the kth year of the planning period;

[0123] The calculation of each expected investment cost is carried out according to the following formula. The calculation formula of the expected investment cost of preventive maintenance is:

[0124]

[0125] Where: θ is the discount rate; Represents a 0-1 indicator variable. If Tn There is a preventive maintenance plan in year k, The value is 1, otherwise it is 0; It represents the probability that the preventive maintenance plan can be implemented, that is, T n The probability that no aging failure will occur before its preventive maintenance time; Indicates that the moisture content of transformer insulation paper is reduced by Δω aver The required costs can be determined through expert experience;

[0126] The expected investment cost of preventive decommissioning is calculated as:

[0127]

[0128] Where: Represents a 0-1 indicator variable. If the transformer T n When there is a preventive decommissioning plan in year k, The value is 1, otherwise it is 0; It indicates the probability that the preventive decommissioning plan can be implemented, that is, the transformer T n The probability that no aging failure will occur before its preventive retirement time; Represents transformer T n decommissioning costs;

[0129] Random failure disposal cost refers to the transformer T n The post-maintenance cost caused by random failures during the planning period, the calculation formula for the disposal cost of random failures is:

[0130]

[0131] Where: C repair represents the cost required to repair a single random failure; and Respectively represent transformer T n The random failure rate and expected number of random failures in the kth year of the future planning period;

[0132] Aging failure disposal cost refers to the transformer T n The ex post replacement cost incurred after aging failure occurs in the future planning period. The calculation formula for the disposal cost of aging failure is:

[0133]

[0134] Where: Represents transformer T n The probability of aging failure occurring in the kth year;

[0135] Reliability cost is the load loss cost caused by load reduction under a given decision plan. and penalty costs for violating the expected reliability requirements of the system Given a decision plan determined by the upper layer of the model, that is, for each particle in the particle swarm algorithm, the calculation formula is:

[0136]

[0137]

[0138] In the formula, EENS k It represents the optimal average load reduction of the system in the kth year under a given decision-making scheme, which is solved by solving the lower model; EENS R represents the optimal average load reduction of the system in the kth year without considering the aging and failure of transformers; u co Represents the load loss factor, which is used to measure the loss cost per unit load loss; EENS pre represents the load reduction that the system can accept each year; u pu Represents the unit load loss penalty factor.

[0139] According to the provisions of the "Power Equipment Maintenance Test Regulations" (PT / CSG1206007-2017), the transformer overhaul cycle is generally more than 10 years, so each aging transformer can only be overhauled once during the study period. Furthermore, there are four types of management strategies for each aging transformer during the planning period: 1) No preventive maintenance or retirement measures; 2) Only preventive maintenance measures are arranged; 3) Only preventive retirement measures are arranged; 4) Both preventive maintenance and retirement measures exist. Therefore, the decision variables of the upper model include the preventive maintenance time and preventive retirement time of each aging transformer.

[0140] Step 3.2, the constraints of the upper-level preventive maintenance and decommissioning plan decision model include maintenance time constraints, decommissioning time constraints, and maintenance-decommissioning coupling constraints, which are expressed as:

[0141] The maintenance time constraint is: if the transformer T n There is a preventive maintenance plan in the future planning period, and its maintenance time d ohl,n Should satisfy: 0 <d ohl,n ≤K;

[0142] The retirement time constraint is: if the transformer T n There is a preventive decommissioning plan in the future planning period, and its decommissioning time is d retire,n Should satisfy: 0 <d retire,n ≤K;

[0143] Generally speaking, the maintenance plan of the transformer should be implemented before it is retired and replaced. At the same time, system operation and maintenance personnel generally do not retire the transformer that has undergone major maintenance in the short term. The maintenance-retirement coupling constraint is that if the transformer has both preventive maintenance and retirement plans, the maintenance time d ohl,n Retirement time retire,n The following should be satisfied: retire,n ≥d ohl,n +d set,val, d set,val Indicates the acceptable preventive maintenance-retirement time interval;

[0144] The maintenance and decommissioning resource constraints are expressed as:

[0145]

[0146] Where: num ohl and num retire The upper limit of the number of transformers that can be subject to preventive maintenance and decommissioning measures each year respectively;

[0147] The system reliability constraint is that the reliability of the system in each year during the future planning period should be maintained within an acceptable level. The expected load loss EENS index is used to construct the system reliability constraint, which is expressed as: EENS k ≤EENS pre ;

[0148] Step 3.3, the construction method of the lower layer reliability verification model is used to calculate the optimal average load reduction (i.e. reliability index) of the system in each year corresponding to the upper layer decision-making scheme, which is expressed as:

[0149]

[0150] Where: Q sam,k represents the total number of simulation cycles performed for reliability assessment in the kth year of the planning period; Θ q represents the set of all system states in the qth evaluation cycle; d q,s represents the state duration of the system state s in the qth evaluation cycle; and represents the optimal load reduction interval [LC q,s The upper and lower limits of ] are obtained by solving the optimal load reduction interval analysis model, and its compact form is expressed as:

[0151] min[LC]=[D][X];

[0152] st[A][X]-[B](≤)(=,≤)0;

[0153] [X]≥0;

[0154] Where: [X] = ([X ij ] n×1 ) represents the n-dimensional decision variables of the objective function, namely, the load reduction and voltage phase angle of each node in the system, the power flow of each branch, and the output of each unit. Its elements D represents the coefficient matrix of [X] in the objective function; A represents the coefficient matrix of the constraint condition; [B] = ([B ij ] m×1 ), whose elements Indicates the interval value of each node load; [LC] = [LC - ,LC + ] represents the optimal value interval of load reduction under a given system state. The constraints in the lower model include system power balance constraints, node power balance constraints, branch flow constraints, node load reduction constraints, unit output constraints, etc., which can all be expressed using existing constraints and will not be repeated here.

[0155] Step 4: Use the particle swarm-simulated annealing algorithm to solve and obtain the optimal decision-making solution. The particle swarm-simulated annealing algorithm for solving the collaborative optimization model of preventive maintenance and decommissioning plan of aging transformers considering system reliability includes the following sub-steps:

[0156] Step 4.1, such as Figure 2 As shown in the figure, input the power system topology parameters, operation parameters, component reliability parameters, and load curve; set the PSO-SA optimization algorithm parameters, including: particle population size, maximum number of iterations R set , inertia weight coefficient, learning factor and temperature attenuation coefficient;

[0157] Step 4.2, initialize the number of iterations r = 1, and randomly generate a particle population that meets the constraints (i.e. N IM Preventive maintenance / retirement time of aging transformers);

[0158] Step 4.3, calculate the system reliability index corresponding to each particle: For each particle in the population (i.e., decision plan), based on the transformer unavailability calculation model, calculate N IM The comprehensive unavailability rate of the aging transformers in each year during the planning period, and the reliability index of the system in each year;

[0159] Step 4.4, calculate the fitness value of each particle: for each particle in the population, calculate the reliability cost of each year in the planning period according to the result obtained in step 3, and then add it to the expected investment cost and fault handling cost to obtain the total cost (i.e., the particle fitness value) to identify the optimal position of the individual in this round of iteration;

[0160] Step 4.5: Determine whether to use the individual optimal position obtained in this round of iteration as the new global optimal position according to the Metropolis criterion; if the new solution is accepted, update the annealing temperature;

[0161] Step 4.6, determine whether r reaches the maximum number of iterations R set : If not reached, set r = r + 1, and update the speed and position of each particle in the population, and then go to step 4.3; if r ≥ R set , then the final decision solution result pg and the corresponding fitness value f(pg) are output.

[0162] Compared with the existing methods, this embodiment takes into account both the preventive maintenance and retirement plans of aging transformers to achieve a balance between system reliability and economy, and achieves the following effects: 1) By collaboratively optimizing the preventive maintenance time and retirement time of aging transformers, the balance between system reliability and economy can be better achieved. 2) The established transformer aging failure model can quantitatively describe the mapping relationship between the transformer maintenance time, retirement time and its aging threshold and expected life, thereby effectively quantifying the impact of preventive maintenance / retirement measures on the reliability of aging transformers. 3) The proposed solution method can effectively improve the efficiency of model solution and is suitable for engineering applications.

[0163] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A collaborative optimization method for preventive maintenance and decommissioning planning of aging transformers, characterized in that: include: Step 1: Combining the structural vulnerability of the transformer and the contribution of aging failure to the transformer risk, the fuzzy entropy weight method is used to obtain the importance index of each aging transformer and rank them; Step 2, establishing a transformer aging failure model based on the calibrable three-parameter Weibull distribution model, and dynamically correcting the two parameters of the aging threshold and expected life in the model according to the preventive maintenance and retirement time; Step 3: According to the two-level collaborative optimization model of preventive maintenance and retirement planning of aging transformers constrained by system reliability, the preventive maintenance time and retirement time of each aging transformer in the future planning period are determined, with the goal of minimizing the total cost expenditure in the entire planning period while satisfying the system reliability constraints; Step 4: Use the particle swarm-simulated annealing algorithm to solve and obtain the optimal decision-making solution.

2. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 1 is characterized in that: In step 1, the electrical intermediate index is used to characterize the structural vulnerability of each aging transformer, and the contribution of each transformer aging failure to the system risk is used as a quantitative index to measure its impact on the system redundancy. The specific steps are: Step 1.1, the formula for calculating the electrical dielectric index of each aging transformer in the system is expressed as: Among them, G Represents the set of unit nodes in the system; Ψ D Represents the set of load nodes in the system; Represents transformer T n The structural vulnerability index of n The first and last nodes of the branch; g and d represent the two node numbers in a single "unit-load" node pair (g, d); Represents the transformer T based on the node pair (g, d) n The flow betweenness index is calculated as follows: Among them, W g W represents the weight of the power generation node g, which is the rated capacity of the generator; d Represents the weight of load node d, which is the peak load of the node; I gd (i, j) indicates that after adding a unit injection current element between the "unit-load" node pair (g, d), the current in the transformer T n The current caused in the branch; Step 1.2, the formula for calculating the contribution of aging failure of each aging transformer in the system to the system risk is expressed as: in, and They are respectively the transformer T before and after aging failure n load loss risk indicator; Step 1.3, the weight formula for calculating the two indicators of structural vulnerability and contribution based on the fuzzy entropy weight method is expressed as: Where ω l represents the fuzzy entropy weight of the lth indicator; Ь1 and Ь2 represent the preference weights of system operators for system structural vulnerability and adequacy respectively; IE1 and IE2 represent the structural vulnerability index H 1 and risk contribution index H 2 The information entropy of nl It represents the normalized value of the lth index of the nth transformer, and the calculation formula is expressed as: Where: N atl Indicates the number of aging transformers in the system; Step 1.4, calculate and output the importance evaluation index of each transformer 3. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 2 is characterized in that: In step 1, after obtaining the importance evaluation index of each transformer, the transformers are sorted according to the index values, and the aging transformers with the highest ranking results are selected as the decision objects for subsequent preventive maintenance and retirement plans.

4. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 2 is characterized in that: The step 2 includes the following sub-steps: Step 2.1, the calibrated three-parameter Weibull distribution model can be used to describe the quantitative relationship between the time-varying aging failure probability of the transformer and its aging threshold and expected life, which can be expressed as: In the formula, γ n With η n Transformer T n The aging threshold and expected life of the device are: c1, c2 and β are constant parameters obtained by fitting the failure statistics data; Step 2.2, according to the execution time of preventive maintenance and decommissioning measures, determine the transformer aging threshold and life update value after preventive maintenance and decommissioning, substitute the updated aging threshold and expected life parameters into the aging failure model of step 2.1, and you can get the transformer aging failure probability function taking into account the impact of preventive maintenance and decommissioning.

5. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 4 is characterized in that: In step 2.2, the transformer T n The expected aging failure probability function after taking into account the impact of preventive maintenance and decommissioning in the future period is expressed as: In the formula, the aging transformer T n T has been running ser,n Year, and They represent the aging threshold and life expectancy evaluated at the current moment respectively; Indicates that in the T ser + D1 year estimated life expectancy after preventive maintenance operations; and Indicates that in the T ser +D2 years of ageing thresholds and life expectancy assessments after preventive decommissioning operations.

6. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 4 is characterized in that: The step 3 includes the following sub-steps: Step 3.1, the calculation formula for the total cost in the future planning period is: Where K represents the total number of years of the planning period; N IM It represents the number of aging transformers to be decided within the planning period selected by importance sorting results; and They represent the transformer T in the kth year of the planning period. n The expected investment costs required to implement preventive maintenance and decommissioning measures; Represents transformer T n The disposal cost of random failures occurring in year k; Represents transformer T n Disposal costs incurred when aging failure occurs in the kth year; represents the reliability cost of the system in the kth year of the planning period; Step 3.2, the constraints of the upper-level preventive maintenance and decommissioning plan decision model include maintenance time constraints, decommissioning time constraints, and maintenance-decommissioning coupling constraints, which are expressed as: If the transformer T n There is a preventive maintenance plan in the future planning period, and its maintenance time d ohl,n Should satisfy: 0 <d ohl,n ≤K; If the transformer T n There is a preventive decommissioning plan in the future planning period, and its decommissioning time is d retire,n Should satisfy: 0 <d retire,n ≤K; If the transformer has both preventive maintenance and decommissioning plans, the maintenance time d ohl,n Retirement time retire,n The following should be satisfied: retire,n ≥d ohl,n +d set,val , d set,val Indicates the acceptable preventive maintenance-retirement time interval; The maintenance and decommissioning resource constraints are expressed as: Where: num ohl and num retire The upper limit of the number of transformers that can be subject to preventive maintenance and decommissioning measures each year respectively; The expected load loss EENS index is used to construct the system reliability constraint, which is expressed as: EENS k ≤EENS pre ; Step 3.3, the construction method of the lower layer reliability verification model is used to calculate the optimal average load reduction of the system in each year corresponding to the upper layer decision-making scheme, which is expressed as: Where: Q sam,k represents the total number of simulation cycles performed for reliability assessment in the kth year of the planning period; Θ q represents the set of all system states in the qth evaluation cycle; d q,s represents the state duration of the system state s in the qth evaluation cycle; and represents the optimal load reduction interval [LC q,s The upper and lower limits of ] are obtained by solving the optimal load reduction interval analysis model, and its compact form is expressed as: min[LC]=[D][X]; st[A][X]-[B](≤)(=,≤)0; [X]≥0; Where: [X] = ([X ij ] n×1 ) represents the n-dimensional decision variable of the objective function, whose elements D represents the coefficient matrix of [X] in the objective function; A represents the coefficient matrix of the constraint condition; [B] = ([B ij ] m×1 ), whose elements Indicates the interval value of each node load; [LC] = [LC - ,LC + ], which represents the optimal value interval of load reduction under a given system state.

7. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 6 is characterized in that: In step 3.1, the expected investment cost calculation formula for the preventive maintenance is: Where: θ is the discount rate; Represents a 0-1 indicator variable. If T n There is a preventive maintenance plan in year k, The value is 1, otherwise it is 0; It indicates the probability that the preventive maintenance plan can be implemented; Indicates that the moisture content of transformer insulation paper is reduced by Δω aver the costs required; The expected investment cost of preventive decommissioning is calculated as: Where: Represents a 0-1 indicator variable. If the transformer T n When there is a preventive decommissioning plan in year k, The value is 1, otherwise it is 0; It indicates the probability that the preventive decommissioning plan can be implemented; Represents transformer T n decommissioning costs; The formula for calculating the disposal cost of random failure is: Where: C repair represents the cost required to repair a single random failure; and Respectively represent transformer T n The random failure rate and expected number of random failures in the kth year of the future planning period; The calculation formula for the disposal cost of aging failure is: Where: Represents transformer T n The probability of aging failure occurring in the kth year; Reliability cost is the load loss cost caused by load reduction under a given decision plan. and penalty costs for violating the expected reliability requirements of the system The calculation formula is: In the formula, EENS k represents the optimal average load reduction of the system in year k under a given decision-making scheme; EENS R represents the optimal average load reduction of the system in the kth year without considering the aging and failure of transformers; u co Represents the load loss factor, which is used to measure the loss cost per unit load loss; EENS pre represents the load reduction that the system can accept each year; u pu Represents the unit load loss penalty factor.

8. The collaborative optimization method for preventive maintenance and decommissioning plan of aging transformers according to claim 6 is characterized in that: In step 4, solving the particle swarm-simulated annealing algorithm for the collaborative optimization model of the preventive maintenance and decommissioning plan of aging transformers considering system reliability includes the following sub-steps: Step 4.1: Input the power system topology parameters, operation parameters, component reliability parameters, and load curve to set the PSO-SA optimization algorithm parameters, including: particle population size, maximum number of iterations R set , inertia weight coefficient, learning factor and temperature attenuation coefficient; Step 4.2, initialize the number of iterations r = 1, and randomly generate a particle population that meets the constraints; Step 4.3, calculate the system reliability index corresponding to each particle: For each particle in the population, calculate N based on the transformer unavailability rate calculation model. IM The comprehensive unavailability rate of each aging transformer in each year during the planning period is calculated, and then the reliability index of the system in each year is calculated; Step 4.4, calculate the fitness value of each particle: for each particle in the population, calculate the reliability cost of each year in the planning period according to the result obtained in step 3, and then add it to the expected investment cost and fault handling cost to obtain the total cost to identify the optimal position of the individual in this round of iteration; Step 4.5: Determine whether to use the individual optimal position obtained in this round of iteration as the new global optimal position according to the Metropolis criterion; if the new solution is accepted, update the annealing temperature; Step 4.6, determine whether r reaches the maximum number of iterations R set : If not reached, set r = r + 1, and update the speed and position of each particle in the population, and then go to step 4.3; if r ≥ R set , then the final decision solution result pg and the corresponding fitness value f(pg) are output.