Multi-party coupling maintenance work negotiation scheduling method for mountain wind turbine generator

Abstract

The invention discloses a mountain wind turbine generator multi-party coupling maintenance work negotiation scheduling method, which comprises the following steps that: at a negotiation decision node, each maintenance work participation main body formulates a plurality of local work scheduling schemes according to scheduling environment information and cost period demands, and combines and integrates the local work scheduling schemes to form a global maintenance work scheduling scheme combination initial solution set; preprocessing the initial solution set by using consistency filtering reduction, and deleting invalid redundant solutions which do not meet constraints or have conflict interference to obtain a feasible global solution set; and a sequential serial game negotiation model is established to assist in obtaining a better decision result considering local and global conditions. Through multi-agent negotiation combined scheduling scheme, invalid solution filtering and game model analysis, the problem of multi-agent scheduling conflict in mountain wind turbine generator maintenance is effectively solved, the collaborative efficiency and decision scientificity of maintenance operation under complex terrains are enhanced, and a systematic scheduling solution is provided for wind power equipment maintenance.

Description

Technical Field

[0001] The present invention relates to a dispatching operation method, and more particularly to a multi-party coupled maintenance operation negotiation dispatching method for mountain wind turbine sets. Background Art

[0002] The maintenance of wind turbines in complex mountainous environments is a dynamic decision-making process involving multiple decision-making nodes and coupled negotiation of scheduling operations among multiple maintenance participants. At some decision nodes, multiple operation participants (responsible for solving detailed maintenance needs at different levels and fields) may have coupled interference in the maintenance operation contents such as cycle, space, object, and target, and need to negotiate and make adjustments to the relevant contents in the comprehensive operation plan (including sub-operation plans for which multiple participants are responsible for their own maintenance contents) (i.e., negotiate and optimize the combination of multiple operation plans). At this time, the multiple parties involved in the negotiation and decision-making are prone to asymmetric and inconsistent cognition of the local and global expected maintenance goals (including different starting points and different understandings). This asymmetric conflict in the cognitive judgment of the various subjects on the predetermined cost, cycle, risk and other goals and the expected "value" or "priority" of different operation plan combinations occurs at each decision node, which not only affects the decision-making results of the decision node, but also affects the evolution direction of the expected quality and efficiency of the final maintenance operation. Summary of the invention

[0003] In view of the deficiencies in the prior art, the object of the present invention is to provide a multi-party coupled maintenance operation negotiation scheduling method for mountain wind turbines that can help promote the expected results of maintenance operations to converge to a better expected quality and efficiency level.

[0004] To achieve the above-mentioned purpose, the present invention provides the following technical solution: a method for negotiating and scheduling maintenance operations of a mountain wind turbine generator set by multiple coupling parties, comprising the following steps: Step 1: At a negotiation decision node, based on the current job scheduling environment information and their respective demands on cost and cycle, each maintenance operation participant participating in the negotiation formulates a variety of acceptable local job scheduling plans for their respective maintenance needs based on their own cognition, and combines and integrates them to form an initial solution set of the global maintenance operation scheduling plan combination; Step 2: Use consistency filtering and reduction to preprocess the initial solution set formed in step 1 to delete the local conflict interference invalid redundant solutions that do not meet the job scheduling constraints or cannot form a feasible global job scheduling solution combination, and obtain a feasible global job scheduling solution combination solution set; Step three, establish a sequential serial game negotiation model, through which the negotiation process among multiple subjects in the combined solution space of multiple feasible job scheduling schemes is characterized and analyzed, so as to assist in obtaining a relatively optimal decision result that takes into account both the local and the global aspects.

[0005] As a further improvement of the present invention, the consistency filtering reduction steps used in the second step are as follows: Step 2-1: Analyze the correlation relationship between sub-maintenance requirements, evaluate and calculate the relative importance of each sub-maintenance requirement, and obtain a new sequence TS of sub-maintenance requirements arranged in descending order of importance, where the relative importance of requirement Ts i is greater than or equal to Ts i +1; establish a requirement coupling correlation matrix C n×n , c ij represents the coupling correlation relationship between the requirement Ts i ranked i and the requirement Ts i ranked j: ; Step 2-2: Select each item Ts i in the sequence TS of sub-maintenance requirements obtained in Step 2-1 in turn, search the matrix C n×n and judge the value of c ij . When c ij =1 and j∈i+1,i+2,…,n, then for each solution combination in the initial solution set of the global maintenance operation scheduling plan combination obtained in Step 1, use the local operation scheduling plan corresponding to Ts i as the current filtering reduction comparison reference item, and perform a conflict interference comparison and judgment on the local operation scheduling plan corresponding to Ts j . When there is a conflict, delete this solution combination from the initial solution set of the global maintenance operation scheduling plan combination; Step 2-3: After the filtering reduction process is completed, output the consistency filtering reduction result of the initial solution set of the global maintenance operation scheduling plan combination to obtain a feasible global operation scheduling plan combination solution set.

[0006] As a further improvement of the present invention, a sequential serial game negotiation model established in the third step includes two serial game negotiation stages, specifically, a stage for obtaining a candidate solution set of Nash equilibrium based on non-cooperative games and a stage for negotiating and preferentially selecting a decision on the solution combination of the global operation scheduling plan based on cooperative games. In the non-cooperative game of the first stage, taking the comprehensive expected realization level of the local maintenance objectives associated with the sub-maintenance requirements undertaken by each negotiation subject as a measure, coordinate the claim conflicts among the subjects, and at the same time consider the fault tolerance ability for fuzzy cognitive evaluation data, and retain the feasible combination solution set with a utility value exceeding a specific threshold as the candidate solution set for the second-stage cooperative game; in the second-stage cooperative game, use the coalition game model of cooperative games to negotiate and preferentially select a maintenance operation scheduling plan solution with an optimal compromise of the comprehensive expected realization level of the global maintenance objective from the candidate solution set output in the first stage.

[0007] As a further improvement of the present invention, the negotiation game in the third step in the first stage is modeled using a non - cooperative game model and characterized as G=(P i ;S i ;U i ), where i ∈ 1, 2, 3, …, n, and it includes three elements: game players, game player strategies, and game player utilities; among them, the game player P i corresponds to each negotiation subject that has undertaken the coupled - associated sub - maintenance and protection requirements; the game player strategy S i corresponds to the filtered and reduced feasible job scheduling plan combination solution corresponding to the subset of sub - maintenance and protection requirements borne by each negotiation subject; the game player utility Ui corresponds to the satisfaction degree of the maintenance job scheduling expectation demands of each subject, and is characterized by measuring the satisfaction and expected realization level of their respective relevant local maintenance goals .

[0008] As a further improvement of the present invention, the comprehensive expected realization level of all associated local maintenance goals of each sub - maintenance and protection requirement borne by the game player P p is measured by the following calculation formula: where dg p is the total number of associated local maintenance goals, is the relative weight of the associated local maintenance goal j, is the expected realization level of the associated local maintenance goal j, is the impact of the job scheduling plan content of the sub - maintenance and protection requirements borne by other game players except P p on the expected realization level of the associated local maintenance goal j of the sub - maintenance and protection requirements borne by the game player P p .

[0009] As a further improvement of the present invention, the measurement calculation formula of is: where dv p is the number of sub - maintenance and protection requirement items borne by the game player P p . is to define the influence degree of the optional job scheduling plan k of the sub - maintenance and protection requirement i in each sub - maintenance and protection requirement borne by the game player P p on the expected realization level of the associated local maintenance goal j. For , the value of ​The evaluation language term set H of odd - numbered item tags is (extremely strong negative impact, strong negative impact, medium negative impact, weak negative impact, no impact, weak positive impact, medium positive impact, strong positive impact, extremely strong positive impact), and cloud model is used for quantitative analysis and calculation. When , the expected realization level of the local maintenance target j associated with the negative impact of the optional job scheduling plan k for the sub - maintenance requirement i. When , the expected realization level of the local maintenance target j associated with the positive impact of the optional job scheduling plan k for the sub - maintenance requirement i.

[0010] As a further improvement of the present invention, the measurement calculation formula is: Among them, m is the total number of tasks, and Cc i,j is the quantitative measurement value of the positive and negative correlation degree between local maintenance targets, representing the dependence or conflict relationship between them.

[0011] As a further improvement of the present invention, the Cc i,j is measured by introducing covariance and correlation coefficient. The positive and negative correlation degree Cc i,j between local maintenance targets i and j is measured by the following calculation formula: Among them, RO i is the correlation degree vector between local target i and all global targets. RO i =(ro i,1 , ro i,2 , …, ro i,k ), and ro i,q represents the correlation degree between local target i and global target q; Cov(.) is used to solve covariance; E[.] is the expectation; is the standard deviation of RO i ; is the mean value of the correlation degrees between local target i and all global targets.

[0012] As a further improvement of the present invention, the process of obtaining the candidate solution set for the two - stage cooperative game through non - cooperative game negotiation analysis in the first stage is as follows: Step 3 - 1: Select each feasible strategy combination of each game player in turn, calculate the utility value of each game player under this strategy, and calculate the corresponding combined utility value using the following formula: ; Step 32: Rely on the combination comparison program to determine whether the selected strategy combination satisfies the Nash equilibrium according to the following formula. When it satisfies the Nash equilibrium, any individual change to this strategy combination by each game player will reduce its own payoff: where, and are the strategies of game player Pi and the strategies of other game players at the Nash equilibrium respectively; Step 33: Obtain all strategy combinations that satisfy the Nash equilibrium, and use the predefined combined utility threshold to select a candidate solution set for the two-stage cooperative game from them; where, and are the maximum and minimum combined utilities respectively, and α takes an empirical value of 0.8.

[0013] As a further improvement of the present invention, in the combination solution negotiation and optimization decision-making stage of the two-stage global job scheduling scheme based on cooperative game in Step 3, it is modeled and characterized by the coalition game model as G=(N,v), where N is the set of game players, and v is the utility function of all possible game coalitions, satisfying and for all and there is , and the negotiation and optimization process of the best combination solution is as follows: Step 34: Calculate the utility vector U(Xi) of each game player under different feasible strategy combinations. The utility of each player is characterized by the expected achievement level ST of the corresponding global maintenance and protection target i and the measurement calculation formula is.

[0014] where, Ncvi is the number of lower-layer targets associated with the upper-layer target i, and there is ; Step 35: Calculate the coalition benefit function v; Step 36: Conduct coalition utility distribution. Among them, the utility distribution method in the coalition is determined by the Shapley value method. The component in the Shapley value vector is calculated as follows: where, the coalition S is any subset of the set of game players N; |S| represents the number of game players included in S, and v(S\{i}) represents the utility of the coalition composed of other game players except game player i in the coalition S; Step 37. According to the solution result Φ of the Shapley value and the utility vectors U(Xi) of each game player corresponding to the strategy combination Xi, the global expected return evaluation value βi corresponding to the strategy combination Xi can be calculated: ; Step 38. Select the strategy combination corresponding to the maximum βi value as the best combination solution of the global maintenance operation scheduling plan for the current negotiation decision node.

[0015] Advantages of the present invention: Aiming at the problems that the existing technical solutions ignore the asymmetric cognitive characteristics among multiple subjects and rarely consider the ambiguity of decision-making environment information in the maintenance negotiation decision-making process, the present invention has the advantages that the analytical model is closer to the actual complex engineering situation, and the sequential decision-making results dynamically take into account both local and global aspects. While taking into account the demands of each negotiation subject under asymmetric cognition, it maximally guarantees the expected quality and efficiency level of the final maintenance operation, and can effectively contribute to the improvement of quality and efficiency in the remote health management and operation and maintenance decision-making response of wind turbine groups in complex mountain environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the framework of the sequential serial game negotiation decision-making method; Figure 2 It is a schematic diagram of the conversion of the non-cooperative game problem model. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The following will further elaborate on the present invention in conjunction with the embodiments given in the drawings.

[0018] Referring to Figures 1 to 2 As shown, a multi-party coupled maintenance operation negotiation scheduling method for mountain wind turbines in this embodiment includes the consistency filtering reduction of the combination solution of the maintenance operation scheduling plan and the sequential serial game negotiation decision-making of the maintenance operation scheduling plan, so as to realize that at a certain negotiation decision node, according to multiple acceptable operation scheduling plans determined by multiple negotiation subjects for their respective responsible sub-maintenance requirements, they are integrated and combined into a solution set of the operation scheduling plan combination, that is, the initial solution space of the negotiation decision-making. Then, in the consistency filtering reduction link of the combination solution of the maintenance operation scheduling plan in the method of the present invention, the initial solution space is reduced to obtain a solution set of multiple feasible operation scheduling plan combinations. Then, in the sequential serial game negotiation decision-making link of the maintenance operation scheduling plan in the method of the present invention, the best global operation scheduling plan combination is preferably selected from the solution set of multiple feasible operation scheduling plan combinations through two-stage game negotiation.

[0019] The specific solution of this embodiment is as follows: 1. Consistency filtering reduction of the combination solution of the maintenance operation scheduling plan At a negotiation decision-making node, according to the current job scheduling environment information and their respective demands for aspects such as cost and cycle, each maintenance operation participant involved in the negotiation formulates multiple acceptable local job scheduling plans for their respective sub-maintenance requirements based on their own cognitions. After combination and integration, an initial solution set of the global maintenance job scheduling plan combination is formed. Due to the problem of asymmetric cognition among the subjects, there are interference conflicts in terms of time, space, resource occupation, etc. in the local scheduling arrangements of some combination solutions in this initial solution set, which belong to the infeasible and inoperable global job scheduling plan solutions.

[0020] The method proposed in the present invention uses consistency filtering reduction as a preprocessing step to delete the invalid redundant solutions of local conflict interference that do not meet the job scheduling constraints or cannot form a feasible global job scheduling plan combination, and obtains a solution set of the feasible global job scheduling plan combination, which is used as the feasible solution search space for subsequent negotiation decision-making optimization after reduction.

[0021] The steps of consistency filtering reduction are as follows: 1) Analyze the correlation relationship between sub-maintenance requirements, evaluate and calculate the relative importance of each sub-maintenance requirement, and obtain a new sequence of sub-maintenance requirements TS arranged in descending order of importance, where the relative importance of requirement Ts i is greater than or equal to that of Ts i+1 ; establish a requirement coupling correlation matrix C n×n , and c ij represents the coupling correlation relationship between the requirement Ts i with the relative importance ranking of i and the requirement Ts j with the ranking of j.

[0022] (1) 2) Initialize, let i = 1, j = i + 1.

[0023] 3) Select the maintenance requirement Ts i as the current filtering reduction comparison reference, search the matrix C n×n and judge the value of c ij . If c ij = 0, then let j = j + 1 and continue to judge the value of c ij ; if c ij = 1, then sequentially select each item in the initial solution set of the global maintenance job scheduling plan combination, give priority to ensuring the content of the acceptable local job scheduling plan corresponding to the requirement Ts i in it, and conduct a conflict interference comparison and judgment on the content of the acceptable local job scheduling plan corresponding to the requirement Ts j . For the plan solutions that cannot form a feasible global job scheduling combination with the content of the acceptable local job scheduling plan corresponding to the requirement Ts i , delete them in the initial solution set, and continue to let j = j + 1 and continue to judge cij value

[0024] 4) When j = n (n is the number of sub-maintenance requirement items in the sequence TS), let i = i + 1 and return to step 3.

[0025] 5) When i = n and j = n, the filtering and reduction process is completed, and the consistency filtering and reduction result of the initial solution set of the global maintenance operation scheduling plan combination is output to obtain the solution set of the feasible global operation scheduling plan combination.

[0026] 2. Sequential serial game negotiation decision-making of maintenance operation scheduling plan Considering the asymmetric cognitive problems of the expected goals of operation scheduling and maintenance goals among negotiating multi-agents, the method proposed in the present invention characterizes and analyzes the negotiation process among multi-agents in the solution space of multiple feasible operation scheduling plan combinations by establishing a sequential serial game negotiation model, and assists in obtaining a relatively optimal decision result that takes into account both local and global aspects.

[0027] The established sequential serial game negotiation model includes two serial game negotiation stages. In the non-cooperative game of the first stage, the comprehensive expected realization level of the local maintenance goals associated with the sub-maintenance requirements undertaken by each negotiation agent is used as a measure to coordinate the claim conflicts among the agents, and at the same time, considering the fault tolerance ability of fuzzy cognitive evaluation data, the feasible combination solution set with utility values exceeding a specific threshold is retained as the candidate solution set for the cooperative game in the second stage; in the cooperative game of the second stage, a maintenance operation scheduling plan solution with the best compromise of the comprehensive expected realization level of the global maintenance goal is negotiated and selected from the candidate solution set output in the first stage using the coalition game model of the cooperative game. The flow schematic framework of the sequential serial game negotiation model is as Figure 1 shown.

[0028] Further, the content of the Nash equilibrium candidate solution set acquisition scheme based on non-cooperative game in the first stage is as follows: Model the negotiation game in the first stage using a non-cooperative game model as G=(P i ;S i ;U i ), i ∈ 1, 2, 3,..., n, which includes three elements: game agents, game agent strategies, and game agent utilities. Among them, the game agent P i corresponds to each negotiation agent that undertakes coupled and associated sub-maintenance requirements; the game agent strategy S i corresponds to the filtered and reduced feasible operation scheduling plan combination solution corresponding to the subset of sub-maintenance requirements undertaken by each negotiation agent; the game agent utility U iCorresponding to the expected demands of the maintenance operation scheduling for each entity, it is carried and characterized by measuring the satisfaction of their respective relevant local maintenance objectives and the expected achievement levels. Considering the influence of different strategic choices of each game entity on each other's expected demands, the comprehensive expected achievement level of the associated local maintenance objectives under the coupled decision defined quantitatively is used here. Measure its U i for measurement.

[0029] The comprehensive expected achievement level of all the associated local maintenance objectives of each sub-maintenance demand borne by a game entity P under the coupled decision p is calculated by the following formula: (2) where, dg is the total number of associated local maintenance objectives. p is the relative weight of the associated local maintenance objective j. is the expected achievement level of the associated local maintenance objective j. is the impact of the content of the operation scheduling plan for the sub-maintenance demands borne by other game entities except P on the expected achievement level of the associated local maintenance objective j of the sub-maintenance demands borne by the game entity P p p p p

[0030] For , the measurement calculation formula is: (3) where, dv p is the number of sub-maintenance demand items borne by the game entity P p is the impact degree of the optional operation scheduling plan k of the sub-maintenance demand i in each sub-maintenance demand borne by the game entity P p on the expected achievement level of the associated local maintenance objective j. For , the value of is calculated by using the evaluation language term set H = (extremely strong negative impact, strong negative impact, medium negative impact, weak negative impact, no impact, weak positive impact, medium positive impact, strong positive impact, extremely strong positive impact) with an odd number of tags and adopting the cloud model for quantitative analysis. When , the optional operation scheduling plan k of the sub-maintenance demand i negatively impacts the expected achievement level of the associated local maintenance objective j. When , the optional operation scheduling plan k of the sub-maintenance demand i positively impacts the expected achievement level of the associated local maintenance objective j.

[0031] For , considering the conflicts and dependencies among the coupled maintenance objectives, when the objectives are positively correlated, the more consistent the expected achievement levels are, the more positive corrections should be made; when they are negatively correlated, the more consistent the expected achievement levels are, the more negative corrections should be made. Therefore, the entropy method can be combined for measurement. Here, the measurement calculation formula is defined as: (4) where m is the total number of tasks. Cc i,j is the quantification metric value of the positive and negative correlation between local maintenance objectives, representing the dependency or conflict relationship between them.

[0032] For Cc i,j , it is measured by introducing covariance and correlation coefficient. The positive and negative correlation degree Cc i,j between local maintenance objectives i and j is measured by the following calculation formula: (5) (6) (7) (8) where RO i is the correlation degree vector between local objective i and all global objectives. RO i = (ro i,1 , ro i,2 , …, ro i,k ), ro i,q represents the correlation degree between local objective i and global objective q; Cov(.) is used to solve the covariance; E[.] is the expectation; is the standard deviation of RO i ; is the mean value of the correlation degrees between local objective i and all global objectives.

[0033] Based on the filtered and reduced set of feasible global job scheduling scheme combinations, relying on the previously proposed game negotiation conversion model and utility value quantification metric calculation method, the process of obtaining the candidate solution set for the second-stage cooperative game through non-cooperative game negotiation analysis in the first stage is as follows: 1) Select each feasible strategy combination of each game player in turn, calculate the utility values of each game player under this strategy, and calculate the corresponding combined utility value using formula (9).

[0034] (9) Non-cooperative game utility matrix of the feasible job scheduling scheme strategy combination *a is the product of the number of strategy items in the respective strategy spaces of n game players 2) Rely on the combination comparison program to determine whether the selected strategy combination satisfies the Nash equilibrium according to formula (10). When it satisfies the Nash equilibrium, any individual change to this strategy combination by each game player will reduce their own payoff.

[0035] (10) Wherein, and are respectively the strategies of game player P i at the Nash equilibrium and the strategies of the other game players.

[0036] 3) Obtain all strategy combinations that satisfy the Nash equilibrium, and use a predefined combined utility threshold to select a candidate solution set for the two-stage cooperative game from them (threshold processing here is used to improve the fault tolerance ability of game negotiation decision-making for the fuzziness of cognitive evaluation information).

[0037] (11) Wherein, and are respectively the maximum and minimum combined utilities, and empirically take α = 0.8.

[0038] Furthermore, the specific solution for the negotiation and optimal selection decision of the combined solution of the global job scheduling scheme based on cooperative game in the second stage is as follows: In the second stage, conduct negotiation and optimal selection among the candidate solution sets obtained in the first stage, and try to improve the global expected quality and efficiency of the selected combined solution of the global job scheduling scheme as much as possible through cooperative game.

[0039] Model the negotiation game in the second stage using the coalition game model as G=(N,v), where N is the set of game players, and v is the utility function of all possible game coalitions (subsets of N), satisfying and for all and there is . Here, select each global maintenance target as a game player, and the candidate solution set output in the first stage as the feasible strategy combination.

[0040] First, calculate the utility vector U(X i ) of each game player under different feasible strategy combinations. The utility of each player is characterized by the expected achievement level ST i of the corresponding global maintenance target, and is measured and calculated through formula (12).

[0041] (12) Wherein, Ncv iis the number of lower-layer objectives associated with the upper-layer objective i, and there are .

[0042] Then, calculate the coalition benefit function v. When calculating, when the members in the coalition only include a single game player, use the "minimax criterion". First, calculate the maximum utility values of each game player under different strategy combinations, and then take the minimum value among them as the measurement value of v; when the members in the coalition include multiple game players, take the minimum value of the sum of the utility values of each game player under different strategy combinations as the measurement value of v.

[0043] Then, perform the coalition utility distribution. The utility distribution method in the coalition uses the Shapley value method to determine. The component in the Shapley value vector is calculated as shown in Equation (13): (13) where the coalition S is any subset of the set N of game players; |S| represents the number of game players included in S, and v(S\{i}) represents the utility of the coalition composed of other game players except the game player i in the coalition S.

[0044] Finally, according to the solution result Φ of the Shapley value and the utility vector U(X i ) of each game player corresponding to the strategy combination X i , the global expected revenue evaluation value β i corresponding to the strategy combination X i can be calculated: (14) Among the candidate strategy combinations obtained in the first stage, the strategy combination corresponding to the maximum β i value is the global optimal maintenance operation scheduling plan combination solution preferably selected through negotiation at this decision node.

[0045] The following example is provided in this embodiment: Here, taking the operation and maintenance scheduling decision of wind turbines in a complex mountain environment as the background, the negotiation decision at the operation plan level of the coupled and associated operation tasks T1 (function inspection of the wind turbine conversion and power generation subsystem) and T2 (function inspection of the wind turbine variable speed drive subsystem) of two maintenance operation participating entities (corresponding maintenance service providers of two wind turbine subsystems) P1 and P2 is taken as an example to illustrate the use of the technical method: Due to the tight coupling between the subsystems of the wind turbine, and the limitations of the terrain, location, and space for the maintenance operations of the wind turbine, the two maintenance tasks T1 and T2 will inevitably have coupling interferences in the operation arrangements such as the arrival time of technicians and special equipment scheduling, the order of maintenance operations, the operation time, the occupation of operation space, and the occupation of maintenance equipment resources. Different operation scheduling arrangements will cause different occupation of operation resources, operation cycles, and operation costs for each maintenance service provider, and will also affect the quality and efficiency level of the overall maintenance operation cycle. Therefore, negotiation and decision-making are required.

[0046] The main body P1 negotiates and customizes the operation scheduling plan around the main sub-maintenance requirement contents Ts1 (pitch jamming maintenance), Ts2 (blade pitch deviation maintenance), Ts3 (over-temperature at the inlet of the internal circulation air cooling maintenance), and Ts4 (yaw brake pressure instability maintenance) of T1. The main body P2 negotiates and customizes the operation scheduling plan around the sub-maintenance requirement contents Ts5 (abnormal wheel box oil temperature maintenance) and Ts6 (wheel shaft wear and scratch maintenance) of T2. The corresponding acceptable operation plan solutions customized by each main body for different sub-maintenance requirements (involving professional resource allocation, arrival time, waiting time, operation order, operation location, operation cycle, etc., which are not related to the technical method of the present invention and will not be listed in detail here) are compared with Table 1 below.

[0047] Table 1 Acceptable solution for coupled maintenance contents Based on the method steps (1. Consistency filtering and reduction of the combined solutions of the maintenance operation scheduling plan) proposed above, experts first evaluate the relative weights of the six maintenance requirements Ts1, Ts2, Ts3, Ts4, Ts5, and Ts6 under the two operation tasks T1 and T2 as 0.1926, 0.2143, 0.1571, 0.1729, 0.1254, and 0.1377 respectively. Arranged in descending order of weight as Ts2, Ts1, Ts4, Ts3, Ts6, Ts5, the operation coupling relationships between the maintenance requirements are shown in Table 2. According to the operation of the consistency reduction steps, experts filter and delete the 864 pairs of combined solutions corresponding to Table 1 without removing the feasible solutions, and finally obtain 20 feasible combined solutions, as shown in Table 3.

[0048] Table 2 Coupling correlation relationship c ij Table 3 Feasible global operation scheduling plan combined solution set For the two job tasks T1 and T2 respectively responsible by the main bodies P1 and P2, and the six sub-maintenance requirements Ts1, Ts2, Ts3, Ts4, Ts5, Ts6 listed below, the overall maintenance goals associated with T1 and T2 are cycle control dtt1, cost control dtt2, and maintenance satisfaction dtt3. The local maintenance goals associated with the sub-maintenance requirements Ts1, Ts2, Ts3, Ts4 listed under T1 include water-cooled cycle detection and risk elimination dst1, loop loose connection detection and risk elimination dst2, sensing feedback detection and risk elimination dst3, hydraulic circuit detection and risk elimination dst4, component deformation detection and risk elimination dst5. The local maintenance goals associated with the sub-maintenance requirements Ts5, Ts6 listed under T2 include oil level and temperature detection and risk elimination dst6, transmission link detection and risk elimination dst7, corrosion and wear detection and risk elimination dst8. Based on the cloud-based language term set evaluation, the positive and negative correlations between the local maintenance goals (see Table 4) and the correlations between the local maintenance goals and the global maintenance goals (see Table 5) are calculated.

[0049] Table 4 Positive and Negative Correlations Cc between Local Goals i,j Table 5 Correlations ro between Local and Global Goals i,q In the first stage of the sequential serial game negotiation decision-making of the maintenance operation scheduling plan, that is, the stage of obtaining the candidate solution set of the Nash equilibrium based on non-cooperative game, according to the model definition of this stage, the game players P1 and P2 are selected. According to Table 3, the strategy space of the game player P1 is {(A1,B1,C2,D1), (A1,B1,C1,D1), (A1,B1,C1,D2), (A1,B1,C3,D1), (A1,B1,C3,D2), (A1,B1,C2,D2), (A2,B1,C2,D2), (A2,B3,C3,D2), (A2,B3,C4,D2), (A3,B2,C1,D1), (A3,B1,C2,D1), (A3,B1,C3,D2), (A3,B2,C1,D1), (A3,B2,C2,D1), (A3,B2,C2,D2)} and the strategy space of the game player P2 is {(E2,F3), (E2,F2), (E2,F1), (E1,F2), (E3,F2), (E1,F3), (E3,F1), (E3,F3)}. Since filtering and reduction have been carried out in the first stage, therefore, among the strategy combinations of the two game players here, there are only 20 feasible strategy combinations in Table 3.

[0050] For each feasible strategy combination, first evaluate the expected achievement levels of the local maintenance objectives under the independent decisions of the two negotiation parties P1 and P2 based on Equation (3), and then calculate their correction values under the coupled decision according to Equation (4), as shown in Table 6.

[0051] Table 6 Example of partial evaluation results of the expected achievement levels of local maintenance objectives for each feasible strategy combination The utilities of the two game players P1 and P2 are measured by the comprehensive expected achievement level of the associated local maintenance objectives under the coupled decision. Calculate according to Equation (2) based on the data in Table 6 to obtain the utility values of the players corresponding to each feasible strategy combination, as shown in Table 7. According to the data in Table 7 and Equation (11), calculate the screening threshold as = 1.1246, and the candidate feasible strategy combinations for the next-stage cooperative game are selected as: s3, s4, s6, s9, s 11 、s 16 、s 18 .

[0052] Table 7 Non-cooperative game utility matrix for each feasible strategy combination In the second stage of the sequential serial game negotiation decision-making of the maintenance operation scheduling plan, that is, the negotiation and optimization decision-making stage of the global operation scheduling plan combination solution based on the cooperative game, pursue the maximization of the overall expected quality and efficiency level of the maintenance combination plan. According to the model definition of this stage, select three global maintenance objectives, dtt1, dtt2, and dtt3, as the game players. The candidate solutions s3, s4, s6, s9, s 11 、s 16 、s 18 selected in the first stage are used as the feasible strategy combinations. Measure and calculate the utilities of each game player under different strategy combinations through Equation (12), and at the same time calculate the coalition utility gains of different coalition combinations. Combine Equation (13) and use the Shapley value method to determine the utility distribution, which represents a contribution degree distribution of each game player to the optimal overall game coalition utility. Then, calculate the global expected revenue evaluation values β 11 、s 16 、s 18 of each feasible combination strategy s3, s4, s6, s9, s i , and the results are β3 = 1.5194, β4 = 1.7483, β6 = 2.6748, β9 = 3.1243, β 11 = 2.1842, β 16 = 1.2164, β 18= 2.1132, where the maximum value β9 = 3.1243, and the corresponding strategy combination plan s9 = {A2, B1, C2, D2, E3, F2} is the optimal solution of the job scheduling plan preferably selected through negotiation at the current decision node.

[0053] In summary, for the multi-party coupled maintenance operation negotiation and scheduling method of the mountain wind turbine unit in this embodiment, compared with the method in the prior art, at each decision node, the method proposed in this embodiment can be used to assist in decision-making according to the current actual task environment, which can help make the expected result of the maintenance operation converge to a better expected quality and efficiency level.

[0054] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the idea of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbine units, characterized in that: It includes the following steps: Step 1: At a negotiation and decision-making node, based on the current job scheduling environment information and their respective demands regarding costs and cycles, each maintenance operation participant involved in the negotiation formulates multiple acceptable local job scheduling plans for their respective sub-maintenance requirements according to their own cognitions, and combines and integrates them to form an initial solution set of the global maintenance job scheduling plan combination. Step 2: Use consistency filtering and reduction to preprocess the initial solution set formed in Step 1, delete the local conflict interference invalid redundant solutions that do not meet the job scheduling constraints or cannot form a feasible global job scheduling plan combination, and obtain a solution set of the feasible global job scheduling plan combination. Step 3: Establish a sequential serial game negotiation model, and through this model, characterize and analyze the negotiation process among multiple subjects in the multi-feasible job scheduling plan combination solution space to obtain a relatively optimal decision result that takes into account both the local and the global.

2. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 1, wherein: The consistency filtering and reduction steps used in Step 2 are as follows: Step 2-1: Analyze the correlation relationships among the sub-maintenance requirements, evaluate and calculate the relative importance of each sub-maintenance requirement, and obtain a new sequence TS of sub-maintenance requirements arranged in descending order of importance, where the relative importance of requirement Ts i is greater than or equal to Ts i +1; Establish the requirement coupling association matrix C n×n , c ij represents the requirement Ts with the relative importance ranking of i i and the requirement Ts with the ranking of j i The coupling association relationship between them is: ； Step 22, sequentially select each item Ts in the sub-maintenance demand sequence TS obtained in Step 21 i , search matrix C n×n and judge the value of c ij . When c ij = 1, j ∈ i + 1, i + 2, …, n, then for each solution combination in the initial solution set of the global maintenance operation scheduling plan obtained in Step 1, using the local operation scheduling plan corresponding to Ts i therein as the current filtering and reduction comparison reference item, conduct a conflict interference comparison and judgment on the local operation scheduling plan corresponding to Ts j therein. When there is a conflict, delete this solution combination from the initial solution set of the global maintenance operation scheduling plan combination; Step 2-3: After the filtering and reduction process is completed, output the consistency filtering and reduction result of the initial solution set of the global maintenance job scheduling plan combination to obtain a solution set of the feasible global job scheduling plan combination.

3. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 1 or 2, characterized in that: A sequential serial game negotiation model established in Step 3 includes two serial game negotiation stages, specifically, the first stage of obtaining a candidate solution set of Nash equilibrium based on non-cooperative game and the second stage of negotiating and preferentially selecting a decision for the global job scheduling plan combination solution based on cooperative game. In the non-cooperative game of the first stage, taking the comprehensive expected realization level of the local maintenance objectives associated with the sub-maintenance requirements undertaken by each negotiation subject as a measure, coordinate the demand conflicts among the subjects, and at the same time consider the fault tolerance ability of the fuzzy cognitive evaluation data, and retain the feasible combination solution set with the utility value exceeding a specific threshold as the candidate solution set for the second-stage cooperative game; in the second-stage cooperative game, use the coalition game model of cooperative game to negotiate and preferentially select a maintenance job scheduling plan solution that is a compromise optimal for the comprehensive expected realization level of the global maintenance objective from the candidate solution set output in the first stage.

4. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 3, wherein: The negotiation game in the first stage of step 3 is modeled using a non - cooperative game model and characterized as G=(P i ;S i ;U i ), where i ∈ 1, 2, 3, …, n; among them, the game player P i corresponds to each negotiation entity that has undertaken the coupled - associated sub - maintenance and protection requirements; the game player strategy S i corresponds to the combined solution of the filtered and reduced feasible job scheduling plans corresponding to the subsets of the sub - maintenance and protection requirements borne by each negotiation entity; the game player utility Ui corresponds to the degree of satisfaction of the maintenance job scheduling expectation demands of each entity, and is characterized by measuring the satisfaction and expected realization levels of their respective relevant local maintenance goals .

5. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 4, characterized in that: The game player P p The comprehensive expected achievement level of all associated local maintenance objectives for each sub-maintenance requirement borne The measurement calculation formula is as follows: Among them, dg p is the total number of associated local maintenance objectives, is the relative weight of the associated local maintenance objective j, is the expected achievement level of the associated local maintenance objective j, is the impact on the expected achievement level of the associated local maintenance objective j of the job scheduling plan content of the sub-maintenance requirements borne by other game players except P p on the sub-maintenance requirements borne by the game player P p ​ 6. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 5, characterized in that: The measurement calculation formula is: ； Among them, dv p is the number of sub-maintenance demand items borne by the game player P p , and is to define the influence degree of the optional job scheduling scheme k of the sub-maintenance demand i in each sub-maintenance demand borne by the game player P p on the expected achievement level of the associated local maintenance target j. For , the value is obtained by using the evaluation language term set H = (extremely strong negative impact, strong negative impact, medium negative impact, weak negative impact, no impact, weak positive impact, medium positive impact, strong positive impact, extremely strong positive impact) with finite odd-term tags and performing quantitative analysis and calculation using the cloud model. When , the optional job scheduling scheme k of the sub-maintenance demand i negatively impacts the expected achievement level of the associated local maintenance target j. When , the optional job scheduling scheme k of the sub-maintenance demand i positively impacts the expected achievement level of the associated local maintenance target j.

7. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 5, wherein: The measurement calculation formula is as follows: ； where m is the total number of tasks, and Cc i,j is the quantification metric value of the positive and negative correlation between local maintenance objectives, representing the dependence or conflict relationship therebetween.

8. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 7, characterized in that: The Cc i,j is measured by introducing covariance and correlation coefficient. The positive and negative correlation degree Cc between local maintenance objectives i and j i,j is measured by the following formula: ； ； ； ； Among them, RO i is the correlation vector between the local target i and all global targets. RO i =(ro i,1 , ro i,2 , …, ro i,k ), where ro i,q represents the correlation between the local target i and the global target q; Cov(.) is used to calculate the covariance; E[.] is the expectation; is the standard deviation of RO i ; is the mean of the correlations between the local target i and all global targets.

9. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 3, wherein: The process of obtaining the candidate solution set for the second-stage cooperative game through non-cooperative game negotiation analysis in the first stage is as follows: Step 3-1: Sequentially select each feasible strategy combination of each game subject, calculate the utility value of each game subject under this strategy, and use the following formula to calculate the corresponding combined utility value: ； Step 3-2: Rely on the combination comparison program and judge whether the selected strategy combination satisfies the Nash equilibrium according to the following formula: Among them, and are the strategies of the game player Pi and the strategies of other game players at the Nash equilibrium, respectively; Step 3: Obtain all strategy combinations that satisfy the Nash equilibrium, and use a predefined combined utility threshold to select a candidate solution set for the two-stage cooperative game from them; Among them, and are the maximum and minimum combined utilities respectively, and α takes an empirical value of 0.

8.

10. The multi-party coupled maintenance operation negotiation and scheduling method for mountain wind turbines according to claim 3, wherein: In the second-stage negotiation and optimal selection decision-making stage of the global job scheduling scheme based on cooperative game in Step 3, the coalition game model is used for modeling and characterized as G=(N,v), where N is the set of game players, and v is the utility function of all possible game coalitions, satisfying and for all and there is , the negotiation and optimal selection process of the best combination solution is as follows: Steps three and four: Calculate the utility vector U(Xi) of each game player under different feasible strategy combinations, and the utility of each player is represented by the expected achievement level ST of the corresponding global maintenance and protection goal. i It is characterized by and the measurement calculation formula is: ； where Ncvi is the number of lower-layer targets associated with the upper-layer target i, and there is ; Step 3-5: Calculate the coalition benefit function v; Step 36, perform the allocation of the coalition utility, where the utility allocation method in the coalition uses the Shapley value method to determine, and the component in the Shapley value vector The calculation is shown in the following formula: where the coalition S is any subset of the game subject set N; |S| represents the number of game subjects included in S, and v(S\{i}) represents the utility of the coalition composed of other game subjects except the game subject i in the coalition S; Step 3-7: According to the solution result Φ of the Shapley value and the utility vector U(Xi) of each game party corresponding to the strategy combination Xi, the global expected revenue evaluation value βi corresponding to the strategy combination Xi can be calculated: Step 3-8: Select the strategy combination corresponding to the maximum βi value as the best combination solution of the global maintenance job scheduling plan at the current negotiation and decision-making node.

Images