Intelligent maintenance order dispatching method and system for industrial production equipment

Abstract

The invention relates to an intelligent maintenance order dispatching method and system for industrial production equipment, and belongs to the technical field of intelligent order dispatching. Obtaining a to-be-assigned repair work order, wherein the to-be-assigned repair work order at least comprises the following information: fault equipment, fault parts and fault phenomena; based on a pre-constructed knowledge graph, determining a plurality of suspected fault reasons according to the fault equipment, the fault part and the fault phenomenon; selecting work orders associated with the plurality of suspected fault reasons from a pre-established historical repair work order set, and constructing an associated work order set; determining the fault maintenance difficulty according to the number of the maintenance workers in the maintenance worker set and the pre-obtained total number of the maintenance workers; and determining an assigned maintainer of the to-be-assigned repair work order by adopting a corresponding order sending rule according to the fault maintenance difficulty. According to the invention, the fault level of the current repair work order is determined according to the number of the maintenance workers and the total number of the maintenance workers, and the corresponding order dispatching rule is determined according to the fault level, so that differential order dispatching is realized.

Description

Technical Field

[0001] The present invention relates to an intelligent maintenance dispatching method and system for industrial production equipment, belonging to the technical field of intelligent dispatching. Background Art

[0002] In an industrial production environment, maintenance management is crucial to the continued efficient operation of production equipment. It occupies a core and frequently used position. It is not only a solid backing for ensuring production efficiency and quality, but also a key driving force for promoting continuous optimization of equipment performance and effective cost control. When equipment fails, an efficient maintenance management system can quickly mobilize resources, including professional maintenance teams, key spare parts, and technical support, to quickly locate and repair the fault, minimize downtime, and reduce direct economic losses caused by production interruptions.

[0003] In traditional equipment management systems, the allocation of repair work orders is often highly dependent on manual operations, especially the subjective judgment of maintenance team leaders, who mainly allocate new tasks based on the number of work orders on hand for maintenance workers. This allocation mechanism is limited by personal subjective judgment and experience accumulation, and is easily affected by the immediate work of the team leader (such as direct participation in maintenance), which may cause delays in the repair work order processing process and reduce overall maintenance efficiency.

[0004] China authorized invention patent CN202011227098.6 A method and system for comprehensive dispatching of distribution network maintenance work orders. This invention discloses a method for comprehensive dispatching of distribution network maintenance work orders. By obtaining the work order dispatching factors, the weights of the maintenance workers corresponding to each type of factor are calculated by classification, and then the comprehensive weights of each maintenance worker are obtained. The maximum weight value is selected as the optimal dispatching solution to achieve accurate dispatching. However, this invention relies on a single non-differentiated dispatching rule, ignores the consideration of the fault difficulty level, limits the flexibility and efficiency of the system, and lacks applicability in the field of industrial production equipment maintenance. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent maintenance dispatching method and system for industrial production equipment, which determines the fault level of the current repair work order by comparing the number of maintenance workers related to the current repair work order and the total number of maintenance workers, and determines the corresponding dispatching rules according to the fault level, so as to achieve the purpose of differentiated dispatching with better applicability.

[0006] To achieve the above object, the present invention is implemented by adopting the following technical solutions.

[0007] In one aspect, the present invention provides an intelligent maintenance dispatching method for industrial production equipment, comprising: Obtaining a repair work order to be assigned, wherein the repair work order to be assigned includes at least the following information: faulty equipment, fault location, and fault phenomenon; Based on the pre-built knowledge graph, multiple suspected fault causes of the repair work order to be assigned are determined according to the faulty equipment, the faulty part and the fault phenomenon; Selecting all work orders associated with the suspected fault cause from a pre-established historical repair work order set, and constructing a related work order set; Obtain the maintenance worker corresponding to each work order in the associated work order set, and construct a maintenance worker set using all the obtained maintenance workers; Determining the difficulty of fault repair according to the number of repairmen in the repairmen set and the total number of repairmen acquired in advance; A dispatching method is determined according to the difficulty of the fault repair, and an assigned maintenance worker for the repair work order to be assigned is determined according to the dispatching method.

[0008] Optionally, determining the difficulty of fault repair according to the number of maintenance workers in the maintenance worker set and the total number of maintenance workers acquired in advance includes: Determine the failure repair response rate according to the number of maintenance workers and the total number of maintenance workers obtained in advance; The calculation formula of the fault maintenance response rate is: ; In the formula, The failure repair response rate; is the total number of maintenance workers, m1 is the number of maintenance workers in the maintenance worker set; If the fault repair response rate is within a preset first range, determining that the fault repair difficulty is at a first level; If the fault repair response rate is within a preset second range, determining that the fault repair difficulty is at a second level; If the fault repair response rate is within a preset third range, determining that the fault repair difficulty is at a third level; The first range, the second range, and the third range decrease in sequence.

[0009] Optionally, when the fault repair difficulty is the first level, determining the assigned repair worker of the repair work order to be assigned according to the dispatching method includes: The backlog work order duration of each maintenance worker in the maintenance worker set is obtained, the maintenance workers are arranged in reverse order according to the backlog work order duration, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker.

[0010] Optionally, when the fault repair difficulty is the second level, determining the assigned repair worker of the repair work order to be assigned according to the order dispatching method includes: Evaluate each associated work order in the associated work order set as a work order to be evaluated in turn; Obtaining evaluation indicators of the work order to be evaluated, wherein the evaluation indicators include positive indicators and negative indicators; The positive indicators include: the probability of the suspected fault cause corresponding to the work order in the repair work order to be assigned, the recall weight of the faulty equipment corresponding to the work order, the pre-acquired repair first pass rate of the work order, and the comprehensive repair evaluation score; The negative indicators include: the pre-acquired backlog work order length of the maintenance worker corresponding to the work order and the maintenance time of the work order; Evaluate each work order in the associated work order set according to the evaluation index to obtain a comprehensive evaluation parameter of each work order; The maintenance worker to be assigned is determined based on the comprehensive evaluation parameters.

[0011] Optionally, evaluating each associated work order in the associated work order set according to the evaluation index to obtain a comprehensive evaluation parameter of each associated work order includes: Step 1: Construct the original indicator data matrix with the associated work orders and evaluation indicators as the number of rows and columns respectively: ;in, represents the original indicator data matrix, Represents the first of the associated work orders The value of the evaluation indicator; Step 2: Normalize the positive and negative indicators: For positive indicators: ; For negative indicators: ;in For the The value of the evaluation index, For the The maximum value of the evaluation index, For the The minimum value of the evaluation index; Step 3: Calculate the corresponding weights of each item in the original indicator data matrix: ; In the formula, p ij Indicates of the associated work orders The weight of the evaluation index, n represents the total number of work orders in the associated work order set; Step 4: Establish a weight matrix corresponding to the original indicator data matrix: ; Step 5: Calculate the Entropy value of evaluation index: ; In the formula, ej Indicates The entropy value of the evaluation index; Step 6: Calculate the The weight coefficients of the evaluation indicators are determined according to the weight coefficients of the evaluation indicators, and the weight coefficient matrix used for interacting with the original indicator data matrix is ​​determined; ; In the formula, w j Indicates The weight coefficient of the evaluation index; Step 7: The original indicator data matrix and the weight coefficient matrix are multiplied to obtain a weighted matrix. The calculation formula of the weighted value is: ; In the formula, represents the weight matrix of the associated work orders The weighted value of each evaluation indicator; Step 8: Calculate the Comprehensive evaluation parameters of associated work orders: ; In the formula, Indicates Comprehensive evaluation parameters of associated work orders; and Respectively represent The distance from the comprehensive evaluation value of each associated work order to the positive ideal solution and the negative ideal solution; and Respectively represent The weighted maximum and minimum values ​​of the evaluation indicators; Determining to assign a maintenance worker according to the comprehensive evaluation parameters includes: According to the comprehensive evaluation parameters, the maintenance workers corresponding to each associated work order are arranged in reverse order, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker.

[0012] Optionally, when the fault repair difficulty is the third level, determining the assigned repair worker of the repair work order to be assigned according to the order dispatching method includes: Obtaining the evaluation parameters of each associated work order in the associated work order set, Determine the evaluation value of each associated work order according to the evaluation parameters; According to the evaluation value, the maintenance workers corresponding to each associated work order are arranged in reverse order, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker; The evaluation parameters include: the recall weight of the faulty equipment corresponding to the work order, the probability of the suspected fault cause corresponding to the work order occurring in the repair work order to be assigned; The product of the possible occurrence probability and the recall weight is calculated as an evaluation value.

[0013] Optionally, The method for determining the possible probability of the suspected fault cause corresponding to the work order in the repair work order to be assigned includes: Based on the pre-built knowledge graph, the probability of occurrence of each suspected fault cause is inferred according to the location nodes of the faulty equipment, fault location and fault phenomenon in the repair work order to be assigned in the knowledge graph; If the work order corresponds to a suspected fault cause, determining the probability of occurrence of the suspected fault cause as the possible occurrence probability; If the work order corresponds to multiple suspected fault causes, determine the one with the highest probability of occurrence among the multiple suspected fault causes as the possible probability; The method for determining the recall weight of the faulty equipment corresponding to the work order includes: The recall weight is determined according to the similarity between the faulty device corresponding to the work order and the faulty device to be assigned the repair work order; the higher the similarity, the higher the recall weight.

[0014] Optionally, if the repair work order to be assigned does not include the fault location; the method for determining the fault location includes: Obtaining the fault picture of the repair work order to be assigned; Comparing the fault picture with each photo in the previously acquired historical fault photo set, respectively, to obtain the image similarity between the fault picture and each photo; Filter out historical photos whose image similarity exceeds a threshold and whose corresponding fault locations are less than 3, and construct a comparison image set; Obtaining the fault location corresponding to each photo in the comparison picture set, and taking all the acquired fault locations as possible fault locations; Determining a comprehensive similarity score for each possible faulty component based on the comparison picture set and the picture similarity; The possible faulty component with the highest comprehensive similarity score is determined as the faulty part of the repair work order to be assigned.

[0015] Optionally, each possible faulty component is scored in turn as a faulty part to be scored; Determining the similarity comprehensive score of each fault location to be scored based on the comparison picture set and the picture similarity includes: Step 1: dividing the comparison picture set into a first set, a second set, and a third set according to the relationship between the fault location corresponding to the historical picture and the fault location to be scored; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored, and the other is in a constitutive relationship with the fault location to be scored, or if one historical photo corresponds to only one fault location and the fault location is the fault location to be scored, the historical photo is added to the first set; wherein the constitutive relationship indicates that one of the fault locations is a part of the other fault location; If one historical photo corresponds to two fault locations, both fault locations are not the fault locations to be scored and the relationship between the two fault locations is a constituent relationship, or if one historical photo corresponds to only one fault location and the fault location is not the fault location to be scored, the historical photo is added to the second set; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored and the relationship between the two fault locations is not constituted, the historical photo is added to the third set; Step 2: The calculation formula for the similarity comprehensive score of the fault part to be scored is: ; In the formula, S a Represents the comprehensive similarity score, a Indicates the fault location to be scored. For the fault picture and the first set of i 1 Image similarity of photos; n 2 is the total number of historical photos in the first set; b Indicates the fault part that has no constitutive relationship with the fault part to be scored; Indicates that the fault picture is the same as the first one in the second set. i 2 Image similarity of photos; m 2 represents the total number of historical photos in the second set; Indicates that the fault image is the same as the third set. u Image similarity of photos; k represents the total number of historical photos in the third set; Repeat the above steps to calculate the comprehensive similarity score of each possible faulty component.

[0016] In a second aspect, the present invention provides an intelligent maintenance management system for industrial production equipment, comprising: An acquisition module is used to acquire a repair work order to be assigned, wherein the repair work order to be assigned includes at least the following information: faulty equipment, fault location, and fault phenomenon; A first determination module is used to determine multiple suspected fault causes based on a pre-built knowledge graph according to the faulty equipment, the faulty location and the fault phenomenon; A first construction module is used to select all work orders associated with the suspected fault cause from a pre-established historical repair work order set, and construct a related work order set; The second construction module obtains the maintenance worker corresponding to each work order in the associated work order set, and constructs a maintenance worker set using all the obtained maintenance workers; A second determination module is used to determine the difficulty of fault repair according to the number of maintenance workers in the maintenance worker set and the total number of maintenance workers obtained in advance; The third determination module is used to determine a dispatching method according to the difficulty of the fault repair, and to assign a maintenance worker according to the dispatching method.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention implements a differentiated dispatching strategy driven by the difficulty of fault repair. Based on the fault repair response rate, the work orders to be assigned are divided into three difficulty levels: the first level (general), the second level (technical), and the third level (professional), and each level corresponds to different dispatching rules. General faults have a lower technical difficulty, and are dispatched directly in reverse order according to the current work order volume of the maintenance worker to simplify the dispatching steps; technical faults take into account technical knowledge and experience, and adopt comprehensive evaluation rules based on multiple evaluation indicators, and the maintenance worker finally determined has better comprehensive adaptability; professional faults focus more on fault matching, and the similarity of the faulty equipment accounts for a higher consideration factor, and the corresponding dispatching rules are adopted to ensure that the order is dispatched to the most suitable maintenance worker, thereby improving maintenance efficiency and accuracy.

[0018] 2. The present invention is a method for identifying fault locations based on the similarity of fault images. When the fault description of the repair work order is incomplete, the fault image is obtained and the similarity is calculated with the historical fault image set, and further analysis and processing are performed based on the similarity calculation results to determine the fault location. In this way, automatic emergency processing can be performed when the fault description cannot directly determine the fault location, without manual intervention, further improving the degree of intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of the intelligent management method for industrial production equipment maintenance in Example 1; Figure 2 is the knowledge graph model of fault maintenance in Example 1; Figure 3 It is a flowchart of a specific embodiment of the intelligent management method for industrial production equipment maintenance in Example 1. DETAILED DESCRIPTION

[0020] It should be noted that: The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. The embodiments of the present invention and the technical features in the embodiments may be combined with each other unless there is a conflict.

[0021] The term "and / or" is only a description of the association relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " generally indicates that the related objects are in an "or" relationship. Example

[0022] Combination Figure 1 This embodiment introduces a method for intelligent maintenance dispatching of industrial production equipment, which includes: Step S1, obtaining a repair work order to be assigned, wherein the repair work order to be assigned includes at least the following information: faulty equipment, fault location, and fault phenomenon; When a repair report needs to be initiated, the repair personnel needs to fill in / modify the equipment number. According to the equipment number, the basic information of the faulty equipment can be automatically obtained, including the equipment number, equipment category, installation location, year and month of use, etc. The repair personnel also fill in the fault information including the urgency, fault type, whether to repair immediately, scheduled repair time, and fault description including the fault location and fault phenomenon, etc., and upload the fault picture to submit the repair report. After the repair report, the work order automatically generates a repair order number. It is recommended to associate one fault with each repair report. The above is a specific embodiment. In actual work, the information of the faulty equipment, fault location, and fault phenomenon must be recorded in the repair work order. Other information can be set as required or not required according to actual requirements.

[0023] In this embodiment, when the repair person fills in the faulty equipment, fault location, and fault phenomenon, there are corresponding preset options for the repair person to choose from, so as to facilitate subsequent dispatch allocation and construction of the knowledge graph.

[0024] In a specific embodiment, when there is no corresponding preset option for the actual fault location or fault phenomenon, NLP / big model is used to perform semantic analysis to extract the fault location and fault phenomenon in the fault description.

[0025] In this embodiment, since there may be multiple repair work orders at the same time, the submitted work orders for review are sorted according to factors such as the urgency of the fault, whether to repair immediately, and the scheduled repair time filled in by the repair personnel. When the work order needs to be repaired immediately, the scheduled repair time is directly set as the submission time of the work order. Combined with Table 1, the priority of the work order repair assignment is determined according to the scheduled repair time and the urgency, and the repair work orders are sorted and displayed according to the priority, so that urgent work orders can be repaired in time.

[0026] Work order reviewers such as section chiefs / production team leaders can approve or reject work orders in this module to eliminate false alarms at the source in a timely manner.

[0027]

[0028] After the work order to be assigned is approved, proceed to step S2.

[0029] Step S2: Based on the pre-built knowledge graph, multiple suspected fault causes are determined according to the faulty equipment, faulty parts and the fault phenomena; In a specific embodiment, the method for constructing the knowledge graph is: Extract historical repair work order information from the production equipment management system and combine Figure 3 The specific implementation flow chart in the figure uses NLP, CNN, RNN and other algorithms to extract entity relationships from work order information, and constructs a knowledge graph of production equipment fault maintenance through knowledge fusion and knowledge reasoning. Figure 2 .

[0030] The types of entities include: faulty equipment, faulty location, fault phenomenon, fault cause, work order number, maintenance method, maintenance worker and related spare parts.

[0031] In conjunction with Table 2, descriptions and examples of the types of each entity are as follows:

[0032] The relationship type table between each entity is as follows:

[0033] In a specific embodiment, based on the knowledge graph, according to the fault location of the work order to be assigned and the location node of the fault phenomenon, the suspected fault cause and the possible probability of occurrence of each suspected fault cause in the current work order to be assigned can be inferred. The reasoning can be performed through Bayesian network, graph neural network and the like.

[0034] Step S3: Select a work order associated with the suspected fault cause from a pre-established historical repair work order set, and construct a set of associated work orders; the association indicates that the suspected fault cause is one of the fault causes of the associated work orders.

[0035] Select all the work orders associated with the suspected fault cause to construct a preliminary set of related work orders , there is a many-to-many relationship between the fault cause and the associated work order. Figure 3 In the specific implementation flow chart, if the set of related work orders (similar historical work orders) is an empty set, manual intervention is performed to dispatch the work orders.

[0036] Since the purpose of this embodiment is to finally determine the assigned maintenance worker, in a specific embodiment, according to whether the repair is immediate, the scheduled repair time and the specific schedule of the corresponding maintenance worker in the repair work order to be assigned, the associated work orders corresponding to the maintenance workers who have other maintenance tasks and do not meet the time requirements are filtered out to obtain the final associated work order set. ,like Empty, waiting to be assigned a repair work order Enter the manual dispatch pool, where the maintenance team leader manually dispatches orders or maintenance workers proactively grab orders.

[0037] This specific embodiment is only a specific way to determine the maintenance workers that can be dispatched. When faced with a very urgent repair work order, it is also possible to determine whether the maintenance worker can be dispatched based only on whether the maintenance worker is on duty, and all on-duty workers are considered dispatchable. In different situations, the standard for whether the maintenance worker can be dispatched can be flexibly adjusted.

[0038] Step S4: Obtain the maintenance worker corresponding to each work order in the associated work order set, and construct a maintenance worker set using all the obtained maintenance workers. ; Step S5, determining the difficulty of fault repair according to the number of repairmen in the repairmen set and the total number of repairmen acquired in advance; Considering that it is more reasonable to use different dispatching rules for different fault levels to achieve differentiated dispatching, this embodiment uses the fault repair response rate The faults reported are divided into three difficulty levels: Level 1 (general), Level 2 (technical) and Level 3 (professional) (the thresholds for the divisions can be configured according to the actual business). is the total number of maintenance workers, 1 for The number of maintenance workers within the Step S6: Determine the assigned maintenance worker for the repair work order to be assigned based on the difficulty of the fault maintenance and using the corresponding dispatching rules.

[0039] In conjunction with Table 4, in a specific embodiment, the first level (general type), the second level (technical type) and the third level (professional type) are divided as follows:

[0040] The first range, second range and third range corresponding to the first level (general), second level (technical) and third level (professional) are 0.85~1, 0.15~0.85 and 0~0.15 respectively.

[0041] When the fault repair difficulty is the first level (general type), the dispatch method is: Obtaining the backlog work order duration of each maintenance worker in the maintenance worker set, sorting the maintenance workers in reverse order according to the backlog work order duration, and determining the first dispatchable maintenance worker in the sequence as the assigned maintenance worker; In a specific embodiment, the calculation formula for the backlog work order duration is:

[0042] In the formula, T is the length of the backlog of work orders, is the current number of work orders for the maintenance worker corresponding to this work order, is the maintenance time of the i3th work order in the maintenance worker’s history, and n1 is the number of historical work orders of the maintenance worker.

[0043] In addition, the duration of the backlog work orders can also be directly obtained by multiplying the current number of work orders of the maintenance worker by a preset coefficient.

[0044] When the fault repair difficulty is the second level (technical), the dispatch method corresponding to Rule 2 is: Step S61: evaluate each work order in the associated work order set as a work order to be evaluated in turn; Step S62: obtaining evaluation indicators of the work order to be evaluated; the evaluation indicators include positive indicators and negative indicators; The positive indicators include: the probability of the suspected fault cause corresponding to the work order occurring in the repair work order to be assigned, the recall weight of the faulty equipment corresponding to the work order, the first-time qualified repair rate of the maintenance worker corresponding to the work order, and the comprehensive evaluation score of the maintenance of the work order; The negative indicators include: the backlog duration and maintenance time of the maintenance worker corresponding to the work order; The determination methods of each evaluation index are: The method for determining the probability of the suspected fault cause corresponding to the work order occurring in the repair work order to be assigned includes: Based on the pre-built knowledge graph, the probability of occurrence of each suspected fault cause is inferred according to the location nodes of the faulty equipment, fault location and fault phenomenon in the repair work order to be assigned; in some specific embodiments, the Bayesian network or graph neural network can be used for reasoning; If the work order corresponds to a suspected fault cause, determining the probability of occurrence of the suspected fault cause as the possible occurrence probability; If there are multiple suspected fault causes corresponding to the work order, the one with the highest probability of occurrence among the multiple suspected fault causes is selected as the possible probability; The recall weight of the faulty device corresponding to the work order is determined by: The recall weight is determined according to the similarity between the faulty device corresponding to the work order and the faulty device of the repair work order to be assigned. In a specific embodiment, the relationship between the faulty device corresponding to the work order and the faulty device of the repair work order to be assigned can be divided into four types: same device, same model device, same type device, and other device. The recall weights corresponding to the four types are They are 1, 0.9, 0.6 and 0.3 respectively. As the similarity of the devices decreases, the recall weight gradually decreases.

[0045] The calculation method for the one-time qualified rate of maintenance is as follows:

[0046] In the formula, F is the first-time qualified rate of maintenance, and k1 is the total number of maintenance times of the work order; As for the comprehensive maintenance evaluation score, it is determined according to the maintenance acceptance results. In a specific embodiment, the maintenance results can be scored according to the degree of functional recovery, appearance recovery, whether the structural strength is reduced, and the parts replacement rate after the maintenance. The specific scoring criteria can be different according to the different requirements of different equipment for different parameters. Different weights are assigned to each scoring parameter, and finally a comprehensive maintenance evaluation score is given manually or by machine. Other scoring methods that can be based on the maintenance acceptance results can also be applied to this embodiment, such as directly scoring based on the satisfaction of the repair personnel, which is not specifically limited here.

[0047] For reverse indicators: As for the duration of the backlog work order, the calculation formula is:

[0048] In the formula, T is the length of the backlog of work orders, is the current number of work orders for the maintenance worker corresponding to this work order, is the maintenance time of the i3th work order in the maintenance worker’s history, and n1 is the number of historical work orders of the maintenance worker.

[0049] As for the maintenance time, from the time the maintenance worker starts the maintenance timing until the maintenance is completed, the maintenance time will be included in the work report to facilitate the use of subsequent related work orders.

[0050] Step S63: Based on the entropy weight method, each work order in the associated work order set is evaluated according to the evaluation index to obtain a comprehensive evaluation parameter of each work order; Step S63 includes: Step 1, constructing an original indicator data matrix with the associated work orders and evaluation indicators as the number of rows and columns respectively: ;in, represents the original indicator data matrix, Represents the first of the associated work orders The value of the evaluation indicator; Step 2: Normalize the positive and negative indicators: For positive indicators: ; For negative indicators: ;in For the The value of the evaluation index, For the The maximum value of the evaluation index, For the The minimum value of the evaluation index; Step 3: Calculate the corresponding weights of each item in the original indicator data matrix: ; In the formula, p ij Indicates of the associated work orders The weight of the evaluation index, n represents the total number of work orders in the associated work order set; Step 4: Establish a weight matrix corresponding to the original indicator data matrix: ; Step 5: Calculate the Entropy value of evaluation index: ; In the formula, e j Indicates The entropy value of the evaluation index; Step 6: Calculate the The weight coefficients of the evaluation indicators are determined according to the weight coefficients of the evaluation indicators, and the weight coefficient matrix used for interacting with the original indicator data matrix is ​​determined; ; In the formula, wj Indicates The weight coefficient of the evaluation index; Step 7: The original indicator data matrix and the weight coefficient matrix are multiplied to obtain a weighted matrix. The calculation formula of the weighted value is: ; In the formula, Represents the weighted matrix of the associated work orders The weighted value of each evaluation indicator; Step 8: Calculate the Comprehensive evaluation parameters of associated work orders: ; In the formula, Indicates Comprehensive evaluation parameters of associated work orders; and Respectively represent The distance from the comprehensive evaluation value of each associated work order to the positive ideal solution and the negative ideal solution; and Respectively represent The weighted maximum and minimum values ​​of the evaluation indicators; Step S64: Arrange the maintenance workers corresponding to the associated work orders in reverse order according to the comprehensive evaluation parameters, and determine the first dispatchable maintenance worker in the sequence as the assigned maintenance worker.

[0051] Here, the dispatchable standard can be to filter out the remaining maintenance workers after filtering out maintenance workers who have other maintenance tasks and do not meet the time requirements based on whether the maintenance worker is on duty, the scheduled maintenance time, and the specific scheduling of the corresponding maintenance workers in the repair work order to be assigned. It can also be determined whether the maintenance worker can be dispatched based solely on whether the maintenance worker is on duty.

[0052] When the fault repair difficulty is the third level, the dispatching method corresponding to rule 1 includes: Step S6.1, obtain the evaluation parameters of each work order in the associated work order set, the evaluation parameters including: the recall weight of the faulty equipment corresponding to the work order; the possible occurrence probability of the suspected fault cause corresponding to the work order in the repair work order to be assigned; the method for determining the possible occurrence probability is consistent with the method for determining the occurrence probability at the second level recorded in the previous article.

[0053] If there are multiple suspected fault causes corresponding to the work order, the maximum value of the occurrence probabilities of the multiple suspected fault causes is taken as the possible occurrence probability; In a specific embodiment, the recall weight of the faulty device is determined by dividing the relationship between the faulty device corresponding to the work order and the faulty device to be assigned the work order into four types: same device, same model device, same type device, and other device; the recall weights corresponding to the four types are They are 1, 0.9, 0.6, and 0.3, respectively. As the similarity of the devices decreases, the recall weight gradually decreases; Step S6.2, calculating the product of the maximum probability in the set of possible occurrence probabilities and the recall weight as an evaluation value, and assigning a priority score to each work order according to the evaluation value; Step S6.3: Arrange the maintenance workers corresponding to each associated work order in reverse order according to the evaluation value, and determine the first dispatchable maintenance worker in the sequence as the assigned maintenance worker.

[0054] Step S7, maintenance execution, the assigned maintenance worker performs the maintenance work and records the maintenance process such as the start time, the end time, the cause of the failure, the maintenance method used to solve the failure, and the spare parts replaced during the maintenance process.

[0055] Step S8, acceptance module after maintenance is completed: the repair person performs acceptance operation on the repair work order after maintenance in this module. If the acceptance fails, the maintenance worker needs to perform the maintenance again. If the acceptance passes, the maintenance result needs to be comprehensively scored. In a specific embodiment, the maintenance result can be scored according to the degree of functional recovery, the degree of appearance recovery, whether the structural strength is reduced, and the parts replacement rate after maintenance. The specific scoring criteria can be different according to the different requirements of different equipment for different parameters. Different weights are given to each scoring parameter, and finally a comprehensive score is given manually or by machine.

[0056] Step S9, after the maintenance result is confirmed and accepted, the maintenance team leader can review and modify the contents of the repair work order, such as fault description, fault cause, and maintenance method, in this module to ensure that the fault description can clearly identify the fault location and fault phenomenon, and the description is complete, clear, and unambiguous, so as to support the update of the fault knowledge graph and the continuous improvement of the corresponding intelligent dispatching function.

[0057] Example 2 The difference between this embodiment and embodiment 1 is that: if the repair work order to be assigned does not include the fault part and the directly related fault part cannot be extracted, the fault part cannot be clearly identified through the fault description, and the fault picture is used to determine the fault part. The specific determination method includes (the fault part mentioned below specifically refers to the directly related fault part, that is, the path from the fault part node to the work order node in the knowledge graph is the shortest): Step Q1, obtaining the fault picture of the repair work order to be assigned; Step Q2: Compare the fault picture with each photo in the previously acquired historical fault photo set to obtain the image similarity between the fault picture and each photo; Photo gallery of the historical failures Extraction is performed based on the entity category of the fault part in Example 1. In some specific embodiments, the image similarity between the fault image and each photo is calculated using algorithms such as SIFT, SURF, Siamese, and CNN.

[0058] Step Q3: Filter out historical photos whose image similarity exceeds a threshold and whose corresponding fault locations are less than 3, and construct a comparison image set ; Obtain the fault location corresponding to each photo in the comparison picture set, and use all the acquired fault locations as possible fault locations. Setting the number of corresponding fault locations to less than 3 is mainly due to the consideration that when there are too many fault locations corresponding to the fault photos, the positioning of the fault location is prone to deviation.

[0059] Step Q4: Determine a comprehensive similarity score for each possible faulty component based on the comparison picture set and the picture similarity; Step Q4 sequentially treats each possible faulty component as a faulty part to be scored, and step Q4 specifically includes: Step 41: According to the relationship between the fault location corresponding to the historical image and the fault location to be scored, the comparison image set Divide into the first set , Second set and the third set ; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored, and the other is in a constituent relationship with the fault location to be scored, or if one historical photo corresponds to only one fault location and the fault location is the fault location to be scored, the historical photo is added to the first set; wherein the constituent relationship indicates that one fault location is a part of the other fault location (refer to Table 3 above); If one historical photo corresponds to two fault locations, both of which are not the fault locations to be scored and the relationship between the two fault locations is a constituent relationship; or if one historical photo corresponds to only one fault location, and the fault location is not the fault location to be scored, the historical photo is added to the second set; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored and the relationship between the two fault locations is not constituted, the historical photo is added to the third set; Step 42: The calculation formula for the similarity comprehensive score of the fault location to be scored is:

[0060] In the formula, S a Represents the comprehensive similarity score, a Indicates the fault location to be evaluated. Indicates that the fault image is the same as the first set. i 1 Image similarity of photos; n 2 Indicates the total number of historical photos in the first set; b Indicates the fault part that has no constitutive relationship with the fault part to be evaluated; Indicates that the fault picture is the same as the first one in the second set. i 2 Image similarity of photos; m 2 represents the total number of historical photos in the second set; Indicates that the fault image is the same as the third set. u Image similarity of photos; k Represents the total number of historical photos in the third set; if there are multiple b, calculate them separately.

[0061] Repeat the above steps to calculate the comprehensive similarity score of each possible faulty component.

[0062] Step Q5: Determine the possible faulty component with the highest comprehensive similarity score as the faulty part of the repair work order to be assigned.

[0063] Through the above method, this embodiment can determine a more accurate possible fault location based on fault picture analysis when the fault location is unclear. Example 3

[0064] This embodiment provides an industrial production equipment intelligent maintenance management system, which includes: Repair module: The repair person fills in / modifies the equipment number, which automatically brings up the basic equipment information including the equipment number, equipment category, installation location, year and month of first use, etc. The person fills in the fault information including the urgency, fault type, whether to repair immediately, scheduled repair time, fault description, etc., and uploads the fault picture to submit the repair report. After the repair report is submitted, the work order automatically generates a repair number. It is recommended to associate one fault with each repair report.

[0065] Work order review module: Sort the submitted work orders for review based on factors such as the urgency of the fault, whether to repair immediately, and the scheduled repair time. When repairing immediately, set the scheduled repair time as the work order submission time, and sort and display the repair work orders according to the scheduled repair time and urgency. Work order reviewers such as section chiefs / production team leaders can approve or reject work orders in this module to eliminate false alarms at the root.

[0066] The third part is the dispatching module: the approved work orders are matched with the fault parts according to the fault description and fault picture, and the historical work orders with the same fault cause are screened out according to the fault phenomenon. The corresponding maintenance worker's shift situation, workload, first-time qualified rate of maintenance, maintenance time, comprehensive maintenance evaluation score and other factors are comprehensively weighed, and the optimal maintenance worker is selected based on the differentiated dispatching strategy to realize automatic dispatching. When the historical work order cannot be matched, the maintenance team leader manually dispatches the order or the maintenance worker grabs the order independently.

[0067] The fourth part is the maintenance execution module: it is used to record the maintenance process of the maintenance worker, such as the start time and end time, the cause of the failure, the maintenance method used to solve the failure, and the spare parts replaced during the maintenance process.

[0068] The fifth part is the acceptance module: the repair person performs acceptance operations on the completed repair work order in this module. If the acceptance fails, the maintenance worker needs to perform the repair again. If the acceptance passes, the maintenance result needs to be comprehensively scored.

[0069] The sixth part is the maintenance result confirmation module: after acceptance, the maintenance team leader can review and modify the content of the repair work order, such as fault description, fault cause, maintenance method, etc. in this module to ensure that the description is complete, clear and unambiguous, and support the update of the fault knowledge graph and intelligent dispatching functions. Example 4

[0070] Based on the same inventive concept as in Example 1, this embodiment provides an intelligent maintenance management device for industrial production equipment, which includes: An acquisition module is used to acquire a repair work order to be assigned, wherein the repair work order to be assigned includes at least the following information: faulty equipment, fault location, and fault phenomenon; A first determination module is used to determine multiple suspected fault causes based on a pre-built knowledge graph according to the faulty equipment, the faulty location and the fault phenomenon; A first construction module is used to select all work orders associated with the suspected fault cause from a pre-established historical repair work order set, and construct a related work order set; The second construction module obtains the maintenance worker corresponding to each work order in the associated work order set, and constructs a maintenance worker set using all the obtained maintenance workers; A second determination module is used to determine the difficulty of fault repair according to the number of maintenance workers in the maintenance worker set and the total number of maintenance workers obtained in advance; The third determination module is used to determine a dispatching method according to the difficulty of the fault repair, and to assign a maintenance worker according to the dispatching method.

[0071] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0072] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0073] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0074] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0075] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A method for intelligent maintenance dispatching of industrial production equipment, characterized in that: include: Obtain a repair work order to be assigned, the repair work order to be assigned at least including the following information: Faulty equipment, faulty location, and fault phenomenon; Based on the pre-built knowledge graph, multiple suspected fault causes of the repair work order to be assigned are determined according to the faulty equipment, the faulty part and the fault phenomenon; Selecting all work orders associated with the suspected fault cause from a pre-established historical repair work order set, and constructing a related work order set; Obtain the maintenance worker corresponding to each work order in the associated work order set, and construct a maintenance worker set using all the obtained maintenance workers; Determining the difficulty of fault repair according to the number of repairmen in the repairmen set and the total number of repairmen acquired in advance; A dispatching method is determined according to the difficulty of the fault repair, and an assigned maintenance worker for the repair work order to be assigned is determined according to the dispatching method.

2. According to claim 1, a method for intelligent maintenance dispatching of industrial production equipment is characterized by: Determining the difficulty of fault repair according to the number of repairmen in the repairmen set and the total number of repairmen acquired in advance includes: Determine the failure repair response rate according to the number of maintenance workers and the total number of maintenance workers obtained in advance; The calculation formula of the fault maintenance response rate is: ; In the formula, The failure repair response rate; is the total number of maintenance workers, m1 is the number of maintenance workers in the maintenance worker set; If the fault repair response rate is within a preset first range, determining that the fault repair difficulty is at a first level; If the fault repair response rate is within a preset second range, determining that the fault repair difficulty is at a second level; If the fault repair response rate is within a preset third range, determining that the fault repair difficulty is at a third level; The first range, the second range, and the third range decrease in sequence.

3. The method for intelligent maintenance dispatching of industrial production equipment according to claim 2 is characterized in that: When the fault repair difficulty is the first level, determining the assigned repair worker of the repair work order to be assigned according to the dispatching method includes: The backlog work order duration of each maintenance worker in the maintenance worker set is obtained, the maintenance workers are arranged in reverse order according to the backlog work order duration, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker.

4. The method for intelligent maintenance dispatching of industrial production equipment according to claim 3 is characterized in that: When the fault repair difficulty is the second level, determining the assigned repair worker of the repair work order to be assigned according to the dispatching method includes: Evaluate each associated work order in the associated work order set as a work order to be evaluated in turn; Obtaining evaluation indicators of the work order to be evaluated, wherein the evaluation indicators include positive indicators and negative indicators; The positive indicators include: the probability of the suspected fault cause corresponding to the work order in the repair work order to be assigned, the recall weight of the faulty equipment corresponding to the work order, the pre-acquired repair first pass rate of the work order, and the comprehensive repair evaluation score; The negative indicators include: the pre-acquired backlog work order length of the maintenance worker corresponding to the work order and the maintenance time of the work order; Evaluate each work order in the associated work order set according to the evaluation index to obtain a comprehensive evaluation parameter of each work order; The maintenance worker to be assigned is determined based on the comprehensive evaluation parameters.

5. The method for intelligent maintenance dispatching of industrial production equipment according to claim 4 is characterized in that: The step of evaluating each associated work order in the associated work order set according to the evaluation index to obtain a comprehensive evaluation parameter of each associated work order includes: Step 1: Construct the original indicator data matrix with the associated work orders and evaluation indicators as the number of rows and columns respectively: ;in, represents the original indicator data matrix, Represents the first of the associated work orders The value of the evaluation indicator; Step 2: Normalize the positive and negative indicators: For positive indicators: ; For negative indicators: ;in For the The value of the evaluation index, For the The maximum value of the evaluation index, For the The minimum value of the evaluation index; Step 3: Calculate the corresponding weights of each item in the original indicator data matrix: ; In the formula, p ij Indicates of the associated work orders The weight of the evaluation index, n represents the total number of work orders in the associated work order set; Step 4: Establish a weight matrix corresponding to the original indicator data matrix: ； Step 5: Calculate the Entropy value of evaluation index: ; In the formula, e j Indicates The entropy value of the evaluation index; Step 6: Calculate the The weight coefficients of the evaluation indicators are determined according to the weight coefficients of the evaluation indicators, and the weight coefficient matrix used for interacting with the original indicator data matrix is ​​determined; ; In the formula, w j Indicates The weight coefficient of the evaluation index; Step 7: The original indicator data matrix and the weight coefficient matrix are multiplied to obtain a weighted matrix. The calculation formula of the weighted value is: ; In the formula, Represents the weighted matrix of the associated work orders The weighted value of each evaluation indicator; Step 8: Calculate the Comprehensive evaluation parameters of associated work orders: ; In the formula, Indicates Comprehensive evaluation parameters of associated work orders; and Respectively represent The distance from the comprehensive evaluation value of each associated work order to the positive ideal solution and the negative ideal solution; and Respectively represent The weighted maximum and minimum values ​​of the evaluation indicators; Determining to assign a maintenance worker according to the comprehensive evaluation parameters includes: According to the comprehensive evaluation parameters, the maintenance workers corresponding to each associated work order are arranged in reverse order, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker.

6. The method for intelligent maintenance dispatching of industrial production equipment according to claim 5, characterized in that: When the fault repair difficulty is the third level, determining the assigned repair worker of the repair work order to be assigned according to the dispatching method includes: Obtaining the evaluation parameters of each associated work order in the associated work order set, Determine the evaluation value of each associated work order according to the evaluation parameters; According to the evaluation value, the maintenance workers corresponding to each associated work order are arranged in reverse order, and the first dispatchable maintenance worker in the sequence is determined as the assigned maintenance worker; The evaluation parameters include: the recall weight of the faulty equipment corresponding to the work order, the probability of the suspected fault cause corresponding to the work order occurring in the repair work order to be assigned; The product of the possible occurrence probability and the recall weight is calculated as an evaluation value.

7. An intelligent maintenance dispatching method for industrial production equipment according to claim 4 or 6, characterized in that: The method for determining the possible probability of the suspected fault cause corresponding to the work order in the repair work order to be assigned includes: Based on the pre-built knowledge graph, the probability of occurrence of each suspected fault cause is inferred according to the location nodes of the faulty equipment, fault location and fault phenomenon in the repair work order to be assigned in the knowledge graph; If the work order corresponds to a suspected fault cause, determining the probability of occurrence of the suspected fault cause as the possible occurrence probability; If the work order corresponds to multiple suspected fault causes, determine the one with the highest probability of occurrence among the multiple suspected fault causes as the possible probability; The method for determining the recall weight of the faulty equipment corresponding to the work order includes: Determining the recall weight according to the similarity between the faulty device corresponding to the work order and the faulty device of the repair work order to be assigned; The higher the similarity, the higher the recall weight.

8. The method for intelligent maintenance dispatching of industrial production equipment according to claim 1, characterized in that: If the repair work order to be assigned does not include the fault location; the method for determining the fault location includes: Obtaining the fault picture of the repair work order to be assigned; Comparing the fault picture with each photo in the previously acquired historical fault photo set, respectively, to obtain the image similarity between the fault picture and each photo; Filter out historical photos whose image similarity exceeds a threshold and whose corresponding fault locations are less than 3, and construct a comparison image set; Obtaining the fault location corresponding to each photo in the comparison picture set, and taking all the acquired fault locations as possible fault locations; Determining a comprehensive similarity score for each possible faulty component based on the comparison picture set and the picture similarity; The possible faulty component with the highest comprehensive similarity score is determined as the faulty part of the repair work order to be assigned.

9. The method for intelligent maintenance dispatching of industrial production equipment according to claim 8, characterized in that: Each possible faulty component is scored in turn as a faulty part to be scored; Determining the similarity comprehensive score of each fault location to be scored based on the comparison picture set and the picture similarity includes: Step 1: dividing the comparison picture set into a first set, a second set, and a third set according to the relationship between the fault location corresponding to the historical picture and the fault location to be scored; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored, and the other is in a constitutive relationship with the fault location to be scored, or if one historical photo corresponds to only one fault location and the fault location is the fault location to be scored, the historical photo is added to the first set; wherein the constitutive relationship indicates that one of the fault locations is a part of the other fault location; If one historical photo corresponds to two fault locations, both fault locations are not the fault locations to be scored and the relationship between the two fault locations is a constituent relationship, or if one historical photo corresponds to only one fault location and the fault location is not the fault location to be scored, the historical photo is added to the second set; If one historical photo corresponds to two fault locations, one of which is the fault location to be scored and the relationship between the two fault locations is not constituted, the historical photo is added to the third set; Step 2: The calculation formula for the similarity comprehensive score of the fault part to be scored is: ; In the formula, S a Represents the comprehensive similarity score, a Indicates the fault location to be scored. The fault picture is the same as the first set i 1 Image similarity of photos; n 2 is the total number of historical photos in the first set; b Indicates the fault part that has no constitutive relationship with the fault part to be scored; Indicates that the fault picture is the same as the first one in the second set. i 2 Image similarity of photos; m 2 represents the total number of historical photos in the second set; Indicates that the fault image is the same as the third set. u Image similarity of photos; k represents the total number of historical photos in the third set; Repeat the above steps to calculate the comprehensive similarity score of each possible faulty component.

10. An intelligent maintenance management system for industrial production equipment, characterized in that: include: The acquisition module is used to acquire a repair work order to be assigned, and the repair work order to be assigned includes at least the following information: Faulty equipment, faulty location, and fault phenomenon; A first determination module is used to determine multiple suspected fault causes based on a pre-built knowledge graph according to the faulty equipment, the faulty location and the fault phenomenon; A first construction module is used to select all work orders associated with the suspected fault cause from a pre-established historical repair work order set, and construct a related work order set; The second construction module obtains the maintenance worker corresponding to each work order in the associated work order set, and constructs a maintenance worker set using all the obtained maintenance workers; A second determination module is used to determine the difficulty of fault repair according to the number of maintenance workers in the maintenance worker set and the total number of maintenance workers obtained in advance; The third determination module is used to determine a dispatching method according to the difficulty of the fault repair, and to assign a maintenance worker according to the dispatching method.

Images