Method for calculating optimal path of nuclear power maintenance schedule

Abstract

The invention particularly relates to a nuclear power maintenance scheduling optimal path calculation method. The method comprises the following steps: step 101, constructing a scheduling algorithm model; step 102, training a scheduling algorithm model through the historical maintenance work order of the equipment to obtain a trained scheduling algorithm model; step 103, scheduling all the work orders to be scheduled through the trained scheduling algorithm model; and step 104, recommending a schedule to a plan engineer. According to the method, the waiting duration under different scheduling conditions is deduced in real time, the waiting duration of the whole overhaul work is optimized and deduced, the cost of the unit overhaul period is reduced, and the efficiency of the unit overhaul period is improved.

Description

Technical Field

[0001] The invention relates to the technical field of nuclear power maintenance scheduling, and in particular to a method for calculating an optimal path for nuclear power maintenance scheduling. Background Art

[0002] During the daily operation and refueling overhaul of nuclear power units, the inspection, repair and maintenance of various equipment in the power plant are involved. Equipment maintenance scheduling is a crucial link because it directly affects production efficiency and cost. Currently, equipment maintenance scheduling is done manually, which is costly and inefficient. Summary of the invention

[0003] The purpose of the present invention is to provide a method for calculating the optimal path for nuclear power maintenance scheduling, to deduce the waiting time under different scheduling conditions in real time, and to optimize and deduce the waiting time for the overall overhaul work, so as to reduce the cost during the overhaul of the unit and improve the efficiency during the overhaul of the unit.

[0004] In order to solve the above problems, the present invention provides the following technical solutions:

[0005] A method for calculating an optimal path for nuclear power maintenance scheduling comprises the following steps:

[0006] Step 101: construct a scheduling algorithm model;

[0007] Step 102: training the scheduling algorithm model through historical equipment maintenance work orders to obtain a trained scheduling algorithm model;

[0008] Step 103: Schedule all work orders to be scheduled using the trained scheduling algorithm model;

[0009] Step 104: recommend a schedule to a planning engineer.

[0010] As one of the achievable methods, the scheduling algorithm model uses three methods to schedule each work order;

[0011] The first method is: calculate the time response coefficient of each work order in the order of duration from small to large; schedule each work order in the order of time response coefficient from small to large, and calculate the start time and completion time of each work order;

[0012] The second method is: calculate the time response coefficient of each work order in the order of equipment influence coefficient from large to small; schedule each work order in the order of time response coefficient from small to large, and calculate the start time and completion time of each work order;

[0013] The third method is to give priority to scheduling work orders that are ready for preventive maintenance; if there are multiple work orders that are ready for preventive maintenance, schedule them according to the first method or the second method, and calculate the average time response coefficient of all work orders, as well as the start time and completion time of each work order.

[0014] As one of the feasible ways, historical maintenance work orders are classified according to equipment level and equipment importance, and the impact coefficient of each equipment is set; equipment level includes nuclear level and non-nuclear level, and equipment importance includes critical sensitive related, general and irrelevant.

[0015] As one of the feasible methods, when a single device has a historical maintenance work order, the impact coefficient of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices at the corresponding device level of the device. The calculation formula is as follows:

[0016]

[0017] Among them, ɑ is the influence coefficient of a single device, n is the historical maintenance times of the device, is the historical average maintenance time of the equipment, T P It is the historical average maintenance time of all equipment at the corresponding equipment level of this equipment.

[0018] As another feasible method, when a single device has a historical maintenance work order, the impact coefficient of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices under the corresponding device importance of the device. The calculation formula is as follows:

[0019]

[0020] Among them, ɑ is the influence coefficient of a single device, n is the historical maintenance times of the device, is the historical average maintenance time of the equipment, T Q It is the average historical maintenance time of all equipment with the corresponding equipment importance.

[0021] As one of the achievable methods, when a single device has no historical maintenance work order, the impact coefficient of the single device is 1.

[0022] As one of the feasible ways, the scheduling algorithm model is trained through the scheduling data of historical maintenance work orders; the scheduling data of all historical maintenance work orders are counted according to the equipment level and equipment importance. The scheduling data of historical maintenance work orders includes ready time, start time, execution time and completion time;

[0023] The historical maintenance work order is a historical overhaul work order or a historical daily maintenance work order, and preferably the scheduling data of all historical overhaul work orders are counted; in the case where there is no scheduling data of historical overhaul work orders, the scheduling data of all historical daily maintenance work orders are counted;

[0024] The scheduling algorithm model is trained by using the scheduling data of historical maintenance work orders to obtain a trained scheduling algorithm model.

[0025] As one of the achievable methods, the work order includes the following field information before scheduling: work order number, ready time and duration; the work order includes the following field information after scheduling: work order number, ready time, duration, start time and completion time;

[0026] When scheduling all work orders, the scheduling algorithm model calculates the time response coefficient of each work order in the order of ready time from the earliest to the latest;

[0027] When the duration of the previous work order is greater than (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order = the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order), and the time response coefficient of the next work order = the waiting time of the next work order / (2 × the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order));

[0028] When the duration of the previous work order is ≤ (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order is 0, and the time response coefficient of the next work order = 0.

[0029] As one possible implementation method, step 103, scheduling all work orders to be scheduled using the trained scheduling algorithm model, includes:

[0030] According to the time interval, all the work order information scheduled for the day is divided, and iterative calculation is performed in a way that minimizes the time response coefficient. The minimum value of the time response coefficient of each work order in the three scheduling methods is calculated respectively.

[0031] As one of the achievable methods, in step 104, the first method is used for scheduling, and the overall waiting time of all work orders is the shortest;

[0032] According to the second scheduling method, equipment types and equipment with high priority will be included in the maintenance work order scheduling in advance, improving the start efficiency of these two types of work orders;

[0033] According to the third scheduling method, the work orders with preventive maintenance attributes will be started in advance, and they have the attributes that meet the preventive maintenance work orders.

[0034] Beneficial technical effects of the present invention:

[0035] The optimal path calculation method for nuclear power maintenance scheduling of the present invention introduces the scheduling calculation logic of key sensitive equipment such as equipment level, equipment importance, preventive maintenance, etc. that nuclear power units are concerned about, and provides scheduling suggestions for planning engineers to improve scheduling efficiency; at the same time, the time response coefficient setting therein can perform big data calculations on historical data and work orders to be scheduled, and provide a suggestion algorithm that optimizes the overall overhaul or maintenance waiting time based on the completion of the inspection and maintenance work order, thereby improving the overall maintenance efficiency of the overhaul based on the key path of the overhaul. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The present invention is a flowchart of an embodiment of a method for calculating an optimal path for nuclear power maintenance scheduling. DETAILED DESCRIPTION

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used in the specification of the application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0038] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0039] The technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings and specific embodiments.

[0040] See also Figure 1 This embodiment provides a method for calculating the optimal path for nuclear power maintenance scheduling, comprising the following steps:

[0041] Step 101: construct a scheduling algorithm model;

[0042] Step 102: training the scheduling algorithm model through historical equipment maintenance work orders to obtain a trained scheduling algorithm model;

[0043] Step 103: Schedule all work orders to be scheduled using the trained scheduling algorithm model;

[0044] Step 104: recommend a schedule to a planning engineer.

[0045] In this embodiment, as one of the achievable methods, the scheduling algorithm model uses three methods to schedule each work order;

[0046] The first method is: calculate the time response coefficient β of each work order in the order of duration from small to large, and calculate the average value of the time response coefficient β of all work orders; schedule each work order in the order of time response coefficient β from small to large, and calculate the start time and completion time of each work order;

[0047] The second method is: calculate the time response coefficient β of each work order in the order of the equipment influence coefficient ɑ from large to small, and calculate the average value of the time response coefficient β of all work orders; schedule each work order in the order of the time response coefficient β from small to large, and calculate the start time and completion time of each work order;

[0048] The third method is to give priority to scheduling work orders that are ready for preventive maintenance; if there are multiple work orders that are ready for preventive maintenance, schedule them according to the first method or the second method, and calculate the average value of the time response coefficient β of all work orders, as well as the start time and completion time of each work order.

[0049] In this embodiment, as one of the feasible ways, historical maintenance work orders are classified according to equipment level and equipment importance, and the influence coefficient ɑ of each equipment is set.

[0050] In this embodiment, as one of the implementable ways, the equipment level includes nuclear level and non-nuclear level, and the equipment importance includes critical sensitive related, general and irrelevant.

[0051] In this embodiment, as one of the achievable methods, when a single device has a historical maintenance work order, the influence coefficient ɑ of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices at the device level corresponding to the device, and the calculation formula is as follows:

[0052]

[0053] Among them, n is the historical maintenance times of the equipment, is the historical average maintenance time of the equipment, T P It is the historical average maintenance time of all equipment at the corresponding equipment level of this equipment.

[0054] In this embodiment, as another achievable method, when a single device has a historical maintenance work order, the influence coefficient ɑ of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices under the device importance corresponding to the device, and the calculation formula is as follows:

[0055]

[0056] Among them, n is the historical maintenance times of the equipment, is the historical average maintenance time of the equipment, T Q It is the average historical maintenance time of all equipment with the corresponding equipment importance.

[0057] In this embodiment, as one of the achievable methods, when there is no historical maintenance work order for a single device, the influence coefficient ɑ of the single device is 1.

[0058] In this embodiment, as one of the achievable methods, the scheduling algorithm model is trained by the scheduling data of historical maintenance work orders; according to the equipment level and equipment importance, the scheduling data of all historical maintenance work orders are counted, and the scheduling data of the historical maintenance work orders includes the ready time Tr, the start time Ts, the execution time Te and the completion time Tf;

[0059] The historical maintenance work order is a historical overhaul work order or a historical daily maintenance work order, and preferably the scheduling data of all historical overhaul work orders are counted; in the case where there is no scheduling data of the historical overhaul work order, the scheduling data of all historical daily maintenance work orders are counted;

[0060] The scheduling algorithm model is trained by using the scheduling data of historical maintenance work orders to obtain a trained scheduling algorithm model.

[0061] In this embodiment, as one of the achievable methods, the work order includes the following field information before scheduling: work order number, ready time and duration; the work order includes the following field information after scheduling: work order number, ready time, duration, start time and completion time;

[0062] When scheduling all work orders, the scheduling algorithm model calculates the time response coefficient β of each work order in the order of ready time from the earliest to the latest;

[0063] When the duration of the previous work order is greater than (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order = the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order), and the time response coefficient β of the next work order = the waiting time of the next work order / (2 × the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order));

[0064] When the duration of the previous work order is ≤ (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order is 0, and the time response coefficient β of the next work order is 0.

[0065] The following example illustrates this: Assume that the ready time of the first work order is T0 and the duration is x, and the ready time of the second work order is T0+n and the duration is y;

[0066] If x is greater than n, the waiting time for the second work order is xn; if x is less than or equal to n, the waiting time for the second work order is 0; and so on for the waiting time of each subsequent work order;

[0067] During the work order scheduling process, set the time response coefficient β of the second work order:

[0068] If x is greater than n, the time response coefficient β of the second work order is (xn) / (2x-n); if x is less than or equal to n, the waiting time coefficient β of the second work order is 0.

[0069] In this embodiment, as one of the achievable methods, step 103, scheduling all the work orders to be scheduled by using the trained scheduling algorithm model, includes:

[0070] According to the time interval, all the work order information scheduled for the day is divided, and iterative calculation is performed in a way that minimizes the time response coefficient β. The minimum value of the time response coefficient β of each work order in the three scheduling methods is calculated respectively.

[0071] In this embodiment, the time slicing window is set to 3, 5, 7 or 15 days.

[0072] In this embodiment, as one of the achievable methods, in step 104, the first method is used for scheduling, and the overall waiting time of all work orders is the shortest;

[0073] According to the second scheduling method, equipment types and equipment with high priority will be included in the maintenance work order scheduling in advance, improving the start efficiency of these two types of work orders;

[0074] According to the third scheduling method, the work orders with preventive maintenance attributes will be started in advance, and they have the attributes that meet the preventive maintenance work orders.

[0075] The three work order scheduling results focus on three scenarios: the shortest waiting time, which means the highest overall efficiency; key equipment is repaired first while ensuring efficiency; preventive maintenance equipment is repaired first while ensuring efficiency; the three work order scheduling results can help planning engineers to schedule.

[0076] Based on historical big data results, the present invention applies deep learning algorithms and a recommended calculation method for scheduling in multiple modes to provide planning engineers with a scientific and efficient solution, solving the problem that nuclear power maintenance scheduling cannot be quickly carried out according to historical maintenance conditions and the lack of intelligent recommendation algorithms.

[0077] The present invention collects and processes data based on the scheduling record information formed by the daily maintenance of the unit, and then provides scheduling suggestions for planning engineers through the optimal path calculation method of maintenance scheduling; the core goal of the optimal path technical method is: under the calculation method of equipment importance factor, preventive maintenance necessity factor and shortest waiting time, by iteratively calculating the scheduling period and waiting period, the recommended value of the overall shortest maintenance period is obtained to minimize production costs and maximize production efficiency; by mining and analyzing the historical scheduling data of the unit, the process rules of equipment maintenance are discovered.

[0078] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for calculating the optimal path for nuclear power maintenance scheduling, characterized in that: The following steps are involved: Step 101: construct a scheduling algorithm model; Step 102: training the scheduling algorithm model through historical equipment maintenance work orders to obtain a trained scheduling algorithm model; Step 103: Schedule all work orders to be scheduled using the trained scheduling algorithm model; Step 104: recommend a schedule to a planning engineer.

2. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 1, characterized in that: The scheduling algorithm model uses three methods to schedule each work order; The first method is: calculate the time response coefficient of each work order in the order of duration from small to large; schedule each work order in the order of time response coefficient from small to large, and calculate the start time and completion time of each work order; The second method is: calculate the time response coefficient of each work order in the order of equipment influence coefficient from large to small; schedule each work order in the order of time response coefficient from small to large, and calculate the start time and completion time of each work order; The third method is to give priority to scheduling work orders that are ready for preventive maintenance; if there are multiple work orders that are ready for preventive maintenance, schedule them according to the first method or the second method, and calculate the average time response coefficient of all work orders, as well as the start time and completion time of each work order.

3. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 2, characterized in that: Historical maintenance work orders are classified according to equipment level and equipment importance, and the impact coefficient of each equipment is set; equipment level includes nuclear level and non-nuclear level, and equipment importance includes critical sensitive related, general and irrelevant.

4. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 3, characterized in that: When a single device has a historical maintenance work order, the impact coefficient of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices at the corresponding device level of the device. The calculation formula is as follows: Among them, ɑ is the influence coefficient of a single device, n is the historical maintenance times of the device, is the historical average maintenance time of the equipment, T P It is the historical average maintenance time of all equipment at the corresponding equipment level of this equipment.

5. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 3, characterized in that: When a single device has a historical maintenance work order, the impact coefficient of the single device is the ratio of the historical average maintenance time of the device to the historical average maintenance time of all devices under the corresponding device importance of the device. The calculation formula is as follows: Among them, ɑ is the influence coefficient of a single device, n is the historical maintenance times of the device, is the historical average maintenance time of the equipment, T Q It is the average historical maintenance time of all equipment with the corresponding equipment importance.

6. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 3, characterized in that: When there is no historical maintenance work order for a single device, the impact coefficient ɑ of the single device is 1.

7. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 3, characterized in that: The scheduling algorithm model is trained through the scheduling data of historical maintenance work orders; the scheduling data of all historical maintenance work orders are counted according to the equipment level and equipment importance. The scheduling data of historical maintenance work orders includes ready time, start time, execution time and completion time; The historical maintenance work order is a historical overhaul work order or a historical daily maintenance work order, and preferably the scheduling data of all historical overhaul work orders are counted; in the case where there is no scheduling data of the historical overhaul work order, the scheduling data of all historical daily maintenance work orders are counted; The scheduling algorithm model is trained by using the scheduling data of historical maintenance work orders to obtain a trained scheduling algorithm model.

8. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 2, characterized in that: Before scheduling, a work order includes the following fields: work order number, ready time, and duration; after scheduling, a work order includes the following fields: work order number, ready time, duration, start time, and completion time; When scheduling all work orders, the scheduling algorithm model calculates the time response coefficient of each work order in the order of ready time from the earliest to the latest; When the duration of the previous work order is greater than (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order = the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order), and the time response coefficient of the next work order = the waiting time of the next work order / (2 × the duration of the previous work order - (the ready time of the next work order - the ready time of the previous work order)); When the duration of the previous work order is ≤ (the ready time of the next work order - the ready time of the previous work order), the waiting time of the next work order is 0, and the time response coefficient of the next work order = 0.

9. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 2, characterized in that: Step 103, scheduling all work orders to be scheduled using the trained scheduling algorithm model, including: According to the time interval, all the work order information scheduled for the day is divided, and iterative calculation is performed in a way that minimizes the time response coefficient. The minimum value of the time response coefficient of each work order in the three scheduling methods is calculated respectively.

10. The method for calculating the optimal path for nuclear power maintenance scheduling according to claim 2, characterized in that: In step 104, according to the first scheduling method, the overall waiting time of all work orders is the shortest; according to the second scheduling method, the equipment types and equipment importance with high priority will be included in the maintenance work order scheduling in advance, thereby improving the start efficiency of these two types of work orders; According to the third scheduling method, the work orders with preventive maintenance attributes will be started in advance, and they have the attributes that meet the preventive maintenance work orders.

Images