Task scheduling method and system for realizing equipment on electric energy meter verification assembly line

Abstract

The invention provides a task scheduling method for realizing equipment on an electric energy meter verification assembly line, which comprises the following steps of: determining equipment and tasks participating in scheduling according to a verification working link of an electric energy meter, and determining the workload, priority and completion time of each task; acquiring state information data of each device, after preprocessing the state information data into feature data of each device, importing the feature data into the trained multifunctional prediction model, and predicting the load capacity and fault risk of each device; calculating the load capacity of each device according to the load capacity and the fault risk of each device, and calculating the load demand of each task in combination with the workload, the priority and the completion time of each task; and constructing a target function, and solving an optimal solution to obtain a task scheduling queue in combination with the load demand of each task so as to perform intelligent scheduling task control on each device. By implementing the method, the task scheduling efficiency and accuracy are improved, the generation efficiency of the electric energy meter verification assembly line is also improved, and the verification period is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of assembly line detection, and in particular to a task scheduling method and system for implementing equipment on an electric energy meter calibration assembly line. Background Art

[0002] Scheduling tasks for equipment on the electricity meter calibration line is a significant challenge in the production and calibration of electricity meters. Currently, meter calibration involves multiple steps, such as equipment preparation, calibration, data collection, and equipment calibration. These steps often require the coordinated operation of multiple devices. However, most of these devices rely on manual intervention and empirical judgment to schedule tasks, which can lead to inefficient utilization of equipment resources, uneven task scheduling, and equipment overload or idleness. This reduces the production efficiency of the meter calibration line and prolongs the calibration cycle.

[0003] Furthermore, in production environments, factors such as the health status, failure risk, and load capacity of these devices significantly impact task scheduling. However, traditional approaches to task scheduling, which rely on manual intervention and empirical judgment, fail to fully account for these factors. This makes task scheduling incapable of adapting to the dynamic state of devices, leading to low efficiency and poor accuracy. For example, after prolonged operation, equipment may experience failures or require maintenance. Without real-time awareness of changes in the device's workload, task allocation cannot be adjusted promptly.

[0004] Therefore, there is an urgent need for a method to realize intelligent task scheduling of equipment on the electricity meter calibration pipeline, which can solve the defects of the traditional method of relying on manual intervention and experience judgment to realize task scheduling, thereby not only improving the efficiency and accuracy of task scheduling, but also improving the generation efficiency of the electricity meter calibration pipeline and reducing the calibration cycle. Summary of the Invention

[0005] The technical problem to be solved by the embodiments of the present invention is to provide a method and system for implementing task scheduling of equipment on an electricity meter calibration pipeline, which can solve the defects of the traditional method of relying on manual intervention and experience judgment to implement task scheduling, thereby not only improving the efficiency and accuracy of task scheduling, but also improving the generation efficiency of the electricity meter calibration pipeline and reducing the calibration cycle.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a method for implementing task scheduling of equipment on an electric energy meter calibration line, the method comprising the following steps:

[0007] S1. Based on the verification work of the electric energy meter on the electric energy meter verification line, determine the equipment and tasks involved in the verification work, and determine the workload, priority and completion time of each task;

[0008] S2. Obtain the status information data of each device and pre-process it into the feature data of the corresponding device. The feature data of each device is further imported into the trained multi-functional prediction model to predict the load capacity and failure risk of each device.

[0009] S3. Calculate the effective load capacity of each device based on its predicted load capacity and failure risk. Also, calculate the load requirements of each task on the device based on the workload, priority, and completion time of each task.

[0010] S4. Construct an objective function that minimizes the total load demand of the device or the total time to complete the task, and find the optimal solution for the objective function based on the load demand of each task on the device, obtain the scheduling queue for each task to be sent to the corresponding device, and further perform intelligent scheduling task control on each device based on the obtained task scheduling queue.

[0011] Wherein, in step S2, the multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, including an adaptive feature selection layer, a time series modeling processing layer and a weighted feedback layer connected in sequence;

[0012] The adaptive feature selection layer is configured to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the load capacity, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the load capacity feature data; and to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the fault risk, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the fault risk feature data;

[0013] The time series modeling processing layer is used to model the feature data of load capacity and failure risk output by the adaptive feature selection layer using time series convolution, and map them into predicted values ​​of load capacity and failure risk;

[0014] The weighted feedback layer is used to introduce an adaptive adjustment method based on a feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they are within a preset error range, and output the final predicted load capacity and failure risk.

[0015] Among them, in step S2, the characteristic data of the equipment currently input into the multifunctional prediction model include average load, load standard deviation, average working cycle, failure frequency, failure interval standard deviation, maintenance interval average, dynamic load ratio and working cycle difference.

[0016] The specific steps of step S3 include:

[0017] S31, through the formula Calculate the effective load capacity of each device; where E j is the effective load capacity of device j; D j is the health status score of device j, and its value is between [0,1], where 0 indicates that device j is in a faulty state and 1 indicates that device j is in a completely healthy state; α1 is a preset weight factor, which is a constant; is the load capacity predicted for equipment j; F j pred is the predicted failure risk of device j; j = 1, 2, …, m, where m is the total number of devices involved in the verification work, which is a positive integer greater than 1;

[0018] S32, through the formula Calculate the load requirements of each task on the equipment; is the load requirement of task i on device j; L i is the workload of task i; T i is the priority of task i; exp() is the exponential function with the natural constant e as the base; T i,time is the time required to complete task i; γ is the preset time adjustment coefficient, which is a constant; α2 is the preset adjustment factor, which is a constant; i = 1, 2, …, n, where n is the total number of tasks.

[0019] Between step S31 and step S32, the method further includes the following steps:

[0020] The effective load capacity of each device is dynamically updated based on the current actual failure risk of the device.

[0021] Among them, through the formula Dynamically update the payload capacity of each device;

[0022] is the effective load capacity of device j after dynamic update; α3 is the preset feedback adjustment factor, which is a constant; is the current actual failure risk of device j, which is calculated through the real-time monitoring data of the device; β1 is the preset feedback sensitivity weight factor, which is a constant.

[0023] Among them, the expression of the objective function with the total load demand of the equipment as the minimum is: Among them, x ij is the i-th task on the j-th device.

[0024] The expression of the objective function that minimizes the total time to complete the task is: in,

[0025] α5 is the preset execution time adjustment factor, which is a constant; is the preset maximum execution time of task i.

[0026] An embodiment of the present invention further provides a task scheduling system for implementing equipment on an electric energy meter calibration production line, comprising:

[0027] The scheduling object determination unit is used to determine the equipment and tasks involved in the calibration work based on the calibration work of the electric energy meters on the electric energy meter calibration production line, and determine the workload, priority and completion time of each task;

[0028] The device status prediction unit is used to obtain the status information data of each device and pre-process it into the feature data of the corresponding device. The feature data of each device is further imported into the trained multi-functional prediction model to predict the load capacity and failure risk of each device;

[0029] The task load demand calculation unit is used to calculate the effective load capacity of each device based on the predicted load capacity and failure risk of each device, and calculate the load demand of each task on the device based on the workload, priority and completion time of each task;

[0030] The task allocation and scheduling unit is used to construct an objective function that minimizes the total load demand of the equipment or the total time to complete the task, and to find the optimal solution for the objective function based on the load demand of each task on the equipment, so as to obtain the scheduling queue for each task to be sent to the corresponding equipment respectively, and further perform intelligent scheduling task control on each equipment based on the obtained task scheduling queue.

[0031] The multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, including an adaptive feature selection layer, a time series modeling processing layer and a weighted feedback layer connected in sequence;

[0032] The adaptive feature selection layer is configured to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the load capacity, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the load capacity feature data; and to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the fault risk, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the fault risk feature data;

[0033] The time series modeling processing layer is used to model the feature data of load capacity and failure risk output by the adaptive feature selection layer using time series convolution, and map them into predicted values ​​of load capacity and failure risk;

[0034] The weighted feedback layer is used to introduce an adaptive adjustment method based on a feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they are within a preset error range, and output the final predicted load capacity and failure risk.

[0035] The implementation of the embodiments of the present invention has the following beneficial effects:

[0036] 1. The present invention builds a multifunctional prediction model (e.g., load capacity prediction and fault risk prediction based on historical status information), combines feature selection, time series modeling, weighted feedback, etc., and can reasonably allocate tasks according to the prediction results in different time periods, thereby improving prediction accuracy and optimizing scheduling decisions;

[0037] 2. The present invention is based on the load capacity and failure risk predicted by each device, which directly affects the effective load capacity of the device, and introduces the workload, priority and completion time of each task to calculate the load demand of each task on the device, and further constructs an objective function with the total load demand of the device or the total time to complete the task as the minimum. The optimal solution of the objective function is obtained to obtain the scheduling queue of each task and send it to the corresponding device respectively, so as to realize intelligent scheduling task control of each device, so that the system can quickly adapt to fluctuations in the external environment such as equipment failure and load changes, thereby ensuring that tasks are always assigned to the most suitable device, which not only improves the efficiency and accuracy of task scheduling, but also improves the generation efficiency of the electricity meter calibration pipeline and reduces the calibration cycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying any creative work, other drawings obtained based on these drawings still fall within the scope of the present invention.

[0039] Figure 1 A flowchart of a method for implementing task scheduling of equipment on an electric energy meter calibration line provided by an embodiment of the present invention;

[0040] Figure 2 A structural diagram of a task scheduling system for implementing equipment on an electric energy meter calibration line provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be described in further detail below with reference to the accompanying drawings.

[0042] like Figure 1 FIG. 1 is a diagram showing a method for implementing task scheduling of equipment on an electric energy meter calibration pipeline in an embodiment of the present invention. The method includes the following steps:

[0043] Step S1: Based on the verification work of the electric energy meter on the electric energy meter verification line, determine the equipment and tasks involved in the verification work, and determine the workload, priority and completion time of each task;

[0044] The specific process is as follows: First, based on the calibration process of the electricity meters on the electricity meter calibration line, the equipment involved in the calibration work is determined. Second, based on the calibration process of the electricity meters on the electricity meter calibration line, the tasks assigned to the equipment are determined, as well as the workload, priority, and completion time required for each task.

[0045] It should be noted that the workload of a task is pre-estimated based on the task type, execution steps, and required resources, and is derived from the task requirements description or actual measurements. Task priority is determined by the business requirements or scheduling strategy, and is assigned based on factors such as the task's urgency and impact. The time required to complete a task is calculated based on the workload and the computing power required.

[0046] It is understandable that each task is dispatched to a designated device, and even the intelligent scheduling of subsequent tasks is based on the scheduling control of the original device.

[0047] Step S2: Acquire the status information data of each device and pre-process it into the characteristic data of the corresponding device. The characteristic data of each device is further imported into the trained multi-functional prediction model to predict the load capacity and failure risk of each device.

[0048] The specific process is as follows: first, the status information data of the equipment (such as temperature, pressure, equipment load, working time of the electricity meter, fault records, etc.) is collected through the data acquisition device deployed on each device (such as temperature sensors, etc.), and the collected status information data is further processed (such as denoising, outlier detection and elimination, data missing processing, data normalization, dimensional processing, etc., which are technical means well known to technical personnel in this field) to obtain the processed data.

[0049] Next, existing feature engineering techniques are used to extract features from the processed status information data, generating characteristic data for each device. The Min-Max normalization method is then used to map each feature's value to the [0, 1] interval to reduce dimensional differences between features. Furthermore, new feature data, such as dynamic load ratio and duty cycle variability, are incorporated into each device's characteristic data to enhance the model's predictive capabilities. Ultimately, the characteristic data for each device includes average load, load standard deviation, average duty cycle, fault frequency, standard deviation of fault intervals, average maintenance interval, dynamic load ratio, and duty cycle variability.

[0050] It should be noted that the dynamic load ratio is the ratio of the difference between the current load and the historical average load to the standard deviation of the historical load. It measures the relationship between the current load and the historical load of the equipment, reflects the changes in the working intensity of the equipment, and can better capture the load fluctuations of the equipment in different time periods, and then determine whether the equipment is in normal working condition.

[0051] The duty cycle variance is the absolute value of the difference between the current duty cycle and the average of the historical duty cycles divided by the standard deviation of the historical cycles. It measures the difference between the current duty cycle of the device and the historical duty cycle, reflects whether the device maintains a stable working mode, and reveals abnormal behavior of the device. If the duty cycle variance is too large, an automatic warning can be issued.

[0052] Next, a multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, consisting of a sequentially connected adaptive feature selection layer, a time series modeling processing layer, and a weighted feedback layer. The multifunctional prediction model takes as input the device's state feature data and outputs the device's load capacity and failure risk.

[0053] Among them, the adaptive feature selection layer is used to dynamically adjust the weight of each feature based on the correlation between each feature in the feature data of the currently input device and the load capacity, using an adaptive feature selection algorithm to sort the feature data in the feature data of the currently input device, and select the top K feature data as the feature data of the load capacity; and, based on the correlation between each feature in the feature data of the currently input device and the fault risk, using an adaptive feature selection algorithm to dynamically adjust the weight of each feature, so as to sort the feature data in the feature data of the currently input device, and select the top K feature data as the feature data of the fault risk.

[0054] For example, by the formula Get the weight of each feature after dynamic adjustment; among them, W adjusted is the i-th feature x i Dynamically adjusted weight; W original is the initial i-th feature x iThe initial weight assigned is a constant, which is initialized randomly or based on certain standards. λ is a preset adjustment factor used to control the adjustment amplitude and determines the degree of influence of feature correlation on weight adjustment. It is determined by experimental method. corr(x i ,y) is the feature x i The correlation between feature x and the target variable y (such as load capacity or failure risk) is a measure of the i An indicator of the influence of the target variable y. The correlation value is usually a number between [-1, 1], indicating the strength and direction of the linear relationship, and is calculated using the Pearson correlation coefficient.

[0055] The time series modeling processing layer uses time series convolution to model the load capacity and failure risk feature data output by the adaptive feature selection layer and maps them into predicted values ​​for load capacity and failure risk. This layer actually models the dependencies at each time point in the time series, helping the network capture device operating patterns and failure trends, i.e., predicted values ​​for load capacity and failure risk.

[0056] The weighted feedback layer is used to introduce an adaptive adjustment method based on the feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they reach the preset error range and output the final predicted load capacity and failure risk. At this time, the weighted feedback layer uses the prediction error of the dynamic feedback model ( Among them, y t is the actual value, taken from the database, is the predicted value;) automatically adjust the network weights ( This allows the model to self-correct in subsequent training until it reaches the error range preset based on expert experience. α is the learning rate, which is determined based on expert experience. is the sign of the weight adjustment direction to ensure that it can quickly adapt to new data and changes; t is the time.

[0057] The multifunctional prediction model is then trained and tested using the sample set and test set to obtain a trained multifunctional prediction model. It should be noted that the acquisition and division of the sample set and test set, as well as the training and testing of the multifunctional prediction model, are all accomplished using conventional techniques in the art and will not be detailed here.

[0058] Finally, the characteristic data of each device is imported into the trained multifunctional prediction model to predict the load capacity and failure risk of each device.

[0059] Step S3: Calculate the effective load capacity of each device based on the predicted load capacity and failure risk of each device, and calculate the load requirements of each task on the device based on the workload, priority, and completion time of each task;

[0060] The specific process is as follows: first, the equipment status is evaluated based on the pre-processed status information data according to the expert experience method, and any equipment status score D is obtained. j , based on the device status score D j , predict the results (load capacity, failure risk) and calculate the effective load capacity of the equipment.

[0061] For example, by the formula Calculate the effective load capacity of each device; where E j is the effective load capacity of device j, which indicates the load capacity that device j can currently bear; D j is the health status score of device j, which ranges from 0 to 1, where 0 indicates that device j is in a faulty state and 1 indicates that device j is in a completely healthy state. α1 is a preset weighting factor used to adjust the sensitivity of the failure risk to the device's payload capacity, and is a constant. is the load capacity predicted for equipment j; F j pred is the predicted failure risk of device j; j = 1, 2, …, m, where m is the total number of devices involved in the verification work, which is a positive integer greater than 1.

[0062] Secondly, the task load demand is not only related to the task workload and priority, but also to the task time and the load capacity of the equipment. Therefore, a nonlinear load demand formula is introduced to calculate the load demand of the task on the equipment.

[0063] For example, by the formula Calculate the load requirements of each task on the equipment; is the load requirement of task i on device j; L i is the workload of task i; T i is the priority of task i; exp() is the exponential function with the natural constant e as the base; T i,time is the time required to complete task i; γ is the preset time adjustment coefficient, which is used to control the impact of task time on load distribution. It is determined by experimental method and is a constant; α2 is the preset adjustment factor, which is used to control the degree of influence of task time on load demand. It is determined by expert experience and is a constant; i = 1, 2, …, n, where n is the total number of tasks.

[0064] Furthermore, a feedback mechanism is introduced to adjust the equipment's payload capacity and task allocation by monitoring the equipment's operating status in real time. and device health status D j Therefore, when the failure risk or health status of the equipment changes, the equipment's payload capacity E j Dynamic updates are required based on this feedback information (i.e., the effective load capacity of each device is dynamically updated based on the current actual failure risk of the device), and the load requirements of each task on the device are updated and calculated based on the updated effective load capacity of each device, combined with the workload, priority and completion time of each task.

[0065] At this time, the dynamic update formula of the device's effective load capacity is: in, is the effective load capacity of device j after dynamic update, that is, the effective load capacity of device j at time t (load capacity after feedback); α3 is the preset feedback adjustment factor, which represents the intensity of the impact of device failure risk on its effective load capacity during the feedback adjustment process. It controls the adjustment range of failure risk on device load capacity and is determined based on expert experience and is a constant. is the current actual failure risk of device j, indicating the actual probability of device failure, which is calculated based on the real-time monitoring data of the device (such as temperature, vibration, and operating time). β1 is the preset feedback sensitivity weight factor, which determines the feedback sensitivity of the device health status to the load capacity. It is determined based on expert experience and is a constant.

[0066] It can be seen that the introduction of the feedback mechanism enables flexible response to changes in equipment status by dynamically adjusting the equipment's payload capacity, thus avoiding the impact of increased equipment failure risk or decreased health status on task progress.

[0067] It is understandable that updating the load requirements of each task on the device based on the updated effective load capacity of each device is just a matter of changing the formula middle Replace E j Other parameters and their values ​​remain unchanged.

[0068] Step S4: Construct an objective function that minimizes the total load demand of the device or the total time to complete the task, and find the optimal solution for the objective function based on the load demand of each task on the device, obtain the scheduling queue for each task to be sent to the corresponding device, and further perform intelligent scheduling task control on each device based on the obtained task scheduling queue.

[0069] The specific process is as follows: first, the corresponding objective function can be constructed based on the total load demand of the equipment or the minimum total time to complete the task.

[0070] At this time, the expression of the objective function with the total load demand of the equipment as the minimum is: Among them, x ij is the i-th task on the j-th device.

[0071] At this time, the expression of the objective function with the minimum total task completion time is Among them, α5 is a preset execution time adjustment factor used to control the impact of task execution time on scheduling optimization. It is determined by experimental method and is a constant; is the preset maximum execution time of task i, which is used to measure the relative difference between the task execution time and the maximum time.

[0072] Next, based on the load requirements of each task on the device, we use existing algorithms such as genetic algorithms and particle swarm optimization to find the optimal solution to the objective function. This optimal solution becomes the scheduling queue for each task, which is then dispatched to the corresponding device. It should be noted that the use of algorithms such as genetic algorithms and particle swarm optimization to find the optimal solution to the objective function is accomplished using conventional techniques in the field and will not be further elaborated here. At this point, the optimal solution output by the objective function is the combined queue of tasks and devices.

[0073] Finally, intelligent scheduling task control is performed on each device based on the obtained task scheduling queue.

[0074] like Figure 2 FIG. 1 is a diagram showing a task scheduling system for implementing equipment on an electric energy meter calibration production line according to an embodiment of the present invention, comprising:

[0075] The scheduling object determination unit 110 is used to determine the equipment and tasks involved in the calibration work based on the calibration work of the electric energy meters on the electric energy meter calibration line, and determine the workload, priority and completion time of each task;

[0076] The device status prediction unit 120 is used to obtain the status information data of each device, pre-process it into the characteristic data of the corresponding device, and further import the characteristic data of each device into the trained multi-functional prediction model to predict the load capacity and failure risk of each device;

[0077] The task load demand calculation unit 130 is used to calculate the effective load capacity of each device based on the predicted load capacity and failure risk of each device, and calculate the load demand of each task on the device based on the workload, priority and completion time of each task;

[0078] The task allocation and scheduling unit 140 is used to construct an objective function that minimizes the total load demand of the device or the total time to complete the task, and combines the load demand of each task on the device to find the optimal solution for the objective function, obtain the scheduling queue for each task to be sent to the corresponding device, and further perform intelligent scheduling task control on each device based on the obtained task scheduling queue.

[0079] The multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, including an adaptive feature selection layer, a time series modeling processing layer and a weighted feedback layer connected in sequence;

[0080] The adaptive feature selection layer is configured to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the load capacity, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the load capacity feature data; and to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the fault risk, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the fault risk feature data;

[0081] The time series modeling processing layer is used to model the feature data of load capacity and failure risk output by the adaptive feature selection layer using time series convolution, and map them into predicted values ​​of load capacity and failure risk;

[0082] The weighted feedback layer is used to introduce an adaptive adjustment method based on a feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they are within a preset error range, and output the final predicted load capacity and failure risk.

[0083] The implementation of the embodiments of the present invention has the following beneficial effects:

[0084] 1. The present invention builds a multifunctional prediction model (e.g., load capacity prediction and fault risk prediction based on historical status information), combines feature selection, time series modeling, weighted feedback, etc., and can reasonably allocate tasks according to the prediction results in different time periods, thereby improving prediction accuracy and optimizing scheduling decisions;

[0085] 2. The present invention is based on the load capacity and failure risk predicted by each device, which directly affects the effective load capacity of the device, and introduces the workload, priority and completion time of each task to calculate the load demand of each task on the device, and further constructs an objective function with the total load demand of the device or the total time to complete the task as the minimum. The optimal solution of the objective function is obtained to obtain the scheduling queue of each task and send it to the corresponding device respectively, so as to realize intelligent scheduling task control of each device, so that the system can quickly adapt to fluctuations in the external environment such as equipment failure and load changes, thereby ensuring that tasks are always assigned to the most suitable device, which not only improves the efficiency and accuracy of task scheduling, but also improves the generation efficiency of the electricity meter calibration pipeline and reduces the calibration cycle.

[0086] It is worth noting that in the above system embodiment, the various system modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the various functional modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0087] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc.

[0088] The above disclosure is merely a preferred embodiment of the present invention and certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for implementing task scheduling of equipment on an electric energy meter calibration line, characterized in that: The method comprises the following steps: S1. Based on the verification work of the electric energy meter on the electric energy meter verification line, determine the equipment and tasks involved in the verification work, and determine the workload, priority and completion time of each task; S2. Obtain the status information data of each device and pre-process it into the feature data of the corresponding device. The feature data of each device is further imported into the trained multi-functional prediction model to predict the load capacity and failure risk of each device. S3. Calculate the effective load capacity of each device based on its predicted load capacity and failure risk. Also, calculate the load requirements of each task on the device based on the workload, priority, and completion time of each task. S4. Construct an objective function that minimizes the total load demand of the device or the total time to complete the task, and find the optimal solution for the objective function based on the load demand of each task on the device, obtain the scheduling queue for each task to be sent to the corresponding device, and further perform intelligent scheduling task control on each device based on the obtained task scheduling queue.

2. The method for implementing task scheduling of equipment on the electric energy meter calibration line according to claim 1, characterized in that: In step S2, the multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, including an adaptive feature selection layer, a time series modeling processing layer and a weighted feedback layer connected in sequence; The adaptive feature selection layer is configured to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the load capacity, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the load capacity feature data; and to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the fault risk, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the fault risk feature data; The time series modeling processing layer is used to model the feature data of load capacity and failure risk output by the adaptive feature selection layer using time series convolution, and map them into predicted values ​​of load capacity and failure risk; The weighted feedback layer is used to introduce an adaptive adjustment method based on a feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they are within a preset error range, and output the final predicted load capacity and failure risk.

3. The method for implementing task scheduling of equipment on an electric energy meter calibration line according to claim 2, characterized in that: In step S2, the characteristic data of the equipment currently input to the multifunctional prediction model include average load, load standard deviation, average working cycle, fault frequency, fault interval standard deviation, maintenance interval average, dynamic load ratio and working cycle difference.

4. The method for implementing task scheduling of equipment on an electric energy meter calibration line according to claim 1, characterized in that: The specific steps of step S3 include: S31, through the formula Calculate the effective load capacity of each device; where E j is the effective load capacity of device j; D j is the health status score of device j, and its value is between [0,1], where 0 indicates that device j is in a faulty state and 1 indicates that device j is in a completely healthy state; α1 is a preset weight factor, which is a constant; is the load capacity predicted for equipment j; F j pred is the predicted failure risk of device j; j = 1, 2, …, m, where n is the total number of devices involved in the verification work, which is a positive integer greater than 1; S32, through the formula Calculate the load requirements of each task on the equipment; is the load requirement of task i on device j; L i is the workload of task i; T i is the priority of task i; exp() is the exponential function with the natural constant e as the base; T i,time is the time required to complete task i; γ is the preset time adjustment coefficient, which is a constant; α2 is the preset adjustment factor, which is a constant; i = 1, 2, …, n, where n is the total number of tasks.

5. The method for implementing task scheduling of equipment on the electric energy meter calibration line according to claim 4, characterized in that: Between step S31 and step S32, the method further includes the following steps: The effective load capacity of each device is dynamically updated based on the current actual failure risk of the device.

6. The method for implementing task scheduling of equipment on an electric energy meter calibration line according to claim 5, characterized in that: By formula Dynamically update the payload capacity of each device; is the effective load capacity of device j after dynamic update; α3 is the preset feedback adjustment factor, which is a constant; is the current actual failure risk of device j, which is calculated through the real-time monitoring data of the device; β1 is the preset feedback sensitivity weight factor, which is a constant.

7. The method for implementing task scheduling of equipment on an electric energy meter calibration line according to claim 6, characterized in that: The expression of the objective function with the minimum total load demand of the equipment is: Among them, x ij is the i-th task on the j-th device.

8. The method for implementing task scheduling of equipment on an electric energy meter calibration line according to claim 6, characterized in that: The expression of the objective function that minimizes the total time to complete the task is: in, α5 is the preset execution time adjustment factor, which is a constant; is the preset maximum execution time of task i.

9. A task scheduling system for implementing equipment on an electric energy meter calibration line, characterized in that: include: The scheduling object determination unit is used to determine the equipment and tasks involved in the calibration work based on the calibration work of the electric energy meters on the electric energy meter calibration production line, and determine the workload, priority and completion time of each task; The device status prediction unit is used to obtain the status information data of each device and pre-process it into the feature data of the corresponding device. The feature data of each device is further imported into the trained multi-functional prediction model to predict the load capacity and failure risk of each device; The task load demand calculation unit is used to calculate the effective load capacity of each device based on the predicted load capacity and failure risk of each device, and calculate the load demand of each task on the device based on the workload, priority and completion time of each task; The task allocation and scheduling unit is used to construct an objective function that minimizes the total load demand of the equipment or the total time to complete the task, and to find the optimal solution for the objective function based on the load demand of each task on the equipment, so as to obtain the scheduling queue for each task to be sent to the corresponding equipment respectively, and further perform intelligent scheduling task control on each equipment based on the obtained task scheduling queue.

10. The task scheduling system for implementing equipment on the electric energy meter calibration line according to claim 9, characterized in that: The multifunctional prediction model is constructed based on a multi-layer weighted adaptive neural network, including an adaptive feature selection layer, a time series modeling processing layer and a weighted feedback layer connected in sequence; The adaptive feature selection layer is configured to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the load capacity, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the load capacity feature data; and to dynamically adjust the weight of each feature based on the correlation between each feature in the currently input device feature data and the fault risk, using an adaptive feature selection algorithm to sort each feature in the currently input device feature data and select the top K feature data as the fault risk feature data; The time series modeling processing layer is used to model the feature data of load capacity and failure risk output by the adaptive feature selection layer using time series convolution, and map them into predicted values ​​of load capacity and failure risk; The weighted feedback layer is used to introduce an adaptive adjustment method based on a feedback mechanism to adjust the predicted values ​​of load capacity and failure risk until they are within a preset error range, and output the final predicted load capacity and failure risk.

Images