A smart manufacturing advanced planning and scheduling system, method and device

Abstract

The present invention discloses an intelligent manufacturing advanced planning and scheduling system, method and device, which relates to the technical field of manufacturing planning and scheduling. After collecting processing task data in a processing area, a recommendation algorithm recommends corresponding processing workers for subtasks from a worker capability database; after scheduling each subtask, a simulation test is performed to select a scheduling scheme as a target scheme, and after collecting production environment data in the processing area, abnormality identification is performed, and adaptive adjustments are made when abnormalities exist. Fault prediction is performed on processing equipment in the processing area, and fault maintenance or performance optimization is performed on the equipment based on the acquired fault prediction data. The corresponding improvement plan is matched according to key factors, and after a time interval that meets the constraint conditions, an improvement report is sent to a data receiving end. The current scheduling scheme is readjusted so that the scheduling scheme can adapt to the current real-time scene and production status.

Description

Technical Field

[0001] The present invention relates to the technical field of manufacturing planning and scheduling, and in particular to an intelligent manufacturing advanced planning and scheduling system, method and device. Background Art

[0002] Manufacturing planning and scheduling is a complex and sophisticated process that involves developing and optimizing a series of detailed production activity arrangements based on the company's overall production goals, existing resource conditions (such as equipment, manpower, raw material inventory, etc.) and changing market demand through in-depth data analysis and future demand forecasting. This process not only needs to consider the production cycle, process path and capacity requirements of a single product or order, but also coordinates the connection and balance between different production lines and different processes, as well as the flexibility to deal with emergencies (such as equipment failure, raw material shortage, etc.).

[0003] Manufacturing planning and scheduling aims to ensure the efficiency and smoothness of the production process, maximize resource utilization, reduce production waiting time and waste, and meet customer delivery requirements and improve customer satisfaction. It uses advanced production planning software or systems, such as ERP (Enterprise Resource Planning) and MES (Manufacturing Execution System), to achieve automated scheduling, real-time monitoring of production progress, dynamic adjustment of production plans, and support cross-departmental collaboration to enhance the transparency and traceability of production management.

[0004] In the Chinese invention patent with application publication number CN109165851A, the technical field of scheduling methods is involved, especially a rolling plan scheduling method for intelligent manufacturing, which includes the following steps: first, receiving the product demand order of the customer, and classifying and processing the product demand order by machine type, counting the customer pick-up time and quantity of each product demand order, and then calculating the product manufacturing completion time node according to the product transportation time, and then calculating the product production time node according to the product production hours, and inputting the concentrated product demand order and the calculated manufacturing completion time into the production scheduling system, and the production scheduling system analyzes and processes the product demand order, and then queries the factory inventory database according to the information of the product demand order, and calculates the inventory model and the corresponding inventory, and if the inventory is greater than the product demand order, it is not included in the scheduling plan. The present invention is simple to operate, easy to use, and the system is automatically scheduled, with high efficiency, which greatly reduces the risk of error.

[0005] Combining the above applications and the prior art:

[0006] When producing door and window products with a large number of parts structures, if the processing orders suddenly increase on a large scale and a large number of new processing tasks are added, in order to improve processing efficiency, the processing tasks are usually decomposed and scheduled. After determining the priorities of different processing sub-tasks, the various processing sub-tasks are rearranged.

[0007] At present, there are many varieties of spare parts to be processed, there are certain differences between the batches of various spare parts, and the corresponding production cycles are long and of varying lengths. Although the existing processing task scheduling method can optimize the scheduling plan and improve the overall production and processing efficiency by reasonably arranging and allocating the processing tasks; however, when there are abnormalities in the environmental conditions in the processing area or process of the spare parts, or when an emergency occurs, the scheduling plan obtained by this optimization often cannot be adjusted and responded to in the first time, resulting in the scheduling plan cannot be effectively executed, and it is difficult to help improve the efficiency of production and processing.

[0008] To this end, the present invention provides an intelligent manufacturing advanced planning and scheduling system, method and device. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides an intelligent manufacturing advanced planning and scheduling system, method and device. After collecting the processing task data in the processing area, the recommendation algorithm recommends the corresponding processing workers for the subtasks from the worker capability database; after scheduling the production of each subtask, a simulation test is performed to select the scheduling plan as the target plan, and after collecting the production environment data in the processing area, abnormalities are identified, and adaptive adjustments are made when abnormalities exist. Fault prediction is performed on the processing equipment in the processing area, and the equipment is maintained or performance optimized based on the acquired fault prediction data. The corresponding improvement plan is matched according to the key factors, and the improvement report is sent to the data receiving end after the time interval that meets the constraint conditions. The current scheduling plan is readjusted so that the scheduling plan can adapt to the current real-time scene and production status, thereby solving the technical problems recorded in the background technology.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0013] An advanced planning and scheduling method for intelligent manufacturing, characterized in that: after collecting processing task data in the processing area, task complexity is generated , if the task complexity If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database;

[0014] After the production scheduling of each subtask is completed, the obtained scheduling scheme is simulated and tested, and the scheduling scheme is screened by the obtained test data. If the scheduling scheme as the target scheme cannot be screened out, one of several scheduling schemes is selected for optimization;

[0015] After collecting the production environment data in the processing area, perform abnormal identification and make adaptive adjustments when abnormalities occur; generate the scene abnormality degree from the corresponding abnormal state data , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside;

[0016] Fault prediction of processing equipment in the processing area, and fault maintenance or performance optimization of equipment based on the acquired fault prediction data, and identification of abnormal events in the processing area. If abnormal events occur, the production scheduling plan is optimized again;

[0017] Generate progress deviation value after real-time monitoring of production progress data of each subtask , if the progress deviation value Exceed expectations, match the corresponding improvement plan according to key factors, and send the improvement report to the data receiver after a time interval that meets the constraints.

[0018] Furthermore, the processing task data of the current stage and the attribute data of the parts to be processed are collected; the similarity analysis is performed between two parts to be processed, and the similarity center of each part is calculated. , to obtain the centrality Mark the spare parts to be processed;

[0019] Generate task complexity from production task data set , if the task complexity When the complexity threshold is exceeded, the work ability data of each processing worker in the processing area is recorded, and the data is summarized to build a worker ability database; after decomposing the current processing task, several subtasks are obtained.

[0020] Furthermore, for similarity centrality , production cycle And the number of spare parts in each batch Perform linear normalization and map the corresponding data values ​​to the interval Then generate the task complexity as follows :

[0021] ;

[0022] Weight coefficient: , , ; For the i The similarity center of the parts to be processed, is the mean of similarity centrality, For the iProduction cycle of spare parts to be processed; is the mean of the production cycle; For the i The number of spare parts to be processed, is the quantity mean.

[0023] Furthermore, equipment data, material data, processing worker data and processing environment data in the processing area are collected as constraint data, and the pre-trained simulated annealing algorithm is used to schedule the current subtasks, and several scheduling plans are obtained. The obtained scheduling plans are used as input, and the spare parts production digital twin model is used to simulate the spare parts processing and production process to obtain the corresponding scheduling test data.

[0024] Furthermore, efficiency values ​​are generated from the scheduling test data , among which, the production efficiency and resource utilization Perform linear normalization and map the corresponding data values ​​to the interval Then, the efficiency value is generated according to the following formula :

[0025] ;

[0026] Weight coefficient: , ,and ;

[0027] The efficiency value The scheduling plan that is higher than the efficiency threshold is used as the target plan; if there is no efficiency value For the scheduling plans that are higher than the efficiency threshold, the scheduling plan closest to the efficiency threshold is taken as the plan to be optimized. A sensitivity analysis is performed on the scheduling test data to determine the sensitivity factors that affect the effect. The obtained plan to be optimized is optimized using a multi-objective optimization algorithm, and the optimized scheduling plan is taken as the target plan.

[0028] Furthermore, the sensor network monitors the operating status of each device and production environment data in real time within the processing area;

[0029] Taking production environment data as input, the trained abnormal data recognition model is used to identify anomalies. When there are anomalies in the production environment data, a reminder instruction is issued to the outside.

[0030] Further, after receiving the reminder instruction, the trained environment control model is used to adjust the environment in the processing area;

[0031] If several reminder instructions are obtained continuously in the current stage, the abnormal time node, abnormal location and corresponding abnormal degree of the environment in the processing area are obtained, and the abnormal status data set is generated after being summarized.

[0032] Furthermore, the abnormal time node in the abnormal state data set T , Abnormal location L and degree of abnormality Generate scene abnormality , as follows:

[0033] ;

[0034] Where: is the time attenuation coefficient, is the comprehensive influence function, is the distance weight coefficient, For the integration domain, specify the analysis range in time, space, and degree of anomaly.

[0035] Furthermore, the equipment operation status data, historical fault and maintenance data, and environmental condition data in the processing area are monitored in real time; after receiving the fault monitoring instruction, the trained fault prediction model is used to predict the potential fault of the equipment, and if a fault exists, an external fault maintenance instruction is issued;

[0036] Taking processing equipment fault maintenance as the target word, a processing equipment fault maintenance knowledge graph is pre-constructed; if a fault maintenance instruction is received, feature extraction is performed on the acquired fault prediction data to obtain the corresponding fault features, and the corresponding fault handling strategy is given by the processing equipment fault maintenance knowledge graph.

[0037] Furthermore, if no fault instruction is received, the equipment control parameters are optimized using a pre-trained genetic algorithm with the goal of reducing the fault rate and improving the processing efficiency;

[0038] The production status data in the processing area is collected in real time. The production status data is used as input and the trained abnormal event recognition model is used to identify abnormal events. If an abnormal event is identified, the pre-trained simulated annealing algorithm is used again to optimize the production scheduling plan and obtain the responsive production scheduling plan.

[0039] Furthermore, the production progress of each subtask is monitored in real time, and the progress deviation value is generated from the acquired progress data. , if the progress deviation value If the deviation threshold is exceeded, a progress adjustment instruction is issued to the outside to generate a progress deviation value The way is as follows:

[0040] ;

[0041] in, , is the weight coefficient, , ,and , n is the number of subtasks, For the i The production progress of each subtask, It is a qualified reference value for production progress.

[0042] Furthermore, principal component analysis is performed on the equipment operation status, production environment data and production status data in the processing area to obtain the key factors affecting the progress of the processing tasks, match the corresponding improvement plan based on the key factors, and generate an improvement report;

[0043] The time node for issuing the improvement report is constrained as follows:

[0044] ;

[0045] Where: is the progress deviation value at time t, is the corresponding qualified reference value, It is i Progress deviation value, is the monitoring interval; weight coefficient, , , ;

[0046] e The value is 2.718, is the time attenuation coefficient, is the distance attenuation coefficient, is the first j The event node of an abnormal event, For the i and j The time interval between event nodes, n is the number of abnormal events, N The expected number.

[0047] An intelligent manufacturing advanced planning and scheduling system includes a task complexity analysis unit that collects processing task data in a processing area and generates a task complexity , if the task complexity If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database;

[0048] The scheduling unit, after scheduling the production of each subtask, performs simulation tests on the obtained scheduling schemes, and screens the scheduling schemes based on the obtained test data. If a scheduling scheme as the target scheme cannot be screened out, one of several scheduling schemes is selected for optimization;

[0049] The abnormality recognition unit collects the production environment data in the processing area and then identifies the abnormality. When an abnormality exists, it makes adaptive adjustments. The scene abnormality degree is generated from the corresponding abnormal state data. , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside;

[0050] The fault prediction unit predicts faults of processing equipment in the processing area, performs fault maintenance or performance optimization on the equipment based on the acquired fault prediction data, identifies abnormal events in the processing area, and re-optimizes the production scheduling plan if abnormal events occur;

[0051] The progress tracking unit monitors the production progress data of each subtask in real time and generates the progress deviation value , if the progress deviation value Exceed expectations, match the corresponding improvement plan according to key factors, and send the improvement report to the data receiver after a time interval that meets the constraints.

[0052] An intelligent manufacturing advanced planning and scheduling device, comprising at least one processor;

[0053] The memory is used to store executable instructions, and the instructions are executed by at least one of the processors to enable at least one of the processors to perform the steps of the intelligent manufacturing advanced planning and scheduling method.

[0054] (III) Beneficial effects

[0055] The present invention provides an intelligent manufacturing advanced planning and scheduling system, method and device, which have the following beneficial effects:

[0056] 1. By performing similarity analysis on different parts to be processed and using the obtained similarity data as a reference, similar subtasks can be merged to improve scheduling efficiency.

[0057] 2. Based on the complexity of the task By judging the complexity of the current processing task and selecting different scheduling strategies, the scheduling can be more targeted; by matching processing workers between subtasks to make the subtasks and processing workers more compatible, the subsequent processing efficiency can be improved.

[0058] 3. Generate efficiency values ​​from the acquired test data , according to the efficiency value Screening several scheduling plans to obtain the most efficient one can improve processing production when used as a reference; optimizing the selected scheduling plans to obtain feasible ones can improve subsequent scheduling efficiency.

[0059] 4. When there are abnormal environmental conditions in the working area, respond to the abnormal environmental conditions quickly, respond to and adjust the abnormal conditions, and avoid the poor environmental conditions in the processing area affecting the execution of the current subtasks and the corresponding processing equipment.

[0060] 5. Generate scene abnormality from abnormal state data , the abnormal conditions in the processing area can be evaluated to verify whether the current abnormal conditions will affect the formation of processing equipment. If there is a large impact, the processing equipment or processing environment conditions can be adaptively adjusted.

[0061] 6. Predict faults for each device under the current scenario conditions. When a fault occurs, the knowledge graph will promptly match the corresponding fault handling plan to avoid affecting the current processing task. By optimizing the equipment, the processing efficiency of the equipment can be improved, so that the current processing scheduling plan can be successfully completed.

[0062] 7. By detecting and identifying current abnormal events, when emergencies occur, the current scheduling plan can be readjusted to make the scheduling plan adapt to the current real-time scenario and production status; the current processing progress of the processing task is comprehensively judged and evaluated based on the progress deviation value. When the current construction progress is significantly different from the expected progress, the current scheduling plan can be optimized and adjusted.

[0063] 8. Continuous improvement of processing status and scheduling plans can realize a dynamic improvement process, improve the overall processing efficiency of each subtask, and adapt the output frequency of improvement reports to the current processing scenario and processing progress, so as to avoid interference of scheduling plan adjustments on the execution of processing tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a flow chart of the advanced planning and scheduling method for intelligent manufacturing of the present invention;

[0065] Figure 2 This is a schematic diagram of the structure of the intelligent manufacturing advanced planning and scheduling system of the present invention;

[0066] Figure 3 It is a structural schematic diagram of the intelligent manufacturing advanced planning and scheduling device of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] See also Figure 1 The present invention provides an intelligent manufacturing advanced planning and scheduling method, comprising:

[0069] Step 1: Collect the processing task data in the processing area and generate the task complexity , if the task complexity If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database;

[0070] The step 1 includes the following contents:

[0071] Step 101: Collect the processing task data of the current stage, including the type of parts to be processed, the production cycle and the quantity of each processing batch, and summarize the production task data set; collect the attribute data of the parts to be processed, including specifications, shapes and processing parameters;

[0072] Perform similarity analysis on the attribute data of two different parts to be processed, and calculate the similarity center of each part through similarity analysis of the similarity data between different parts to be processed. , the specific method is as follows:

[0073] ;

[0074] in, is the number of spare parts to be processed, Spare parts to be processed i And spare parts to be processed j The similarity difference of Mark the spare parts to be processed;

[0075] When in use, when scheduling different production tasks, similarities between different parts to be processed are analyzed to determine the similarities between different subtasks. The similarity data obtained is used as a reference to merge more similar subtasks, which can improve scheduling efficiency.

[0076] Step 102: The similarity center of the parts to be processed in the production task data set , production cycle And the number of spare parts in each batch Generate task complexity , where the similarity centrality , production cycle And the number of spare parts in each batch Perform linear normalization and map the corresponding data values ​​to the interval Then follow the steps below:

[0077] ;

[0078] Weight coefficient: , , which can be obtained by analytic hierarchy process; ;

[0079] For the i The similarity center of the parts to be processed, is the mean of similarity centrality, For the i Production cycle of spare parts to be processed; is the mean of the production cycle; For the i The number of spare parts to be processed, is the quantity mean;

[0080] It should be noted that the analytic hierarchy process (AHP) is a structured and systematic decision support technology. It decomposes complex multi-criteria decision problems into a series of interrelated and hierarchical sub-problems or factors, uses expert judgment to compare the importance and quantitatively evaluate the factors at each level, and finally obtains the relative advantages and disadvantages of each decision plan through synthetic calculation, providing decision makers with a clear and quantitative decision-making basis. This method combines qualitative and quantitative analysis, effectively improving the scientificity and accuracy of decision-making.

[0081] Based on historical data and management expectations for processing tasks, the complexity threshold is set in advance; if the complexity of the task obtained is If the complexity threshold is exceeded, it means that the work tasks in the current processing area are relatively complex. In order to improve the corresponding processing efficiency, it may be necessary to optimize the allocation of processing tasks. At this time, an allocation instruction is issued to the outside.

[0082] When using, obtain the corresponding task complexity through analysis , according to the complexity of the task The complexity of the current processing task can be judged. When the task complexity is different, different scheduling strategies can be selected to improve the targetedness of the scheduling.

[0083] Step 103: after receiving the allocation instruction, record the work ability data of each processing worker in the processing area, including processing skills, qualifications, experience, work interest preference data, etc., and build a worker ability database after summarizing;

[0084] After decomposing the current processing task, several subtasks are obtained, and the content-based recommendation algorithm recommends corresponding processing workers for the subtasks from the worker capability database to facilitate the reasonable allocation of work tasks;

[0085] When using, combine the contents in steps 101 to 103:

[0086] As a further content, when there are many sub-tasks and the processing tasks are more complicated, after collecting the work data of each processing worker in the processing area, the processing workers can be matched between the sub-tasks to make the sub-tasks and the processing workers more adaptable, which can improve the subsequent processing efficiency.

[0087] At present, there are many varieties of spare parts to be processed, there are certain differences between the batches of various spare parts, and the corresponding production cycles are long and of varying lengths. Although the existing processing task scheduling method can optimize the scheduling plan and improve the overall production and processing efficiency by reasonably arranging and allocating the processing tasks; however, when there are abnormalities in the environmental conditions in the processing area or process of the spare parts, or when an emergency occurs, the scheduling plan obtained by this optimization often cannot be adjusted and responded to in the first time, resulting in the scheduling plan cannot be effectively executed, and it is difficult to help improve the efficiency of production and processing.

[0088] Step 2: After scheduling production for each subtask, simulate and test the obtained scheduling scheme, and screen the scheduling scheme based on the obtained test data. If the scheduling scheme that serves as the target scheme cannot be screened out, select one from several scheduling schemes for optimization;

[0089] The step 2 includes the following contents:

[0090] Step 201: After determining each subtask and the corresponding processing worker, collect the equipment data, material data, processing worker data and processing environment data in the processing area as constraint condition data, use the pre-trained simulated annealing algorithm to schedule the current subtasks, and obtain several scheduling plans;

[0091] When in use, after determining the current processing workers and task complexity, the simulated annealing algorithm can quickly complete the processing task scheduling, so that the obtained scheduling plan can be more adapted to the current processing tasks and the allocation of processing workers;

[0092] Step 202: After training the machine learning algorithm with the collected sample data, a digital twin model for spare parts production is built;

[0093] Taking the obtained scheduling scheme as input, the spare parts production digital twin model is used to simulate the spare parts processing and production process, and the corresponding scheduling test data is obtained, including order processing time, scheduling accuracy, resource utilization, production efficiency and production cost, etc. The efficiency value is generated from the scheduling test data , the effects of different scheduling schemes are evaluated based on the efficiency value, as follows:

[0094] On production efficiency and resource utilization Perform linear normalization and map the corresponding data values ​​to the interval Then according to the following formula:

[0095] ;

[0096] Weight coefficient: , ,and ; Among them, the value of the weight coefficient can be consistent with the previous value;

[0097] Based on historical data and production task scheduling management expectations, efficiency thresholds are pre-set; efficiency values ​​are The scheduling scheme that is higher than the efficiency threshold is used as the target scheme, so as to find the best production and processing task arrangement based on the efficiency value;

[0098] As a further step, simulation tests are performed based on the obtained scheduling schemes, and efficiency values ​​are generated from the obtained test data. , according to the efficiency value Screen several scheduling plans and obtain the most efficient one, which can be used as a reference to increase processing production.

[0099] Step 203: If there is no efficiency value For the scheduling plans that are higher than the efficiency threshold, the scheduling plan closest to the efficiency threshold is used as the plan to be optimized, and a sensitivity analysis is performed on the scheduling test data to determine the sensitivity factors that affect the effect; the obtained plan to be optimized is optimized using a multi-objective optimization algorithm, such as optimizing cost, time, and quality, and the optimized scheduling plan is used as the target plan;

[0100] It should be noted that sensitivity analysis is a comprehensive and in-depth evaluation technique that aims to explore how small changes in input variables can significantly affect the output of a model or system. By systematically examining the changes in each variable within a given range, this technique can accurately identify variables that have a key impact on model predictions or system performance. This analysis not only helps us understand the operating mechanism of the model or system more deeply, but also provides valuable insights for optimizing decision-making processes, improving prediction accuracy, and developing effective risk management strategies. Sensitivity analysis plays a vital role in many fields, such as finance, environmental science, engineering, and medical research.

[0101] When using, combine the contents in steps 201 to 203:

[0102] As a further content, when several scheduling plans are currently given, the selected scheduling plan is optimized in a targeted manner to obtain a feasible scheduling plan, which is used as a target plan and applied in scheduling, which can improve the subsequent scheduling efficiency.

[0103] Step 3: Collect the production environment data in the processing area and identify the anomalies. If anomalies exist, make adaptive adjustments. Generate the scene anomaly degree from the corresponding abnormal state data. , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside;

[0104] The step three includes the following contents:

[0105] Step 301: deploy a sensor network in the processing area, including temperature sensors, pressure sensors, and vibration sensors; the sensor network monitors the operating status of various equipment and production environment data in the processing area in real time, such as environmental temperature and humidity data and vibration data;

[0106] The machine learning algorithm is trained with the labeled sample data to obtain the trained abnormal data recognition model. The production environment data is used as input, and the trained abnormal data recognition model is used to perform abnormality recognition. When there are abnormalities in the production environment data, a reminder instruction is issued to the outside.

[0107] Among them, the impact of environmental temperature and humidity data on processing equipment:

[0108] Temperature: Processing equipment usually needs to operate within a certain temperature range to ensure its accuracy and stability. For example, precision CNC lathes generally require the use of a constant temperature environment to ensure the working accuracy of the machine tool, usually requiring a constant temperature of around 20 degrees Celsius. Too high or too low a temperature may cause equipment performance degradation, such as deformation of circuit boards, accelerated equipment aging, too fast evaporation of coolant, or excessive wear of mechanical parts.

[0109] Temperature changes can also affect the thermal expansion and contraction of metal materials, thereby affecting processing accuracy. For example, during the processing, due to changes in ambient temperature, the temperature rise of various parts of the machine tool is uneven, which may lead to changes in the machine tool's shape accuracy and processing accuracy.

[0110] Humidity: Excessive humidity may cause electrical short circuits, metal corrosion and other problems. When the humidity is too high, the moisture in the air will adhere to the surface of the insulating material, reducing the insulation strength of the equipment and increasing the risk of electrical failure. At the same time, moisture will also accelerate the corrosion process of metal materials, especially at high temperatures, the rust rate will be faster.

[0111] Too low humidity may cause unstable operation, fire risk, etc. In an overly dry environment, the generation and accumulation of static electricity may increase, posing a threat to the normal operation of the equipment.

[0112] Impact of vibration data on processing equipment:

[0113] Reduced processing quality: Vibration will reduce the quality of the processed surface, increase the surface roughness, and affect the dimensional accuracy and shape accuracy of the workpiece. At the same time, vibration will also affect the relative displacement between the workpiece and the tool, destroying the normal motion trajectory of the CNC lathe.

[0114] Reduced tool life: Vibration will increase the cutting wear of the tool, and in severe cases will cause cracks on the edge of the tool, thereby reducing the service life of the tool.

[0115] Reduced productivity: To avoid severe vibration, the cutting volume must be reduced, which will lead to reduced productivity. At the same time, due to the short tool life, frequent tool replacement and sharpening will also increase downtime, further reducing productivity.

[0116] Impact on equipment accuracy and life: Vibration can cause the displacement components of the machine tool to loosen, increase the wear rate of the machine tool components, and thus reduce the cutting quality and expected life of the machine tool. In addition, vibration may also have an adverse effect on electronic components, resulting in poor contact or failure.

[0117] When in use, after the current schedule has been completed, considering that external conditions will bring a certain degree of interference to the current schedule, therefore, as a further content, the environmental condition data in the processing area is collected and abnormality detection is performed, and timely reminders can be given when abnormalities exist;

[0118] Step 302: train a machine learning algorithm with the labeled sample data to obtain a trained environment control model; after receiving the reminder instruction, set corresponding qualified reference values ​​for the production environment parameters in the processing area, and use the trained environment control model to adjust the environment in the processing area;

[0119] If several reminder instructions are obtained continuously in the current stage, it means that the adjustment of the production environment may be difficult to achieve the expected effect, and there are persistent abnormalities in the processing area. The abnormal time node, abnormal location and corresponding abnormal degree of the environment in the processing area are obtained, and the abnormal state data set is generated after aggregation;

[0120] When in use, when there are abnormal environmental conditions in the working area, respond to the abnormal environmental conditions promptly and quickly, respond to and adjust the abnormal conditions, and avoid the poor environmental conditions in the processing area from affecting the execution of the current subtasks and the corresponding processing equipment;

[0121] Step 303: Under dimensionless conditions, the abnormal time nodes in the abnormal state data set T , Abnormal location L and degree of abnormality Generate scene abnormality , as follows:

[0122] ;

[0123] Where: is the time attenuation coefficient, and its value falls between 0 and 1. is the comprehensive influence function, is the distance weight coefficient, and its value falls between 0 and 1. Define the domain for the integral, specifying the analysis range in time, space, and degree of anomaly;

[0124] Based on historical data and abnormal management of environmental conditions in the processing area, corresponding abnormal thresholds are pre-set;

[0125] If the scene abnormality obtained If the preset abnormality threshold is exceeded, it means that the environmental conditions in the processing area have a high degree of environmental abnormality, which may have a certain degree of impact on the current production schedule. In order to reduce this impact, a fault monitoring instruction is issued to the outside;

[0126] When using, combine the contents in steps 301 to 303:

[0127] As a further content, on the basis of adaptively adjusting the environmental conditions in the processing area, continue to collect abnormal state data, and generate scene abnormality degree from the abnormal state data. , the abnormal conditions in the processing area can be evaluated to verify whether the current abnormal conditions will affect the formation of processing equipment. If there is a large impact, the processing equipment or processing environment conditions can be adaptively adjusted.

[0128] Step 4: Fault prediction is performed on the processing equipment in the processing area. Fault maintenance or performance optimization is performed on the equipment based on the acquired fault prediction data. Abnormal events in the processing area are identified. If abnormal events occur, the production scheduling plan is optimized again.

[0129] The step 4 includes the following contents:

[0130] Step 401: train a machine learning algorithm with the labeled sample data to obtain a trained fault prediction model; monitor equipment operation status data, historical fault and maintenance data, and environmental condition data in the processing area in real time;

[0131] After receiving the fault monitoring instruction, use the trained fault prediction model to predict the potential fault of the equipment. If a fault exists, send an external fault maintenance instruction;

[0132] Step 402: Taking processing equipment fault maintenance as the target word, after deep retrieval and entity relationship building, pre-build a processing equipment fault maintenance knowledge graph; if a fault maintenance instruction is received, extract features from the acquired fault prediction data to obtain corresponding fault features; according to the correspondence between the fault features and the fault maintenance plan, the processing equipment fault maintenance knowledge graph gives the corresponding fault handling strategy;

[0133] If no fault command is received, the equipment is temporarily not faulty. The optimization goal is to reduce the fault rate and improve the processing efficiency. The pre-trained genetic algorithm is used to optimize the equipment control parameters to achieve the optimization of equipment performance.

[0134] During use, when the current scene conditions continue to be abnormal, in order to extend the service life of the processing equipment and improve the performance status, fault prediction is performed on each device under the current scene conditions. When a fault may exist, the knowledge graph promptly matches the corresponding fault handling plan to avoid the impact of the sudden event of the fault on the current processing task. At the same time, when there is no fault at present, the processing efficiency of the equipment can be improved by optimizing the equipment, so that the current processing scheduling plan can be completed smoothly.

[0135] It should be noted that door and window processing equipment can be divided into many types according to the types of doors and windows being processed and the different processes. Some of them are as follows: cutting equipment, including double-head saws, profile cutting saws and V-mouth saws; welding and angle assembly equipment, including plastic steel profile welding machines, angle assembly machines, and presses; milling and punching equipment, including end milling, copy milling machines, and punching machines; other auxiliary equipment, including punching machines, corner cleaning equipment, rolling compound machines, and toothed strip threading machines.

[0136] Step 403: collect production status data in the processing area in real time, including material and inventory data, product quality data, personnel operation data, and energy consumption data; train a machine learning algorithm with the labeled sample data to obtain a trained abnormal event recognition model;

[0137] Taking production status data as input, the trained abnormal event recognition model is used to identify abnormal events. If abnormal events are identified, such as equipment failure and material shortage, it is necessary to quickly respond to the emergency in production. The pre-trained simulated annealing algorithm is used again to search and optimize the production scheduling plan to obtain the production scheduling plan after the response.

[0138] When using, combine the contents in steps 401 to 403:

[0139] As a further content, when the current scheduling plan continues to be executed, by collecting and monitoring the current production and processing status, and detecting and identifying the current abnormal events, emergencies can be determined and discovered in real time. When an emergency occurs, this is used as an optimization node to readjust the current scheduling plan so that the scheduling plan can adapt to the current real-time scenario and production status.

[0140] Step 5: Generate progress deviation value after real-time monitoring of production progress data of each subtask , if the progress deviation value Exceed expectations, match the corresponding improvement plan according to key factors, and send the improvement report to the data receiving end after a time interval that meets the constraints;

[0141] The step five includes the following contents:

[0142] Step 501: monitor the production progress of each subtask in real time, and generate a progress deviation value from the acquired progress data under dimensionless conditions. , determine the gap between the current progress and the expected progress in the following way:

[0143] ;

[0144] in, , is the weight coefficient, , ,and , n is the number of subtasks, For the i The production progress of each subtask, It is the qualified reference value for production progress;

[0145] Pre-set deviation thresholds based on historical data and production progress management of production tasks;

[0146] If the progress deviation value If the deviation threshold is exceeded, it means that the current production progress is lower than expected, and the production plan needs to be adjusted according to the actual situation, such as giving priority to urgent orders and adjusting equipment scheduling. At this time, a progress adjustment instruction is issued to the outside.

[0147] When in use, based on the real-time update of the scheduled tasks, the progress status data of each subtask is collected and monitored in real time, and the progress deviation value is constructed based on this. The current processing progress of the processing task is comprehensively judged and evaluated based on the progress deviation value, so as to evaluate the effectiveness of the current scheduled task. When the current construction progress is significantly different from the expected progress, the current scheduling plan can be further optimized and adjusted.

[0148] Step 502: Perform principal component analysis on the equipment operation status, production environment data and production status data in the processing area to obtain key factors affecting the progress of the processing task, match corresponding improvement plans based on the key factors, including improvement measures, responsibility allocation, timetable and expected results, and generate an improvement report;

[0149] Constrain the time node for issuing the improvement report, and after a time interval that meets the constraint conditions After that, the improvement report is sent to the data receiving end, and the constraints are as follows:

[0150] ;

[0151] Where: is the progress deviation value at time t, is the corresponding qualified reference value, It is i Progress deviation value, is the monitoring interval; weight coefficient, , , ;

[0152] e The value is 2.718, is the time attenuation coefficient, ranging from 0 to 1. is the distance attenuation coefficient, ranging from 0.01 to 0.1; is the first j The event node of an abnormal event, For the i and j The time interval between event nodes, n is the number of abnormal events, N is the expected number;

[0153] When using, combine the contents in steps 501 and 502:

[0154] As a further content, when the current scheduling plan needs to be optimized, the corresponding key factors are determined through principal component analysis, and the processing status and scheduling plan are continuously improved. This can realize a dynamic improvement process and improve the overall processing efficiency of each subtask. In addition, by sending improvement reports to the outside as a reference, the output frequency of the improvement report is adapted to the current processing scenario and processing progress, avoiding the interference of the scheduling plan adjustment on the execution of the processing task.

[0155] Building a knowledge graph for processing equipment failure maintenance is a complex but valuable process. The following is a detailed method for building a knowledge graph for processing equipment failure maintenance:

[0156] Clarify goals and requirements - First, it is necessary to clarify the goals and requirements of building a knowledge graph. This includes determining the application scenarios of the knowledge graph, user groups, types of processing equipment to be covered, and their fault types, etc. This will help guide the subsequent knowledge modeling and knowledge acquisition process.

[0157] Knowledge modeling - Knowledge modeling is one of the core steps in building a knowledge graph. It involves defining an ontology, which is a conceptual framework that describes entities, attributes, and relationships. In the knowledge graph of processing equipment failure maintenance, entities may include equipment type, failure type, maintenance method, etc.; attributes may include equipment model, production date, failure occurrence time, etc.; relationships may include the correspondence between equipment type and failure type, the correspondence between failure type and maintenance method, etc.

[0158] Knowledge acquisition - Knowledge acquisition is to collect data related to processing equipment failure maintenance from various data sources. These data may come from equipment manuals, maintenance records, expert experience, the Internet, etc. When collecting data, attention should be paid to the accuracy, completeness and consistency of the data. At the same time, the data needs to be pre-processed, such as cleaning, conversion, integration, etc., to facilitate subsequent knowledge extraction and fusion.

[0159] Knowledge extraction - Knowledge extraction is the process of extracting entity, attribute and relationship information from text data. In the knowledge graph of processing equipment failure maintenance, named entity recognition (NER), relationship extraction and other technologies can be used to extract relevant information. For example, entity information such as equipment type and model can be extracted from the equipment manual; entity and relationship information such as fault type and maintenance method can be extracted from the maintenance record.

[0160] Knowledge fusion - Knowledge fusion is the process of integrating knowledge obtained from different data sources. In the knowledge graph of processing equipment fault maintenance, there may be multiple heterogeneous knowledge systems or entity representations. Therefore, they need to be fused to eliminate contradictions and ambiguities. This can be achieved through technologies such as entity linking and knowledge system mapping.

[0161] Knowledge storage and query - After building the knowledge graph, it needs to be stored in a suitable database for efficient query and reasoning. Currently, common knowledge graph storage methods include graph databases (such as Neo4j) and RDF databases (such as Jena). At the same time, it is also necessary to design a suitable query interface and query language (such as SPARQL) so that users can easily query and obtain the required knowledge.

[0162] Knowledge reasoning and updating - Knowledge reasoning is the process of using existing knowledge to derive new knowledge. In the knowledge graph of processing equipment failure maintenance, logical reasoning, association rules and other technologies can be used to discover new fault types, maintenance methods, etc. At the same time, since equipment types and fault types may be continuously updated with the development of technology, it is also necessary to regularly update the knowledge graph to maintain its timeliness and accuracy.

[0163] See also Figure 2 The present invention provides an intelligent manufacturing advanced planning and scheduling system, including:

[0164] The task complexity analysis unit collects the processing task data in the processing area and generates the task complexity , according to the complexity of the task If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database;

[0165] The scheduling unit, after scheduling the production of each subtask, performs simulation tests on the obtained scheduling schemes, and screens the scheduling schemes based on the obtained test data. If a scheduling scheme as the target scheme cannot be screened out, one of several scheduling schemes is selected for optimization;

[0166] The abnormality recognition unit collects the production environment data in the processing area and then identifies the abnormality. When an abnormality exists, it makes adaptive adjustments. The scene abnormality degree is generated from the corresponding abnormal state data. , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside;

[0167] The fault prediction unit predicts faults of processing equipment in the processing area, performs fault maintenance or performance optimization on the equipment based on the acquired fault prediction data, identifies abnormal events in the processing area, and re-optimizes the production scheduling plan if abnormal events occur;

[0168] The progress tracking unit monitors the production progress data of each subtask in real time and generates the progress deviation value , if the progress deviation value Exceed expectations, match the corresponding improvement plan according to key factors, and send the improvement report to the data receiver after a time interval that meets the constraints.

[0169] See also Figure 3 ,The present invention provides an intelligent manufacturing advanced planning and scheduling device, comprising, at least one processor;

[0170] The memory is used to store executable instructions, where the instructions are executed by at least one of the processors to enable at least one of the processors to execute the scheduling method.

[0171] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0172] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0173] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0174] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0175] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An advanced planning and scheduling method for intelligent manufacturing, characterized in that: include, Generate task complexity after collecting processing task data in the processing area , if the task complexity If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database; After the production scheduling of each subtask is completed, the obtained scheduling scheme is simulated and tested, and the scheduling scheme is screened by the obtained test data. If the scheduling scheme as the target scheme cannot be screened out, one of several scheduling schemes is selected for optimization; After collecting the production environment data in the processing area, abnormality identification is performed, and adaptive adjustments are made when abnormalities occur. The scene abnormality degree is generated from the corresponding abnormal state data. , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside; Fault prediction of processing equipment in the processing area, and fault maintenance or performance optimization of equipment based on the acquired fault prediction data; Identify abnormal events in the processing area, and if there are abnormal events, re-optimize the production scheduling plan; in ； The production status data in the processing area is collected in real time. The production status data is used as input and the trained abnormal event recognition model is used to identify abnormal events. If an abnormal event is identified, the pre-trained simulated annealing algorithm is used again to optimize the production scheduling plan and obtain the production scheduling plan after the response. Generate progress deviation value after real-time monitoring of production progress data of each subtask , if the progress deviation value Exceeding expectations, matching corresponding improvement plans based on key factors, and sending improvement reports to the data receiving end after a time interval that meets the constraints. Principal component analysis is performed on the equipment operating status, production environment data, and production status data in the processing area to obtain key factors that affect the progress of processing tasks, matching corresponding improvement plans based on key factors, and generating improvement reports; Constrain the time node for issuing the improvement report, and after a time interval that meets the constraint conditions After that, the improvement report is sent to the data receiving end, and the constraints are as follows: ； Where: is the progress deviation value at time t, is the corresponding qualified reference value, is the ith progress deviation value, is the monitoring interval; weight coefficient, , , ; The value is 2.718, is the time attenuation coefficient, is the distance attenuation coefficient, is the event node of the jth abnormal event at time t, is the time interval between the i-th and j-th event nodes, n is the number of abnormal events, The expected number.

2. The intelligent manufacturing advanced planning and scheduling method according to claim 1, characterized in that: Collect the processing task data of the current stage and the attribute data of the parts to be processed, perform similarity analysis between two parts to be processed, and calculate the similarity center of each part ; Generate task complexity from production task data set , if the task complexity When the complexity threshold is exceeded, the work ability data of each processing worker in the processing area is recorded, and the data is summarized to build a worker ability database; after decomposing the current processing task, several subtasks are obtained.

3. The intelligent manufacturing advanced planning and scheduling method according to claim 2 is characterized in that: Similarity Centrality , production cycle And the number of spare parts in each batch Perform linear normalization and map the corresponding data values ​​to the interval Then generate the task complexity as follows : ； Weight coefficient: , , ; For the The similarity center of the parts to be processed, is the mean of similarity centrality, For the Production cycle of spare parts to be processed; is the mean of the production cycle; For the The number of spare parts to be processed, is the quantity mean.

4. The intelligent manufacturing advanced planning and scheduling method according to claim 3 is characterized in that: Collect equipment data, material data, processing worker data and processing environment data in the processing area as constraints, use the pre-trained simulated annealing algorithm to schedule the current subtasks, and obtain several scheduling plans; Taking the obtained scheduling plan as input, the spare parts production digital twin model is used to simulate the spare parts processing and production process to obtain the corresponding scheduling test data.

5. The intelligent manufacturing advanced planning and scheduling method according to claim 4 is characterized in that: Generate efficiency values ​​from schedule test data , among which, the production efficiency and resource utilization Perform linear normalization and map the corresponding data values ​​to the interval Then according to the following formula: ； Weight coefficient: , ,and ; The efficiency value The scheduling scheme that is higher than the efficiency threshold is used as the target scheme; If there is no efficiency value For the scheduling plans that are higher than the efficiency threshold, the scheduling plan closest to the efficiency threshold is used as the plan to be optimized, and the obtained plan to be optimized is optimized using a multi-objective optimization algorithm, and the optimized scheduling plan is used as the target plan.

6. The intelligent manufacturing advanced planning and scheduling method according to claim 5 is characterized by: The sensor network monitors the operating status of each device and production environment data in real time within the processing area; Taking production environment data as input, the trained abnormal data recognition model is used to identify anomalies. When there are anomalies in the production environment data, a reminder instruction is issued to the outside.

7. The intelligent manufacturing advanced planning and scheduling method according to claim 6 is characterized by: After receiving the reminder command, the trained environment control model is used to adjust the environment in the processing area; If several reminder instructions are obtained continuously in the current stage, the abnormal time node, abnormal location and corresponding abnormal degree of the environment in the processing area are obtained, and the abnormal status data set is generated after being summarized.

8. The intelligent manufacturing advanced planning and scheduling method according to claim 7 is characterized in that: The abnormal time node in the abnormal state data set , Abnormal location and degree of abnormality Generate scene abnormality , as follows: ； Where: is the time attenuation coefficient, is the comprehensive influence function, is the distance weight coefficient, For the integration domain, specify the analysis range in time, space, and degree of anomaly.

9. The intelligent manufacturing advanced planning and scheduling method according to claim 8, characterized in that: Real-time monitoring of equipment operating status data, historical fault and maintenance data, and environmental conditions data within the processing area; After receiving the fault monitoring instruction, use the trained fault prediction model to predict the potential fault of the equipment. If a fault exists, send an external fault maintenance instruction; Taking processing equipment failure maintenance as the target word, a knowledge graph of processing equipment failure maintenance is pre-built; If a fault maintenance instruction is received, feature extraction is performed on the acquired fault prediction data to obtain the corresponding fault features, and the corresponding fault handling strategy is given by the processing equipment fault maintenance knowledge graph.

10. The intelligent manufacturing advanced planning and scheduling method according to claim 9, characterized in that: If no fault command is received, the equipment control parameters are optimized using a pre-trained genetic algorithm with the goal of reducing the fault rate and improving the processing efficiency.

11. The intelligent manufacturing advanced planning and scheduling method according to claim 10, characterized in that: Monitor the production progress of each subtask in real time, and generate progress deviation values ​​based on the acquired progress data , if the progress deviation value If the deviation threshold is exceeded, a progress adjustment instruction is issued to the outside to generate a progress deviation value The way is as follows: ； in, is the weight coefficient, , ,and , , is the number of subtasks, For the The production progress of each subtask, It is a qualified reference value for production progress.

12. An intelligent manufacturing advanced planning and scheduling system, applying the scheduling method according to any one of claims 1 to 11, characterized in that: include, The task complexity analysis unit collects the processing task data in the processing area and generates the task complexity , if the task complexity If it exceeds expectations, the recommendation algorithm will recommend corresponding processing workers for the subtask from the worker capability database; The scheduling unit, after scheduling the production of each subtask, performs simulation tests on the obtained scheduling schemes, and screens the scheduling schemes based on the obtained test data. If a scheduling scheme as the target scheme cannot be screened out, one of several scheduling schemes is selected for optimization; The abnormality recognition unit collects the production environment data in the processing area and then identifies the abnormality. When an abnormality exists, it makes adaptive adjustments. The scene abnormality degree is generated from the corresponding abnormal state data. , if the obtained scene abnormality If the preset abnormal threshold is exceeded, a fault monitoring instruction will be issued to the outside; The fault prediction unit predicts faults of processing equipment in the processing area, performs fault maintenance or performance optimization on the equipment based on the acquired fault prediction data, identifies abnormal events in the processing area, and re-optimizes the production scheduling plan if abnormal events occur; The progress tracking unit monitors the production progress data of each subtask in real time and generates the progress deviation value , if the progress deviation value Exceed expectations, match the corresponding improvement plan according to key factors, and send the improvement report to the data receiver after a time interval that meets the constraints.

13. An intelligent manufacturing advanced planning and scheduling device, characterized in that: include, at least one processor; A memory for storing executable instructions, wherein the instructions are executed by at least one of the processors to enable at least one of the processors to perform the steps of the method according to any one of claims 1 to 11.

Images