Production task dynamic adjustment method and system based on industrial vertical large model

Abstract

The present application relates to the field of data processing, more particularly, the present application relates to a production task dynamic adjustment method and system based on an industrial vertical large model, the method comprising: taking a production task scheduling as a time node, collecting workpiece process parameters from the completion of scheduling to the current time of processing completion in real time, and obtaining historical workpiece process parameters at each process, preprocessing the collected data to obtain historical parameter sequences and current parameter sequences respectively; calculating the influence coefficient between real-time production processes based on the current parameter sequence; calculating the influence coefficient cumulative value corresponding to the processing of all workpieces from the completion of scheduling to the current time, and when the influence coefficient cumulative value is greater than a preset scheduling threshold, triggering a rescheduling instruction to adjust the pre-set production task. The present application realizes the self-adaptive optimization of the production process through real-time data analysis and historical experience fusion.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing. More particularly, the present application relates to a production task dynamic adjustment method and system based on an industrial vertical large model. BACKGROUND

[0002] Job shop is a common production mode in modern manufacturing industry, covering multiple processing links. In actual production, the workshop environment is full of dynamic changes, such as machine failure or emergency orders and other emergency events. In order to ensure production efficiency, these events must be responded quickly, and production plans must be adjusted in real time to ensure that workpieces are completed on time.

[0003] At present, the processing technology for emergency events is relatively mature. However, when dealing with these emergency situations, low-priority disturbances such as minor device fluctuations or slight delays in materials are often overlooked. Although the impact is small, these disturbances can trigger a chain reaction through the interaction between processes, thereby potentially affecting the entire production process. Therefore, while dealing with emergency events, attention should also be paid to the mutual influence between processes, and production tasks should be adjusted in real time to maintain the stability of the production process and the timely completion of workpieces. SUMMARY

[0004] To solve the technical problem that the existing production scheduling technology ignores the chain reaction caused by low-priority disturbances through process coupling when dealing with emergency dynamic events, thereby affecting the stability of the entire production process and the timely completion of workpieces, the present application provides solutions in the following aspects.

[0005] In a first aspect, a production task dynamic adjustment method based on an industrial vertical large model, comprising:

[0006] Taking a production task scheduling as a time node, real-time collection of workpiece process parameters from the completion of scheduling to the current time of processing completion, and acquisition of historical workpiece process parameters at each process, preprocessing of the collected data to obtain historical parameter sequences and current parameter sequences respectively;

[0007] Based on the historical parameter sequences, the influence coefficients between the historical production processes are calculated; based on the current parameter sequences, the influence coefficients between the real-time production processes are calculated;

[0008] The influence coefficient cumulative value corresponding to the processing of all workpieces from the completion of scheduling to the current time is calculated, and when the influence coefficient cumulative value is greater than a preset scheduling threshold, a rescheduling instruction is triggered to adjust the pre-set production task.

[0009] Preferably, the process parameters include workpiece processing time, processing device temperature, and processing device pressure.

[0010] Preferably, the calculation of the influence coefficient between the historical production processes based on the historical parameter sequence comprises:

[0011] Defining the standard value of the process parameter of the workpiece in any process, constructing the standard parameter sequence of the workpiece in each process; taking the absolute value of the difference between each parameter of the historical parameter sequence of any workpiece in each process and the corresponding standard parameter to form the deviation difference sequence;

[0012] Calculating the mutual information between the deviation difference sequences of the previous and next processes of all workpieces, calculating the standard deviation of the deviation difference sequence of any workpiece in all processes, adding the product of the mutual information and the standard deviation of all historical workpieces, and then dividing by the total number of historical workpieces and performing normalization processing to obtain the influence coefficient between adjacent processes.

[0013] Preferably, the calculation of the influence coefficient between the historical production processes based on the historical parameter sequence comprises:

[0014] Defining the standard value of the process parameter of the workpiece in any process, constructing the standard parameter sequence of the workpiece in each process; taking the absolute value of the difference between each parameter of the historical parameter sequence of any workpiece in each process and the corresponding standard parameter to form the deviation difference sequence;

[0015] Calculating the Pearson correlation coefficient between the deviation difference sequences of the previous and next processes of all workpieces, calculating the standard deviation of the deviation difference sequence of any workpiece in all processes, and then adding the product of the Pearson correlation coefficient and the standard deviation of all historical workpieces, and then dividing by the total number of historical workpieces and performing normalization processing to obtain the influence coefficient between adjacent processes.

[0016] Preferably, the calculation method of the influence coefficient between the real-time production processes is the same as that of the influence coefficient between the historical production processes.

[0017] Preferably, the setting of the scheduling threshold comprises:

[0018] For all historical production, calculating the average of the cumulative sum of the difference between the actual value and the standard value of the process parameter of any process as the deviation average of the process, and further calculating the standard deviation value of the deviation average of the process;

[0019] The deviation baseline threshold value of the single process is obtained by weighted summing the deviation average and the standard deviation value; the scheduling threshold value is obtained by combining the deviation baseline threshold value of the single process with the influence coefficient between the process and the next process.

[0020] Preferably, the setting of the scheduling threshold further comprises adjusting the scheduling threshold to obtain a final scheduling threshold, and the adjustment process comprises:

[0021] The difference between the deviation of each process in the current production process from the mean value and the deviation of the corresponding process in the history from the mean value is calculated, and the adjustment factor is obtained by summing all processes.

[0022] In a second aspect, a production task dynamic adjustment system based on an industrial vertical large model includes a processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement any of the production task dynamic adjustment methods based on an industrial vertical large model.

[0023] The present application has the following advantages:

[0024] The present application can dynamically adjust production tasks according to actual conditions in the production process by collecting data in real time and calculating the cumulative value of the influence coefficient, triggering a rescheduling instruction when the scheduling threshold is exceeded, avoiding the inability to respond to unexpected situations or production deviations due to fixed plans, and thus improving the accuracy and adaptability of production plans and ensuring the smooth completion of production tasks.

[0025] In calculating the influence coefficient between processes, historical and real-time data are fully utilized. By defining a standard parameter sequence and constructing a deviation difference sequence, whether using a mutual information-based calculation method or a Pearson correlation coefficient calculation method, the mutual influence relationship and correlation degree between processes can be accurately revealed from different dimensions and angles. Second, the scheduling threshold is not only based on the historical deviation mean value and standard deviation, but also dynamically adjusted (through an adjustment factor) based on the actual deviation of the current production, making the threshold setting more accurate and avoiding excessive or insufficient scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 The present application is based on the production task dynamic adjustment method of the industrial vertical large model, and the method flowchart of steps S1-S3 is shown in the figure. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all.

[0028] Referring to Figure 1 The production task dynamic adjustment method based on the industrial vertical large model includes steps S1-S3, as follows:

[0029] S1: Taking a production task scheduling as the time node, collect the workpiece process parameters in real time from the completion of scheduling to the current time of processing completion, and obtain the process parameters of historical workpieces in each process. Preprocess the collected data to obtain the historical parameter sequence and the current parameter sequence respectively.

[0030] In one embodiment, parameter data for at least 100 completed workpieces during various processes (including turning, milling, grinding, etc.) are collected, including parameters such as workpiece processing time, equipment temperature, and equipment pressure. The parameter data for each workpiece in any process are then combined to form a corresponding historical parameter sequence.

[0031] Using a single production task scheduling as a time node, the process parameters of at least 100 completed workpieces are collected in real time from the completion of the scheduling to the current moment. These process parameters are identical to the parameter data acquired above. Similarly, the parameter data of each workpiece in any process are used to form the corresponding current parameter sequence.

[0032] Further preprocessing of the historical and current parameters collected above is performed, such as using Gaussian filtering to remove noise and standardizing each parameter to ensure that the data are comparable on the same scale.

[0033] S2: Calculate the influence coefficients between historical production processes based on historical parameter sequences; calculate the influence coefficients between real-time production processes based on current parameter sequences.

[0034] In modern manufacturing workshops, urgent dynamic events such as machine malfunctions and emergency work order interruptions are frequently encountered. These events require timely response and handling; otherwise, the smooth progress of the entire production plan will be affected. Besides urgent events, there are also low-priority disturbances in the production process, such as minor equipment fluctuations and slight material delays. Although these disturbances may have small individual impacts, they can create a chain reaction through process coupling, gradually accumulating and causing a significant impact on the production process. Therefore, while responding promptly to urgent dynamic events, it is necessary to simultaneously pay attention to the cumulative impact of low-priority disturbances.

[0035] By analyzing historical data and quantifying the deviation of workpiece parameters in each processing step, the degree of influence between different processes can be clearly identified. This quantitative analysis helps to better understand the inherent relationships within the production process. Calculating the influence coefficient between adjacent processes can more accurately reflect the impact of the preceding process on the following process. This impact may manifest as a change in the processing quality of the following process due to parameter deviations in the preceding process.

[0036] First, define the standard values ​​of the process parameters of the workpiece in any process, such as the standard processing time being 10 minutes, and then construct the standard parameter sequence of the workpiece in each process.

[0037] Further, the absolute value of the difference between each parameter in the historical parameter sequence of any workpiece in each process and the corresponding standard parameter is taken to form a deviation difference sequence.

[0038] In one embodiment, the method of mutual information is used to quantify the influence coefficient of the previous process on the next process, so as to better understand the mutual relationship between the processes in the production process.

[0039] The mutual information between the deviation difference sequences of all workpieces in the previous and next processes is calculated. In short, the mutual information represents how much information in the deviation difference sequences of the two processes is shared with each other, or how strong their correlation is. The greater the mutual information value, the more relevant the deviation degree of the two processes, that is, the higher the influence coefficient of the previous process on the next process.

[0040] At the same time, the standard deviation of the deviation difference sequence of any workpiece in all processes is also considered. The greater the standard deviation, the more dispersed the parameter deviation of the workpiece, and the greater the possibility of abnormality.

[0041] Finally, the product of the mutual information and the standard deviation of all historical workpieces is added, and then divided by the total number of historical workpieces and normalized to obtain the influence coefficient between adjacent processes.

[0042] Exemplarily, the above influence coefficient is expressed by a relationship as follows:

[0043]

[0044] In the formula, is the influence coefficient between process and process , is the deviation difference sequence of the historical th workpiece in process , is the deviation difference sequence of the historical th workpiece in process , represents mutual information, is the total number of historical workpieces, is the standard deviation of the deviation difference sequence of the historical th workpiece in all processes, represents normalization processing.

[0045] In another embodiment, the method of Pearson correlation coefficient is used to quantify the influence coefficient of the previous process on the next process, so as to better understand the mutual relationship between the processes in the production process.

[0046] First, the Pearson correlation coefficient between the deviation value sequences of all workpieces in the front and rear processes is calculated, the standard deviation of the deviation value sequence of any workpiece in all processes is calculated, then the product of the Pearson correlation coefficient and the standard deviation of all historical workpieces is added, and then divided by the total number of historical workpieces and normalized to obtain the influence coefficient between adjacent processes.

[0047] The above influence coefficient satisfies the relationship:

[0048]

[0049] In the formula, is the influence coefficient between the process and the process , is the deviation value sequence of the historical first workpiece in the process , is the deviation value sequence of the historical first workpiece in the process , is the Pearson correlation coefficient, is the total number of historical workpieces, is the standard deviation of the deviation value sequence of the historical first workpiece in all processes, and denotes normalization processing. The Pearson correlation coefficient value approaches 1, indicating that the two sequences are highly positively correlated in terms of change trend, that is, when the parameter deviation degree of one process increases, the parameter deviation degree of the next process also tends to increase; if the Pearson correlation coefficient approaches -1, it indicates that the two sequences are highly negatively correlated, that is, when the parameter deviation degree of one process increases, the parameter deviation degree of the next process tends to decrease; if the Pearson correlation coefficient value is close to 0, it indicates that there is almost no linear correlation between the two sequences.

[0050] In general, if the PPMCC value is high (approaching 1 or -1), it means that the parameter change of the previous process has a significant impact on the next process, which may mean that the slight fluctuations of the previous process will be transmitted to the next process through process coupling; if the PPMCC value is low (close to 0), it indicates that the linear correlation between the two processes is weak, which may mean that the fluctuations of the previous process have little effect on the next process.

[0051]

[0052] ​​The standard deviation is used as a weight factor. The greater the standard deviation, the greater the deviation of the workpiece in the process parameter, the greater the possibility of abnormality, and thus a higher weight is given when calculating the influence coefficient between adjacent processes. The smaller the standard deviation, the smaller the deviation of the workpiece in the process parameter, the smaller the possibility of abnormality, and thus a lower weight is given when calculating the influence coefficient between adjacent processes.

[0053] By calculating the deviation of the historical workpiece and the influence coefficient between the historical processes, a reference can be established for the entire production process, and the approximate range of the deviation of the workpiece parameter in each process from the standard and the mutual influence law of the parameter deviation between adjacent processes under normal production conditions can be understood.

[0054] Since the production process is dynamic and may be affected by various factors such as equipment wear and tear, workpiece raw material quality fluctuations, etc., attention needs to be turned to the workpiece parameter data in the current production process.

[0055] In one embodiment, the influence coefficient between processes in real-time production is calculated according to the calculation formula for calculating the influence coefficient between processes in historical production based on the historical parameter sequence.

[0056] S3: Calculate the influence coefficient cumulative value of all workpiece processing from the completion of scheduling to the current time. When the influence coefficient cumulative value is greater than the preset scheduling threshold, trigger a rescheduling instruction to adjust the pre-set production task.

[0057] After obtaining the influence coefficient between processes in real-time production, the mean of the deviation value sequence of any workpiece in any process from the completion of scheduling to the current time and the product of the influence coefficient between the process and the next process are calculated, and the sum of all workpieces and all processes is summed to obtain the influence coefficient cumulative value of all workpiece processing from the completion of scheduling to the current time. Further analyze the impact of the influence coefficient cumulative value on the production progress.

[0058] It needs to be considered that in actual production, the deviation of the workpiece processing parameter in each process from the parameter standard value has a certain tolerance. By setting a reasonable deviation baseline threshold and cumulative influence coefficient threshold, the production strategy can be flexibly adjusted according to the actual production situation, the flexibility and adaptability of production are improved under the premise of ensuring quality, and the actual production demand is better met.

[0059] In one embodiment, the deviation baseline threshold of each process is calculated according to the historical processing parameter data.

[0060] First, acquire all processing parameter data for any process in each historical production run, and simultaneously determine the standard or ideal values ​​of the processing parameters for that process. For all historical production runs, calculate the average of the cumulative sums of the differences between the actual and standard values ​​of the processing parameters for that process, using this as the deviation from the mean for that process. Then, calculate the standard deviation of the deviation from the mean for that process.

[0061] The deviation from the mean and standard deviation is weighted and summed to obtain the baseline deviation threshold for this process. That is, it satisfies the following relationship:

[0062]

[0063] In the formula, For process Deviation from baseline threshold, Historical production process The deviation from the mean, Historical production process The standard deviation of the deviation from the mean. For safety, the initial value is set to 3.

[0064] Considering the influence between adjacent processes, after obtaining the deviation threshold of each process from the baseline, the deviation threshold of a single process is combined with the influence coefficient between adjacent processes to obtain the scheduling threshold, which satisfies the following relationship:

[0065]

[0066] In the formula, For scheduling threshold, For process Deviation from baseline threshold, For real-time production processes With process The influence coefficient between them This represents the total number of processes.

[0067] The aforementioned scheduling thresholds reflect the cumulative impact of deviations in each process from historical processing data on the overall production task.

[0068] When the disturbances in the production process are small, frequently triggering the rescheduling of production tasks according to a fixed threshold will increase scheduling costs, including the workload of scheduling personnel and the downtime for production equipment. Therefore, it is necessary to further dynamically adjust the scheduling threshold.

[0069] In one embodiment, the difference between the deviation from the mean of each process in the current production process and the deviation from the mean of the corresponding historical process is calculated, and an adjustment factor is obtained by summing all processes.

[0070] The adjustment factor is multiplied by the above-mentioned scheduling threshold to obtain the final scheduling threshold. By dynamically adjusting the scheduling threshold, it is ensured that the scheduling decision can adapt to the current production situation and is not adjusted too frequently to avoid resource waste.

[0071] In one embodiment, the cumulative value of the influence coefficient of the entire machining process caused by completing all machining workpieces to the current time is compared with the final scheduling threshold. If the cumulative value of the influence coefficient is less than or equal to the final scheduling threshold, it indicates that the cumulative influence coefficient in the current production process has not exceeded the final scheduling threshold, and the production task is still within the controllable range. The production task can continue to be arranged according to the last scheduling result. Otherwise, if the cumulative value of the influence coefficient is greater than the final scheduling threshold, it indicates that the disturbance in the production process has accumulated to a certain extent and exceeds the final scheduling threshold. At this time, it is determined that the production process needs to be adjusted, and the production task is automatically triggered to re-schedule instructions to arrange the production task for new scheduling.

[0072] When the re-scheduling instruction is triggered, the target and principle of re-scheduling are determined, such as prioritizing key processes, minimizing the impact on the overall production schedule, and reasonably utilizing existing resources. For example, after issuing the re-scheduling instruction, the production scheduling department is required to develop a new production plan within 2 hours to ensure that the sequence of the current production task and the allocation of resources are adjusted without affecting other production delivery.

[0073] The system includes a processor and a memory, and the memory stores computer program instructions. When the computer program instructions are executed by the processor, the method for dynamically adjusting the production task based on the industrial vertical large model according to the first aspect of the present application is realized.

[0074] The system also includes a communication bus, a communication interface, and other components familiar to those skilled in the art. Their settings and functions are known in the art, and therefore will not be described here.

[0075] It should be noted that for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are all within the scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A method for dynamically adjusting production tasks based on a large-scale industrial vertical model, characterized in that: include: Using a single production task scheduling as a time node, the process parameters of the workpieces from the completion of scheduling to the current time are collected in real time, and the process parameters of historical workpieces in each process are obtained. The collected data is preprocessed to obtain the historical parameter sequence and the current parameter sequence. Calculate the influence coefficients between historical production processes based on historical parameter sequences; calculate the influence coefficients between real-time production processes based on current parameter sequences; Calculate the cumulative value of the influence coefficients corresponding to the processing of all workpieces from the completion of scheduling to the current time. When the cumulative value of the influence coefficients is greater than the preset scheduling threshold, trigger a rescheduling instruction to adjust the pre-set production task. The setting of the scheduling threshold includes: For all historical production, calculate the average of the cumulative sum of the differences between the actual values ​​and the standard values ​​of the processing parameters for any process, and use this as the deviation from the mean for that process. Then, calculate the standard deviation of the deviation from the mean for that process. The deviation from the baseline threshold of a process is obtained by weighted summation of the deviation from the mean and the standard deviation; the deviation from the baseline threshold of a single process is combined with the influence coefficient between the process and its next process to obtain the scheduling threshold. Calculate the difference between the deviation of each process in the current production process and the deviation of the corresponding historical process, and sum the differences for all processes to obtain the adjustment factor; multiply the adjustment factor by the above scheduling threshold to obtain the final scheduling threshold.

2. The method for dynamically adjusting production tasks based on a large industrial vertical model according to claim 1, characterized in that, The process parameters include workpiece processing time, processing equipment temperature, and processing equipment pressure.

3. The method for dynamically adjusting production tasks based on a large industrial vertical model according to claim 1, characterized in that, The calculation of the influence coefficients between various processes in historical production based on historical parameter sequences includes: Define the standard values ​​of the process parameters of a workpiece in any process, and construct the standard parameter sequence of the workpiece in each process; take the absolute value of the difference between each parameter in the historical parameter sequence of any workpiece in each process and the corresponding standard parameter to form a deviation difference sequence; Calculate the mutual information between the deviation difference sequences of all workpieces in the preceding and following processes, calculate the standard deviation of any workpiece in all processes, add the products of the mutual information and standard deviation of all historical workpieces, divide by the total number of historical workpieces and normalize to obtain the influence coefficient between adjacent processes.

4. The method for dynamically adjusting production tasks based on a large industrial vertical model according to claim 1, characterized in that, The calculation of the influence coefficients between various processes in historical production based on historical parameter sequences includes: Define the standard values ​​of the process parameters of a workpiece in any process, and construct the standard parameter sequence of the workpiece in each process; take the absolute value of the difference between each parameter in the historical parameter sequence of any workpiece in each process and the corresponding standard parameter to form a deviation difference sequence; Calculate the Pearson correlation coefficient between the deviations of all workpieces in the preceding and following processes, calculate the standard deviation of any workpiece in all processes, then add the products of the Pearson correlation coefficients and standard deviations of all historical workpieces, divide by the total number of historical workpieces, and normalize to obtain the influence coefficient between adjacent processes.

5. The method for dynamically adjusting production tasks based on a large industrial vertical model according to claim 1, characterized in that, The method for calculating the influence coefficient between processes in real-time production is the same as the method for calculating the influence coefficient between processes in historical production.

6. A dynamic adjustment system for production tasks based on a large-scale industrial vertical model, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the method for dynamically adjusting production tasks based on a large industrial vertical model according to any one of claims 1-5.

Images