Mold whole-process quality monitoring system and device and storage medium

Abstract

The invention relates to the technical field of comprehensive factory control, in particular to a mold whole-process quality monitoring system and device and a storage medium, and the system comprises a data collection and synchronization module which is connected with a plurality of mold manufacturing machine tool connection networks, collects machining data and operation states of mold manufacturing machine tools, generates a real-time data flow, and carries out the real-time processing of the mold manufacturing machine tools based on the real-time data flow. And analyzing the working efficiency and state of the mold manufacturing machine tool to obtain an equipment state result. According to the method, the machining data and the operation state of the mold manufacturing machine tool are collected in real time, the real-time data flow is formed, the machine tool operation efficiency and the equipment state are dynamically analyzed, state monitoring and production quality prediction in the mold manufacturing process are achieved, and the transparency and the management and control capacity of the mold machining link are improved; and according to the quality prediction result obtained through analysis, the production line performance and the equipment efficiency are dynamically evaluated, the production tasks are scheduled in real time, the equipment load distribution is optimized, a dynamic production scheduling system is formed, and the production efficiency and the resource utilization rate are improved.

Description

Technical Field

[0001] The present invention relates to the field of comprehensive factory control technology, and in particular to a mold full-process quality monitoring system, device and storage medium. Background Art

[0002] The mold full-process quality monitoring system is a system that performs real-time quality monitoring and control on the entire mold manufacturing process. It aims to implement continuous monitoring and analysis by collecting data from the entire process of mold design, production, assembly to use.

[0003] In actual operation, the existing technology only relies on data collection and basic monitoring and analysis. The data monitoring and analysis methods for the entire mold manufacturing process are relatively simple, lacking dynamic analysis and real-time optimization of equipment operating status and production efficiency, and it is difficult to quickly respond to problems such as unbalanced equipment load and unreasonable task scheduling in actual production. For example, in the mold manufacturing process, when a certain equipment has reduced operating efficiency or abnormally increased load, it is difficult to make timely adjustments, which may cause problems such as partial blockage of the production line, unstable mold processing quality, and excessive wear of equipment. Therefore, improvements are needed. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a mold full-process quality monitoring system.

[0005] In order to achieve the above purpose, the present invention adopts the following technical solution: The mold full process quality monitoring system includes: A data acquisition and synchronization module, wherein a plurality of mold manufacturing machine tools are connected to a network, collects processing data and operation status of the mold manufacturing machine tools, generates a real-time data stream, analyzes the working efficiency and status of the mold manufacturing machine tools based on the real-time data stream, and obtains equipment status results; A quality analysis and prediction module performs real-time quality monitoring on the production process of the mold manufacturing machine tool based on the equipment status results to obtain quality prediction results; The production scheduling and optimization module evaluates the quality performance and production efficiency of each production line according to the quality prediction results, performs dynamic task scheduling, generates a scheduling optimization plan, and adjusts the production tasks and mold manufacturing machine tool load based on the scheduling optimization plan to obtain an optimized production process; The mold life management and maintenance module monitors the usage status and wear data of each mold based on the optimized production process, performs mold life analysis, and generates mold usage analysis results.

[0006] Preferably, the steps of acquiring the real-time data stream are: Several mold manufacturing machine tools are connected to the database through a network interface. The mold manufacturing machine tools store data in the database in real time, including the current processing speed, operating pressure and temperature of the machine tools, to obtain a preliminary real-time data stream; Based on the preliminary real-time data stream, data verification and error checking are performed, and then the data is formatted to obtain a real-time data stream.

[0007] Preferably, the steps of obtaining the device status result are: Based on the real-time data stream, the working efficiency of the mold manufacturing machine tool is calculated using the following formula: ; in, To improve the working efficiency of mold manufacturing machine tools, For the performance values ​​of operating parameters, is the average of all performance values, is the minimum value of the machine tool operation rate, is the total number of parameters; Based on the working efficiency of the mold manufacturing machine tool and combined with the temperature and pressure data of the mold manufacturing machine tool, the equipment status result is obtained.

[0008] Preferably, the steps of obtaining the quality prediction result are: Based on the equipment status results, the quality prediction index in the production process is calculated, and the calculation formula is: ; in, is a quality prediction indicator. To improve the working efficiency of mold manufacturing machine tools, is the device temperature, is the equipment pressure, is the maximum vibration amplitude, is the energy consumption per unit time; Based on the quality prediction index, the operating status of the mold manufacturing machine tool production line is monitored in real time, and the data fluctuation and trend of the quality prediction index are analyzed to obtain the quality prediction result.

[0009] Preferably, the steps for obtaining the scheduling optimization plan are: Based on the quality prediction results, the dynamic task load adaptation value is calculated using the following formula: ; in, is the dynamic task load adaptation value, is a quality prediction indicator. is the average downtime of a single device in the production line. is the standard processing time of the current task, is the number of task switching times per hour on the current production line, is the energy consumption level per unit task of the current production line; According to the dynamic task load adaptation value, a scheduling optimization solution is obtained by matching the production line with the optimal dynamic task load adaptation value.

[0010] Preferably, the steps for obtaining the optimized production process are: Based on the scheduling optimization scheme, the production efficiency maximization evaluation value is calculated, and the calculation formula is: ; in, To maximize the evaluation value of production efficiency, is the dynamic task load adaptation value, The mold change time for a single task, The number of tasks processed per minute for mold manufacturing machines, is the number of operation steps for the current task, is the length of the current machine tool's real-time idle period, The starting sequence position of the task within the current production cycle; Based on the production efficiency maximization evaluation value, the mold manufacturing machine tool load balancing is performed by reconstructing the relationship between the task sequence and the execution order of the mold manufacturing machine tool to obtain an optimized production process.

[0011] Preferably, the steps for obtaining the mold usage analysis result are: Filter out the task information assigned to each mold manufacturing machine tool, read the number and usage cycle of the bound mold, collect the current operating frequency and accumulated processing time of each mold, and generate mold usage status information; Based on the mold usage status information, the vibration waveform and thermal imaging map of the mold end face are read, the wear area is identified in the grayscale change area of ​​the image, and the mold usage analysis result is generated.

[0012] The present invention provides a mold full-process quality monitoring device, comprising: a processor and a memory, the memory is used to store a computer program, the processor is used to call and run the computer program stored in the memory, so that the mold full-process quality monitoring device executes the mold full-process quality monitoring system.

[0013] The present invention provides a computer-readable storage medium, in which a computer program or instruction is stored. When the computer program or instruction is executed, a mold full-process quality monitoring system is implemented.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by real-time collection of mold manufacturing machine processing data and operating status, a real-time data stream is formed, the machine tool operating efficiency and equipment status are dynamically analyzed, the status monitoring of the mold manufacturing process and the production quality prediction are realized, and the transparency and control capabilities of the mold processing link are improved; and based on the quality prediction results obtained by analysis, the production line performance and equipment efficiency are dynamically evaluated, production tasks are scheduled in real time and equipment load distribution is optimized to form a dynamic production scheduling system to improve production efficiency and resource utilization; at the same time, continuous monitoring is implemented for the actual use and wear data of the mold, and through the analysis of the mold life, the mold service life and reliability are extended, further reducing production costs, and improving the overall manufacturing quality level and market competitiveness of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] See also Figure 1 The present invention provides a technical solution: the mold full process quality monitoring system includes: Data acquisition and synchronization module: several mold manufacturing machine tools are connected to the network, the processing data and operation status of the mold manufacturing machine tools are collected, and real-time data streams are generated. Based on the real-time data streams, the working efficiency and status of the mold manufacturing machine tools are analyzed to obtain the equipment status results; The quality analysis and prediction module conducts real-time quality monitoring of the production process of mold manufacturing machine tools based on the equipment status results and obtains quality prediction results; The production scheduling and optimization module evaluates the quality performance and production efficiency of each production line based on the quality prediction results, performs dynamic task scheduling, generates scheduling optimization plans, adjusts production tasks and mold manufacturing machine tool loads based on the scheduling optimization plans, and obtains the optimized production process; The mold life management and maintenance module monitors the usage status and wear data of each mold based on the optimized production process, conducts mold life analysis, and generates mold usage analysis results.

[0018] The steps to obtain real-time data stream are: Several mold manufacturing machine tools are connected to the database through a network interface. The mold manufacturing machine tools store data in the database in real time, including the current processing speed, operating pressure and temperature of the machine tools, to obtain a preliminary real-time data stream; Based on the preliminary real-time data stream, data verification and error checking are performed, and then the data is formatted to obtain the real-time data stream.

[0019] Specifically, multiple mold manufacturing machine tools are connected to the network and establish communication with the database respectively, and the processing speed, operating pressure, temperature and other information of the machine tools are collected in real time and range comparison is performed in sequence. For example, the processing speed is compared with the range of 0m / min to 400m / min, the operating pressure is compared with the range of 0MPa to 2MPa, and the temperature is compared with the range of 0℃ to 200℃. If any value is higher or lower than the range, it is recorded as suspicious data and the corresponding mark is retained in the database. Then, the average value and standard deviation of speed, pressure and temperature under different materials and different tool usage scenarios are calculated by referring to experience or through statistics, and the speed threshold is calculated based on this. , Pressure threshold , Temperature threshold ,in , , , , , are the average values ​​of the three parameters, are their standard deviations, It is a coefficient selected after on-site verification, for example, set it to 1.5 and correct it based on multiple measured results. When the data actually collected by the machine tool deviates significantly, these thresholds can be used to quickly detect anomalies and add marks to the database. After that, all speed, pressure and temperature data that meet the normal range and have not triggered the threshold are summarized and collected to obtain a preliminary real-time data stream.

[0020] Based on the preliminary real-time data stream obtained above, first compare all values ​​with the aforementioned speed threshold , Pressure threshold , Temperature threshold The data is compared to screen out suspicious records with significant deviations and retain the corresponding indexes. Then, an abnormality detection method based on a neural network is used to check the remaining data for errors. The neural network is set to a three-layer structure. The input layer receives the time series data of processing speed, pressure, and temperature and the corresponding machine tool number information. The hidden layer is composed of several neurons and uses the activation function ReLU. The output layer returns a scalar used to determine whether it is abnormal. The training process selects normal machine tool operation data within the past three months as positive samples, and data that greatly exceeds the threshold or has multiple fluctuation abnormalities as negative samples. All samples are first normalized and randomly split into training sets and validation sets. The back propagation algorithm is used to gradually update the weights of each connection. The model training is completed after the loss function converges to a certain range on the validation set. During reasoning, new real-time data is input into the neural network, and the output judgment value is calculated and compared with the preset judgment benchmark. For example, a value higher than 0.7 is considered an abnormal judgment. If the detection result is abnormal, the record is summarized again and marked together with the aforementioned suspicious record. Finally, the normal data is formatted, the data type and decimal precision are unified, and rearranged in chronological order to obtain a real-time data stream.

[0021] The steps to obtain the device status results are: Based on real-time data flow, the working efficiency of mold manufacturing machine tools is calculated using the following formula: ; in, To improve the working efficiency of mold manufacturing machine tools, For the performance values ​​of operating parameters, is the average of all performance values, is the minimum value of the machine tool operation rate, is the total number of parameters; Based on the working efficiency of the mold manufacturing machine tools, combined with the temperature and pressure data of the mold manufacturing machine tools, the equipment status results are obtained.

[0022] Specifically, the formula is beneficial in that by introducing and The combination of the deviation degree of multiple operating parameters and the minimum value of the machine tool operation rate is comprehensively reflected, so as to obtain the minimum value of the single efficiency value. Incorporate multi-angle information at the same time, including As a normalization factor, the overall result is controlled within a comparable range. In practice, this structure can identify the performance differences of different machine tools and give the degree of difference, so that the subsequent steps can perform further production scheduling or operation analysis based on the results; The steps to obtain the parameters are: Indicates The performance value of an operating parameter includes quantifiable data such as machine vibration amplitude, tool consumption, feed rate, etc. Converted into tool consumption The process can be referred to ,in is the maximum reference value corresponding to the full wear area, for example, 10 square millimeters is selected. When the number of records reaches a certain scale, a The array is used for subsequent calculations; The steps to obtain the parameters are: It represents the average of all performance values. This parameter is used to measure the central trend of the overall operating parameters. When calculating, all Add them one by one and then divide by When collecting data, the values ​​of all operating parameters can be summarized according to a complete production cycle or a fixed period of time, and the data collected in a fixed time window can be summed up to obtain Then divide by , such as in the above example ,make ,therefore ; The steps for obtaining the parameter are as follows: This parameter represents the machine tool operation rate and belongs to a sequence that changes with time. In the implementation, the speed value at each moment is extracted through the speed detector built into the machine tool, and then all the speed values ​​are numbered and obtained at each recording point. Form a velocity array, where It is the minimum rate value during the entire acquisition period. This minimum rate is regarded as an important benchmark for status assessment. The acquisition process can be referred to as follows: In real-time monitoring, the speed is stored in a 2-second cycle, and all speed values ​​are read out in sequence within a complete cycle of 6000 records. The smallest value in the sequence is searched as , for example, ,but ; The steps of obtaining the parameters are as follows: This parameter is the total number of operating parameters, which can be determined according to the actual number of types monitored by the machine tool during implementation. When the operating parameters monitored by the machine tool are vibration amplitude, tool consumption, feed rate, load current and spindle speed, a total of 5 types, record .

[0023] Calculation process: First calculate , the previously collected as well as , square the difference and substitute it into the cosine function to get: ; in , , , , , The other two items are similarly and , add all the results together to get 4.9983, then take the cube root ,and Add, where , , so the sum is 1.709+22.1979=23.9069, and the natural logarithm of this value is taken , the denominator is = , and finally get ; The result shows that the current working efficiency of the machine tool is about 1.973. The higher the numerical range, the more stable and better the overall state of the machine tool is measured by the monitored operating parameters and the minimum rate. When more time period data is generated in the future, the same formula can be used for recalculation and compared with the previous results to reflect the fluctuation of working efficiency and the differences in different time periods. The subsequent process can combine this efficiency value to carry out the correlation analysis of temperature and pressure and obtain the corresponding equipment status results.

[0024] Based on the working efficiency of the mold manufacturing machine tool and combining it with the temperature and pressure data of the machine tool, it is necessary to record the efficiency value and the temperature and pressure values ​​of the corresponding machine tool in the same time window. For example, compare the temperature with 0℃ to 200℃. If the temperature value is outside this range, record a suspicious mark. Similarly, compare the pressure with 0MPa to 2MPa. If it exceeds this range, also record a suspicious mark. Then, group all the extracted values ​​according to the machine tool identification. After grouping, ensure that the working efficiency, temperature and pressure of the same machine tool can be synchronized in time. For example, the efficiency, temperature and pressure collected within one minute are regarded as the same record. Then, verify these grouped records one by one and make multi-dimensional comparisons according to the temperature range, pressure range and working efficiency value range, such as comparing the efficiency values ​​1.0 to 3.0. If the efficiency value is lower than 1.0, it is recorded as a low record, and if it is higher than 3.0, it is recorded as a high record. When the record is completed, further step judgment can be made based on these multi-dimensional evaluation tags. For example, the conditions where the temperature is higher than 200°C and the efficiency value is lower than 1.0 in the same time window are summarized, and a centralized query entry is added in the system. These records can be used as a basis for judgment when the machine tool is maintained in the future. If a machine tool has a temperature exceeding 200°C for many times in different collection periods, the working condition of its cooling system will be checked in the maintenance link. If the same machine tool has been maintained in the efficiency range of 1.5 to 2.5, it means that its operating status is temporarily stable. Therefore, these relevant records will still be retained for tracing the process. When all steps are completed, the equipment status results including temperature, pressure and working efficiency can be obtained.

[0025] The steps to obtain the quality prediction results are: Based on the equipment status results, the quality prediction index in the production process is calculated using the following formula: ; in, is a quality prediction indicator. To improve the working efficiency of mold manufacturing machine tools, is the device temperature, is the equipment pressure, is the maximum vibration amplitude, is the energy consumption per unit time; Based on the quality prediction indicators, the operating status of the mold manufacturing machine tool production line is monitored in real time, the data fluctuations and trends of the quality prediction indicators are analyzed, and the quality prediction results are obtained.

[0026] Specifically, the formula: The benefit of the formula is that it combines the exponential function, logarithmic function and hyperbolic function to express the working efficiency of mold manufacturing machine tools. , Maximum vibration amplitude , Equipment temperature , Equipment Pressure And energy consumption per unit time Multiple factors are unified into the same indicator calculation, which can not only highlight the impact of work efficiency and vibration amplitude on the overall quality, but also take into account the periodic changes of temperature and pressure, and combine energy consumption to conduct a more comprehensive evaluation. This design idea structurally couples multi-dimensional factors and attributes them to quantifiable quality prediction indicators. , through the exponential and logarithmic coupling of the numerator, the sensitivity to changes in the operating environment is improved, and the hyperbolic function in the denominator is used to constrain the overall numerical scale, and finally a clear The value provides a basis for subsequent analysis and monitoring.

[0027] The steps for obtaining the parameter are as follows: the parameter represents the working efficiency of the mold manufacturing machine tool, and the working efficiency of the mold manufacturing machine tool is calculated by the previous formula.

[0028] The steps of obtaining the parameter are as follows: This parameter represents the device temperature. In actual acquisition, the temperature sensor records the measurement value once every 10 seconds to form a temperature time series, and then selects the relevant The temperature data corresponding to the same time is The value of must be converted into radians before it can be used in trigonometric calculations. The conversion formula is: ,in Refers to the actual measured temperature in degrees Celsius, Take 3.1415926, if the temperature sensor detects , then substitute it into the above formula to get .

[0029] The steps to obtain the parameter are as follows: This parameter represents the equipment pressure. In common die-casting or injection molding processes, a pressure sensor is used to detect the pressure in the machine cavity and generate a numerical sequence. The data is recorded in a cycle of 5 seconds or less, and the recorded results are marked in the database item corresponding to the machine tool number. Finally, the data is compared with the temperature. When comparing synchronously, select the pressure value under the same period In this formula, if the pressure exceeds 2MPa, it is registered as a high event and counted. If it does not exceed this range, the original recorded value is directly used in subsequent calculations. When a value of 1.60MPa is obtained, the conversion ratio of pressure to angle is 1MPa:1rad. .

[0030] The steps for obtaining the parameter are as follows: This parameter represents the maximum vibration amplitude, which refers to the peak value quantified result of the vibration of the machine tool during operation. When obtaining it, the real-time vibration waveform can be collected through an accelerometer or a vibration meter, and the amplitude is detected at a frequency of once per second. The maximum value within a certain time window is recorded as For example, if the maximum vibration amplitude detected is 0.2 mm, then in this formula .

[0031] The steps to obtain the parameter are as follows: This parameter represents the energy consumption per unit time. Usually, the energy consumption of the machine tool per unit hour is captured by the electric energy metering device, and a scalar is measured against the actual working conditions. For example, when the machine tool is working continuously, the power consumption per hour may reach 2.1kWh, then Set to 2.1.

[0032] Calculation process: make ; First calculate the numerator: ; ; Adding the two together gives

[0033] ; Then calculate : ; Add the two together and you get ; ; ; Adding these two terms gives ; The molecule as a whole is ; Then calculate the denominator : ; ; ; The denominator totals ; Divide the numerator by the denominator: The result shows that the quality prediction index at this time is about 14.88. A higher value indicates that under the comprehensive influence of the current equipment temperature, equipment pressure, maximum vibration amplitude and energy consumption per unit time, combined with work efficiency The quality performance obtained after the calculation is maintained in the upper range. When the measured data is continuously monitored every hour and the formula is recalculated, if the obtained The value is continuously high, indicating that the production process is stable as a whole and there are no significant abnormalities in the performance parameters. If the value drops significantly to below 1.0, it means that fluctuations that are unfavorable to quality have occurred in the equipment temperature, pressure, vibration, energy consumption and other aspects. At the corresponding stage, it may be necessary to further check the specific links to locate potential problems.

[0034] Based on the quality prediction indicators and the real-time operation status of the production line, first retrieve the quality prediction indicators that have been continuously obtained before The machine tool number label and timestamp information in the sequence and corresponding period are compared one by one with the pre-established indicator validity range, for example Compare the interval from 0 to 20. If a record is detected If the value is higher than 20, it is recorded as an excessive item and assigned a flag. If the value of is lower than 1, it is marked as a too low project. The thresholds of these ranges can be calculated based on the historical data of the past three months. The specific method is to first calculate all the Sort and select the 1% and 99% percentiles as the upper and lower limits, so that in subsequent inspections, data above or below the corresponding percentiles can be marked and the index can be retained. When all data are marked, they are grouped and summarized according to the machine tool number, and each machine tool is compared one by one to see if it has appeared multiple times in the same time period. In the case of exceeding the normal range, the time continuity check is also introduced during the comparison. For example, when a machine tool is If the temperature of a machine tool is between 0℃ and 200℃ and the pressure is between 0MPa and 2MPa in a large range of time, and the temperature sampling value and pressure sampling value of the machine tool are associated based on these summary information. However, if it is always lower than 2, it means that it is necessary to focus on whether the vibration amplitude and energy consumption are also in an acceptable range. This process is repeated on all machine tools, and the daily or weekly quality prediction indicator distribution is viewed in combination with timestamps. Finally, the query results aggregated for different machine tool numbers are classified, recorded and integrated to obtain the quality prediction results.

[0035] The steps to obtain the scheduling optimization plan are: Based on the quality prediction results, the dynamic task load adaptation value is calculated using the following formula: ; in, is the dynamic task load adaptation value, is a quality prediction indicator. is the average downtime of a single device in the production line. is the standard processing time of the current task, is the number of task switching times per hour on the current production line, is the energy consumption level per unit task of the current production line; According to the dynamic task load adaptation value, the scheduling optimization solution is obtained by matching the production line with the optimal dynamic task load adaptation value.

[0036] Specifically, the formula: The formula is useful in combining quality predictors The downtime is calculated by trigonometric, inverse trigonometric and radical operations with multiple production line characteristic parameters. , Standard processing time , Number of task switches per hour And unit task energy consumption level The integration of unified metrics can measure the compatibility between task loads and production line elements at a numerical level and output a quantifiable , so as to make more targeted arrangements for the scheduling of the production line in the subsequent steps.

[0037] The steps for obtaining the parameter are: This parameter represents a quality prediction index and can be calculated using the aforementioned formula.

[0038] The steps to obtain the parameter are as follows: This parameter represents the average fault downtime of a single device in the production line. In actual collection, it is necessary to record the downtime of the equipment during the entire observation period, accumulate the duration of each downtime and divide it by the number of downtimes to get the average downtime. For example, within a month, the equipment has 10 downtimes, lasting for 1.2 hours, 1.6 hours, 1.1 hours, etc., respectively. The sum of these durations and divided by 10 can get an average value. If 1.35 hours is obtained, then record it. .

[0039] The steps to obtain the parameter are as follows: This parameter represents the standard processing time of the current task, which refers to the time required to complete a standard task from start to finish. When obtaining it, first define the process flow and processing steps of each type of task in the production process, and quantify the total time in several actual productions, and then take the average of multiple observations as the standard time. For example, when processing a certain specific component, 30 complete production cycles are collected, and the time consumed each time is measured separately. Then, they are added up and divided by 30 to get an average value. If it is obtained Hour.

[0040] The steps to obtain the parameter are as follows: This parameter represents the number of task switches per hour on the production line, which refers to the number of times one task switches to another task within a 1-hour time frame on the same production line. During the collection, the timestamp of each task switch is first recorded in the production management system, and then the total number of times the timestamp appears in each hour window is counted, and finally the specified segment is averaged. If a total of 24 switches occur within 8 hours, the average number of switches per hour is 3, and the number of times the timestamp appears in each hour window is counted. .

[0041] The steps to obtain the parameter are as follows: This parameter represents the energy consumption level per unit task of the current production line. The energy consumption of the production line when executing a single task is counted. The electric energy or other energy consumption generated by completing a unit task is digitized. Based on the total energy consumption recorded and the total number of tasks produced, the average energy consumption value corresponding to each task is calculated. For example, when processing 100 identical parts, 210 kWh of electricity is consumed, then Determined as .

[0042] Calculation process: shilling ; Molecular part: ; have to .

[0043] ; Adding the above two together, we get ; ; Calculate first , and then calculate ,therefore , ; The total numerator = 1.566 × 1.418 = 2.221; Denominator: ; Add 44.64 and 1.8225 to get 46.4625, and then take the square root, which is about 6.813; Finally, divide the numerator by the denominator: ; The result shows that the dynamic task load adaptation value of this production line under given conditions is 0.326. When this value is low, it means that the current fault downtime and task switching times make the load difficulty of the production line high. If the monitoring finds that a production line The value continues to increase, which can be considered to be easier to match a variety of task orders. When the value is below 0.3, the specific process conditions can be combined to evaluate whether to change the allocation strategy. At the application level, the allocation strategies of different production lines at the same time are usually The values ​​are compared horizontally to select the production line with the highest value or within a reasonable range to enter the task scheduling process. At the same time, the production line with abnormally low values ​​can be marked to limit its high-intensity load distribution in subsequent scheduling. The sequence can be directly used to screen the production line that matches the optimal dynamic task load adaptation value and perform subsequent scheduling.

[0044] According to the dynamic task load adaptation value and matching the production line with the optimal dynamic task load adaptation value, firstly obtain the The sequences are searched centrally, read one by one and Compare with the pre-set effective range, for example, set the effective range to between 0.3 and 1.5. If it exceeds this range, it will be marked as an option that is not suitable for the current production needs. After the marking is completed, the production lines within the range will be marked as The values ​​are sorted, and the ones with higher values ​​are prioritized and indexed. Then, the average energy consumption data corresponding to each production line and the number of task switches per hour are read. If the energy consumption of a production line is found to be too high, it will be placed in the suboptimal list. For example, a suboptimal label is set for a production line with an energy consumption higher than 5.0kWh. Then, the average fault downtime of the equipment in each production line and the actual available processing time period are retrieved to see if they meet the start and end times of subsequent production tasks. If the average fault downtime of the production line is greater than 2 hours, it is assigned to the end list. If it is less than 2 hours, it is retained in the candidate list. The number of task switches of the retained production line is queried again, and the number of suboptimal b Production lines with values ​​greater than 7 are marked with frequent switching and additional instructions are given when the results are returned. After confirming that the parameters of all production lines have been compared, one or two production lines with the highest ranking are selected, and they are inspected on site and the production efficiency data and quality prediction indicators obtained previously are read for a second comparison. When the second comparison still remains in a relatively optimal range and the temperature records in the range of 0℃ to 200℃ and the pressure records in the range of 0MPa to 2MPa are continuously stable, these production lines are included in the short-term scheduling objects. Finally, the selected production line numbers and the schedulable time periods are merged to generate a dynamic task load allocation list and integrated into the scheduling optimization plan.

[0045] The steps to obtain the optimized production process are: Based on the scheduling optimization plan, the maximum evaluation value of production efficiency is calculated. The calculation formula is: ; in, To maximize the evaluation value of production efficiency, is the dynamic task load adaptation value, The mold change time for a single task, The number of tasks processed per minute for mold manufacturing machines, is the number of operation steps for the current task, is the length of the current machine tool's real-time idle period, The starting sequence position of the task within the current production cycle; Based on the maximum evaluation value of production efficiency, the load balancing of mold manufacturing machine tools is carried out by reconstructing the relationship between task sequences and the execution order of mold manufacturing machine tools, and the optimized production process is obtained.

[0046] Specifically, the formula: The formula is useful in adapting the dynamic task load to the value , mold change time for a single task , machine tool processing speed 、Number of current task operation steps , Real-time idle period length of the machine tool And the starting order of the tasks Multiple factors are coupled, and through the combination of various operations such as integral operations, fourth roots and trigonometric functions, an evaluation value can be obtained. It reflects the production process's requirements for mold replacement, machine tool processing efficiency, idle resources, etc. The numerator takes into account the processing time and machine tool efficiency-related quantities in the integral and radical terms, and the denominator records the machine tool status and beat position in the cosine and fractional terms. In this way, the dynamic elements and task sequence factors in production are woven together into a compact formula structure, allowing subsequent links to measure the production efficiency potential through a clear numerical indicator.

[0047] The steps for obtaining the parameter are: This parameter represents the dynamic task load adaptation value, which is calculated in the previous formula.

[0048] The steps to obtain the parameter are as follows: This parameter represents the mold replacement time of a single task. It is obtained by actually recording the start and end time of mold disassembly or switching during the production process. For example, the same type of mold is replaced 10 times, and the replacement periods are 0.35 hours, 0.42 hours, 0.38 hours, etc. These data are counted and averaged to determine a stable value.

[0049] The steps to obtain the parameter are as follows: This parameter represents the number of tasks processed per minute by the mold manufacturing machine tool. By detecting the actual processing output, the number of unit tasks processed by the machine tool in a fixed period of time is counted, and then divided by the total time, the result can be obtained. For example, if 480 standard parts are processed in one hour, the processing volume per minute is 8, so in the formula .

[0050] The steps to obtain the parameter are as follows: This parameter represents the number of operation steps of the current task, which is used to quantify the subdivided steps that need to be taken when the task is executed on the machine tool. In the specific production process, a step-by-step operation list can be listed and numbered through process documents or work instructions, and the number of workstations or process nodes involved in each step can be combined and counted to form an integer or decimal. For example, when processing a certain special-shaped part, 20 process steps need to be completed first, then , convert every 5 operation steps into 1 radian, so that Connect to trigonometric operations. If the final conversion result is around 10 radians, then when entering the formula .

[0051] The steps to obtain the parameter are as follows: This parameter represents the length of the current real-time idle period of the machine tool, which is determined by collecting the total time that the machine tool is actually not running in the daily schedule. For example, in an 8-hour shift, the machine tool is turned on for 6.6 hours and is actually processing, and the remaining 1.4 hours are counted as idle periods, so .

[0052] The steps to obtain the parameter are as follows: This parameter indicates the starting order position of the task in the current production cycle. The tasks need to be arranged in the execution order according to the production schedule. For example, if 5 tasks are planned in a cycle, the first task corresponds to , the second task corresponds to , and so on. If the current task is the third priority, it can be recorded .

[0053] Calculation process: set up ; First calculate the integral ,make , then the integrand is , the integral can be solved analytically or numerically. When the analytical formula is used, , here, let , , perform a definite integral from 0 to 0.326, and substitute at x=0.326 to get , x=0 is substituted as 0, so the integral result is approximately ; calculate ,make ,but ,make As input, , so .; Numerator = integral value ; Denominator = ,make ,but: , the absolute value is still 0.638, let ,but , the sum of the two is .

[0054] Finally get The results show that the current maximum production efficiency evaluation value is about 4.625. When this value is greater than 4, it means that within the calculated period, combined with the dynamic task load adaptation value, mold change time, machine tool processing rate, task operation steps, machine tool idle time and task sequence position, the production process shows good execution potential. If the subsequent comparison of different shifts or different time periods in the same factory is carried out It remains at a high level, indicating that the existing scheduling and load balancing are reasonable. If it is lower than 1 or significantly drops below 1, it means that factors that are unfavorable to efficiency have accumulated in the process. This result can be reviewed with the various parameters obtained previously to further reconstruct the task sequence and machine tool execution order, and make appropriate production scheduling and load balancing decisions.

[0055] Based on the production efficiency maximization evaluation value and by reconstructing the execution order relationship between the task sequence and the mold manufacturing machine tool, all the calculated production efficiency maximization evaluation values ​​are first The data is sorted into a data comparison table with the machine tool number and timestamp. The upper and lower limits of the Z value in each time period are retrieved in the comparison table, and compared with the pre-established intervals such as 1 to 5. When the Z value of a machine tool in a certain time period exceeds 5, it is marked and the dynamic task load adaptation value M of the machine tool and the number of operation steps h of the corresponding task are recorded in turn. At the same time, a parallel query is performed on the mold change time r. If r is in the range of 0.2 hours to 1 hour, it is archived together with the original record. If r is higher than 1 hour, a special mark is added. Then check whether there are continuous vacancies of more than 8 hours in the idle period of this machine tool. If the continuous vacancies are less than 8 hours, continue to record. The summarized data are grouped again according to the machine tool number and the tasks are sorted in ascending order according to the s value in the same group. After the sorting is completed, the corresponding starting sequence position and the available machine tool are read one by one. Use time periods and perform time comparisons. If a task is scheduled to a time period with a usage conflict within the scheduled beat, add an identifier to the conflicting row in the record table and remove it before scheduling. Then detect the f values ​​of all tasks. For example, if a row of records shows that the number of tasks processed per minute is f=8 and the corresponding s=0, then this row is used as the priority allocation target, and it is inserted into a new sequence and attention is paid to connecting with possible idle segments g of other tasks. If the g value is too different from the M value, for example, g exceeds 10 and M is only 0.3, consider enabling additional sequence adjustments in subsequent links to balance the machine tool. By repeatedly traversing each indexed task, all records that meet the time and sequence are sequentially assembled into a new execution list. Finally, this execution list is regarded as a new production beat queue and corresponds to the machine tool number one by one, so that the optimized production process can be obtained.

[0056] The steps to obtain mold usage analysis results are as follows: Filter out the task information assigned to each mold manufacturing machine tool, read the number and usage cycle of the bound mold, collect the current operating frequency and accumulated processing time of each mold, and generate mold usage status information; Based on the mold usage status information, the vibration waveform and thermal imaging map of the mold end face are read, the wear area is identified in the grayscale change area of ​​the image, and the mold usage analysis results are generated.

[0057] Specifically, the task information assigned to each mold manufacturing machine tool is filtered out. First, the specific mold number currently associated with each machine tool and the corresponding usage cycle are extracted by comparing the existing production scheduling records with the mold number list. The usage cycle can be counted through the previously accumulated data. For example, the number of mold changes and replacement intervals are recorded within 90 days. The replacement time of each mold can be compared with each other to determine its usage cycle, and these mold numbers and cycle information are summarized. Then, the current operating frequency and accumulated processing time of each mold are collected. The operating frequency can be measured by reading once per minute when the machine tool is running, and the measured frequency value is compared with the pre-established effective range, such as 0Hz to 20 The 00Hz interval is used as the reasonable vibration frequency range for this type of machine tool. Records that exceed this range are marked as data to be checked. The cumulative processing time can be calculated through the machine tool's work records. For example, the time difference between the machine tool startup time and the shutdown time is accumulated and matched with the mold to obtain the total working time of the mold on the machine tool. Then, for each mold, the operating frequency data and the cumulative processing time are matched one by one to form a complete status entry. For example, if the use cycle of a mold is 14 days and the cumulative processing time has reached 200 hours, and the frequency value remains stable within the default range, it will be written into the status summary table for centralized viewing. After the above data of all molds are collected, they are retrieved and merged to finally obtain the mold usage status information.

[0058] Based on the mold usage status information and reading the mold end face vibration waveform and thermal imaging map, first install a vibration sensor device in the machine tool operation room and read the vibration value every 2 seconds, match these original values ​​with the current operating frequency data in the mold usage status information, and retrieve the vibration waveform curve at the same time for each corresponding record. If the waveform peak value of a certain period of time exceeds the vibration amplitude threshold originally set by historical statistics, for example, the threshold is determined to be 1.2g through an average analysis of the normal processing status in the past 30 days, then mark the suspected wear clues after the data, and then read the thermal imaging map according to the shooting interval and the mold number. Compare the temperature distribution information with the vibration peak time. If there is a high temperature area exceeding 500°C in the map and there is a sudden change in the gray value distribution, it will be classified as a possible wear concentration area. In order to more accurately identify the wear area in the image, it can be A convolutional neural network is used for classification. The neural network sets the image input layer as a three-channel temperature distribution map, where each channel corresponds to a pixel value under a specific range. The hidden layer contains several convolution kernels to extract image features related to the thermal image and the end face vibration waveform. The output layer returns the coordinate set and area estimation of the wear area. Through the training process, 500 sets of grayscale and temperature data of normal mold end faces and 300 sets of worn mold end faces are collected, the wear boundaries are marked and the errors are calculated using cross entropy loss, and then back propagation is repeatedly iterated until the accuracy of the verification set is stable to obtain fixed weight parameters. During inference, the aforementioned convolution operation is performed on the end face image taken in real time and the coordinates of the wear area are output according to the activation map. These coordinates are then cross-compared with the high temperature distribution area in the same image for multiple times. Finally, the identified wear information is summarized and associated with the mold usage status information to generate the mold usage analysis results.

Claims

1. The whole process quality monitoring system of mold is characterized by: The system comprises: A data acquisition and synchronization module, wherein a plurality of mold manufacturing machine tools are connected to a network, collects processing data and operation status of the mold manufacturing machine tools, generates a real-time data stream, analyzes the working efficiency and status of the mold manufacturing machine tools based on the real-time data stream, and obtains equipment status results; A quality analysis and prediction module performs real-time quality monitoring on the production process of the mold manufacturing machine tool based on the equipment status results to obtain quality prediction results; The production scheduling and optimization module evaluates the quality performance and production efficiency of each production line according to the quality prediction results, performs dynamic task scheduling, generates a scheduling optimization plan, and adjusts the production tasks and mold manufacturing machine tool load based on the scheduling optimization plan to obtain an optimized production process; The mold life management and maintenance module monitors the usage status and wear data of each mold based on the optimized production process, performs mold life analysis, and generates mold usage analysis results.

2. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps of acquiring the real-time data stream are: Several mold manufacturing machine tools are connected to the database through a network interface. The mold manufacturing machine tools store data in the database in real time, including the current processing speed, operating pressure and temperature of the machine tools, to obtain a preliminary real-time data stream; Based on the preliminary real-time data stream, data verification and error checking are performed, and then the data is formatted to obtain a real-time data stream.

3. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps for obtaining the device status result are: Based on the real-time data stream, the working efficiency of the mold manufacturing machine tool is calculated using the following formula: ； in, To improve the working efficiency of mold manufacturing machine tools, For the performance values ​​of operating parameters, is the average of all performance values, is the minimum value of the machine tool operation rate, is the total number of parameters; Based on the working efficiency of the mold manufacturing machine tool and combined with the temperature and pressure data of the mold manufacturing machine tool, the equipment status result is obtained.

4. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps for obtaining the quality prediction result are: Based on the equipment status results, the quality prediction index in the production process is calculated, and the calculation formula is: ； in, is a quality prediction indicator. To improve the working efficiency of mold manufacturing machine tools, is the device temperature, is the equipment pressure, is the maximum vibration amplitude, is the energy consumption per unit time; Based on the quality prediction index, the operating status of the mold manufacturing machine tool production line is monitored in real time, and the data fluctuation and trend of the quality prediction index are analyzed to obtain the quality prediction result.

5. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps for obtaining the scheduling optimization plan are: Based on the quality prediction results, the dynamic task load adaptation value is calculated using the following formula: ； in, is the dynamic task load adaptation value, is a quality prediction indicator. is the average downtime of a single device in the production line. is the standard processing time of the current task, is the number of task switching times per hour on the current production line, is the energy consumption level per unit task of the current production line; According to the dynamic task load adaptation value, a scheduling optimization solution is obtained by matching the production line with the optimal dynamic task load adaptation value.

6. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps for obtaining the optimized production process are: Based on the scheduling optimization scheme, the production efficiency maximization evaluation value is calculated, and the calculation formula is: ； in, To maximize the evaluation value of production efficiency, is the dynamic task load adaptation value, The mold change time for a single task, The number of tasks processed per minute for mold manufacturing machines, is the number of operation steps for the current task, is the length of the current machine tool's real-time idle period, The starting sequence position of the task within the current production cycle; Based on the production efficiency maximization evaluation value, the mold manufacturing machine tool load balancing is performed by reconstructing the relationship between the task sequence and the execution order of the mold manufacturing machine tool to obtain an optimized production process.

7. The mold full-process quality monitoring system according to claim 1 is characterized in that: The steps for obtaining the mold usage analysis result are as follows: Filter out the task information assigned to each mold manufacturing machine tool, read the number and usage cycle of the bound mold, collect the current operating frequency and accumulated processing time of each mold, and generate mold usage status information; Based on the mold usage status information, the vibration waveform and thermal imaging map of the mold end face are read, the wear area is identified in the grayscale change area of ​​the image, and the mold usage analysis result is generated.

8. A mold full-process quality monitoring device, characterized in that: include: A processor and a memory, the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory, so that the mold full-process quality monitoring device executes the system as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program or instruction, and when the computer program or instruction is executed, the system according to any one of claims 1 to 7 is implemented.

Images