A method for controlling the cycle time of carton processing based on process status perception

Abstract

This invention discloses a method for controlling the cycle time of carton processing based on process status perception, relating to the field of carton processing control technology. The method includes: Step 1: According to a preset sampling period for various data at different process equipment in the production line, during the carton processing process, data on equipment speed, vibration, temperature, inlet accumulation, outlet accumulation, and process processing time at different process equipment locations are collected; Step 2: By calculating the target capacity coefficient of the entire line based on the bottleneck process capacity, the cycle time of non-bottleneck processes is adjusted according to health status to ensure coordinated operation among processes; Collaborative constraint verification ensures that the cycle time difference between adjacent processes does not exceed 20% of the standard cycle time to avoid downstream material waiting; Smoothing and limiting processing ensures that the single adjustment amplitude does not exceed 10% of the standard cycle time to avoid the impact of sudden cycle time changes on equipment and products; Boundary protection verification ensures that the final cycle time is within the equipment's capacity range to guarantee safe operation.

Description

Technical Field

[0001] This invention belongs to the field of carton processing control technology, specifically a carton processing cycle control method based on process status perception. Background Technology

[0002] A cardboard box processing production line typically consists of multiple processes connected in series, including printing, die-cutting, slotting, and gluing. The equipment in each process must coordinate and cooperate according to a specific processing rhythm to ensure the efficient operation of the entire production line. The production rhythm, as a core indicator for measuring the processing speed of each process, directly determines the overall capacity and product quality of the production line when set appropriately. However, existing methods for controlling the cycle time in carton processing mainly rely on fixed cycle times. When the production line is put into operation, a fixed processing cycle time is set for each process based on the equipment's rated parameters. This cycle time remains unchanged throughout the entire production cycle and cannot adapt to dynamic changes in equipment status. When the processing capacity of a process decreases due to wear or malfunction, continuing to operate at the original cycle time can easily lead to product quality defects or further equipment damage. Furthermore, existing methods are typically reactive, adjusting only after a bottleneck has formed or a malfunction has occurred. They lack the ability to predict the trend of bottleneck changes and cannot take smooth transition measures in advance to prevent production rhythm disruptions caused by bottleneck shifts. This results in problems such as excessively large differences in cycle times between adjacent processes, leading to material supply interruptions or severe accumulation between processes. Therefore, a carton processing cycle time control method based on process status perception is proposed. Summary of the Invention

[0003] The purpose of this invention is to provide a method for controlling the cycle time of carton processing based on process status perception, so as to solve the problems mentioned in the background art.

[0004] A method for controlling the cycle time of carton processing based on process status perception includes the following steps: Step 1: According to the preset sampling cycle for various data at different process equipment in the production line, during the carton processing, collect data on equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process processing time at different process equipment in the production line, and mark each collected data as normal or missing; Step 2: Set 1 second as the reference period. For each type of data with a sampling period shorter than the reference period, take the average of all sampled values ​​within the reference period as the reference value. For each type of data with a sampling period greater than or equal to the reference period, calculate the reference value using a linear interpolation method. Then, obtain the reference value for each type of data. Based on the data status label, determine the confidence level of each type of data within each reference period to obtain the process status time series dataset for each process equipment. Step 3: Use equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process time as health assessment indicators. Analyze the data of each indicator to obtain the attenuation weighted average and fluctuation stability index. Combine the fluctuation stability index and the reliability coefficient to calculate the comprehensive health and health change rate of the equipment in each process. Step 4: Calculate the stacking rate based on the inlet and outlet stacking amounts of each process equipment, and compare the stacking rate with the preset warning threshold to calculate the balance index of each process equipment; Step 5: Calculate the upstream transmission coefficient based on the comprehensive health and balance index of the upstream process equipment corresponding to each process equipment, determine the downstream transmission coefficient based on the balance index of each process equipment, and obtain the comprehensive capacity coefficient of each process equipment based on the comprehensive health, upstream transmission coefficient and downstream transmission coefficient of each process equipment. Step 6: Mark the process equipment with the lowest overall capacity coefficient as the current bottleneck process, and calculate the bottleneck degree index; Step 7: Calculate the predicted health and predicted capacity coefficients based on the health change rate of each process equipment. The process equipment with the smallest predicted capacity coefficient is designated as the predicted bottleneck process. If the predicted bottleneck process is different from the current bottleneck process, a bottleneck transfer warning message is generated, and the predicted bottleneck process is marked as a warning process. Step 8: Calculate the overall target capacity coefficient based on the bottleneck level index. Calculate the initial target cycle time of each process based on the overall target capacity coefficient and the comprehensive health status of each process's equipment. Adjust the cycle time of the early warning process based on the bottleneck transfer early warning information. After collaborative constraint verification, smoothing and limiting processing, and boundary protection verification, obtain the final target cycle time of each process's equipment. Send the final target cycle time to the controller of each process's equipment for execution.

[0005] As a further aspect of the present invention, the specific method for marking each collected data as normal or missing is as follows: If the sensor returns a valid value that is within the range of the corresponding sensor, it is marked as normal. If the sensor fails to return data or the returned value is outside the range, it is marked as missing.

[0006] As a further aspect of the present invention: the specific method for calculating the reference value using a linear interpolation method for various types of data with a sampling period greater than or equal to the reference period is as follows: First, the end time of the reference period for each type of data is used as the timestamp of the corresponding reference period and also as the target time. For a single type of data, the two adjacent actual sampling points of the reference period are first determined. The sampling point earlier in time is defined as the pre-sampling point, and the sampling point later in time is defined as the post-sampling point. The time of the pre-sampling point is marked as T1 and the sampled value at that time is marked as V1. The time of the post-sampling point is marked as T2 and the sampled value at that time is marked as V2. Then, the value of the pre-sampling point is subtracted from the value of the post-sampling point V1 to obtain the change in value. Then, the target time of the data type is subtracted from the time of the pre-sampling point T1 to obtain the target time interval. Next, the time of the post-sampling point T2 is subtracted from the time of the pre-sampling point T1 to obtain the total time interval. Then, the change in value is multiplied by the target time interval and divided by the total time interval to obtain the intermediate product. Finally, the intermediate product is divided by the total time interval to obtain the increment value. The value of the pre-sampling point V1 is added to the increment value to obtain the reference value of the reference period at the target time. When the target time is exactly equal to the actual sampling time, the reference value is V1 or V2.

[0007] As a further aspect of the present invention: the specific method for determining the credibility level of various types of data within each benchmark period based on data status markers is as follows: For all types of data with a sampling period shorter than the baseline period, the confidence level is Level 1 when all sampling points within the baseline period are in a normal state. When the proportion of missing sampling points to the total number of sampling points does not exceed 30%, the confidence level is Level 2 after filling the missing points with the average value of the two adjacent valid sampling points. When the missing proportion exceeds 30%, the confidence level is Level 3 after replacing the data with the data of the corresponding time period in the most recent complete baseline period in which all sampling points of the equipment are in a normal state. For all types of data with a sampling period greater than or equal to the baseline period, the confidence level is Level 1 when both the previous and subsequent sampling points are in a normal state. When one of the previous or subsequent sampling points is missing, the confidence level is Level 2 after replacing the value of the missing sampling point with the value of the nearest valid sampling point of the same type and performing interpolation calculation. When both the previous and subsequent sampling points are missing, the confidence level is Level 3 after replacing the data with the interpolation result within the same sampling interval in which both the previous and subsequent sampling points of the equipment are normal.

[0008] As a further aspect of the present invention, the specific method for calculating the overall health status and health status change rate of equipment in each process is as follows: First, the health assessment index data are normalized to convert them into dimensionless values ​​in the range of 0 to 1. Then, 60 baseline periods are set as the assessment backtracking window, with the current time being time zero, and the preceding times being time -1 to -60. The attenuation weight at time -K is equal to the attenuation base raised to the power of K, where K is an integer from 0 to 60. The attenuation base is set to 0.95. The normalized value at each time within the assessment window is multiplied by the corresponding attenuation weight and summed to obtain the weighted sum. The 61 attenuation weights are summed to obtain the total weight. The weighted sum is divided by the total weight to obtain the attenuation weighted mean. Then, using the attenuation weighted mean as the center value, the absolute value of the difference between the normalized value at each time within the assessment window and the center value is calculated to obtain the absolute value of the deviation. The absolute value of the deviation at each time is multiplied by the corresponding attenuation weight and summed to obtain the weighted deviation sum. The weighted deviation sum is divided by the total weight. The attenuation-weighted average deviation is obtained, and multiplied by 2 to obtain the fluctuation penalty value. The fluctuation stability index is obtained by subtracting the fluctuation penalty value from 1. If the calculated result is less than 0, it is set to 0. The credibility coefficient is 1 for level 1, 0.8 for level 2, and 0.5 for level 3. The fluctuation stability index of each indicator data is multiplied by the corresponding credibility coefficient to obtain the effective weight. The effective weights of each indicator data are added together to obtain the sum of effective weights. The effective weights of each indicator data are divided by the sum of effective weights to obtain the normalized dynamic weight. The attenuation-weighted average of each indicator data is multiplied by the corresponding normalized dynamic weight and then added together to obtain the comprehensive health score. The comprehensive health score at the current moment is subtracted from the comprehensive health score 30 seconds ago to obtain the health score difference. The health score difference is divided by the comprehensive health score 30 seconds ago to obtain the health score change rate. When the comprehensive health score 30 seconds ago is 0, the health score change rate is 1.

[0009] As a further aspect of the present invention, the specific method for calculating the balance index of equipment in each process is as follows: First, subtract the outlet accumulation from the inlet accumulation of each process equipment to obtain the net accumulation. Divide the net accumulation by the design capacity of the process inlet buffer to obtain the accumulation rate. The preset positive warning threshold is 0.7 and the negative warning threshold is -0.3. If the accumulation rate is between -0.3 and 0.7, multiply the absolute value of the accumulation rate by 2 to obtain the accumulation penalty value. Subtract the accumulation penalty value from 1 to obtain the balance index. If the accumulation rate is greater than or equal to 0.7, the balance index is equal to 0.7 minus the accumulation rate. If the accumulation rate is less than or equal to -0.3, the balance index is equal to -0.3 minus the accumulation rate.

[0010] As a further aspect of the present invention, the specific method for obtaining the comprehensive capacity coefficient of each process equipment based on the comprehensive health status, upstream transmission coefficient, and downstream transmission coefficient of each process equipment is as follows: The upstream transmission coefficient is obtained by multiplying the overall health of the upstream equipment corresponding to each process equipment by the balance index. For the first process equipment, since there is no upstream equipment, its upstream transmission coefficient is its own overall health. The downstream transmission coefficient is the balance index corresponding to each process equipment. For the last process equipment, since there is no downstream equipment, its downstream transmission coefficient is 1. The overall health of the process equipment is multiplied by 0.5 to obtain the health contribution value of this process. The upstream transmission coefficient is multiplied by 0.3 to obtain the upstream transmission contribution value. The downstream transmission coefficient is multiplied by 0.2 to obtain the downstream transmission contribution value. The overall capacity coefficient is obtained by adding the health contribution value of this process, the upstream transmission contribution value, and the downstream transmission contribution value.

[0011] As a further aspect of the present invention, the specific method for calculating the bottleneck degree index is as follows: Sort all process equipment according to their comprehensive capacity coefficient from smallest to largest. Mark the process equipment with the smallest comprehensive capacity coefficient as the current bottleneck process. If there are multiple process equipment with the same comprehensive capacity coefficient and all of them are the smallest, mark the process that is first in the production flow as the current bottleneck process. Add up the comprehensive capacity coefficients of all processes and divide by the number of processes to get the average capacity coefficient. Subtract the smallest capacity coefficient from the average capacity coefficient to get the capacity gap value. Divide the capacity gap value by the average capacity coefficient to get the bottleneck degree index.

[0012] As a further aspect of the present invention: the specific method for calculating and predicting the health status and the predicted production capacity coefficient based on the health status change rate of equipment in each process is as follows: Multiply the health change rate of each process equipment by the prediction duration coefficient 3 to obtain the health change trend value. Add 1 to the health change trend value to obtain the prediction adjustment coefficient. Multiply the current comprehensive health of the process equipment by the prediction adjustment coefficient to obtain the initial predicted health. If the initial predicted health is less than 0.1, the predicted health is 0.1; if the initial predicted health is greater than 1, the predicted health is 1; otherwise, the predicted health equals the initial predicted health. Substitute the predicted health of each process equipment into the calculation process of the comprehensive capacity coefficient. Replace the comprehensive health used in the calculation of the health contribution value of this process with the predicted health. Replace the comprehensive health of the upstream process in the calculation of the upstream transmission coefficient with the predicted health of the upstream process. Keep other calculation methods unchanged to obtain the predicted capacity coefficient. The bottleneck transfer early warning information includes the equipment number of the early warning process and the expected time to become a bottleneck. The expected time to become a bottleneck is the current time plus 90 seconds.

[0013] As a further aspect of the present invention, the specific method for obtaining the final target cycle time of each process equipment is as follows: First, obtain the standard cycle time of each process equipment. Multiply the standard cycle time by 0.6 to obtain the minimum allowable cycle time, multiply the standard cycle time by 1.5 to obtain the maximum allowable cycle time, and multiply the standard cycle time by 0.1 to obtain the cycle time adjustment step. Multiply the bottleneck degree index by 0.5 to obtain the bottleneck impact value. Subtract the bottleneck impact value from 0.1 to obtain the adjustment factor. Add 1 to the adjustment factor to obtain the capacity adjustment coefficient. Multiply the comprehensive capacity coefficient of the bottleneck process by the capacity adjustment coefficient to obtain the overall target capacity coefficient. For non-bottleneck processes, divide the standard cycle time by the overall target capacity coefficient. The adjusted cycle time is obtained by dividing the adjusted cycle time by the overall health of the equipment in that process to obtain the initial target cycle time. For bottleneck processes, the larger of the current actual cycle time and the standard cycle time is used as the initial target cycle time. If a bottleneck transfer warning is generated, the current capacity coefficient of the warning process is subtracted from the predicted capacity coefficient to obtain the capacity reduction amount. The capacity reduction amount is divided by the current capacity coefficient to obtain the expected capacity reduction percentage. The expected capacity reduction percentage is multiplied by 0.5 to obtain the response adjustment amount. 1 is added to the response adjustment amount to obtain the warning response coefficient. The initial target cycle time of the equipment in the warning process is then calculated. The initial target cycle time is multiplied by the warning response coefficient to obtain the adjusted target cycle time. The collaborative constraint verification method is as follows: for two adjacent process equipment, the target cycle time of the downstream process equipment is subtracted from the target cycle time of the upstream process equipment to obtain the cycle time difference. If the cycle time difference is negative and its absolute value is greater than 20% of the standard cycle time of the upstream process equipment, then the target cycle time of the downstream process equipment is adjusted to the value obtained by multiplying the target cycle time of the upstream process equipment by 0.8. The smoothing and limiting processing method is as follows: the final target cycle time of the previous control cycle of each process equipment is obtained as the historical cycle time, and the current target... The beat change is obtained by subtracting the historical beat from the beat. If the absolute value of the beat change is greater than the beat adjustment step size, then when the beat change is positive, the historical beat plus the beat adjustment step size is used as the smoothed target beat. When the beat change is negative, the historical beat minus the beat adjustment step size is used as the smoothed target beat. The boundary protection verification method is as follows: the target beat is compared with the minimum allowable beat and the maximum allowable beat. If it is less than the minimum allowable beat, the minimum allowable beat is taken as the final target beat. If it is greater than the maximum allowable beat, the maximum allowable beat is taken as the final target beat.

[0014] Compared with the prior art, the beneficial effects of the present invention are: (1) This invention uses the fluctuation stability index to reflect the stability of data and penalize indicators with drastic fluctuations, and uses the credibility coefficient to reflect the credibility of data and reduce the weight contribution of low-quality data. By combining the two, the weight of each indicator is dynamically determined, avoiding evaluation bias caused by fixed weight allocation and low-quality data. The health change rate reflects the trend of equipment status change, providing a quantitative basis for bottleneck prediction. (2) In this invention, the net accumulation amount is obtained by calculating the difference between the inlet and outlet accumulation amounts and then divided by the buffer design capacity to obtain the accumulation rate, thereby quantifying the material flow balance status of each process equipment; by setting a positive warning threshold of 0.7 and a negative warning threshold of -0.3 to distinguish between three situations: normal fluctuation, severe backlog, and material shortage, the balance index is calculated to comprehensively reflect the direction and degree of imbalance in material flow; by weighting and summing the health of this process, the upstream transmission coefficient, and the downstream transmission coefficient with weights of 0.5, 0.3, and 0.2 to calculate the comprehensive capacity coefficient, comprehensively considering the condition of this process itself and the material transmission coupling relationship between upstream and downstream processes, the actual capacity level of each process equipment is fully reflected, and the full-link comprehensive evaluation of the capacity of each process equipment is realized, avoiding the problem that when the cycle control is based solely on single process indicators, the material coupling relationship between processes is ignored, resulting in the cycle time of local processes appearing reasonable but the rhythm of the entire line being unbalanced; (3) This invention identifies the current bottleneck process by comparing the comprehensive capacity coefficient of each process equipment, and calculates the bottleneck severity index to quantify the severity of the bottleneck, i.e. the difference between the worst process and the average level, providing a quantitative basis for the intensity of cycle adjustment; by predicting the capacity coefficient change trend in the next 90 seconds based on the health change rate, it identifies potential bottleneck transfer risks and generates early warning information in advance, providing sufficient decision lead time for cycle adjustment, avoiding passive response and drastic cycle adjustment when the bottleneck transfers. (4) In this invention, the target capacity coefficient of the entire line is calculated based on the capacity of the bottleneck process. The cycle time of non-bottleneck processes is adjusted according to the health status to make each process coordinate and cooperate. The bottleneck process is protected by taking the larger value between the current actual cycle time and the standard cycle time to avoid putting further acceleration pressure on the bottleneck. The cycle time of the early warning process is pre-adjusted in advance according to the expected capacity reduction ratio to achieve a smooth transition. The collaborative constraint verification ensures that the difference in cycle time between adjacent processes does not exceed 20% of the standard cycle time to avoid downstream waiting for materials. The smoothing and limiting process ensures that the single adjustment range does not exceed 10% of the standard cycle time to avoid the impact of sudden changes in cycle time on equipment and products. The boundary protection verification ensures that the final cycle time is within the equipment capacity range to ensure safe operation. The triple verification mechanism jointly ensures the coordination, stability and safety of cycle time adjustment. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method framework structure of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: Please refer to Figure 1 This application provides a method for controlling the cycle time of carton processing based on process status perception, including the following steps: Step 1: According to the preset sampling cycle for various data at different process equipment in the production line, collect equipment and product data at different process equipment in the carton processing process. The specific method is as follows: The equipment data includes the real-time speed of the main motor, the vibration amplitude of key components, the temperature of the main bearing, and the cumulative runtime of the equipment during the current operating cycle. All data is collected by sensors installed on the equipment at each stage of the process. Key components refer to those that play a decisive role in the normal operation of the equipment and are prone to failure. In the carton processing production line, these specifically include: the ink roller bearing of the printing press, the main drive gearbox of the die-cutting machine, the cutter shaft assembly of the slotting machine, and the rubber roller support of the gluing machine. The equipment speed is collected by an encoder with a sampling period of 100 milliseconds; equipment vibration is collected by an acceleration sensor with a sampling period of 10 milliseconds; equipment temperature is collected by a thermocouple with a sampling period of 30 seconds; and the cumulative runtime is recorded by a timer built into the equipment control system with a sampling period of 1 second. Product data includes the quantity of products awaiting processing at the entrance of each process equipment, the quantity of finished products at the exit of each process equipment, the actual processing time of products at each process equipment, and the transmission time of products between adjacent process equipment. All data is collected by detection devices installed at the entrance and exit of each process. The quantity of products awaiting processing is collected by a photoelectric counter with a sampling period of 500 milliseconds. Processing and transmission times are calculated using the time difference of trigger signals. Specifically, photoelectric sensors are installed at the entrance and exit of each process equipment as trigger detection points. When a product passes the entrance trigger detection point, the photoelectric sensor outputs a rising edge signal, and the controller records the system timestamp at that moment. As the entry trigger moment, when the same product passes through the exit trigger detection point, the photoelectric sensor outputs a rising edge signal, and the controller records the system timestamp of this moment as the exit trigger moment. The difference between the exit trigger moment and the entry trigger moment is the processing time of the product at that process equipment, in milliseconds. Similarly, the difference between the entry trigger moment of the downstream process equipment and the exit trigger moment of the upstream process equipment is the transmission time of the product between the two adjacent process equipment, in milliseconds. Product identification and matching are achieved by barcode or QR code labels pre-attached to the product. Each trigger detection point is equipped with a barcode scanner to read the product identification code at the same time as detecting the trigger signal to ensure the accuracy of the timestamp correspondence.

[0018] During data acquisition, for each type of data, if the sensor returns a valid value within its range, the data status is marked as normal. If the sensor fails to return data or returns a value outside its range, the data status is marked as missing. The sensor's range is determined by the measurement range specified in the sensor's manufacturer's technical specifications. For example, the encoder's rotational speed measurement range is 0 to 3,000 revolutions per minute; a negative value or a value exceeding 3,000 revolutions per minute indicates it is outside its range. The accelerometer's vibration amplitude measurement range is 0 to 50 millimeters per second squared; a negative value or a value exceeding 50 millimeters per second squared indicates it is outside its range. The thermocouple's temperature measurement range is -40 degrees Celsius to 200 degrees Celsius; a value outside this range indicates it is outside its range.

[0019] Step Two: Based on the preset sampling period for various data at different process equipment in the production line, analyze the equipment, product, and environmental data at different process equipment. Determine the reliability level of each type of data according to the data status label, and obtain the process status time series dataset for each process equipment. The specific method is as follows: The baseline period is set to 1 second. For each type of data with a sampling period shorter than the baseline period, the mean of all sampled values ​​within the baseline period is used as the baseline value for that type of data in the baseline period. For example, if the sampling period of the equipment vibration sensor is ten milliseconds, that is, one hundred data points are collected per second, then the sum of these one hundred vibration amplitude values ​​and the result divided by one hundred are taken as the representative value of vibration amplitude for that second; if the sampling period of the equipment speed sensor is one hundred milliseconds, that is, ten data points are collected per second, then the sum of these ten speed values ​​and the result divided by ten are taken as the representative value of equipment speed for that second. For data types with a sampling period greater than or equal to the reference period, the reference value for each data type in that reference period is calculated using a linear interpolation method based on the values ​​of the data at two adjacent actual sampling points. The specific method is as follows: First, the end time of the reference period for each type of data is used as the timestamp of the corresponding reference period and also as the target time. Then, the timestamps and corresponding target times of each type of data with a sampling period greater than or equal to the reference period are obtained. For example, if the first reference period covers the time period from second 0 to second 1, its timestamp is second 1; if the second reference period covers the time period from second 1 to second 2, its timestamp is second 2, and so on. In the subsequent linear interpolation calculation, the timestamp of the reference period, i.e., its end time, is used as the target time for interpolation. For a single type of data, first determine the two adjacent actual sampling points of the reference period. Define the sampling point earlier in time as the pre-sampling point and the sampling point later in time as the post-sampling point. Mark the time of the pre-sampling point as T1 and the sampled value at that time as V1. Mark the time of the post-sampling point as T2 and the sampled value at that time as V2. Then, subtract the value of the pre-sampling point V1 from the value of the post-sampling point V2 to obtain the change in value. Next, subtract the time of the pre-sampling point T1 from the target time of the data type to obtain the target time interval. Then, subtract the time of the pre-sampling point T1 from the time of the post-sampling point T2 to obtain the total time interval. Next, multiply the change in value by the target time interval and divide by the total time interval to obtain the intermediate product. Finally, divide the intermediate product by the total time interval to obtain the increment value. Add the increment value to the value of the pre-sampling point V1 to obtain the reference value of the reference period at the target time. When the target time is exactly equal to the actual sampling time, that is, when the target time interval is 0 or the target time interval is equal to the total time interval, the increment value is 0 or equal to the change in value, respectively. The reference value is V1 or V2 itself. For example, the sampling period of the device's temperature sensor is 30 seconds. Assume the temperature value collected at time 0 is 45 degrees Celsius, and the temperature value collected at time 30 is 48 degrees Celsius. We need to calculate the baseline value at time 10. Following the calculation process above, first, calculate the change in value: 48 minus 45 equals 3 degrees Celsius. Then, calculate the target time interval: 10 minus 0 equals 10 seconds. Next, calculate the total time interval: 30 minus 0 equals 30 seconds. Then, calculate the intermediate product: 3 multiplied by 10 equals 30. The increment value: 30 divided by 30 equals 1 degree Celsius. Finally, calculate the baseline value: 45 plus 1 equals 46 degrees Celsius. Therefore, the temperature baseline value at time 10 is 46 degrees Celsius. By calculating in the same way, we can obtain the baseline values ​​corresponding to various data within each baseline period. Each type of data in the equipment, product, and environmental data of different process equipment is analyzed to obtain the baseline values ​​corresponding to each type of data in different baseline periods. Using the timestamp of the baseline period as an index, the baseline values ​​of various types of data of the same process equipment in the same baseline period are used as single status data of the corresponding process equipment in the corresponding baseline period. The single status data of each process equipment includes equipment speed value, equipment vibration value, equipment temperature value, equipment cumulative running time, inlet accumulation amount, outlet accumulation amount, process processing time, inter-process transmission time, ambient temperature value, ambient humidity value, and dust concentration value. The multiple single status data of each process equipment are arranged in the order of timestamp to form the process status time series dataset of each process equipment. Based on the data status markings in Step 1, the confidence level of each type of data within each baseline period is determined. For each type of data with a sampling period shorter than the baseline period, if all sampling points of this type of data are in a normal state within the baseline period, the confidence level of this type of data within the baseline period is marked as Level 1, indicating that the data is completely reliable. If some sampling points of this type of data are missing within the baseline period, but the number of missing sampling points does not exceed 30% of the total number of sampling points, the missing points are filled by the average value of the two adjacent valid sampling points. After filling, the confidence level of this type of data within the baseline period is marked as Level 2, indicating that the data is basically reliable. If the missing proportion of this type of data within the baseline period exceeds 30%, the data from the corresponding time period in the most recent complete acquisition cycle of the equipment for this process is used for overall replacement. The most recent complete acquisition cycle refers to the most recent baseline period before the current baseline period where all sampling points of this type of data are in a normal state. After replacement, the confidence level of this type of data within the baseline period is marked as Level 3, indicating that the data has low confidence. For all types of data with a sampling period greater than or equal to the baseline period, if the data status of both the previous and subsequent sampling points is normal, the reliability level of this type of data within the baseline period is marked as Level 1; if the data status of either the previous or subsequent sampling point is missing, the value of the missing sampling point is replaced by the value of the nearest valid sampling point of the same type before interpolation calculation, and the reliability level is marked as Level 2; if the data status of both the previous and subsequent sampling points is missing, the interpolation result within the same sampling interval where both the previous and subsequent sampling points of the equipment in this process are normal is used as a replacement, and the reliability level is marked as Level 3. Finally, using the timestamp of the baseline period as an index, the baseline values ​​and their reliability levels of various data for the same process equipment within the same baseline period are combined into a single status data entry for the corresponding process equipment within the corresponding baseline period. Each single status data entry for a process equipment contains eleven data items and their corresponding reliability levels: equipment rotation speed and its reliability level, equipment vibration value and its reliability level, equipment temperature value and its reliability level, equipment cumulative running time and its reliability level, inlet accumulation amount and its reliability level, outlet accumulation amount and its reliability level, process processing time and its reliability level, inter-process transmission time and its reliability level, ambient temperature value and its reliability level, ambient humidity value and its reliability level, and dust concentration value and its reliability level. The multiple single status data entries for each process equipment are arranged in order of timestamp to form the process status time-series dataset for each process equipment.

[0020] By unifying data with different sampling periods onto a time axis with a one-second base period, for high-frequency data with sampling periods shorter than the base period, the arithmetic mean of all sampled values ​​within a base period is taken as the base value for that base period, thus reducing data density while preserving information. For low-frequency data with sampling periods greater than or equal to the base period, a linear interpolation method is used to generate a base value at the target time of each base period using the values ​​of two adjacent actual sampling points, thereby filling data gaps on the base period scale. The specific process of linear interpolation is to multiply the numerical change by the target time interval, divide by the total time interval to obtain the increment value, and add the previous sampling point value to obtain the base value. Based on this, the confidence level is determined for each type of data within each base period according to the data status label in step one. For high-frequency data, it is divided into three levels according to the proportion of missing sampling points within the base period: all normal data is labeled as level one, and when the missing proportion does not exceed 30%, adjacent data with... After filling in the mean of the validity values, they were marked as Level 2. When the missing rate exceeded 30%, the most recent complete period data was used as the replacement and marked as Level 3. For low-frequency data, they were divided into three levels according to the missing status of the sampling points before and after sampling: all normal were marked as Level 1; when one was missing, the most recent valid value was used as the replacement and marked as Level 2; when all were missing, the most recent normal interval interpolation result was used as the replacement and marked as Level 3. Finally, the eleven data benchmark values ​​and their credibility levels of the same process equipment were arranged in the order of timestamps to form a process status time series dataset. This solved the problem of time series alignment and quality classification of multi-source heterogeneous data. By unifying the benchmark period, the time misalignment between data with different sampling frequencies was eliminated, enabling subsequent analysis to perform multi-dimensional data fusion on the same time scale. The credibility level mechanism quantified the uncertainty introduced in the data processing process, enabling subsequent health assessments to dynamically adjust the weight contribution of each indicator according to the data credibility, avoiding the excessive influence of low-quality data on the assessment results.

[0021] Step 3: Analyze the time-series datasets of the process status of equipment in each process and the reliability level of each data item to obtain the comprehensive health status and health status change rate of equipment in each process. The specific method is as follows: First, 60 baseline cycles are set as the evaluation backtracking window, i.e., the evaluation backtracking window is 60 seconds. Then, six data items are selected from the process status time series dataset: equipment speed value, equipment vibration value, equipment temperature value, inlet accumulation amount, outlet accumulation amount, and process processing time as health evaluation indicators. For a single type of data, 61 status data are read from the corresponding process status time series dataset, including the current time and the previous 60 baseline cycles. After normalizing each data item, the original data with different dimensions are converted into dimensionless values ​​in the range of 0 to 1. The specific method is as follows: The lower and upper limits of the normal operating range are pre-defined for each type of data. The lower and upper limits of the normal operating range for each type of data are determined with reference to the carton processing industry standards and equipment factory parameters and can be read directly, so they will not be elaborated here. When the parameter baseline value is between the lower and upper limits, the lower limit value is first subtracted from the baseline value to obtain the first difference value, then the lower limit value is subtracted from the upper limit value to obtain the second difference value, and finally the first difference value is divided by the second difference value to obtain the normalized value. If the baseline value is lower than the lower limit value, the normalized value is 0; if the baseline value is higher than the upper limit value, the normalized value is 1. Then, let the current time be time zero, the previous second be time -1, the previous two seconds be time -2, and so on up to time -60. The decay weight at time -K is equal to the decay base raised to the power of K, where K is an integer from 0 to 60, and the decay base is set to 0.95. The basis for choosing 0.95 is that 0.95 raised to the power of 60 is approximately equal to 0.046, which means that the weight of the data 60 seconds ago is about one-twentieth of the weight of the current data. This way, the reference value of historical data is preserved while recent data takes the lead. For example, the decay weight at time zero (the current time) is 0.95 raised to the power of 0, which equals 1; the decay weight at time -1 is 0.95 raised to the power of 1, which equals 0.95; the decay weight at time -2 is 0.95 raised to the power of 2, which equals 0.9025; the decay weight at time -10 is 0.95 raised to the power of 10, which equals approximately 0.599; and the decay weight at time -60 is 0.95 raised to the power of 60, which equals approximately 0.046. Next, the normalized value at each time step within the evaluation window is multiplied by the corresponding decay weight to obtain 61 weighted values. These 61 weighted values ​​are then summed to obtain a total weighted value. The sum of the 61 decay weights is then summed to obtain a total weight. Finally, the total weighted value is divided by the total weight to obtain the decay-weighted mean of this data type. Using this decay-weighted mean as the center value, the difference between the normalized value and the center value at each time step within the evaluation window is calculated, and the absolute value is taken to obtain 61 absolute deviation values. The absolute deviation value at each time step is multiplied by the corresponding decay weight and summed to obtain a weighted deviation sum. Finally, the absolute deviation value at each time step is multiplied by... The weighted deviations are summed after applying the corresponding attenuation weights at each time point. Then, the weighted deviation sum is divided by the weight sum to obtain the attenuation weighted average deviation. Next, the attenuation weighted average deviation is multiplied by 2 to obtain the fluctuation penalty value. Finally, the fluctuation penalty value is subtracted from 1 to obtain the fluctuation stability index. If the calculation result is less than 0, it is taken as 0. The value of this index ranges from 0 to 1, and the larger the value, the more stable the data. The same analysis is performed on the data of six health assessment indicators: equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process processing time, to obtain the fluctuation stability index corresponding to each of the six health assessment indicators. When the confidence level is level 1, the confidence coefficient is 1; when the confidence level is level 2, the confidence coefficient is 0.8; and when the confidence level is level 3, the confidence coefficient is 0.5. The effective weight of each health assessment indicator is obtained by multiplying the fluctuation stability index of each health assessment indicator data by its corresponding confidence coefficient. The effective weights of the six health assessment indicators data (equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process processing time) are added together to obtain the total effective weights. The normalized dynamic weight of each health assessment indicator data is obtained by dividing the effective weight of each data item by the total effective weights. Next, the attenuation-weighted average of the health assessment indicators of each process equipment is multiplied by the corresponding normalized dynamic weight, and then the six products are added together to obtain the comprehensive health of the corresponding process equipment. The comprehensive health value ranges from 0 to 1. The higher the comprehensive health value, the better the operating status of the process equipment. The same analysis is performed on each process equipment to obtain the comprehensive health of each process equipment. Finally, the overall health status of each process equipment at the current moment is subtracted from the overall health status 30 seconds ago to obtain the health status difference. Then, the health status difference is divided by the overall health status 30 seconds ago to obtain the health status change rate of each process equipment. The 30-second interval is chosen as the basis because 30 seconds is about half of the evaluation backtracking window, which can effectively reflect the mid-term change trend. A positive change rate indicates that the health status is rising, that is, the equipment status is improving, while a negative change rate indicates that the health status is falling, that is, the equipment status is deteriorating. When the overall health status 30 seconds ago is 0, in order to avoid division by zero error, the health status change rate is set to positive 1, indicating that the equipment status is improving from the worst state. To illustrate with an example, suppose the overall health of a certain process equipment is 0.78 at the current moment, and 0.80 30 seconds ago. The difference in health is 0.78 minus 0.80, which equals -0.02. The rate of change in health is -0.02 divided by 0.80, which equals -0.025. This means that the health of the equipment has decreased by 2.5% in the past 30 seconds, and the equipment condition is slightly deteriorating.

[0022] By using exponential decay weighting, the evaluation results can quickly respond to changes in equipment status while maintaining a certain degree of smoothness. The fluctuation stability index penalizes indicators with drastic data fluctuations to prevent them from interfering too much with the overall evaluation. The credibility coefficient reduces the weight contribution of low-quality data to improve the reliability of the evaluation results. The introduction of the health change rate enables the system not only to perceive the current status of the equipment, but also to capture the trend of status changes, providing a basis for bottleneck prediction.

[0023] Step 4: Based on the overall health status and health status change rate of the equipment in each process and the product stacking data of each process, obtain the balance index of the equipment in each process. The specific method is as follows: First, subtract the outlet accumulation from the inlet accumulation of each process equipment to obtain the net accumulation. Then, divide the net accumulation by the design capacity of the process inlet buffer to obtain the accumulation rate of each process equipment. The design capacity of the process inlet buffer refers to the maximum number of products that the area at the inlet of the process equipment used to temporarily store products to be processed can accommodate. This value is determined by the specifications of each process buffer marked in the production line design document. A positive accumulation rate indicates inlet backlog, that is, the processing speed of this process is slower than the upstream feeding speed. A negative accumulation rate indicates outlet backlog, that is, the processing speed of this process is faster than the downstream digestion speed. Then, preset positive and negative warning thresholds. The positive warning threshold is set to 0.7 and the negative warning threshold is set to -0.3. The basis for selecting these two thresholds is: when the inlet accumulation reaches 70% of the buffer capacity, continued accumulation will quickly lead to buffer overflow; when the outlet accumulation is 30% higher than the inlet accumulation, it indicates that the upstream supply may be insufficient. For a single process equipment, if the stacking rate of that equipment is between -0.3 and 0.7, which is within the normal range, first calculate the absolute value of the stacking rate, then multiply the absolute value by 2 to obtain the stacking penalty value, and finally subtract the stacking penalty value from 1 to obtain the balance index of that process equipment. If the stacking rate is greater than or equal to 0.7, it indicates severe backlog, then the balance index is equal to 0.7 minus the stacking rate. In this case, the balance index is negative or zero, and the smaller the value, the more severe the backlog. If the stacking rate is less than or equal to -0.3, it indicates a material shortage, then the balance index is equal to -0.3 minus the stacking rate, thus obtaining the balance index of that process equipment. The same analysis is performed on each process equipment to obtain the balance index of each process equipment. The balance index, calculated based on the product accumulation at the inlet and outlet of each process equipment, reflects the balance of material flow. First, the net accumulation is obtained by subtracting the outlet accumulation from the inlet accumulation. Then, this net accumulation is divided by the design capacity of the inlet buffer zone to obtain the accumulation rate. A positive accumulation rate indicates inlet backlog, meaning the processing speed of this process is slower than the upstream supply speed. A negative accumulation rate indicates outlet backlog, meaning the processing speed of this process is faster than the downstream digestion speed. Then, positive warning thresholds of 0.7 and negative warning thresholds of -0.3 are set as judgment boundaries. Within the normal range, the balance index is calculated through a backlog penalty mechanism, making the balance index closer to 1 as the accumulation rate approaches zero. When the positive threshold is exceeded, the balance index becomes negative, and the smaller the value, the more severe the inlet backlog. When the negative threshold is exceeded, the balance index also becomes negative, reflecting the degree of insufficient upstream supply. By comparing and analyzing the accumulation situation at the inlet and outlet through the accumulation rate, the directional imbalance of material flow can be accurately determined. By setting bidirectional warning thresholds and calculating the balance index in segments, normal fluctuations, slight imbalances, and severe imbalances can be distinguished and handled, providing a quantitative indicator reflecting the material flow status for subsequent comprehensive capacity coefficient calculations.

[0024] Step 5: Based on the production flow sequence of each process equipment in the production line, and according to the comprehensive health and balance index of the upstream process equipment corresponding to each piece of equipment, and the balance index corresponding to each process equipment, calculate the upstream transfer coefficient and downstream transfer coefficient of each process equipment. Then, based on the upstream transfer coefficient, downstream transfer coefficient, and comprehensive health of each process equipment, obtain the comprehensive capacity coefficient of each process equipment. The specific method is as follows: The product of the overall health of the upstream process equipment corresponding to each process equipment and the balance index is used as the upstream transmission coefficient of each process equipment, and the balance index corresponding to each process equipment is used as the downstream transmission coefficient of each process equipment. For the first process equipment, since there is no corresponding upstream process equipment, its upstream transmission coefficient is taken as the overall health of the first process itself; for the last process, since there is no corresponding downstream process equipment, 1 is used as its downstream transmission coefficient, thus obtaining the upstream transmission coefficient and downstream transmission coefficient of each process equipment. The specific process for calculating the comprehensive capacity coefficient of a single process equipment is as follows: First, multiply the comprehensive health status of the process equipment by 0.5 to obtain the health status contribution value of this process. Then, multiply its corresponding upstream transmission coefficient by 0.3 to obtain the upstream transmission contribution value. Next, multiply its corresponding downstream transmission coefficient by 0.2 to obtain the downstream transmission contribution value. Finally, add the health status contribution value of this process, the upstream transmission contribution value, and the downstream transmission contribution value together to obtain the comprehensive capacity coefficient of the process equipment. Perform the same calculation on each process equipment to obtain the comprehensive capacity coefficient of each process equipment. The weighting is based on the fact that the health of this process itself is the primary factor determining the production capacity, so it is given the highest weight of 0.5. The upstream supply capacity has a direct impact on this process, so it is given a weight of 0.3. The downstream digestion capacity has a relatively indirect impact, so it is given a lower weight of 0.2.

[0025] First, the upstream transmission coefficient is obtained by multiplying the overall health of the upstream equipment by the balance index. This coefficient reflects the upstream's ability to stably supply materials to this process. The balance index of the equipment in this process is used as the downstream transmission coefficient, reflecting the process's ability to stably output materials to the downstream. Then, the overall health of this process is multiplied by 0.5, the upstream transmission coefficient by 0.3, and the downstream transmission coefficient by 0.2. The three are then added together to obtain the overall capacity coefficient. For the first process, since there is no upstream, its upstream transmission coefficient is taken as the health of this process. For the last process, since there is no downstream, its downstream transmission coefficient is taken as 1. The overall capacity coefficient not only reflects the operational health of the equipment in each process but also incorporates the impact of upstream supply capacity and downstream digestion capacity on the actual capacity of this process.

[0026] Step Six: Mark the current bottleneck process based on the comprehensive capacity coefficient of each process's equipment. The specific method is as follows: First, sort all process equipment according to their comprehensive capacity coefficient from smallest to largest. Mark the process equipment with the smallest comprehensive capacity coefficient as the current bottleneck process. If multiple process equipment have the same comprehensive capacity coefficient and are all the minimum, mark the process that is first in the production flow sequence of these process equipment as the current bottleneck process, because the upstream bottleneck has a more direct impact on the overall capacity. Then, add up the comprehensive capacity coefficients of all processes and divide by the number of processes to get the average capacity coefficient. Then, subtract the minimum capacity coefficient from the average capacity coefficient to get the capacity gap value. Finally, divide the capacity gap value by the average capacity coefficient to get the bottleneck degree index. The larger the bottleneck degree index value, the more serious the bottleneck is, that is, the greater the gap between the worst process and the average level. Based on the comprehensive capacity coefficient of each process's equipment, the bottleneck process currently restricting the overall production line's capacity is identified and its severity is quantified. First, all process equipment is sorted by comprehensive capacity coefficient from smallest to largest. The process with the smallest coefficient is marked as the current bottleneck process. If multiple processes have the same smallest coefficient, the process with the earliest production flow is taken as the bottleneck, as upstream bottlenecks have a more direct impact on overall capacity. Then, the average comprehensive capacity coefficient of all processes is calculated as the average capacity coefficient. Subtracting the smallest capacity coefficient from the average capacity coefficient yields the capacity gap value. Dividing the capacity gap value by the average capacity coefficient yields the bottleneck severity index. The larger this index value, the greater the gap between the bottleneck process and the average level, indicating a more severe bottleneck. The beneficial effect of this step is that it achieves automatic identification of bottleneck processes and quantitative assessment of bottleneck severity. Compared to existing technologies that rely on manual observation of accumulation phenomena to determine bottlenecks, comparing and sorting based on the comprehensive capacity coefficient of each process can accurately locate the weak links restricting overall efficiency from the perspective of capacity contribution; it also provides a quantitative basis for determining the subsequent adjustment of cycle time.

[0027] Step 7: Based on the health status change rate and comprehensive capacity coefficient of equipment in each process, predict bottleneck trends and obtain bottleneck transfer early warning information. The specific method is as follows: For a single process piece of equipment, the health change rate of that equipment is multiplied by a prediction duration coefficient of 3 to obtain the health change trend value. The selection of 3 as the prediction duration coefficient is based on the fact that 3 times the health change rate measurement cycle (30 seconds multiplied by 3 equals 90 seconds) provides sufficient advance warning for cycle time adjustment. Then, 1 is added to the health change trend value to obtain the prediction adjustment coefficient. Next, the current overall health of the equipment in that process is multiplied by the prediction adjustment coefficient to obtain the initial predicted health. To avoid unreasonable prediction values, the lower limit of the predicted health is set to 0.1 and the upper limit to 1. If the initial predicted health is less than 0.1, the predicted health is set to 0.1; if the initial predicted health is greater than 1, the predicted health is set to 1; otherwise, the predicted health equals the initial predicted health. The same calculation is performed for each process piece of equipment to obtain the predicted health of each process piece of equipment. Substitute the predicted health of each process equipment into the calculation of the comprehensive capacity coefficient in step five. Replace the comprehensive health used in the calculation of the health contribution value of this process with the predicted health of this process. Replace the comprehensive health of the upstream process in the calculation of the upstream transmission coefficient with the predicted health of the upstream process. Keep other calculation methods unchanged, and then obtain the predicted capacity coefficient of each process equipment. Next, all process equipment is sorted in ascending order of predicted capacity coefficient. The process equipment with the smallest predicted capacity coefficient is designated as the predicted bottleneck process. If the predicted bottleneck process is different from the current bottleneck process marked in step six, it is determined that the bottleneck is about to shift and an early warning message needs to be generated. The predicted bottleneck process is then designated as the early warning process equipment. The bottleneck shift early warning message includes the early warning process equipment number and the expected time to become a bottleneck. The expected time to become a bottleneck is the current time plus 90 seconds, which is obtained by multiplying the predicted duration coefficient of 3 by the health change rate measurement period of 30 seconds. The current capacity coefficient is the current comprehensive capacity coefficient of the process, and the predicted capacity coefficient is the predicted capacity coefficient of the process. If the process equipment with the smallest predicted capacity coefficient is the same as the current bottleneck process, no bottleneck shift early warning message is generated, indicating that the bottleneck position is not expected to change. By predicting future health and capacity coefficient trends based on the health change rate of each process's equipment and identifying potential bottleneck transfer risks, the predicted health of each process is substituted into the comprehensive capacity coefficient calculation process in step five to obtain the predicted capacity coefficient of each process. Then, all processes are sorted according to the predicted capacity coefficient, and the process with the smallest predicted coefficient is identified as the predicted bottleneck process. If the predicted bottleneck process is different from the current bottleneck process, it is determined that the bottleneck is about to transfer, and an early warning message containing the warning process equipment number and the expected time of becoming the bottleneck is generated. The beneficial effect is that it realizes early warning of bottleneck transfer, and can identify which process may become the new bottleneck about 90 seconds in advance. This warning time is sufficient for the system to adjust the cycle time of the warning process in advance to achieve a smooth transition and avoid the impact on the production line caused by drastic cycle time adjustment when the bottleneck transfers.

[0028] Step 8: Calculate the overall target capacity coefficient based on the bottleneck severity index. Then, calculate the initial target cycle time for each process based on the overall health of the equipment in each process and the overall target capacity coefficient. Adjust the cycle time of the warning processes based on the bottleneck transfer early warning information obtained in Step 7. Finally, calculate the final target cycle time for each process. The specific method is as follows: First, the design processing cycle of each process equipment under rated operating conditions is obtained as the standard cycle time of each process equipment. This value is obtained by converting the rated production capacity marked in the equipment's technical documents. The conversion method is to divide 60 seconds by the rated number of pieces processed per minute to obtain the standard processing time of each product. The standard cycle time of each process equipment is multiplied by 0.6 to obtain the minimum allowable cycle time of each process equipment, and multiplied by 1.5 to obtain the maximum allowable cycle time of each process equipment. The basis for determining this range is: when the cycle time is lower than 60% of the standard value, the equipment will be overloaded and the risk of failure will be significantly increased; when the cycle time is higher than 150% of the standard value, the production efficiency is too low and does not meet the economic requirements. The standard cycle time of each process equipment is multiplied by 0.1 to obtain the cycle time adjustment step size of each process equipment, that is, the single adjustment range shall not exceed 10% of the standard cycle time, so as to avoid the impact of drastic changes in cycle time on equipment and product quality. First, multiply the bottleneck severity index by 0.5 to get the bottleneck impact value. Then, subtract the bottleneck impact value from 0.1 to get the adjustment factor. Next, add 1 to the adjustment factor to get the capacity adjustment coefficient. Finally, multiply the comprehensive capacity coefficient of the bottleneck process by the capacity adjustment coefficient to get the target capacity coefficient for the entire line. Next, the initial target cycle time of each process equipment is calculated. For non-bottleneck processes, the standard cycle time of the corresponding process equipment is first divided by the overall target capacity coefficient to obtain the capacity-adjusted cycle time of the corresponding process equipment. Then, the capacity-adjusted cycle time is divided by the overall health of the corresponding process equipment to obtain the initial target cycle time of the corresponding process equipment. For bottleneck processes, the current actual cycle time of the process equipment is first compared with the standard cycle time, and the larger value is taken as the initial target cycle time of the corresponding process equipment to avoid putting further acceleration pressure on the bottleneck. If step seven generates a bottleneck transfer warning, the cycle time of the warning process needs to be adjusted in advance to achieve a smooth transition; the specific method is as follows: First, subtract the predicted capacity coefficient from the current capacity coefficient of the warning process to obtain the capacity reduction amount. Then, divide the capacity reduction amount by the current capacity coefficient to obtain the expected capacity reduction ratio. Next, multiply the expected capacity reduction ratio by 0.5 to obtain the response adjustment amount. Then, add 1 to the response adjustment amount to obtain the warning response coefficient. Finally, multiply the initial target cycle time of the equipment in the warning process by the warning response coefficient to obtain the adjusted target cycle time of the equipment in the warning process. Keep the initial target cycle time of other process equipment as the target cycle time. If no bottleneck transfer warning information is generated in step seven, no processing is performed. At the same time, the target cycle time of each process equipment is kept as the initial target cycle time. After obtaining the target cycle time of each process equipment, three verification operations are performed sequentially: collaborative constraint verification, smoothing and limiting processing, and boundary protection verification. The specific method is as follows: For each pair of adjacent process equipment in the production line, the target cycle time of the downstream process equipment is subtracted from the target cycle time of the upstream process equipment to obtain the cycle time difference. If the cycle time difference is negative and its absolute value is greater than 20% of the standard cycle time of the upstream process equipment, it indicates that the cycle time of the downstream process is significantly faster than that of the upstream process, which will cause the downstream to wait for materials. In this case, the target cycle time of the downstream process equipment is adjusted to the value obtained by multiplying the target cycle time of the upstream process equipment by 0.8 to ensure that the cycle time of the downstream process is not too fast than that of the upstream process. After performing the same verification and adjustment on each pair of adjacent process equipment, the adjusted target cycle time of each process equipment is obtained; thus, the collaborative constraint verification is completed.

[0029] Then, after the collaborative constraint verification, the target cycle time of each process equipment is smoothed and limited. Specifically, the final target cycle time of the previous control cycle of each process equipment is obtained as the historical cycle time. The current target cycle time is subtracted from the historical cycle time to obtain the cycle time change. If the absolute value of the cycle time change is greater than the cycle time adjustment step size of the process equipment, then when the cycle time change is positive, the historical cycle time is added to the cycle time adjustment step size to obtain the smoothed target cycle time. When the cycle time change is negative, the historical cycle time is subtracted from the cycle time adjustment step size to obtain the smoothed target cycle time. If the absolute value of the cycle time change is less than or equal to the cycle time adjustment step size, the current target cycle time is kept unchanged. This process ensures that the cycle time adjustment is smooth and gradual, avoiding sudden changes that could impact the equipment, thus completing the smoothing and limiting process.

[0030] Finally, the target cycle time of each process equipment after smoothing and limiting is compared with the minimum and maximum allowable cycle time of the corresponding process equipment. If it is less than the minimum allowable cycle time, the minimum allowable cycle time value is taken as the final target cycle time. If it is greater than the maximum allowable cycle time, the maximum allowable cycle time value is taken as the final target cycle time. This ensures that the final target cycle time is always within the equipment's capacity range, thereby completing the boundary protection verification and finally obtaining the final target cycle time of the process equipment.

[0031] The final target cycle time of each process equipment is sent to the controller of each process equipment. After receiving the final target cycle time, the controller of each process equipment adjusts the running speed of the main motor of the equipment according to the cycle time value, thereby completing the real-time control of the processing cycle time of each process equipment. When the next reference cycle arrives, steps one to eight are repeated to achieve continuous dynamic adjustment of the cycle time.

[0032] First, the standard cycle time of each process is determined through the equipment's factory technical documents, and three constraint parameters—minimum allowable cycle time, maximum allowable cycle time, and cycle time adjustment step size—are calculated accordingly. Then, the capacity adjustment coefficient and the overall target capacity coefficient are calculated based on the bottleneck severity index. For non-bottleneck processes, the initial target cycle time is obtained by dividing the standard cycle time by the overall target capacity coefficient and then by the process's health status. For bottleneck processes, the larger value between the current actual cycle time and the standard cycle time is used as the initial target cycle time to avoid further pressure on the bottleneck. If a bottleneck transition warning exists, the cycle time of the warning process is pre-adjusted to achieve a smooth transition. Finally, a collaborative constraint verification is performed to ensure that the cycle time difference between adjacent processes does not exceed 20% of the standard cycle time, avoiding downstream material waiting. Smoothing and limiting processes ensure that a single adjustment does not exceed 10% of the standard cycle time, avoiding abrupt changes in cycle time. Boundary protection checks ensure that the final cycle time is within the equipment's capacity. Finally, the final target cycle time is sent to the motion controllers of each process's equipment for execution. Overall coordination based on the bottleneck process is achieved through the target capacity coefficient of the entire line. Differentiated calculations that distinguish between bottleneck and non-bottleneck processes avoid blindly increasing pressure on the bottleneck. A warning response mechanism ensures a smooth transition before the bottleneck shifts. Collaborative constraint checks guarantee cycle time matching between adjacent processes. Smoothing and limiting processes prevent the impact of drastic cycle time changes on equipment and products. Boundary protection checks ensure that the equipment always operates within its safe capacity.

[0033] First, multiple types of sensors deployed on equipment at each process stage continuously collect equipment operation status data, product flow data, and workshop environment data. This raw data is then processed through a baseline period to form a time-aligned process status time-series dataset. Next, the overall health of each process's equipment is calculated based on attenuation-weighted average and fluctuation stability analysis. A balance index is calculated based on a comparison of inlet and outlet accumulation amounts. Finally, a comprehensive capacity coefficient is calculated by considering the health of each process, upstream supply capacity, and downstream processing capacity. Then, by comparing the comprehensive capacity coefficients of each process's equipment, the current bottleneck process is identified and its severity is quantified. Simultaneously, future capacity change trends are predicted based on the rate of change in health status. Potential bottleneck transfer risks; finally, the target capacity coefficient of the entire line is calculated based on the capacity of the bottleneck process, and the initial target cycle time is calculated separately for bottleneck and non-bottleneck processes. The early warning processes are pre-adjusted. After triple verification of collaborative constraints, smoothing limits and boundary protection, a safe and executable final target cycle time is obtained and issued to the equipment for execution. When the next control cycle arrives, the above process is repeated to achieve continuous dynamic optimization of the cycle time. It can adapt to changes in equipment status and fluctuations in material flow in real time, effectively avoid local accumulation or material shortage problems caused by cycle time mismatch, prevent production fluctuations caused by bottleneck transfer in advance, and ensure stable operation of equipment within the safe capacity range through multiple safety mechanisms.

[0034] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0035] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for controlling the cycle time of carton processing based on process status perception, characterized in that, Includes the following steps: Step 1: According to the preset sampling cycle for various data at different process equipment in the production line, during the carton processing, collect data on equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process processing time at different process equipment in the production line, and mark each collected data as normal or missing; Step 2: Set 1 second as the reference period. For each type of data with a sampling period shorter than the reference period, take the average of all sampled values ​​within the reference period as the reference value. For each type of data with a sampling period greater than or equal to the reference period, calculate the reference value using a linear interpolation method. Then, obtain the reference value for each type of data. Based on the data status label, determine the confidence level of each type of data within each reference period to obtain the process status time series dataset for each process equipment. Step 3: Use equipment speed, equipment vibration, equipment temperature, inlet accumulation, outlet accumulation, and process time as health assessment indicators. Analyze the data of each indicator to obtain the attenuation weighted average and fluctuation stability index. Combine the fluctuation stability index and the reliability coefficient to calculate the comprehensive health and health change rate of the equipment in each process. Step 4: Calculate the stacking rate based on the inlet and outlet stacking amounts of each process equipment, and compare the stacking rate with the preset warning threshold to calculate the balance index of each process equipment; Step 5: Calculate the upstream transmission coefficient based on the comprehensive health and balance index of the upstream process equipment corresponding to each process equipment, determine the downstream transmission coefficient based on the balance index of each process equipment, and obtain the comprehensive capacity coefficient of each process equipment based on the comprehensive health, upstream transmission coefficient and downstream transmission coefficient of each process equipment. Step 6: Mark the process equipment with the lowest overall capacity coefficient as the current bottleneck process, and calculate the bottleneck degree index; Step 7: Calculate the predicted health and predicted capacity coefficients based on the health change rate of each process equipment. The process equipment with the smallest predicted capacity coefficient is designated as the predicted bottleneck process. If the predicted bottleneck process is different from the current bottleneck process, a bottleneck transfer warning message is generated, and the predicted bottleneck process is marked as a warning process. Step 8: Calculate the overall target capacity coefficient based on the bottleneck level index. Calculate the initial target cycle time of each process based on the overall target capacity coefficient and the comprehensive health status of each process's equipment. Adjust the cycle time of the early warning process based on the bottleneck transfer early warning information. After collaborative constraint verification, smoothing and limiting processing, and boundary protection verification, obtain the final target cycle time of each process's equipment. Send the final target cycle time to the controller of each process's equipment for execution.

2. The method for controlling the cycle time of carton processing based on process status perception according to claim 1, characterized in that, The specific method for marking each collected data as normal or missing is as follows: If the sensor returns a valid value that is within the range of the corresponding sensor, it is marked as normal. If the sensor fails to return data or the returned value is outside the range, it is marked as missing.

3. The method for controlling the cycle time of carton processing based on process status perception according to claim 1, characterized in that, The specific method for calculating the reference value using linear interpolation for various types of data with a sampling period greater than or equal to the reference period is as follows: First, the end time of the baseline period for each type of data is used as the timestamp of the corresponding baseline period and also as the target time. For a single type of data, the two adjacent actual sampling points of the baseline period are first determined. The sampling point that is earlier in time is defined as the pre-sampling point, and the sampling point that is later in time is defined as the post-sampling point. The time of the pre-sampling point is marked as T1 and the sampled value at that time is marked as V1. The time of the post-sampling point is marked as T2 and the sampled value at that time is marked as V2. Then, subtract the value of the previous sampling point V1 from the value of the subsequent sampling point V2 to obtain the change in value; then subtract the time of the previous sampling point T1 from the target time of this type of data to obtain the target time interval; then subtract the time of the previous sampling point T1 from the time of the subsequent sampling point T2 to obtain the total time interval; then multiply the change in value by the target time interval and divide by the total time interval to obtain the intermediate product; finally, divide the intermediate product by the total time interval to obtain the increment value; add the increment value to the value of the previous sampling point V1 to obtain the reference value of the reference period at the target time. When the target time is exactly equal to the actual sampling time, the reference value is V1 or V2.

4. The method for controlling the cycle time of carton processing based on process state perception according to claim 1, characterized in that, The specific method for determining the credibility level of various types of data within each baseline period based on data status labels is as follows: For all types of data with a sampling period shorter than the baseline period, the confidence level is Level 1 when all sampling points within the baseline period are in a normal state. When the proportion of missing sampling points to the total number of sampling points does not exceed 30%, the confidence level is Level 2 after filling the missing points with the average value of the two adjacent valid sampling points. When the missing proportion exceeds 30%, the confidence level is Level 3 after replacing the data with the data of the corresponding time period in the most recent complete baseline period in which all sampling points of the equipment are in a normal state. For all types of data with a sampling period greater than or equal to the baseline period, the confidence level is Level 1 when both the previous and subsequent sampling points are in a normal state. When one of the previous or subsequent sampling points is missing, the confidence level is Level 2 after replacing the value of the missing sampling point with the value of the nearest valid sampling point of the same type and performing interpolation calculation. When both the previous and subsequent sampling points are missing, the confidence level is Level 3 after replacing the data with the interpolation result within the same sampling interval in which both the previous and subsequent sampling points of the equipment are normal.

5. The method for controlling the cycle time of carton processing based on process status perception according to claim 1, characterized in that, The specific method for calculating the overall health status and health status change rate of equipment in each process is as follows: First, the health assessment index data are normalized to convert them into dimensionless values ​​in the range of 0 to 1. Then, 60 baseline periods are set as the assessment backtracking window, with the current time being time zero, and the preceding times being time -1 to -60. The attenuation weight at time -K is equal to the attenuation base raised to the power of K, where K is an integer from 0 to 60. The attenuation base is set to 0.

95. The normalized value at each time within the assessment window is multiplied by the corresponding attenuation weight and summed to obtain the weighted sum. The 61 attenuation weights are summed to obtain the total weight. The weighted sum is divided by the total weight to obtain the attenuation weighted mean. Then, using the attenuation weighted mean as the center value, the absolute value of the difference between the normalized value at each time within the assessment window and the center value is calculated to obtain the absolute value of the deviation. The absolute value of the deviation at each time is multiplied by the corresponding attenuation weight and summed to obtain the weighted deviation sum. The weighted deviation sum is divided by the total weight. The attenuation-weighted average deviation is obtained, and multiplied by 2 to obtain the fluctuation penalty value. The fluctuation stability index is obtained by subtracting the fluctuation penalty value from 1. If the calculated result is less than 0, it is set to 0. The credibility coefficient is 1 for level 1, 0.8 for level 2, and 0.5 for level 3. The fluctuation stability index of each indicator data is multiplied by the corresponding credibility coefficient to obtain the effective weight. The effective weights of each indicator data are added together to obtain the sum of effective weights. The effective weights of each indicator data are divided by the sum of effective weights to obtain the normalized dynamic weight. The attenuation-weighted average of each indicator data is multiplied by the corresponding normalized dynamic weight and then added together to obtain the comprehensive health score. The comprehensive health score at the current moment is subtracted from the comprehensive health score 30 seconds ago to obtain the health score difference. The health score difference is divided by the comprehensive health score 30 seconds ago to obtain the health score change rate. When the comprehensive health score 30 seconds ago is 0, the health score change rate is 1.

6. The method for controlling the cycle time of carton processing based on process state perception according to claim 1, characterized in that, The specific method for calculating the balance index of equipment in each process is as follows: First, subtract the outlet accumulation from the inlet accumulation of each process equipment to obtain the net accumulation. Divide the net accumulation by the design capacity of the process inlet buffer to obtain the accumulation rate. The preset positive warning threshold is 0.7 and the negative warning threshold is -0.

3. If the accumulation rate is between -0.3 and 0.7, multiply the absolute value of the accumulation rate by 2 to obtain the accumulation penalty value. Subtract the accumulation penalty value from 1 to obtain the balance index. If the accumulation rate is greater than or equal to 0.7, the balance index is equal to 0.7 minus the accumulation rate. If the accumulation rate is less than or equal to -0.3, the balance index is equal to -0.3 minus the accumulation rate.

7. The method for controlling the cycle time of carton processing based on process state perception according to claim 1, characterized in that, The specific method for obtaining the comprehensive capacity coefficient of each process's equipment based on the overall health of each process's equipment, the upstream transmission coefficient, and the downstream transmission coefficient is as follows: The upstream transmission coefficient is obtained by multiplying the overall health of the upstream equipment corresponding to each process equipment by the balance index. For the first process equipment, since there is no upstream equipment, its upstream transmission coefficient is its own overall health. The downstream transmission coefficient is the balance index corresponding to each process equipment. For the last process equipment, since there is no downstream equipment, its downstream transmission coefficient is 1. The overall health of the process equipment is multiplied by 0.5 to obtain the health contribution value of this process. The upstream transmission coefficient is multiplied by 0.3 to obtain the upstream transmission contribution value. The downstream transmission coefficient is multiplied by 0.2 to obtain the downstream transmission contribution value. The overall capacity coefficient is obtained by adding the health contribution value of this process, the upstream transmission contribution value, and the downstream transmission contribution value.

8. The method for controlling the cycle time of carton processing based on process state perception according to claim 1, characterized in that, The specific method for calculating the bottleneck level index is as follows: Sort all process equipment according to their comprehensive capacity coefficient from smallest to largest. Mark the process equipment with the smallest comprehensive capacity coefficient as the current bottleneck process. If there are multiple process equipment with the same comprehensive capacity coefficient and all of them are the smallest, mark the process that is first in the production flow as the current bottleneck process. Add up the comprehensive capacity coefficients of all processes and divide by the number of processes to get the average capacity coefficient. Subtract the smallest capacity coefficient from the average capacity coefficient to get the capacity gap value. Divide the capacity gap value by the average capacity coefficient to get the bottleneck degree index.

9. The method for controlling the cycle time of carton processing based on process state perception according to claim 7, characterized in that, The specific method for calculating and predicting health and capacity coefficients based on the health change rate of equipment in each process is as follows: Multiply the health change rate of each process equipment by the prediction duration coefficient 3 to obtain the health change trend value. Add 1 to the health change trend value to obtain the prediction adjustment coefficient. Multiply the current comprehensive health of the process equipment by the prediction adjustment coefficient to obtain the initial predicted health. If the initial predicted health is less than 0.1, the predicted health is 0.1; if the initial predicted health is greater than 1, the predicted health is 1; otherwise, the predicted health equals the initial predicted health. Substitute the predicted health of each process equipment into the calculation process of the comprehensive capacity coefficient. Replace the comprehensive health used in the calculation of the health contribution value of this process with the predicted health. Replace the comprehensive health of the upstream process in the calculation of the upstream transmission coefficient with the predicted health of the upstream process. Keep other calculation methods unchanged to obtain the predicted capacity coefficient. The bottleneck transfer early warning information includes the equipment number of the early warning process and the expected time to become a bottleneck. The expected time to become a bottleneck is the current time plus 90 seconds.

10. The method for controlling the cycle time of carton processing based on process state perception according to claim 1, characterized in that, The specific method for obtaining the final target cycle time of each process equipment is as follows: First, obtain the standard cycle time of each process equipment. Multiply the standard cycle time by 0.6 to obtain the minimum allowable cycle time, multiply the standard cycle time by 1.5 to obtain the maximum allowable cycle time, and multiply the standard cycle time by 0.1 to obtain the cycle time adjustment step. Multiply the bottleneck degree index by 0.5 to obtain the bottleneck impact value. Subtract the bottleneck impact value from 0.1 to obtain the adjustment factor. Add 1 to the adjustment factor to obtain the capacity adjustment coefficient. Multiply the comprehensive capacity coefficient of the bottleneck process by the capacity adjustment coefficient to obtain the overall target capacity coefficient. For non-bottleneck processes, divide the standard cycle time by the overall target capacity coefficient to obtain the adjusted cycle time. Divide the adjusted cycle time by the comprehensive health of the equipment in that process to obtain the initial target cycle time. For bottleneck processes, take the larger value between the current actual cycle time and the standard cycle time as the initial target cycle time. If a bottleneck transfer early warning is generated, the current capacity coefficient of the warning process is subtracted from the predicted capacity coefficient to obtain the capacity reduction amount. The capacity reduction amount is divided by the current capacity coefficient to obtain the expected capacity reduction ratio. The expected capacity reduction ratio is multiplied by 0.5 to obtain the response adjustment amount. 1 is added to the response adjustment amount to obtain the early warning response coefficient. The initial target cycle time of the equipment in the warning process is multiplied by the early warning response coefficient to obtain the adjusted target cycle time. The collaborative constraint verification method is as follows: For two adjacent equipment processes, the target cycle time of the downstream equipment is subtracted from the target cycle time of the upstream equipment to obtain the cycle time difference. If the cycle time difference is negative and its absolute value is greater than 20% of the standard cycle time of the upstream equipment, then the target cycle time of the downstream equipment is adjusted to that of the upstream equipment. The target beat is multiplied by 0.8 to obtain the value; the smoothing and limiting processing method is as follows: obtain the final target beat of the previous control cycle of each process equipment as the historical beat, subtract the historical beat from the current target beat to obtain the beat change amount. If the absolute value of the beat change amount is greater than the beat adjustment step size, then when the beat change amount is positive, add the historical beat to the beat adjustment step size to obtain the smoothed target beat; when the beat change amount is negative, subtract the beat adjustment step size from the historical beat to obtain the smoothed target beat; the boundary protection verification method is as follows: compare the target beat with the minimum allowable beat and the maximum allowable beat. If it is less than the minimum allowable beat, take the minimum allowable beat as the final target beat; if it is greater than the maximum allowable beat, take the maximum allowable beat as the final target beat.

Images