Method and system for controlling energy consumption data of an automotive lock catch production line

Abstract

The application discloses a kind of energy consumption data control method and system of automobile lock catch production line, and the present application relates to the technical field of energy consumption data, the lock catch production yield data of current production cycle is collected, and the yield energy consumption benchmark value of each lock catch production process is determined in combination with lock catch product type, and process level energy consumption data set is normalized to the same yield dimension with yield energy consumption benchmark value, and then the normalized energy consumption deviation data set is determined, after the normalization processing of yield dimension, the determination of energy consumption deviation is not interfered by the difference of production scale, improve the accuracy and fairness of energy consumption deviation determination. Energy consumption deviation data set is processed, and after removing the deviation component caused by yield fluctuation, the abnormal category of energy consumption anomaly and its corresponding process position are determined according to the correlation between the time distribution characteristics of residual deviation and the corresponding equipment operating state data, and the corresponding energy consumption control instruction is generated based on the abnormal category and its corresponding process position.

Description

Technical Field

[0001] This invention relates to the technical field of energy consumption data, and in particular to a method and system for controlling energy consumption data in an automotive latch production line. Background Technology

[0002] As a key component of the vehicle body, automotive latches are typically produced using a continuous production line. This line consists of multiple workstations along the production logistics path, including stamping, bending, punching and trimming, surface treatment, assembly and riveting, and online testing. The equipment at each workstation utilizes various energy sources, such as electric drive, hydraulic power, and compressed air drive. The energy consumption of the entire production line is complex, influenced by numerous factors. With the increasing demands for green manufacturing and cost reduction and efficiency improvement, implementing refined energy consumption monitoring and control for automotive latch production lines has become a crucial requirement for manufacturing enterprises.

[0003] In existing technologies, the output of automotive latch production lines fluctuates frequently with orders, order insertions, and batch switching: during low-production cycles, the fixed energy consumption per batch is distributed across fewer workpieces, passively increasing the energy consumption per unit output; during high-production cycles, the opposite is true. Existing technologies lack the ability to normalize the fluctuations in unit energy consumption caused by these differences in production scale, leading to a systematic misjudgment where high-production cycles are mistakenly judged as meeting energy-saving standards and low-production cycles are mistakenly judged as having abnormal energy consumption. This results in distorted energy consumption assessment results, and the accuracy and fairness of deviation judgment cannot be guaranteed. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides a method and system for controlling energy consumption data in an automotive latch production line.

[0005] This invention provides a method for controlling energy consumption data in an automotive latch production line, comprising:

[0006] Mark multiple workstations on the automotive latch production line, obtain energy consumption data for each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process.

[0007] Collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latch based on the product type of the latch. Then, normalize the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, and then determine the normalized energy consumption deviation dataset.

[0008] The energy consumption deviation dataset is processed, and after removing the deviation component caused by production fluctuations, the abnormality category of energy consumption anomaly and its corresponding process location are determined based on the correlation between the time distribution characteristics of the remaining deviation and the corresponding equipment operating status data.

[0009] Based on the anomaly category and its corresponding process location, a corresponding energy consumption control command is generated and sent to the control unit corresponding to the process location.

[0010] This invention provides a control system for energy consumption data of an automotive latch production line. The control system is applied to the aforementioned method for controlling energy consumption data of an automotive latch production line. The control system includes:

[0011] The energy consumption dataset module is used to label multiple workstations on the automotive latch production line, obtain energy consumption data for each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process.

[0012] The deviation module is used to collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latches in combination with the product type of the latches. It also normalizes the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, thereby determining the normalized energy consumption deviation dataset.

[0013] The anomaly module is used to process the energy consumption deviation dataset and, after removing the deviation components caused by production fluctuations, determine the anomaly category of energy consumption anomalies and their corresponding process locations based on the correlation between the time distribution characteristics of the remaining deviations and the corresponding equipment operating status data.

[0014] The instruction module is used to generate corresponding energy consumption control instructions based on the exception type and its corresponding process position, and to send the energy consumption control instructions to the control unit corresponding to the process position.

[0015] Compared with the prior art, the beneficial effects of the present invention are:

[0016] (1) Mark multiple workstations on the automotive latch production line, obtain energy consumption data of each workstation in the current production cycle, and combine the time sequence relationship between latch production processes to trigger time alignment and process classification of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process; collect latch production data of the current production cycle, and determine the production energy consumption benchmark value of each latch production process in combination with the latch product type, and normalize the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, thereby determining the normalized energy consumption deviation dataset. After normalization processing in the production dimension, the determination of energy consumption deviation is not affected by the difference in production scale, avoiding the systematic misjudgment that high production cycles are misjudged as energy-saving and low production cycles are misjudged as energy consumption abnormal, thus improving the accuracy and fairness of energy consumption deviation determination.

[0017] (2) Process the energy consumption deviation dataset, and after removing the deviation component caused by production fluctuation, determine the abnormal category of energy consumption abnormality and its corresponding process position based on the correlation between the time distribution characteristics of the remaining deviation and the corresponding equipment operating status data. Generate the corresponding energy consumption control instruction based on the abnormal category and its corresponding process position, and send the energy consumption control instruction to the control unit corresponding to the process position, so that the energy consumption management of the automotive latch production line is transformed from post-event statistics to real-time detection and active control. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method for controlling energy consumption data of an automotive latch production line according to an embodiment of the present invention.

[0019] Figure 2 This is a simplified schematic diagram illustrating a specific application scenario of the energy consumption data control method for an automotive latch production line according to an embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram of the structure of the control system for energy consumption data of the automotive latch production line in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0022] Please see Figures 1 to 3 A method for controlling energy consumption data in an automotive latch production line, applied to energy consumption data scenarios; the method for controlling energy consumption data in an automotive latch production line includes:

[0023] Step S11: Mark multiple workstations on the automotive latch production line, obtain energy consumption data for each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process.

[0024] Step S12: Collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latch in combination with the product type of the latch. Then, normalize the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, and then determine the normalized energy consumption deviation dataset.

[0025] Step S13: Process the energy consumption deviation dataset, and after removing the deviation component caused by production fluctuations, determine the abnormality category of energy consumption anomaly and its corresponding process location based on the correlation between the time distribution characteristics of the remaining deviation and the corresponding equipment operating status data.

[0026] Step S14: Generate corresponding energy consumption control instructions based on the anomaly category and its corresponding process position, and send the energy consumption control instructions to the control unit corresponding to the process position.

[0027] In step S11, the specific steps are as follows:

[0028] S111: Mark the stamping station, bending and forming station, punching and trimming station, surface treatment station, assembly and riveting station and online inspection station in sequence along the production logistics path of the latch. Configure an independent energy consumption acquisition terminal for each station to collect the corresponding energy consumption data synchronously. The energy consumption data includes current, active power, compressed air flow and hydraulic pressure signal.

[0029] S112: Construct process cycle time sequence content based on the standard cycle time of the locking product. The process cycle time sequence content records the process lag time, workpiece transfer time, and work-in-process buffer time between adjacent processes. The time axis of the energy consumption flow data is phase-shifted according to the process lag time and workpiece transfer time, so that the energy consumption data generated by the same locking workpiece at different workstations is aligned to the virtual processing time window of the locking workpiece, so as to output the time-aligned energy consumption data. The time-aligned energy consumption data is assigned to the corresponding locking production process. The energy consumption data during idle waiting for materials, mold changing, and standby periods are assigned to the auxiliary energy consumption category and aggregated by process to form a process-level energy consumption dataset.

[0030] In step S111, the specific steps are as follows:

[0031] At this point, along the production logistics path of the automotive latch production line, based on the actual flow sequence of the workpieces, the stamping station, bending and forming station, punching and trimming station, surface treatment station, assembly and riveting station, and online inspection station are physically marked and logically coded in sequence. Each station is assigned a unique station identification code, which is simultaneously mapped to the latch production process carried by that station, forming a station-process mapping table. This mapping table is stored in the configuration database of the data acquisition system as the retrieval basis for subsequent energy consumption data attribution.

[0032] During the marking process, the station boundaries also need to be defined: the physical boundaries are the workpiece arrival sensor at the station entrance and the workpiece departure sensor at the station exit, and the time interval between the two trigger signals is the effective working range of the station; for continuous through-type stations such as surface treatment stations, the boundaries are the inlet grating and outlet grating of the treatment tank; thus, when a car latch workpiece enters and leaves a station, the system can accurately determine which process the energy consumption generated by the station during that time period should be attributed to based on the trigger records of the boundary sensors.

[0033] After the workstation marking is completed, an independent energy consumption acquisition terminal is configured for each workstation. The meaning of independent metering is that the energy consumption flow data of each workstation is measured separately by the acquisition terminal dedicated to this workstation, and does not share the metering loop with other workstations, thereby ensuring the separability of process-level energy consumption data at the physical level.

[0034] The specific components of the data acquisition terminal are as follows:

[0035] A three-phase current transformer and a multi-functional power monitoring instrument are connected in series in the power distribution circuit of each workstation to collect current signals and active power signals.

[0036] A temperature and pressure compensated mass flow meter is installed on the compressed air branch line of this workstation to collect compressed air flow signals;

[0037] A pressure transmitter is installed on the hydraulic station oil supply line at this workstation to collect hydraulic pressure signals. The sampling frequency of the above four types of signals is not less than once per second. Among them, the current and active power signals are used to characterize electrical energy consumption, the compressed air flow signal is used to calculate pneumatic energy consumption, and the hydraulic pressure signal is used to characterize the working state of the hydraulic system.

[0038] Each data acquisition terminal is connected to the workstation-level edge gateway via industrial Ethernet. The gateway performs unified clock synchronization and adopts a precision time protocol to ensure that the energy consumption data of all workstations are under the same time reference, with synchronization errors controlled within milliseconds, thus providing a prerequisite for subsequent time alignment.

[0039] After the data acquisition terminal is in place, the system enters the synchronous acquisition phase. The edge gateways of each workstation read the current, active power, compressed air flow and hydraulic pressure signals of their respective workstations in parallel at a unified sampling rate. Each data point is appended with a timestamp, workstation identification code and signal type identifier, and encapsulated into a standardized energy consumption stream data frame, which is then uploaded to the data processing layer in real time via the message middleware.

[0040] During the acquisition process, the gateway performs preprocessing on the raw signals: low-pass filtering is applied to the current and power signals to suppress harmonic glitches caused by the frequency converter; the compressed air flow signal is recorded in dual channels for instantaneous flow and cumulative flow; the hydraulic pressure signal retains the peak hold value to capture the power impact during the pressing moment; when a communication interruption occurs, the gateway caches the data locally and retransmits it according to the timestamp after the link is restored, ensuring the continuity and integrity of the energy consumption flow data.

[0041] In step S112, the specific steps are as follows:

[0042] At this point, after obtaining the synchronous energy consumption flow data output by S111, the process cycle time sequence is constructed based on the standard cycle time of the automotive latch product. The process cycle time sequence takes the standard cycle time as the base period and unfolds the operation intervals of the six processes of stamping, bending, punching and trimming, surface treatment, assembly and riveting, and online inspection on the time axis according to the production logistics sequence, forming a time sequence table describing the start and end times of each process and their interrelationships.

[0043] This timeline focuses on recording three types of time elements between adjacent processes:

[0044] Process lag time refers to the waiting time that must be retained between the completion of the preceding process and the start of the following process due to process curing requirements, such as the surface drying time of the coating after surface treatment, or the interval time for the workpiece to be demolded and removed after stamping.

[0045] Workpiece transfer time refers to the time required for a workpiece to travel from the current station exit to the next station entrance via a "transfer mechanism: conveyor belt, robot, or gantry".

[0046] Work-in-process buffer time refers to the length of time a workpiece stays in the buffer channel or silo between processes. This length of time changes dynamically with the amount of buffer stock. Among the three types of time elements mentioned above, process lag time and workpiece transfer time are determined by the process specifications and the speed of the transfer mechanism. They are relatively fixed calibration parameters, which are obtained by on-site stopwatch calibration or PLC signal time difference statistics and pre-loaded into the configuration database. Work-in-process buffer time, on the other hand, is a dynamic quantity, which is measured in real time by the trigger signals of the buffer inlet and outlet sensors.

[0047] After the cycle timing content is constructed, phase shift is performed on the energy consumption flow data time axis of each station. The purpose of phase shift is to eliminate the time misalignment caused by transmission and lag when the same workpiece is processed at different stations, so that the energy consumption data scattered at each station and at different times can be re-aligned on the time axis to the unified time coordinate of the workpiece itself.

[0048] For the energy consumption data of a certain workstation, its time axis is shifted forward by the cumulative offset of that workstation relative to the first process. This cumulative offset is equal to the sum of the process lag time and workpiece transfer time experienced by the workpiece as it moves from the stamping station to that workstation. The work-in-process buffer time is not included in the shift, as it does not affect the workpiece processing energy consumption itself, but only the timing of energy consumption. Therefore, the triggering time is corrected by the buffer entry and exit signal. The shifted data falls into a virtual processing time window starting from the stamping start time of the workpiece. The length of this virtual processing time window covers the processing range of the workpiece in all six workstations. Thus, although the actual occurrence times of the energy consumption data of the same automotive latch workpiece in the six workstations are staggered, they all converge in the same virtual processing time window after phase shifting, achieving energy consumption data alignment at the workpiece level.

[0049] Optionally, to ensure that the energy consumption data of the same locking workpiece generated at different workstations can be collected and compared under a unified time coordinate, this embodiment constructs a virtual processing time window covering the entire processing process of each workpiece. The time window takes the actual start time of the first process of the workpiece at the stamping station as the starting reference time. The starting reference time is determined as follows: when the position sensor at the entrance of the stamping station detects that the workpiece blank is sent in and triggers a signal, the system records the trigger time as the starting time of the virtual processing time window of the workpiece.

[0050] Based on the pre-built process cycle time sequence, the system sequentially determines the virtual processing time after alignment when the workpiece is transferred to each subsequent station. For bending, punching and trimming, assembly and riveting, and online inspection processes, the virtual processing time after alignment is determined as follows: taking the start time of the virtual processing time window as the reference, the system accumulates all process lag times and all workpiece transfer times involved in the workpiece moving from the stamping station through the intermediate transfer mechanisms to the target station. The resulting time is used as the alignment reference time for the capacity consumption data of the target station within the virtual processing time window.

[0051] In the above calculations, process lag time refers to the waiting time that must be retained due to process solidification requirements from the completion of the preceding process to the start of the following process. Workpiece transfer time refers to the time required for a workpiece to move from the current station exit to the next station entrance via a conveyor belt, robot, or gantry. Both of these time parameters are fixed values ​​that are pre-calibrated and stored in the configuration database.

[0052] For surface treatment processes, the method for determining the alignment reference time is the same as that for the aforementioned processes, but the actual transit time of the workpiece in the treatment tank is additionally added. The work-in-process buffer time is not included in the above cumulative calculation; its function is only to correct the actual occurrence time of energy consumption data: the system monitors the trigger signal of the arrival sensor at the station entrance in real time. If the difference between the actual timestamp of a certain energy consumption data point and the theoretical alignment reference time of the station is within plus or minus one cycle, then the data point is identified as belonging to the virtual processing time window of the current workpiece. If the actual arrival time is delayed by more than one cycle due to queuing in the buffer, the system uses the actual trigger time of the arrival sensor at the station entrance as the new alignment reference, and from this starting point, realigns all subsequent energy consumption data belonging to the workpiece to the actual trigger time, thereby dynamically adapting to the time delay fluctuations caused by changes in buffer stock.

[0053] When a workpiece passes through all processes sequentially and is ultimately determined to be qualified or unqualified at the online inspection station and leaves the station's exit sensor, the system records the trigger time of the exit sensor as the termination time of the virtual processing time window. In this way, each workpiece has a virtual processing time window with a clearly defined alignment reference time for each process, from the start of the stamping inlet trigger to the end of the online inspection outlet trigger. When assigning data, the system assigns energy consumption data generated at each station to the corresponding alignment period within the workpiece's virtual processing time window, based on its actual timestamp. This ensures clear time alignment rules and well-defined boundaries for energy consumption data, and allows for adaptation to actual time delays caused by fluctuations in buffer stock.

[0054] After time alignment is completed, the energy consumption data is assigned to a specific location. The assignment is determined based on the correspondence between the station boundary sensor trigger records and the virtual processing time window. Energy consumption data that falls within the virtual processing time window of a workpiece and originates from the station where the workpiece is located is assigned to the locking production process corresponding to that station and is marked with the process identifier of that process and the traceability number of the workpiece.

[0055] Meanwhile, energy consumption data that does not fall within any workpiece processing range is categorized into auxiliary energy consumption; the auxiliary energy consumption category specifically covers three time periods:

[0056] The first period is the idle waiting period, which is the period when the workstation is in a state of no workpiece due to insufficient upstream material supply. It is determined based on the fact that the workstation entrance arrival sensor has not been triggered within a set threshold time.

[0057] The second period is the mold change period, which is the period when the production line switches product models and the mold is changed and the equipment is adjusted. It is determined according to the start and end time of the mold change work order issued by MES.

[0058] The third period is the standby period, which is the period of downtime maintenance outside of the production plan. It is identified based on the "status signal: running, standby, and stopped" of the equipment master control. The energy consumption of the above three periods is separated from the process energy consumption and classified separately into the auxiliary energy consumption category, so as to avoid the interference of non-process energy consumption on the process energy consumption level.

[0059] After the attribution is completed, the time-aligned energy consumption data is aggregated according to the process. The aggregation method is as follows: using the process identifier as the grouping key, all energy consumption data belonging to the process in the current production cycle are accumulated separately according to the signal type. That is, the active power signal in each virtual processing time window of the process is integrated over time to obtain the process's electrical energy consumption, the compressed air flow signal is accumulated to obtain the process's air consumption, the peak value and mean value of the hydraulic pressure signal are statistically analyzed to characterize the process's work intensity characteristics, and the current signal retains its statistical characteristics for waveform comparison for anomaly diagnosis. At the same time, the data of auxiliary energy consumption categories are also accumulated separately in the above manner to form an auxiliary energy consumption dataset.

[0060] Thus, a process-level energy consumption dataset is finally formed, corresponding one-to-one with the six production processes of the car latch. Each dataset contains the amount of electrical energy consumed, the amount of compressed air consumed, the hydraulic work characteristics, and the current waveform characteristics, and is associated with the workpiece traceability sequence and auxiliary energy consumption dataset within that cycle.

[0061] In step S12, the specific steps are as follows:

[0062] S121: Based on the process-level energy consumption dataset, collect the latch production data of the current production cycle, combine the latch product type with the historical production cycles retrieved from the historical energy consumption database, extract the unit output energy consumption distribution of the corresponding process, and determine the output energy consumption benchmark value of each process after removing abnormal samples.

[0063] S122: Obtain the cumulative energy consumption value in the process-level energy consumption dataset, and determine the actual unit output energy consumption in combination with the actual output of the current cycle. Express the output energy consumption benchmark value in the same dimension, and introduce a batch size correction factor for mold-dependent processes, and introduce a fixed energy consumption amortization correction factor for bath heating and heat preservation for time-dependent processes. Based on the actual unit output energy consumption and the corresponding output energy consumption benchmark value, determine the corresponding energy consumption deviation dataset.

[0064] In step S121, the specific steps are as follows:

[0065] At this point, based on the process-level energy consumption dataset, the system enters the production data acquisition stage. The production data acquisition primarily uses the pass / fail judgment signals from the online inspection station (ST06) as the data source: when a car latch workpiece passes through the online inspection station and is determined to be a qualified product by visual inspection and dimensional measurement, the inspection system outputs a pass / fail signal carrying the workpiece's traceability number. All pass / fail signals within the "current production cycle: for example, a shift or a batch" are counted to obtain the qualified production output of car latches for the current production cycle. Simultaneously, the system synchronously obtains the input quantity, defective product quantity, and production records by model from the MES system within this cycle for verification and correction of production data, preventing misuse of production data when there is a discrepancy between the inspection count and the planned production output.

[0066] After the production data is collected, it is cross-checked with the workpiece traceability number sequence in the process-level energy consumption dataset: check whether the workpiece sets covered by the two are consistent. If there are missing traceability numbers due to data loss or missed detection, the smaller set is used to converge the production caliber to ensure that the production and energy consumption data cover the same batch of workpieces and avoid the benchmark deviation caused by the difference between the numerator and denominator.

[0067] After obtaining the production data for the current cycle, and combining the product type of the latch, historical production cycles that can be used as a benchmark are retrieved from the historical energy consumption database. The historical energy consumption database pre-stores process-level energy consumption datasets for each previous production cycle, as well as corresponding operating condition tags such as output, product model, shift time, ambient temperature and humidity, and number of times the mold is used.

[0068] During the search, the product model is used as the first search condition, and only historical cycles of automobile latch production are selected. On this basis, further screening conditions of similar working conditions are applied: First, the production scale is similar to the current cycle, for example, the deviation does not exceed ±20%, so as to exclude the difference in unit energy consumption caused by large differences in production volume; Second, the equipment condition is similar, for example, the number of times the mold is used is within a similar lifespan range and there are no major maintenance records; Third, the season and temperature and humidity conditions are similar, so as to exclude the influence of environmental factors on the energy consumption of surface treatment drying and other processes.

[0069] The search results are sorted in reverse chronological order, prioritizing several historical periods that are closer to the current time, such as the last thirty periods that meet the conditions, as the candidate sample set. If the number of samples that meet all the conditions is less than the set lower limit, such as less than ten, then the secondary screening conditions, such as temperature and humidity conditions, are relaxed in order of priority from high to low according to the similarity of working conditions, until the number of samples reaches the lower limit, so as to ensure the stability of the baseline statistics.

[0070] After obtaining candidate historical cycle samples, the energy consumption distribution per unit output is extracted for each process. For each historical cycle sample, the centralized electricity consumption and compressed air consumption of the process within that cycle are converted into total energy consumption, and then divided by the automobile latch production of that cycle to obtain the energy consumption per unit output of that process in that cycle, including electricity consumption and air consumption per unit output. This calculation is performed on all candidate samples to form the energy consumption per unit output sample sequence for that process.

[0071] Anomaly removal is performed on the sample sequence. The removal uses a statistical method: the median and absolute median difference of the sample sequence are calculated. The median is used as the positional benchmark, and a "set multiple: for example, three times" of the absolute median difference is used as the dispersion tolerance. Any sample exceeding this tolerance range is judged as an anomaly and removed. The median and absolute median difference are chosen instead of the mean and standard deviation to reduce the contamination of position and dispersion estimation by the anomaly samples themselves. The removal process can be performed iteratively: after one round of removal, the median and absolute median difference of the remaining samples are recalculated, and it is checked again whether there are still samples exceeding the limit, until all remaining samples are within the tolerance range. The causes of the removed sample corresponding to the period are archived separately. For example, if the period has experienced material jamming, machine stoppage, frequent mold changes, or measurement abnormalities, it will be used as a priori cases for subsequent S13 anomaly diagnosis.

[0072] After the abnormal samples are removed, the energy consumption benchmark value for each process is determined based on the energy consumption distribution per unit output of the remaining samples. The energy consumption benchmark value is determined as follows: the median of the remaining sample distribution is taken as the benchmark value. The median is not sensitive to the distribution shape and can robustly represent the energy consumption level per unit output of the process under normal operating conditions. At the same time, the quartile interval of the distribution is recorded as the tolerance band of the benchmark for subsequent steps to determine whether the deviation is significant. The benchmark value of each process includes two components: the energy consumption benchmark value per unit output and the gas consumption benchmark value per unit output. The hydraulic work characteristics and current waveform characteristics are archived as accompanying characteristics of the benchmark and are used for comparison in the abnormal diagnosis stage without participating in the benchmark value itself.

[0073] Optionally, to ensure the comparability and repeatability of the benchmark values ​​for production energy consumption across different production cycles, this embodiment adopts the following standardized process for determining the benchmark values: The system uses the model of the latch product produced in the current production cycle as the first search condition, extracts all historical production cycle records of the same model from the historical energy consumption database within the past ninety days, and further filters out records where the actual output deviates from the planned output of the current cycle by no more than ±20%, and the number of times the mold is used is within the middle 60% range of the mold's entire life cycle as a candidate sample set; if the number of candidate samples is less than twenty, the output deviation threshold is successively relaxed to 30%, 40%, until the number of samples reaches twenty.

[0074] For each process in each candidate sample, calculate its unit output power consumption (in kilowatt-hours per piece) and unit output gas consumption (in standard cubic meters per piece). Sort the unit output energy consumption values ​​of all candidate samples under the same process in ascending order of numerical value, and take the value in the middle position after sorting as the output energy consumption benchmark value of the process. At the same time, the numerical range between 25% and 75% is taken as the normal tolerance zone of the benchmark value.

[0075] During the calculation process, if the energy consumption per unit output of a sample exceeds the absolute median difference of the median of all samples by more than or equal to three times, the sample is considered an outlier and is removed. The sorting and median calculation process is repeated until there are no more records exceeding this range in the remaining samples. Finally, the median obtained from the last calculation is used as the final benchmark value for this process. This benchmark value and its tolerance band are stored in the benchmark library for subsequent comparison of energy consumption deviations of the same model of product, thereby ensuring that the rules for determining the benchmark value are clear, reproducible, and do not produce substantial differences due to different operators.

[0076] Therefore, for the six processes of automotive latches—stamping, bending, punching and trimming, surface treatment, assembly and riveting, and online inspection—each has its own production energy consumption benchmark value. This benchmark value, along with the tolerance zone, sample size, and operating condition label, is stored in the benchmark library as a comparison benchmark in S122 to normalize the process-level energy consumption dataset and the benchmark value to the same production dimension.

[0077] In step S122, the specific steps are as follows:

[0078] At this point, after determining the production energy consumption baseline value for each process in S121, the system obtains the cumulative energy consumption value in the process-level energy consumption dataset and, in conjunction with the actual output locked in the current cycle, determines the actual unit output energy consumption. The cumulative energy consumption value in the process-level energy consumption dataset includes two parts: "electrical energy consumption: obtained by time integration of the active power signal in each virtual processing time window after S112 assignment" and "compressed air equivalent energy consumption: converted from the cumulative air consumption to the equivalent energy consumption using the compressed air unit conversion factor, or retained separately as a volume quantity."

[0079] When determining the actual unit output energy consumption, the cumulative energy consumption value of a certain process in the current cycle is divided by the actual output of the car latches in that cycle to obtain the actual unit output electricity consumption and actual unit output gas consumption of that process; the output here adopts the qualified output after calibration in S121 to ensure that the energy consumption and output cover the same batch of workpieces.

[0080] The actual unit output energy consumption and the benchmark value of output energy consumption must be under the same dimension to be comparable. Among them, the benchmark value determined by S121 is derived from the statistics of historical period samples, and its dimension is unit output electricity consumption and unit output gas consumption. If the benchmark value is expressed as electricity consumption converted to gas or as a comprehensive energy equivalent such as standard coal equivalent, then both sides need to be uniformly converted to the same dimension system before comparison. This scheme adopts a classification and separate processing: the electricity consumption component and the gas consumption component maintain independent dimensions and are compared separately to avoid the caliber drift introduced by the difference in the selection of different energy equivalent conversion factors. Only when conducting comprehensive comparison between processes, the comprehensive energy consumption index is synthesized according to the preset conversion factor.

[0081] Considering that the energy consumption of the two types of processes has different dependence characteristics on output and time, directly comparing the actual value with the benchmark value will result in a systematic mismatch. Therefore, correction factors are introduced for the two types of processes to correct the benchmark value.

[0082] A batch size correction factor is introduced for mold-dependent processes. Mold-dependent processes refer to processes whose energy consumption structure contains a significant component that varies with mold status and the number of batch start-ups and shutdowns. Typical examples include stamping, bending, and punching / trimming processes. The energy consumed by mold preheating, trial stamping, and first-piece verification at the start of each batch is a fixed input for the batch. The "share" allocated to a single piece varies with the batch size: i.e., the number of consecutive production pieces in the cycle. The smaller the batch size, the greater the fixed energy consumption allocated to a single piece. The batch size correction factor is determined based on the ratio of the actual batch size of each mold-dependent process in the current cycle to the average batch size of historical samples when the baseline value is statistically analyzed. This is used to correct the fixed energy consumption allocation assumption implicit in the baseline value, so that small batch cycles are not misjudged as having high energy consumption when compared with the baseline value.

[0083] For time-dependent processes, a fixed energy consumption amortization correction factor is introduced for bath heating and heat preservation. Time-dependent processes refer to processes where energy consumption includes a fixed component that is unrelated to output and accumulates only over time, such as surface treatment processes: when the treatment bath is started, it needs to be heated from ambient temperature to process temperature, and the energy consumption for heating is unrelated to the output of that cycle; the energy consumption for maintaining the bath temperature during the heat preservation stage also accumulates mainly over time rather than output. This fixed energy consumption component is amortized to the unit along with the total energy consumption of the cycle, and its unit amortization amount is inversely proportional to the output. The amortization correction factor is determined based on the relationship between the current cycle length and the average cycle length of historical benchmark samples, as well as the relationship between the current cycle output and the average output of historical samples, and is used to correct the benchmark value to the output amortization caliber of the current cycle.

[0084] Optionally, for the three mold-dependent processes of stamping, bending, and punching / trimming, the batch size correction factor is determined as follows: the actual number of consecutive production batches of the process and the actual number of pieces produced in each batch are read from the manufacturing execution system in the current cycle, and the average batch size of the cycle is calculated in pieces per batch; at the same time, the arithmetic mean of the average batch size of the process in each historical period is calculated from all historical sample periods used in the historical benchmark statistics, and this is used as the reference batch size.

[0085] The ratio obtained by dividing the reference batch size by the average batch size of the current period is used as the batch size correction factor for this process in the current period. When the ratio is greater than one, it indicates that the average batch size of the current period is less than the historical reference level, and the corrected benchmark value will be increased proportionally to reflect the reasonable expectation that the fixed energy consumption of the batch will be distributed to fewer workpieces, resulting in an increase in the energy consumption per piece. When the ratio is less than one, the corrected benchmark value will be decreased proportionally. When making the correction, the original output energy consumption benchmark value of this process is multiplied by the batch size correction factor to obtain the corrected output energy consumption benchmark value.

[0086] For the time-dependent process of surface treatment, the correction factor for the fixed energy consumption amortization of the bath heating and heat preservation is determined as follows: The actual total working time of the treatment tank in the current cycle is read from the on-site process parameter acquisition system, in hours. At the same time, the arithmetic mean of the total working time of the treatment tank in each historical cycle is calculated from all historical sample cycles used in the statistical analysis of the historical benchmark value of this process, and used as the reference working time.

[0087] The ratio of the actual total working time in the current cycle to the reference working time is used as the time correction factor for the process in the current cycle. Then, the actual output of the process in the current cycle is divided by the average output in the reference cycle, and the ratio is used as the output correction factor. The time correction factor and the output correction factor are divided to obtain the fixed energy consumption amortization correction factor for the process. When correcting, the original output energy consumption benchmark value of the process is multiplied by the fixed energy consumption amortization correction factor to obtain the corrected output energy consumption benchmark value.

[0088] Both types of correction factors are calculated independently for each process step. Correction calculations are only initiated when the average batch size of the current cycle changes by more than ±5% relative to the reference batch size, or when the processing tank working time of the current cycle changes by more than ±5% relative to the reference working time. If the change does not exceed these ranges, the correction factor is set to one, meaning no correction is made to the baseline value. After correction, the system compares the actual unit output energy consumption with the corrected output energy consumption baseline value process by process and energy component by component. The difference between the two values ​​determines the energy consumption deviation dataset, ensuring that the deviation results reflect real energy consumption anomalies caused by factors other than output, batch size, and time duration.

[0089] After the correction is completed, the actual unit output energy consumption of each process is compared with the corrected output energy consumption benchmark value to determine the corresponding energy consumption deviation dataset. Among them, for each process and each energy component, electricity consumption and gas consumption, the percentage deviation of the actual value from the corrected benchmark value is used as the deviation amount to form a deviation dataset organized by process and energy component. A positive deviation value indicates that the actual energy consumption is higher than the benchmark, and a negative value indicates that it is lower than the benchmark. The deviation dataset also includes the "hydraulic work characteristics: peak sequence" and current waveform characteristic index of each process, as well as the benchmark tolerance band determined by S121, for reference in S13 when distinguishing the significance of anomalies and the anomaly category.

[0090] In step S13, the specific steps are as follows:

[0091] S131: Take the fluctuation of actual output relative to planned output as input and obtain the predicted value of the deviation component caused by output fluctuation. Subtract the predicted value of the deviation component from the energy consumption deviation dataset to obtain the corresponding residual deviation sequence. Extract the deviation duration, deviation change time, periodicity and phase relationship with the process operation cycle of the residual deviation sequence, and collect the equipment operation status data of the same time window at the same time.

[0092] S132: Perform time-series correlation analysis between the time distribution characteristics and the equipment operating status data, and determine the anomaly category of energy consumption anomaly. The anomaly category of energy consumption anomaly includes equipment idling standby anomaly, equipment performance degradation anomaly, process parameter deviation anomaly, and energy leakage anomaly; at the same time, mark the remaining deviation sequence of the anomaly, and determine the corresponding process position based on the process-level data to which the remaining deviation sequence belongs.

[0093] In step S131, the specific steps are as follows:

[0094] At this point, based on the energy consumption deviation dataset output by S122, the system determines the output fluctuation amount. Specifically, the planned output of the current production cycle is obtained from the MES system and compared with the actual qualified output after verification by S121. The difference between the two constitutes the fluctuation amount of the actual output relative to the planned output. A positive value of the fluctuation amount indicates underproduction, and a negative value indicates overproduction. This fluctuation amount is used as input to the pre-constructed output fluctuation-energy consumption deviation mapping model.

[0095] The mapping model is constructed as follows: production fluctuations affect process energy consumption deviations through two pathways. For the batch start-stop pathway: underproduction means that the planned processing range within the cycle is not operating at full capacity, and the fixed energy consumption of batches in mold-dependent processes is distributed to fewer workpieces, passively increasing the energy consumption per unit output. For the idle operation pathway: when production is lower than expected, downstream processes experience material waiting, and the maintenance power consumption of equipment in low-load or idle states still occurs, while the denominator production decreases. Based on the above two pathways, the mapping model combines the fixed energy consumption component ratio of each process registered in the S121 benchmark library, such as the fixed batch input ratio of the stamping process and the fixed heat preservation component ratio of the surface treatment process, to output the predicted value of the deviation component caused by production fluctuations in each process, that is: the part of the deviation of each process that is expected to exist reasonably and is not abnormal under the current fluctuation amount.

[0096] After obtaining the predicted values ​​of the deviation components of each process, they are deducted from the energy consumption deviation dataset. Specifically, for each process and each energy component, the deviation value obtained in S122 for that process is subtracted from the predicted value of the corresponding deviation component, and the difference constitutes the residual deviation of that process. The residual deviation represents the net abnormality deviation of the process energy consumption from the baseline after deducting the reasonable factor of output fluctuation. The residual deviations of each process in the current cycle are organized according to the process-energy component to obtain the residual deviation sequence.

[0097] During the deduction process, the system retains the original deviation value and the predicted component value in the same data structure to make the remaining deviation traceable. If the remaining deviation of a certain process falls within the reference tolerance band determined by S121, the process is marked as normal and will not proceed to the subsequent feature extraction. Only processes with remaining deviations exceeding the tolerance band will proceed to the next stage of time distribution feature extraction.

[0098] For the remaining deviations, the system further extracts their distribution characteristics in the time dimension; in this process, the current production cycle is divided into several time segments of equal length, for example, a group of several standard beats, and the values ​​of the remaining deviations of each process in each time segment are counted to form a deviation curve that evolves over time; four types of features are extracted from the deviation curve.

[0099] The first type of characteristic is the duration of deviation, which refers to the length of time during which the remaining deviation continuously exceeds the tolerance zone: instantaneous deviations only last for a few time segments, drifting deviations persist from a certain moment until the end of the cycle, and intermittent deviations appear and disappear repeatedly.

[0100] The second type of feature is the moment of deviation change, which refers to the time point when the deviation curve changes significantly: the system detects the position of the slope change of the deviation curve and compares this moment with the "production event log: mold change, batch switching, parameter adjustment record" in the same time window to determine whether the change coincides with the time of human operation or batch event.

[0101] The third type of characteristic is the periodic characteristic, which refers to the regularity of the deviation curve repeating over time: whether the system detects whether the deviation has periodic fluctuations that are the same frequency or multiples of the production cycle, such as the deviation occurring regularly once in each cycle, or occurring regularly every few cycles.

[0102] The fourth type of feature is the phase relationship with the process cycle, which refers to the degree of phase alignment between the peak time of the periodic deviation and the "standard process cycle, such as the downward-holding-returning process of the stamping slide". If the peak deviation stably appears in a specific phase segment within the cycle, such as the holding segment or the unloading segment, then this phase relationship becomes an important basis for identifying the physical causes of abnormal energy consumption.

[0103] Synchronous with the extraction of time distribution features, the system collects equipment operating status data within the same time window as the remaining deviation. The operating status data is read from the main control PLC of the equipment via the edge gateway of each workstation. The collected content includes: main drive current curve, hydraulic system pressure curve and oil temperature, servo load rate, equipment alarm code, air pressure curve, mold usage count, and equipment status signals, including running, standby, and fault. The timestamps of the operating status data and the energy consumption data are under the same clock reference to ensure strict alignment of the time window. The acquisition frequency is consistent with that of the energy consumption data, so that the status features and deviation features can be compared point by point on the time axis.

[0104] In step S132, the specific steps are as follows:

[0105] At this point, after obtaining the four types of time distribution characteristics of the residual deviation sequence and the equipment operating status data of the same time window in S131, the system performs a time-series correlation analysis between the two. The time-series correlation analysis is based on time alignment and uses feature comparison as a means: the residual deviation curve of a certain process is placed under the same time coordinate with the main drive current curve, hydraulic pressure and oil temperature curve, servo load rate curve, air pressure curve and alarm record of that station, and the shape and time correspondence of the two are compared segment by segment.

[0106] The criteria for correlation analysis include the degree of overlap at the start time, consistency of continuous morphology, consistency of periodic phase, and corroboration by accompanying alarms. For the degree of overlap at the start time, it refers to whether the moment of the deviation's sudden change coincides with the start time of the abnormal state data. For the consistency of continuous morphology, it refers to whether the "trend of the deviation drift" evolves synchronously with the trend of state quantities, such as average current and oil temperature. For the consistency of periodic phase, it refers to whether the periodic peak phase of the deviation is locked with the phase of the power-operating phase within the equipment's operating cycle. For corroboration by accompanying alarms, it refers to whether the alarm codes in the state data can provide equipment-side evidence for the cause of the deviation. The system calculates the correlation strength for each of the above four criteria and records the chain of evidence in the form of a time-series correlation graph.

[0107] Based on the time-series correlation analysis results, the system classifies energy consumption anomalies into one of the following four categories, with the following characteristics for each category:

[0108] Equipment idle standby anomalies; their time distribution characteristics are as follows: the deviation occurs during non-processing periods, does not coincide with the virtual processing time window, but coincides with the period when the station boundary sensor is not triggered by a workpiece; the corresponding status data characteristics are that the equipment status signal is in standby while the main drive current or air pressure remains at a high level; the physical causes are the lack of equipment standby strategy and excessive idle maintenance power consumption.

[0109] Optionally, the criteria for determining equipment idling / standby anomalies are as follows: the overlap between the positive period of the remaining deviation and the period without workpieces recorded by the boundary sensor of the workstation reaches more than 80%, meaning the deviation mainly occurs during the idle period when there are no workpieces being processed at the workstation; simultaneously, during this period, the status signal of the equipment's main control is in standby mode, but the main drive current or air pressure value still remains above 40% of the normal processing current value or normal working pressure under no-load operation, and this high state continues for more than three time segments; and the ratio of the positive value of the remaining deviation to the average energy consumption value during the normal processing period of the workstation does not exceed 0.3. When all three conditions are met, the equipment is determined to be idling / standby anomaly.

[0110] Equipment performance degradation anomalies; their time distribution characteristics are: the deviation shows a continuous and slow drift, with no clear abrupt change or the abrupt change does not coincide with the production event; the corresponding status data characteristics are a slow increase in the average value of the main drive current, a synchronous increase in oil temperature or bearing temperature, and no batch event to corroborate it; the physical causes are mechanical wear, lubrication deterioration, and decreased working efficiency caused by hydraulic internal leakage.

[0111] Optionally, the criteria for determining equipment performance degradation anomalies are as follows: the residual deviation shows a unidirectional drift trend over at least twenty consecutive time segments, i.e., the deviation value increases or decreases monotonically; the determination of this drift trend is based on linear fitting of the deviation sequence using the least squares method, with the absolute value of the fitting slope greater than 2% per time segment and the goodness of fit greater than 0.7; there are no abrupt change points during this drift process that coincide with any production event log record; and the average value of the main drive current within the same time window shows a unidirectional trend, i.e., when the residual deviation of power consumption increases, the average value of the main drive current increases synchronously, or when the residual deviation of gas consumption increases, the average value of the gas path pressure increases synchronously, and the Pearson correlation coefficient between the two is greater than 0.36. When all four conditions are met, it is determined to be an equipment performance degradation anomaly.

[0112] Process parameter deviation anomalies; their time distribution characteristics are as follows: the deviation changes in a step manner, and the sudden change time coincides with the mold change, batch switching or parameter adjustment event in time; the corresponding status data characteristics are that the servo load rate or pressure set value jumps at the sudden change time and then remains stable; the physical cause is that the parameter is not reset to the process specification value or the raw material batch difference.

[0113] Optionally, the criteria for determining process parameter deviation anomalies are as follows: the residual deviation curve experiences a step jump within two consecutive time segments, with the jump amplitude exceeding twice the width of the process's baseline tolerance band; the time difference between the jump's occurrence and the mold change start time or batch switching time recorded in the manufacturing execution system is within plus or minus one time segment; the residual deviation after the jump remains stable at a level not exceeding 10% of the jump value for at least five subsequent time segments; and the servo load rate setting or pressure setting value corresponding to the jump time shows a change of at least 5%. If all four conditions are met, the anomaly is determined to be a process parameter deviation anomaly.

[0114] Energy leakage anomalies; their temporal distribution characteristics are: persistently high gas consumption deviations, with periodic or continuous components unrelated to production cycle time; the corresponding status data characteristics are periodic pressure replenishment sawtooth waves in the gas circuit, frequent start-stop compensation of the air compressor, and basic leakage of flow even when there is no downstream operation; the physical causes are gas circuit leakage, loose pipe joints, or sealing failure.

[0115] Optionally, the criteria for determining an energy leakage anomaly are as follows: The residual deviation in gas consumption is positive for more than ten consecutive time segments, and the proportion of this positive value exceeding the upper limit of the baseline tolerance band for the process exceeds 20%; simultaneously, the gas pressure curve within the same time window exhibits a periodic pressure-replenishing sawtooth wave pattern, meaning the pressure value drops below the set lower limit and then rapidly rises back to the upper limit, with an average interval of less than sixty seconds for this pressure-replenishing cycle; and during periods when no workpiece passes through the workstation, the compressed air flow meter still records a continuous base flow rate, which exceeds the upper limit of the normal standby flow rate for the process by more than 10%. If all three conditions are met, the anomaly is determined to be an energy leakage anomaly.

[0116] During the judgment process, the system evaluates the matching degree of the four types of anomalies based on the matching of time characteristics and state characteristics, and takes the category with the highest matching degree as the judgment result; when the matching degrees of multiple categories are close, the second best category is retained as a candidate for subsequent steps to confirm in conjunction with the handling feedback.

[0117] After the anomaly category is determined, the system marks the remaining deviation sequences that have anomalies. The marking content is arranged as follows: the process identifier corresponding to the remaining deviation sequence, the "energy component type: electricity consumption or gas consumption", the anomaly category, the "evidence chain of the judgment basis: time feature summary, state feature summary, correlation strength", deviation amplitude and occurrence time window, together with the workpiece traceability number sequence registered in S112 for the process, are uniformly packaged into an anomaly mark record. This record is also written back to the data entry of the corresponding period in the historical energy consumption database as a priori case for S121 to remove anomaly samples when retrieving historical periods in the future, so as to achieve self-purification of the benchmark library.

[0118] The residual deviation sequence already carries the process identifier during the attribution and classification stage. The mapping relationship between the process identifier and the workstation identifier is uniquely determined by the workstation-process mapping table of S111. Therefore, by looking up the mapping table from the process identifier of the abnormal sequence, the location of the abnormal process and its corresponding physical workstation can be directly determined. For time-dependent processes such as surface treatment, the relative relationship between the abnormal time window and the bath heating and heat preservation period is also recorded in the marking record for subsequent handling.

[0119] In step S14, the specific steps are as follows:

[0120] S141: Load the abnormal category of energy consumption abnormality and its corresponding process position into the same data space, mark the abnormal handling measures corresponding to the energy consumption abnormality in the data space, and generate the corresponding energy consumption control instructions in combination with the current working status of the automotive latch production line.

[0121] S142: Before issuing the energy consumption control command, a conflict check is performed between the command and the production cycle and safety interlock conditions. If there is a timing conflict, the execution time is postponed to the process interval, and then the energy consumption control command is issued to the control unit corresponding to the process position.

[0122] In step S141, the specific steps are as follows:

[0123] At this point, after the S132 outputs the abnormal marker records, the system loads the abnormal categories of each energy consumption abnormality and their corresponding process positions into the same data space. This data space is a structured abnormality handling data space deployed in the production line control layer. Its data structure is organized with abnormal events as the basic unit: each abnormal event unit includes an abnormality category field, process position field, energy component field, deviation magnitude field, evidence chain field, and time window field. During loading, each abnormal marker record output by S132 is written into the data space one by one after field verification, forming a set view of all abnormalities to be handled in the current cycle.

[0124] The data space simultaneously mounts three supporting mapping tables as static configurations: the first is an anomaly category-handling measure mapping table, which registers the candidate handling measure sets corresponding to each of the four anomalies; the second is a process location-control unit routing table, which registers the control unit addresses of each physical workstation and the set of acceptable instruction types; and the third is an instruction priority definition table, which registers the response timeliness level of various instructions. The above mapping tables are preset during system configuration and can be expanded as handling experience accumulates. After loading, the data space has a complete context with the three elements of "anomaly-measure-routing", providing a retrieval basis for subsequent measure marking.

[0125] Within the data space, the system marks the corresponding anomaly handling measures for each abnormal event unit. The marking of these measures is not a simple lookup; instead, it is based on the deviation magnitude and time window information carried by the abnormal event unit, and is tiered within a set of candidate measures. The candidate measure set for each anomaly category is arranged from light to heavy intervention intensity. For example, equipment performance degradation is tiered as "generating inspection work orders, switching to reduced-load operation mode, limiting production, and shutting down for maintenance"; energy leakage is tiered as "online leak detection work orders, isolating gas branch valves, and switching to backup branches." During selection, measures with a deviation magnitude falling within the mild range are chosen, while measures with a larger deviation magnitude or duration spanning multiple time segments are chosen. If alternative anomaly categories exist in the evidence chain, they are retained as remarks in the measure marking, pending confirmation after handling feedback.

[0126] After the marking is completed, each abnormal event unit is upgraded from "abnormal description" to a complete record of "abnormal plus handling plan", and the content of the measures is limited to the acceptable instruction types registered in the routing table of the corresponding process position, ensuring that the control instructions generated subsequently are executable at the execution end.

[0127] After the measures are marked, the system generates energy consumption control instructions based on the current working status of the automotive latch production line. The current working status is read in real time by the production line main controller and MES, including: the running / standby status of each workstation, the inventory in the work-in-process buffer, the remaining planned quantity of the current batch, the remaining time until the end of the current cycle, and whether there is an urgent order insertion mark. The energy consumption control instructions are generated in accordance with three status constraint rules, covering production continuity constraints, cycle timing constraints, and order insertion protection constraints.

[0128] To address production continuity constraints, if the target process is in continuous processing and the work-in-process buffer is below the set lower limit, instructions that would cause the process to stop immediately should be avoided. Instead, a gradual unloading or delay should be implemented until the buffer is restored. If the buffer is sufficient, stronger intervention measures are allowed.

[0129] Regarding the timing constraints of the cycle, highly interventionist "instructions, such as shutdown for maintenance and gas path isolation", should be prioritized for execution at the end of the cycle or between batches, unless the abnormal magnitude has threatened equipment safety; regarding the order insertion protection constraints: if there is an urgent order insertion, the strong intervention instruction will be further delayed and the alarm will be upgraded to notify manual intervention simultaneously.

[0130] The generated energy consumption control instructions are encapsulated in standardized instruction frames. The frame content includes: target process location, target control unit address, instruction type, "instruction parameters, such as load reduction target value, valve closing time period", priority, "effective condition: immediate effective or timed effective", and abnormal event number for traceability. After "multiple instructions in the same data space undergo conflict checking to avoid contradictory instructions targeting the same control unit", the instruction set for the current cycle is formed.

[0131] In step S142, the specific steps are as follows:

[0132] At this point, after S141 generates the energy consumption control instruction set but before issuing it, the system performs conflict verification on each instruction against production cycle time and safety interlock conditions. The verification is performed within the instruction issuing engine, which reads two types of verification benchmarks from the production line main controller in real time: one is the production cycle time benchmark, which is the phase position within the standard cycle time at the current moment, the current stage of the action cycle of each workstation, and the inventory in the work-in-process buffer; the other is the safety interlock benchmark, which is the set of safety interlock conditions for each workstation, including the motion state of the slider or pressure arm, the closing state of the safety light curtain and protective door, the safety pressure range of the hydraulic and pneumatic systems, and the interlock relationship table, which actions are prohibited from being triggered under what conditions.

[0133] Conflict checking is performed on two dimensions. The first dimension is cycle time conflict checking: it determines whether the expected execution action of the instruction will interrupt the ongoing workpiece processing action. If the arrival time of the instruction falls within the virtual processing time window of a certain workpiece, and the instruction action will interrupt the forming, riveting, or inspection action of that station, then a timing conflict is determined to exist. In addition, if the instruction action does not directly interrupt processing, but causes the action response delay of that station in the following several cycles, such as the need for a short pause in equipment adjustment for load reduction switching, causing downstream processes to lose cycle time due to insufficient material supply, a timing conflict is also determined to exist.

[0134] The second dimension is safety interlock conflict verification: determining whether the command action violates safety interlock conditions. For example, commands involving air valve closure or hydraulic pressure reduction need to verify whether the target actuator is in a safe position at the moment of command action. If the slider is in the downward pressurization stroke when the hydraulic pressure adjustment command arrives, the pressure reduction will lead to insufficient forming force, resulting in defective products or even equipment vibration, thus determining a safety conflict. Similarly, for air isolation commands, it needs to verify whether a pneumatic clamping device is suspended on that branch and clamping the workpiece. Forcibly closing the valve will cause clamping failure and the risk of the workpiece falling, also determining a safety conflict. The verification results, along with "judgment basis: conflict type, conflict time, and involved interlock items," are added to the verification field of the command frame to form a verification record for each command.

[0135] For instructions that are determined to have timing or safety conflicts after verification, the system does not cancel the instruction, but postpones its execution time to the process gap period. The process gap period is determined as follows: the system retrieves the gap event sequence within the subsequent predetermined time range based on the process cycle timing content constructed in S112. The gap events include the workstation waiting window, the gap period between the arrival of two adjacent workpieces when the workstation entrance position sensor arrives, the mold change preparation period before batch switching, the unplanned period at the end of the cycle, and the fixture gap period of the surface treatment station.

[0136] In the intermittent event sequence, the system selects a time point that meets the following conditions as the time to postpone the execution of the instruction: First, the time point is compatible with the safe position state of the equipment involved in the instruction, such as the slider being at the top dead center and the fixture being in the released position; Second, the time delay between the time point and the current time does not exceed the maximum tolerable time delay specified by the instruction priority. High-priority instructions select the nearest qualified gap, while regular instructions can select a later gap to reduce the disturbance to the cycle time; Third, the duration of the delayed instruction action does not exceed the upper limit of the gap duration, ensuring that the instruction action is completed before the next workpiece arrives and that the equipment is restored to a ready state.

[0137] After the delayed execution time is determined, the system rewrites the effective condition field of the instruction frame: changing the immediate effective or original timed effective to the delayed effective time, and writing "Delay reason: conflict type and conflict time" into the traceability field; the rewritten instruction is re-executed for a quick review of conflict verification to confirm that the delayed time no longer conflicts with the gap event and safety interlock status; if the production line conditions change during the waiting period, causing the delayed time to become invalid, such as the gap sequence being changed by inserting an order, the gap retrieval and delayed time positioning are re-executed.

[0138] After completing conflict verification and delayed location, the system sends energy consumption control commands to the control units corresponding to the process positions. The sending is based on the process position-control unit routing table mounted in the S141 data space: the communication address of the control unit at the target process position of the command frame is looked up, and the command is delivered via the industrial Ethernet bus using standard industrial protocols, such as OPCUA-based command channels or fieldbus messages. Commands are executed according to priority, with safety-related strong intervention commands occupying the bus first, followed by regular load reduction and energy-saving commands. A confirmation mechanism is used: after receiving the command, the control unit sends back a confirmation frame containing the command number, parsing result, and expected execution time. If no confirmation is received within the set time limit, the sending engine resends the command; if the resend fails, a manual alarm notification is issued.

[0139] When the instruction reaches its effective time, the control unit executes the instruction action and sends back an execution feedback frame after the action is completed. This frame includes the actual execution time, the execution result status, and snapshots of key state quantities before and after execution, such as the actual value of hydraulic pressure after execution and the branch pressure drop curve after valve closure. The feedback frame is written back to the corresponding abnormal event unit in the data space, forming a complete closed-loop record of "abnormality - measure - instruction - execution". This record also serves as a known handling event for the deviation feature comparison of the next production cycle S131, enabling subsequent diagnosis to correctly attribute the energy consumption change after instruction execution from the deviation analysis and avoid confusion between the handling effect and the new abnormality.

[0140] Please see Figure 2 The specific application scenarios and corresponding embodiments of this application are as follows:

[0141] In step S111, for the automotive latch production line, the stamping station is marked as ST01, which corresponds to the stamping process; the bending and forming station is marked as ST02, which corresponds to the bending process; and so on, the online detection station is marked as ST06; when the stamping completion signal of the automotive latch is triggered, all energy consumption flow data recorded within the boundary of ST01 is marked as stamping process data and bound to the "unique traceability number, such as batch number plus serial number" of the workpiece.

[0142] Both the stamping station ST01 and the bending and forming station ST02 are equipped with hydraulic presses, so in addition to the power monitor and air flow meter, each station is also equipped with a hydraulic pressure transmitter; the assembly and riveting station ST05 mainly uses a pneumatic riveting gun, and the range of its compressed air flow meter is selected according to the peak air consumption of the riveting gun; the online detection station ST06 has low energy consumption and is only equipped with a power monitor to collect current and active power.

[0143] Assuming a standard cycle time of 45 seconds per piece, the hydraulic pressure signal at the stamping station exhibits a process of depressurization to peak value and then falling back within each cycle time. Correspondingly, the current signal exhibits an impact peak value. These periodic characteristics are fully preserved in the energy consumption flow data, allowing subsequent steps to verify the validity of the data based on waveform characteristics and providing physical evidence for time alignment.

[0144] In step S112, the standard cycle time is set to 45 seconds per piece. After the stamping process is completed, the workpiece needs 2 seconds to be demolded and removed, which is the process lag time. It takes 5 seconds for the robot to transport the workpiece to the bending and forming station, which is the workpiece transfer time. Therefore, the cycle time reference of the bending process relative to the stamping process is 7 seconds. Similarly, the cycle time content of the six processes constitutes a time sequence chain that is connected end to end and shifted station by station. Among them, the surface treatment process has the largest shift due to the drying and curing requirement.

[0145] A car latch with the serial number "Piece 100: Batch Serial Number" begins processing at the stamping station at a certain moment. Its stamping energy consumption occurs within the cycle time interval starting from that moment. The workpiece is processed at the bending station seven seconds later, and the bending energy consumption occurs accordingly. After phase shifting, the time axis of the energy consumption data at the bending station is shifted forward by seven seconds, aligning with the stamping energy consumption to the same virtual processing time window. The energy consumption data at the surface treatment station is shifted forward by its corresponding cumulative offset, also falling into the same time window. Each workpiece within the production cycle is processed in this way, that is, the time-aligned energy consumption data is output.

[0146] When the automotive latch production line switches from one automotive latch to another model, the stamping station performs a mold change operation. The energy consumed by starting the hydraulic pump and adjusting the slide during the mold change is classified as auxiliary energy consumption according to the time interval of the mold change work order, and is not included in the energy consumption of the stamping process. Similarly, when the bending station is waiting for materials due to the temporary limitation of the surface treatment drying speed, the basic power consumption of its servo system to maintain power is also classified as auxiliary energy consumption.

[0147] The current production cycle has completed 800 automotive latches. The system outputs a dataset of energy consumption for the stamping process, including cumulative energy consumption values ​​and hydraulic peak sequences for 800 virtual processing time windows, energy consumption datasets for bending and forming processes, punching and trimming processes, surface treatment processes, assembly and riveting processes, online inspection processes, and independent auxiliary energy consumption datasets. This provides a clear and well-defined data foundation for subsequent production volume normalization and deviation analysis after S112.

[0148] In step S121, the current production cycle is set to an eight-hour shift. The online inspection station has determined that 800 car latches are qualified. The MES records that 820 are put into production and 20 are defective. Cross-checking confirms that the traceability number of the 800 qualified products is completely consistent with the set of workpieces covered by the process-level energy consumption dataset. Therefore, the output of the current cycle is set to 800.

[0149] The system retrieves production records of automotive latches from the historical database, filtering out cycles during the high-temperature summer and low-temperature winter periods to avoid the impact of the environment on drying energy consumption, and filtering out cycles where the molds are nearing the end of their service life. Finally, it obtains the most recent thirty historical cycle samples that are similar to the current working conditions. Each sample contains energy consumption datasets for six processes and the corresponding automotive latch production output.

[0150] For the stamping process, after iterative elimination of the unit output power consumption of the thirty historical cycle samples, twenty-eight samples remained. The two eliminated samples corresponded to the two cycles in which hydraulic system pressure relief failure and electricity meter communication packet loss occurred, respectively. For the surface treatment process, because the drying energy consumption is sensitive to the influence of environmental humidity, some cycle samples with abnormally high humidity were eliminated. The unit output energy consumption distribution after cleaning was obtained for each process.

[0151] After the above processing, the benchmark value of unit output power consumption for the stamping process is the median of twenty-eight cleaning samples, and the benchmark value for the surface treatment process is the median of its cleaned samples with an attached humidity condition label. The benchmark values ​​of the six processes together constitute the output energy consumption benchmark system of the automotive latch production line, which, together with the process-level energy consumption dataset of the current cycle and the current output of eight hundred pieces, serves as the input for entering the normalization and deviation calculation stage.

[0152] In step S122, the cumulative power consumption of the stamping process in the current cycle is a certain value, for example, equivalent to the total power consumption of all hydraulic work and equipment maintenance at the stamping station for 800 workpieces. Dividing this value by the output of 800 workpieces gives the actual power consumption per unit output of the stamping process. The total energy consumption of the drying process in the surface treatment process is divided by 800 workpieces to obtain its power consumption per unit output. The actual power consumption and gas consumption per unit output of the six processes are obtained in this way, and compared with the benchmark value.

[0153] The actual unit output power consumption of the stamping process is directly compared with the power consumption benchmark value of its historical sample statistics using the power consumption dimension; the gas consumption component of the assembly and riveting process is compared with its gas consumption benchmark value using the gas consumption volume dimension; when the surface treatment process is weighted and synthesized into a comprehensive index with electricity as the main component and gas as the auxiliary component, the conversion factor calibrated by the production line is used.

[0154] If the current cycle only produces 500 automotive latches due to a rush order, and the batch size is smaller than the historical average of 800 pieces, then the fixed energy consumption per piece in the stamping process will increase. The system will introduce a batch size correction factor to adjust the stamping benchmark value upwards before comparison, avoiding misjudging the reasonable high energy consumption of small batches as abnormal. The electrical energy consumed by the treatment tank in the surface treatment process during the start-up heating phase is unrelated to the output. When the cycle length is comparable to the historical average, the system will introduce an amortization correction factor to amortize and correct the heating and heat preservation energy consumption according to the current standard, so that the comparison standard is consistent.

[0155] The actual unit output power consumption of the stamping process is higher than the corrected benchmark by a certain percentage; the air consumption of the assembly and riveting process deviates from the benchmark; and the power consumption of the online detection process is within the tolerance zone. The power consumption and air consumption deviation values ​​of all six processes, together with the accompanying characteristics and tolerance zones, constitute the energy consumption deviation dataset of the current cycle of the automotive latch production line, which serves as the direct input for entering the S13 deviation processing and anomaly diagnosis stage.

[0156] In step S131, the current cycle plan is to produce 800 car latches. Due to a temporary slowdown at the surface treatment station, the actual output is 760 pieces, resulting in a shortfall of 40 pieces. The system inputs the shortfall into the mapping model. Based on the fixed energy consumption ratio of the stamping process and the heat preservation ratio of the surface treatment process, the model predicts that the reasonable deviation component of the stamping process due to the increase in batch distribution is a certain small percentage, and the reasonable deviation component of the surface treatment process due to the waiting for idle operation is another percentage. The predicted components of the other processes are determined accordingly based on their energy consumption structure.

[0157] The original deviation of the surface treatment process is higher than a certain percentage. After deducting the predicted component caused by the production shortfall of forty pieces, the remaining deviation falls back to within the tolerance zone. Therefore, the surface treatment process is judged to be caused by production fluctuation rather than an anomaly. However, the original deviation of the stamping process, after deducting the predicted component, still has a significant deviation exceeding the tolerance zone. Therefore, the stamping process is retained in the remaining deviation sequence as a candidate for anomaly diagnosis.

[0158] The remaining deviation curve of the stamping process shows a continuous drift starting from a certain moment in the middle of the cycle, and the abrupt change moment does not coincide with the mold change event; a periodic component with the same frequency as the 45-second cycle was detected in the deviation curve, and its peak value appeared stably in the slide holding pressure phase segment, indicating abnormal work in the holding pressure stage.

[0159] For candidates of anomalies in the stamping process, the system simultaneously retrieves the main drive current curve, hydraulic pressure and oil temperature curves, and alarm records. It is found that the average current slowly rises from the moment of deviation change, the hydraulic oil temperature rises synchronously, and intermittent low pressure alarms occur. This state characteristic coincides with the time of deviation drift, which serves as direct evidence for the next step of correlation analysis and "anomaly category determination: for example, determining it as an equipment deterioration type anomaly".

[0160] In step S132, the remaining deviation of the stamping process drifts continuously from the middle of the cycle. Its starting time coincides with the starting time of the rise in the average value of the main drive current. The drift direction is synchronized with the rise in hydraulic oil temperature. The periodic peak of the deviation is locked in the slider pressure holding phase segment, and it is accompanied by intermittent low pressure alarm. All four criteria are met, and the correlation strength is high.

[0161] Based on four pieces of evidence—continuous drift, synchronous rise in current and oil temperature, no overlap in batch events, and alarm confirmation—the stamping process was determined to be an anomaly of equipment performance degradation. The gas consumption deviation at the assembly and riveting station was found to have a continuous component unrelated to the cycle time, and the gas pressure showed a periodic pressure-replenishing sawtooth wave, which was determined to be an anomaly of energy leakage.

[0162] In the stamping process anomaly marking record, the process location is found to stamping station ST01 through reverse lookup in the mapping table. The anomaly category is equipment performance degradation. The evidence chain includes deviation drift range, correlation graph of synchronous rise of current and oil temperature, and periodic characteristics of pressure holding phase lock. The energy leakage anomaly is located to the air circuit branch of assembly and riveting station ST05 through the mapping table. The above two anomaly marking records, together with the process location, constitute the input for issuing control commands to S14.

[0163] In step S141, the two output anomaly records—the equipment performance degradation anomaly of stamping station ST01 and the energy leakage anomaly of assembly and riveting station ST05—are written into the data space after field verification. The data space then matches the two records with the anomaly category-handling measure mapping table, retrieves the candidate measure set corresponding to the performance degradation category and the candidate measure set corresponding to the leakage category, and confirms the address and instruction capability of the control units of ST01 and ST05 respectively through the routing table.

[0164] The performance degradation anomaly at stamping station ST01 is moderate and accompanied by an increase in oil temperature. The measure is marked as "switching to a reduced load operation mode and generating a lubrication point inspection work order" instead of directly shutting down the machine. The air leakage anomaly at assembly and riveting station ST05 is continuous. The measure is marked as "online leak detection work order plus intermittent valve closure isolation for this branch". Both measures are within the command capability range of the corresponding station control unit.

[0165] The stamping station ST01 is currently in continuous processing, and the inventory in each downstream buffer zone is higher than the lower limit. However, there is still a large amount of planned work remaining in this cycle. Therefore, the load reduction operation mode instruction is generated immediately, and the instruction parameter is that the hydraulic system pressure target value is reduced to the load reduction level. The air circuit branch valve isolation instruction for the assembly and riveting station ST05 is changed to a timed activation mode because this station is pneumatic riveting and the buffer inventory is low. It is specified that the valve will be closed within the waiting window during the interval of this batch to avoid affecting the normal riveting cycle. Both instructions are accompanied by corresponding abnormal event numbers, which constitute the instruction set to be issued to S142.

[0166] In step S142, the unloading command of the stamping station ST01 was checked at the time of generation and it was found that the slider was in the downward pressurization stroke. If it was executed immediately, it would result in insufficient forming force of the car latch of this part, and it was determined that there was a safety interlock conflict. Although the air branch valve closing command of the assembly and riveting station ST05 was set to be effective for the waiting window, the second check found that the pneumatic fixture on the branch may still be holding the previous workpiece at the edge of the window, and it was determined that there was a potential timing conflict.

[0167] The unloading command of the stamping station ST01 is determined to be delayed until a window of about two seconds after the next workpiece is stamped, the slide returns to the top dead center, and the next workpiece is still in transit. During this window, the pressure switching of the hydraulic system does not affect any forming action. The valve closing command of the assembly and riveting station ST05 is delayed until the pneumatic clamp release signal is detected and the next workpiece has not yet arrived. After execution, the air circuit isolation action is completed before the next workpiece arrives.

[0168] Within a two-second interval, the ST01 control unit completes the load reduction switch and sends back execution feedback. The feedback shows that the hydraulic pressure has stabilized at the load reduction level and the slider movement is normal. The ST05 control unit completes the branch valve closure during the material waiting period, and the feedback pressure drop curve confirms that the leakage section has been isolated. Both feedbacks are written back to the corresponding abnormal event unit. During the periodic review, it is confirmed that the air consumption deviation of the assembly and riveting process has fallen back to the tolerance zone in the next cycle, verifying that the handling is effective and the abnormal closed loop is completed.

[0169] Please see Figure 3 A control system for energy consumption data of an automotive latch production line is applied to the aforementioned method for controlling energy consumption data of the automotive latch production line; the control system for energy consumption data of the automotive latch production line includes:

[0170] The energy consumption dataset module 21 is used to mark multiple workstations on the automotive latch production line, obtain the energy consumption data of each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger the time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process.

[0171] Deviation module 22 is used to collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latch in combination with the product type of the latch. It also normalizes the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, and then determines the normalized energy consumption deviation dataset.

[0172] The anomaly module 23 is used to process the energy consumption deviation dataset and, after removing the deviation component caused by production fluctuations, determine the anomaly category of energy consumption anomaly and its corresponding process location based on the correlation between the time distribution characteristics of the remaining deviation and the corresponding equipment operating status data.

[0173] The instruction module 24 is used to generate corresponding energy consumption control instructions based on the exception type and its corresponding process position, and to send the energy consumption control instructions to the control unit corresponding to the process position.

[0174] It should be noted that although multiple modules are mentioned in the detailed description above, this division is not mandatory; the above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the protection scope of the present invention; therefore, the protection scope of the present invention should be determined by the protection scope of the claims.

Claims

1. A method for controlling energy consumption data in an automotive latch production line, characterized in that, include: Mark multiple workstations on the automotive latch production line, obtain energy consumption data for each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process. Collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latch based on the product type of the latch. Then, normalize the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, and then determine the normalized energy consumption deviation dataset. The energy consumption deviation dataset is processed, and after removing the deviation component caused by production fluctuations, the abnormality category of energy consumption anomaly and its corresponding process location are determined based on the correlation between the time distribution characteristics of the remaining deviation and the corresponding equipment operating status data. Based on the anomaly category and its corresponding process location, a corresponding energy consumption control command is generated and sent to the control unit corresponding to the process location.

2. The method for controlling energy consumption data of an automotive latch production line according to claim 1, characterized in that, The process involves marking multiple workstations on the automotive latch production line, acquiring energy consumption data for each workstation within the current production cycle, and combining this data with the cycle time sequence relationship between latch production processes to trigger time alignment and process attribution of energy consumption data at the process dimension. This determines the process-level energy consumption dataset corresponding to each latch production process, including: Along the production logistics path of the latch, the stamping station, bending and forming station, punching and trimming station, surface treatment station, assembly and riveting station, and online inspection station are marked in sequence. Each station is equipped with an independent energy consumption acquisition terminal to collect the corresponding energy consumption data synchronously. The energy consumption data includes current, active power, compressed air flow and hydraulic pressure signal.

3. The method for controlling energy consumption data of an automotive latch production line according to claim 2, characterized in that, The process involves marking multiple workstations on the automotive latch production line, acquiring energy consumption data for each workstation within the current production cycle, and combining this data with the cycle time sequence relationship between latch production processes to trigger time alignment and process attribution of energy consumption data at the process dimension. This determines the process-level energy consumption dataset corresponding to each latch production process, and also includes: Based on the standard cycle time of the locking product, the process cycle time sequence is constructed. The process cycle time sequence records the process lag time, workpiece transfer time, and work-in-process buffer time between adjacent processes. The time axis of the energy consumption data is phase-shifted according to the process lag time and workpiece transfer time, so that the energy consumption data generated by the same locking workpiece at different workstations is aligned to the virtual processing time window of the locking workpiece, so as to output the time-aligned energy consumption data. The time-aligned energy consumption data is assigned to the corresponding locking production process. The energy consumption data during idle waiting for materials, mold changing, and standby periods are assigned to the auxiliary energy consumption category and aggregated by process to form a process-level energy consumption dataset.

4. The method for controlling energy consumption data of an automotive latch production line according to claim 1, characterized in that, The process involves collecting the production output data of the current production cycle for the latches, determining the production energy consumption benchmark value for each latch production process based on the latch product type, and normalizing the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension. This results in the determination of the normalized energy consumption deviation dataset, including: Based on the process-level energy consumption dataset, the production output data of the latches in the current production cycle is collected. Combined with the historical production cycles retrieved from the historical energy consumption database by the latch product type, the unit output energy consumption distribution of the corresponding process is extracted, and after removing abnormal samples, the output energy consumption benchmark value of each process is determined.

5. The method for controlling energy consumption data of an automotive latch production line according to claim 4, characterized in that, The process of collecting the production output data of the current production cycle of the latches, determining the production energy consumption benchmark value for each latch production process based on the latch product type, and normalizing the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension to determine the normalized energy consumption deviation dataset also includes: Obtain the cumulative energy consumption value in the process-level energy consumption dataset, and determine the actual unit output energy consumption in combination with the actual output of the current cycle. Express the output energy consumption benchmark value in the same dimension, and introduce a batch size correction factor for mold-dependent processes, and introduce a fixed energy consumption amortization correction factor for bath heating and heat preservation for time-dependent processes. Based on the actual unit output energy consumption and the corresponding output energy consumption benchmark value, determine the corresponding energy consumption deviation dataset.

6. The method for controlling energy consumption data of an automotive latch production line according to claim 1, characterized in that, The process of processing the energy consumption deviation dataset, and after removing the deviation components caused by production fluctuations, determines the anomaly category of energy consumption anomalies and their corresponding process locations based on the correlation between the temporal distribution characteristics of the remaining deviations and the corresponding equipment operating status data, including: The fluctuation of actual output relative to planned output is used as input, and the predicted value of the deviation component caused by output fluctuation is obtained. The predicted value of the deviation component is subtracted from the energy consumption deviation dataset to obtain the corresponding residual deviation sequence. The deviation duration, deviation change time, periodicity and phase relationship with the process operation cycle of the residual deviation sequence are extracted, and the equipment operation status data of the same time window are collected at the same time.

7. The method for controlling energy consumption data of an automotive latch production line according to claim 6, characterized in that, The process of processing the energy consumption deviation dataset, and after removing the deviation components caused by production fluctuations, determining the anomaly category of energy consumption anomalies and their corresponding process locations based on the correlation between the temporal distribution characteristics of the remaining deviations and the corresponding equipment operating status data, further includes: The time distribution characteristics are correlated with the equipment operating status data in a time series analysis to determine the anomaly category of energy consumption anomalies. The anomaly category of energy consumption anomalies includes equipment idling and standby anomalies, equipment performance degradation anomalies, process parameter deviation anomalies, and energy leakage anomalies. At the same time, the remaining deviation sequence of the anomaly is marked, and the corresponding process position is determined according to the process-level data to which the remaining deviation sequence belongs.

8. The method for controlling energy consumption data of an automotive latch production line according to claim 7, characterized in that, The process of generating corresponding energy consumption control commands based on the anomaly type and its corresponding process location, and issuing the energy consumption control commands to the control unit corresponding to the process location, includes: The abnormal categories of energy consumption anomalies and their corresponding process positions are loaded into the same data space, and the anomaly handling measures corresponding to the energy consumption anomalies are marked in the data space. In addition, the corresponding energy consumption control instructions are generated in combination with the current working status of the automotive latch production line.

9. The method for controlling energy consumption data of an automotive latch production line according to claim 8, characterized in that, The step of generating corresponding energy consumption control commands based on the anomaly type and its corresponding process position, and issuing the energy consumption control commands to the control unit corresponding to the process position, further includes: Before issuing energy consumption control commands, conflict checks are performed between the commands and production cycle time and safety interlock conditions. If a timing conflict exists, the execution time is postponed to the interval between processes, and then the energy consumption control commands are issued to the control unit corresponding to the process position.

10. A control system for energy consumption data of an automotive latch production line, characterized in that, The control system for the energy consumption data of the automotive latch production line is applied to the control method for the energy consumption data of the automotive latch production line as described in any one of claims 1-9; the control system for the energy consumption data of the automotive latch production line includes: The energy consumption dataset module is used to label multiple workstations on the automotive latch production line, obtain energy consumption data for each workstation in the current production cycle, and combine the cycle time relationship between latch production processes to trigger time alignment and process attribution of energy consumption data in the process dimension, thereby determining the process-level energy consumption dataset corresponding to each latch production process. The deviation module is used to collect the production output data of the current production cycle, and determine the production energy consumption benchmark value of each production process of the latches in combination with the product type of the latches. It also normalizes the process-level energy consumption dataset and the production energy consumption benchmark value to the same production dimension, thereby determining the normalized energy consumption deviation dataset. The anomaly module is used to process the energy consumption deviation dataset and, after removing the deviation components caused by production fluctuations, determine the anomaly category of energy consumption anomalies and their corresponding process locations based on the correlation between the time distribution characteristics of the remaining deviations and the corresponding equipment operating status data. The instruction module is used to generate corresponding energy consumption control instructions based on the exception type and its corresponding process position, and to send the energy consumption control instructions to the control unit corresponding to the process position.

Images