Method for dynamic adjustment of energy consumption in production of electric porcelain insulator based on carbon footprint feedback

Abstract

The application provides a kind of based on carbon footprint feedback electric porcelain insulator production energy consumption dynamic adjustment method, comprising: with batch identification establishment batch process record, collect electric porcelain insulator batch in the process of drying after forming, preheating before glazing, sintering preparation, sintering keeps, cooling recovery and process transfer Process running data and quality feedback data, generate batch carbon footprint control data set;According to the reference carbon footprint of each process calculated by historical qualified batch, and the actual carbon footprint increment of current batch is obtained Process carbon footprint deviation, the deviation part transferred from the previous state shortage to the subsequent process is registered as the pending carbon footprint debt value, and the carbon footprint debt account is formed;Again identify the attribution type of energy consumption increase in current process, determine the controllable node and control parameter deviation of previous sequence, generate debt attribution record;According to the energy consumption dynamic adjustment instruction generated by debt attribution record and executes.

Description

Technical Field

[0001] This invention relates to the field of production process control technology, and in particular to a method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback. Background Technology

[0002] The production of porcelain insulators involves a series of closely coupled thermal processes, including drying after molding, preheating before glazing, and firing. Deviations in any process can easily transfer to subsequent processes, leading not only to additional energy consumption but also to quality problems such as drying cracks and glaze defects.

[0003] The existing energy consumption control methods for the production of porcelain insulators have the following shortcomings: First, it is limited to local energy saving in a single process and lacks global coordination. Existing control is mostly based on the operating parameters of a single process. This method is difficult to identify the hidden energy consumption burden transferred to subsequent processes due to insufficient conditions in the preceding process (such as water content and insufficient preheating). It is easy to cause the preceding process to forcibly reduce energy consumption, resulting in the subsequent process needing to make large compensations, which leads to the total energy consumption and carbon footprint of the whole process increasing instead of decreasing. Secondly, carbon footprint feedback is delayed and lacks dynamic attribution and linkage. Existing solutions typically treat carbon footprint as a post-batch static statistic, making it difficult to link it with process parameters in real time. When energy consumption increases in a certain process, the system struggles to distinguish whether it's due to normal process heat absorption, current control deviation, or compensation from previous states. This lack of attribution can lead the control system to blindly implement energy reduction measures for currently high-energy-consuming processes, which not only fails to truly save energy but may also induce product quality risks or exacerbate the shift of energy consumption to later stages.

[0004] Therefore, this invention proposes a method for dynamically adjusting energy consumption in the production of porcelain insulators based on carbon footprint feedback. The information disclosed in the background section is only for enhancing understanding of the background of this disclosure and may therefore contain prior art information that is not common knowledge to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a method for dynamically adjusting energy consumption in the production of porcelain insulators based on carbon footprint feedback, thereby solving the technical problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback includes the following steps: S1. Establish batch process records with batch identifiers, collect process operation data and quality feedback data of electric porcelain insulator batches in drying after molding, preheating before glazing, firing preparation, firing holding, cooling recovery and process transfer, and generate batch carbon footprint control datasets. S2. Based on the batch carbon footprint control dataset, screen historical qualified batches, calculate the baseline carbon footprint, actual carbon footprint increment and process carbon footprint deviation for each process, and register the deviation portion transferred from insufficient previous state to subsequent process as carbon footprint debt value to be confirmed, forming a carbon footprint debt account. S3. Identify the attribution type of the current process energy consumption increase based on the carbon footprint debt account, determine the controllable nodes and control parameter deviations of the preceding process, and form a debt attribution record; S4. Generate a set of candidate actions for dynamic energy consumption adjustment based on debt attribution records. After verification of quality constraint boundaries and carbon footprint carrying capacity of subsequent processes, obtain a set of executable adjustment actions and generate dynamic energy consumption adjustment instructions. S5. Execute the dynamic energy consumption adjustment command to form an adjustment execution record, recalculate the carbon footprint debt value and mark the effective adjustment action or the restriction adjustment action, and feed the updated carbon footprint debt account back to the subsequent batch production process control model.

[0007] S1 specifically includes: After the batch of porcelain insulators enters the production process control system, a batch process record is established with the batch identifier as the main index and the process identifier and control time as auxiliary indexes. The product specifications, number of blanks, total weight of blanks, loading position, loading density, process flow sequence and planned process curve are written in. According to the batch process record, the blank status, process execution, energy consumption, waste heat utilization, process waiting and quality inspection data of each process are collected and collected into process operation data according to the process control cycle. The batch process record, process operation data and quality feedback data are aligned by batch identifier, process identifier and collection time, and the downtime, sensor failure and non-production waiting data are removed to generate a batch carbon footprint control dataset.

[0008] S2 specifically includes: reading product specifications, batch loading density, planned process curve, energy consumption field, and quality feedback field from the batch carbon footprint control dataset; screening historical qualified batches that match the product specifications, loading density, process curve, and quality results; calculating the baseline carbon footprint for each process; calculating the actual carbon footprint increment based on the actual electricity consumption, actual gas consumption, and actual waste heat utilization of each process in the current batch; comparing the actual carbon footprint increment with the baseline carbon footprint to obtain the process carbon footprint deviation; marking the deviation corresponding to insufficient previous states as carbon footprint debt values ​​to be confirmed; and registering the carbon footprint debt values ​​to be confirmed according to batch identifier, debt source process, subsequent inheriting process, and control time to form a carbon footprint debt account.

[0009] S3 specifically includes: reading the actual carbon footprint increment, baseline carbon footprint, process carbon footprint deviation, and carbon footprint debt account; calculating the relative deviation rate of the current process's carbon footprint; identifying whether the increase in energy consumption in the current process is due to normal process heat absorption, current control deviation compensation, or repayment of debts left over from previous processes; when it is not determined to be normal process heat absorption, determining the preceding controllable nodes based on the debt items pointing to the current process, the time of energy consumption increase, response delay, quality risk performance, and the carbon footprint debt value to be confirmed; reading the target parameters and actual execution parameters of the preceding controllable nodes, determining the control parameter deviation, and writing the preceding controllable nodes, debt attribution strength, control parameter deviation, state insufficiency type, energy consumption compensation method, and attribution conclusion into the debt attribution record.

[0010] S4 specifically includes: reading the preceding controllable nodes, control parameter deviations, state insufficiency types, energy consumption compensation methods, and carbon footprint debt values ​​to be confirmed from the debt attribution record, mapping and generating a set of candidate actions for dynamic energy consumption adjustment, and combining historical adjustment execution records and reinforcement learning to call the priority of candidate actions; The set of candidate actions for dynamic energy consumption adjustment is verified by inputting them into the quality constraint boundary and the carbon footprint bearing capacity of subsequent processes. Candidate actions that lead to increased quality risk or transfer of carbon footprint debt are eliminated to obtain the set of executable adjustment actions. Based on the predicted reduction in carbon footprint debt, the predicted change in quality risk, the predicted change in pressure on subsequent processes, and the penalty value for energy-saving actions with high debt, a comprehensive priority is calculated to generate dynamic energy consumption adjustment instructions.

[0011] S5 specifically includes: before the energy consumption dynamic adjustment command arrives at the execution time, verifying the batch identifier, execution process, adjustment object, adjustment range, quality constraint boundary, and prohibited action conditions. After verification, the command is executed, and actual control parameters, energy consumption, process response status, and quality feedback data are collected to form an adjustment execution record. Based on the adjustment execution record, the adjusted carbon footprint debt value is recalculated using the same calculation method as S2, the debt change before and after adjustment is compared, and effective adjustment actions or restricted adjustment actions are marked. Effective adjustment actions, restricted adjustment actions, updated carbon footprint debt accounts, adjustment execution records, and quality feedback results are fed back to the subsequent batch production process control model, and the action selection priority is updated through reinforcement learning.

[0012] The beneficial effects of this invention are as follows: This invention establishes a batch carbon footprint control dataset, unifying and linking batch status, process operation, energy consumption, and quality feedback data. This provides a clear data foundation for energy consumption regulation in porcelain insulator production, avoiding reliance on single-process energy consumption or empirical parameters for crude control. By calculating the baseline carbon footprint, actual carbon footprint increment, and process carbon footprint deviation, and forming a carbon footprint debt account, it can identify the hidden energy consumption burden transferred from insufficient preceding processes to subsequent processes, preventing localized energy conservation from increasing the overall carbon footprint.

[0013] This invention, by distinguishing between normal process heat absorption, current control deviation compensation, and repayment of debts left over from previous processes, can accurately pinpoint the controllable preceding nodes and control parameter deviations that cause carbon footprint debt. This avoids quality risks arising from directly adjusting the energy consumption of current high-energy-consuming processes. Based on debt attribution records, candidate actions for dynamic energy consumption adjustment are generated and verified through quality constraint boundaries and the carbon footprint carrying capacity of subsequent processes. This allows for the reduction of carbon footprint debt while suppressing quality risks such as drying cracks, glaze defects, and firing shrinkage deviations.

[0014] This invention recalculates the carbon footprint debt value by adjusting execution records and marks effective or restrictive adjustment actions. This enables the production process control model to continuously adjust action selection priorities based on actual execution results, improving the accuracy of dynamic energy consumption adjustment in subsequent batches. By forming a closed-loop process control system integrating carbon footprint feedback, debt attribution, adjustment command execution, and reinforcement learning updates, the production of porcelain insulators can prioritize energy consumption control parameters with low carbon footprint debt under different loading densities, moisture content, and process cycle times. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback, as described in this invention. Detailed Implementation

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

[0017] Example: Figure 1 As shown, this embodiment provides a method for dynamically adjusting the energy consumption of porcelain insulator production based on carbon footprint feedback, including the following steps: S1. Establish batch process records with batch identifiers, collect process operation data and quality feedback data of electric porcelain insulator batches in drying after molding, preheating before glazing, firing preparation, firing holding, cooling recovery and process transfer, and generate batch carbon footprint control datasets. S2. Based on the batch carbon footprint control dataset, screen historical qualified batches, calculate the baseline carbon footprint, actual carbon footprint increment and process carbon footprint deviation for each process, and register the deviation portion transferred from insufficient previous state to subsequent process as carbon footprint debt value to be confirmed, forming a carbon footprint debt account. S3. Identify the attribution type of the current process energy consumption increase based on the carbon footprint debt account, determine the controllable nodes and control parameter deviations of the preceding process, and form a debt attribution record; S4. Generate a set of candidate actions for dynamic energy consumption adjustment based on debt attribution records. After verification of quality constraint boundaries and carbon footprint carrying capacity of subsequent processes, obtain a set of executable adjustment actions and generate dynamic energy consumption adjustment instructions. S5. Execute the dynamic energy consumption adjustment command to form an adjustment execution record, recalculate the carbon footprint debt value and mark the effective adjustment action or the restriction adjustment action, and feed the updated carbon footprint debt account back to the subsequent batch production process control model.

[0018] The dynamic adjustment method described in this embodiment can be programmed and packaged into industrial control software and deployed on an industrial computer in the field or on a cloud server. This industrial control software reads sensor data and outputs control commands to the actuators to complete closed-loop control of the porcelain insulator thermal production line. The dynamic energy consumption adjustment method for porcelain insulator production based on carbon footprint feedback provided in this embodiment is essentially an automated control scheme that relies on the collaborative efforts of an industrial control system and the industrial control software running on that system.

[0019] S1 specifically includes the following sub-steps: S110. After the batch of porcelain insulators enters the production process control system (industrial control system), the production process control system establishes a batch process record with the batch identifier as the main index and the process identifier and control time as auxiliary indexes, so that the data collected later can be accurately classified into the same batch and its corresponding process.

[0020] Batch identifiers are derived from batch work orders generated by the production execution system; product specifications, billet quantity, and planned process curves are derived from the process formula library and batch work orders; total billet weight is derived from weighing records or batch process sheets; loading location is derived from loading barcode records, kiln car layer records, or confirmed electronic loading tables; and process flow sequence is derived from the process configuration records preset by the production process control system. Batch process records must include at least the batch identifier, product specifications, billet quantity, total billet weight, loading location, loading density, process flow sequence, planned process curve, batch start time, and estimated process start time.

[0021] Loading density is used to characterize the degree of heat load concentration within a unit loading space of the same batch. The production process control system calculates the loading density based on the total weight of the billet and the effective loading volume, and writes the calculation results into the batch process record.

[0022] in, This refers to the batch loading density, expressed in kg / m³. 3 ; The total weight of porcelain insulator blanks from the same batch is expressed in kg. This refers to the effective loading volume occupied by this batch, in m³. 3 The effective loading volume is determined by the kiln car level record, loading tool specifications, and loading location.

[0023] In one embodiment, the total weight of a batch of billets is 1200 kg, occupying an effective loading volume of 6 m³. 3 The batch loading density is 200 kg / m³. 3 This value is used in subsequent step S210 to determine the baseline energy consumption range corresponding to this batch. Through the above processing, the batch process record is not only used to record the batch identity, but also to serve as the basis for the loaded heat load that affects subsequent energy consumption calculations and carbon footprint debt assessments.

[0024] S120 The production process control system records batch process data and collects process operation data during drying after molding, preheating before glazing, firing preparation, firing holding, cooling recovery and process transfer. Each process operation data is bound to the batch identifier, process identifier and collection time.

[0025] During the drying stage after molding, the moisture content of the green body, air volume adjustment, process duration, and power consumption are collected; during the preheating stage before glazing, the temperature change of the green body, preheating intensity, and residual heat utilization are collected; during the firing preparation and firing holding stages, the heating power, gas consumption, power consumption, heating slope, and holding time are collected; during the cooling and recovery stage, the residual heat utilization, cooling duration, and subsequent residual heat utilization status are collected; during the process transfer process, the process waiting time, batch transfer time, and manual pause records are collected.

[0026] The above data are derived from control command records of the production process control system, energy consumption metering devices, gas metering devices, temperature and humidity acquisition devices, waste heat recovery records, batch transfer records, and quality inspection systems. For data with different sampling periods, the process control cycle of the production process control system is used as the alignment benchmark; energy consumption data is collected according to the cumulative amount within the control cycle, and status data is collected according to the interval average or the most recent valid sample value within the control cycle.

[0027] In one implementation, the process control cycle is 60 seconds, and the billet temperature is collected once every 10 seconds. The average value of the six temperature samples within the 60 seconds is then taken and bound to the same collection time along with the power consumption, gas consumption, and air volume adjustment within the cycle.

[0028] To ensure a unified calculation standard for different energy data, the production process control system aggregates the effective energy consumption of each process in the following manner:

[0029] in, The effective energy consumption of the i-th process is expressed in kWh; i is the process number. This represents the actual power consumption of the i-th process, in kWh. The actual gas consumption of the i-th process after conversion according to the uniform calorific value is expressed in kWh. This represents the actual waste heat utilization amount for the i-th process, expressed in kWh.

[0030] Effective energy consumption serves as the energy consumption basis for subsequent calculations of the baseline carbon footprint in S210 and the actual carbon footprint increment in S220.

[0031] S130 The production process control system performs time alignment and validity screening on batch process records, process operation data and quality feedback data to generate a batch carbon footprint control dataset.

[0032] Quality feedback data comes from online inspection records, spot check records, rework records, and scrap records, including drying cracks, glaze defects, firing shrinkage deviations, and rework results. Among them, drying cracks are linked to the drying stage after molding, glaze defects are linked to the preheating and firing preparation stages before glazing, firing shrinkage deviations are linked to the firing holding stage, and rework results are linked to the final batch quality feedback.

[0033] When aligning time, the start time, end time and acquisition time of the process in the batch process record are used as the benchmark. Status data, process execution data, energy consumption data and quality feedback data under the same batch identifier and the same process identifier are written into the same data segment.

[0034] For invalid data caused by downtime, sensor failure, or non-production waiting, the production process control system will discard the following: downtime data will be identified based on the equipment operating status being down, heating power being 0, and no batch flow records being available; sensor failure data will be identified based on sensor fault codes, data exceeding the range, or no change for multiple consecutive control cycles; non-production waiting data will be identified based on the process not starting, batch not entering the current process, equipment being idle, or manual pause records being available.

[0035] Data validity is marked according to the following rules:

[0036] in, This is a data validity marker for the i-th process at acquisition time t; t is the acquisition time. This is used to mark the equipment's operating status. A value of 1 indicates that the equipment is in production operation, while a value of 0 indicates that it is not in production operation. This is a non-production waiting flag; a value of 1 indicates that there is a non-production waiting, and a value of 0 indicates that there is no non-production waiting. This is a sensor failure flag; a value of 1 indicates that a sensor failure exists, and a value of 0 indicates that no sensor failure exists. This indicates a logical AND operation. Indicates logical OR. Only if At that time, the corresponding data is written into the batch carbon footprint control dataset.

[0037] In one implementation, if a process is shut down and has no batch flow record between 10:00 and 10:20, even if the energy consumption meter shows a small amount of standby power consumption, this period will be marked as non-production waiting data and removed. The batch carbon footprint control dataset includes at least batch identifier, process identifier, collection time, batch status field, process execution field, energy consumption field, quality feedback field, and validity label field, serving as a unified data source for subsequent steps: S210 calculating the baseline carbon footprint, S220 identifying process carbon footprint deviations, and S530 updating energy consumption control parameters through reinforcement learning.

[0038] S2 specifically includes the following sub-steps: S210 The production process control system reads the current batch's product specifications, batch loading density, planned process curve, process identifier, energy consumption field, quality feedback field, and validity mark field from the batch carbon footprint control dataset generated in S130, and filters historical qualified batches from historical production data.

[0039] Historically qualified batches should simultaneously meet the following conditions: the product specifications are consistent with the current batch, or belong to the same specification group pre-divided by the production process control system; the batch loading density is within the allowable deviation range of the current batch loading density; the planned process curve is consistent; the data validity is marked as valid; the final quality feedback is qualified, and no invalid data caused by downtime anomalies, sensor failures, or non-production waiting is recorded.

[0040] The same specification group is pre-divided by the production process control system according to product specifications, blank weight range, loading method, and planned process curve. In one embodiment, the current batch is a certain specification of rod-shaped porcelain insulator, and the batch loading density is 200 kg / m³. 3 The production process control system selects batches of the same specifications with a loading density of 180 kg / m³. 3 Up to 220kg / m 3 Historical batches that are consistent with the planned process curves and whose final quality feedback is qualified are used as benchmark samples.

[0041] After screening, the production process control system calculates the baseline power consumption, baseline gas consumption, and baseline waste heat utilization for each process according to the process identifiers such as post-forming drying, preheating before glazing, firing preparation, firing holding, and cooling recovery. It then calculates the baseline carbon footprint for each process by combining the power carbon factor, gas carbon factor, and waste heat substitution carbon factor.

[0042] in, The baseline carbon footprint for the i-th process is expressed in kgCO2e. This represents the baseline power consumption of the i-th process in historical qualified batches, in kWh. This represents the baseline gas consumption for the i-th process in historical qualified batches, in kWh. The baseline waste heat utilization amount for the i-th process in the historical qualified batches, in kWh; The carbon factor for electricity is expressed in kgCO2e / kWh. The carbon factor of fuel gas is expressed in kgCO2e / kWh. The waste heat replacement carbon factor is expressed in kgCO2e / kWh.

[0043] The carbon factor for electricity comes from the company's energy management system, publicly available carbon factors from the local power grid, or energy accounting sheets confirmed by the company. The carbon factor for natural gas comes from the standard conversion factor corresponding to the natural gas metering system. The carbon factor for waste heat substitution is determined based on the caliber of electricity or natural gas energy that can be substituted by waste heat. The baseline carbon footprint is used in subsequent S220 checks to determine whether the current batch deviates from the carbon footprint requirements of the corresponding process in the corresponding step beyond the normal production process control requirements.

[0044] S220: The production process control system calculates the actual carbon footprint increment of the current batch according to the process identifier based on the energy consumption field in the batch carbon footprint control data set.

[0045] The actual electricity consumption comes from the electricity metering device, the actual gas consumption comes from the gas metering device and is converted to kWh according to a unified calorific value, and the actual waste heat utilization comes from the waste heat utilization record. When the energy metering device outputs the cumulative energy consumption of the cycle, the process duration is used to determine the cumulative time period. When the energy metering device outputs the instantaneous power, the production process control system performs cycle accumulation according to the process duration to obtain the actual electricity consumption and actual gas consumption of the corresponding process.

[0046] The actual increase in carbon footprint is calculated as follows:

[0047] in, This represents the actual carbon footprint increment for the i-th process, expressed in kgCO2e. This represents the actual power consumption of the i-th process, in kWh. This represents the actual gas consumption of the i-th process, in kWh. The actual waste heat utilization amount for the i-th process is expressed in kWh. , , The meaning is the same as in S210.

[0048] Subsequently, the production process control system compares the actual carbon footprint increment of the same process with the baseline carbon footprint to obtain the process carbon footprint deviation:

[0049] in, This represents the deviation of the carbon footprint of the i-th process, expressed in kgCO2e.

[0050] like If the value is less than or equal to 0, it indicates that the current process has not deviated from the benchmark carbon footprint and will not be included in the carbon footprint debt assessment; if If the value is greater than 0, the production process control system will further read the status field of the preceding process to determine if there are any deficiencies in the preceding status. Deficiencies in the preceding status include: the moisture content of the billet at the end of the preceding drying process is higher than the upper limit allowed by the planned process curve; the billet temperature at the end of the preceding preheating process is lower than the lower limit allowed by the planned process curve; the temperature difference between different loading positions of the same batch at the end of the preceding process is higher than the allowed temperature difference; and the actual duration of the preceding process is shorter than the minimum duration in the planned process curve.

[0051] When the process carbon footprint deviation is positive and there is at least one insufficiency in the preceding state, the production process control system will mark the deviation corresponding to the insufficiency in the preceding state as a carbon footprint debt value to be confirmed:

[0052] in, The unconfirmed carbon footprint debt value transferred from the j-th preceding process to the i-th subsequent process is expressed in kgCO2e; j is the preceding process number. is the debt attribution coefficient for the carbon footprint deviation of the j-th preceding process to the i-th subsequent process, with a value ranging from 0 to 1.

[0053] If there is only one source of insufficient preceding state, then Set to 1; if there are multiple sources of insufficient preceding states, then allocate based on the normalized results of the residual moisture content amplitude, insufficient preheating amplitude, unresolved temperature difference amplitude, and cycle compression amplitude. Furthermore, the sum of multiple debt attribution coefficients is 1.

[0054] S230: The production process control system uses batch identifier as the primary index and debt-originating process, subsequent receiving process, and control time as secondary indexes to register the unconfirmed carbon footprint debt value obtained in S220, forming a carbon footprint debt account.

[0055] The carbon footprint debt account is not simply a storage of carbon footprint values, but a record of the complete chain of energy consumption compensation transferred from insufficient preceding processes to subsequent processes. It includes at least the batch identifier, the process from which the debt originates, the subsequent process taking over, the control time, the carbon footprint debt value to be confirmed, the type of insufficient preceding process, the magnitude of the insufficient process, the energy consumption compensation method, the associated quality risk performance, the debt confirmation status, and subsequent adjustment markers. If multiple sources of insufficient preceding processes exist for the same subsequent process, the production process control system should establish separate debt entries for each, or split and register them under the same debt entry according to the debt attribution coefficient, to avoid commingling carbon footprint debts from different sources.

[0056] In one implementation, if a batch of green bodies has a moisture content higher than the allowable upper limit at the end of post-forming drying, and a delay in temperature response occurs after entering the firing preparation stage, resulting in a positive carbon footprint deviation, then the portion of this deviation attributable to residual moisture in the preceding process is marked as a carbon footprint debt value to be confirmed, transferred from the post-forming drying stage to the firing preparation stage. The carbon footprint debt account records the batch identifier, the debt source process as post-forming drying, the subsequent process as firing preparation, the type of insufficient preceding state as residual moisture, the energy consumption compensation method as temperature compensation, the associated quality risk as drying crack risk or firing shrinkage deviation risk, and the debt confirmation status as pending confirmation.

[0057] The resulting carbon footprint debt account is used by S310 to read and determine whether the increase in energy consumption in the current process is due to normal process heat absorption, compensation for current control deviation, or repayment of debts left over from previous processes, and provides a data basis for S320 to determine the controllable nodes in the preceding process.

[0058] S3 specifically includes the following sub-steps: S310: The production process control system reads the actual carbon footprint increment, baseline carbon footprint, and process carbon footprint deviation obtained from S220, and reads the carbon footprint debt account formed in S230 to make a preliminary classification of the reasons for the increase in energy consumption in the current process.

[0059] The current process is one that is currently being executed or has just completed carbon footprint calculation. An increase in energy consumption in the current process is not immediately considered abnormal; instead, it is first determined whether it falls under normal process heat absorption. Normal process heat absorption refers to a reasonable increase in energy consumption in the current process due to product specifications, batch loading density, and planned process curve requirements during heating, temperature equalization, heat preservation, or cooling, and this increase in energy consumption does not correspond to the carbon footprint debt left over from previous processes.

[0060] The production process control system calculates the relative deviation rate of the carbon footprint of the current process:

[0061] in, Let be the relative deviation rate of the carbon footprint of the i-th process; The deviation of the carbon footprint of the i-th process is expressed in kgCO2e. This represents the baseline carbon footprint of the i-th process, expressed in kgCO2e.

[0062] If the relative deviation rate of the carbon footprint is not greater than the allowable deviation threshold preset in the process recipe library by the production process control system, and there are no pending carbon footprint debt values ​​pointing to the current process in the carbon footprint debt account, and the current process does not exhibit temperature and humidity response delays or heating response delays exceeding the allowable range of the planned process curve, then the current increase in energy consumption is marked as normal process heat absorption. If the relative deviation rate of the carbon footprint is greater than the allowable deviation threshold, or there are pending carbon footprint debt values ​​pointing to the current process in the carbon footprint debt account, then further identification of current control deviation compensation and repayment of debts left over from previous processes will proceed.

[0063] Current control deviation compensation refers to the additional energy consumption incurred to restore the temperature and humidity state, heating state, or heat preservation state after the execution value of the current process's own control parameters deviates from the planned process curve; repayment of debts left over from previous processes refers to the additional energy consumption of the current process to compensate for the previous process's failure to eliminate residual water, insufficient preheating, temperature difference, or insufficient cycle compression state.

[0064] In one implementation, if a batch has a higher loading density than a regular batch, and the actual carbon footprint increase during the firing and holding stage is slightly higher than the baseline carbon footprint, but the relative deviation rate of the carbon footprint does not exceed the allowable deviation threshold, and there are no debt entries in the carbon footprint debt account pointing to the firing and holding stage, then the increase in energy consumption is determined to be normal process heat absorption.

[0065] S320. When S310 does not determine the current energy consumption increase as normal process heat absorption, the production process control system reads the debt entries pointing to the current process in the carbon footprint debt account, and performs correlation analysis on the current process's energy consumption increase time, temperature and humidity response delay, temperature rise response delay, quality risk performance, and unconfirmed carbon footprint debt value with the air volume adjustment, preheating intensity, temperature rise slope, heat preservation time, waste heat utilization ratio, and process cycle time in the preceding process to determine the preceding controllable nodes.

[0066] A preceding controllable node refers to a process position that occurs before the current process, can be corrected by adjusting the process cycle time, air volume, preheating intensity, heating slope, heat preservation time or waste heat utilization ratio, and has an attribution relationship to the carbon footprint debt value of the current process.

[0067] The correlation analysis is performed according to the time sequence, state insufficiency, consistency of compensation method, and debt value: 1. Determine whether the debt source process is earlier than the current process; 2. Determine whether the debt source process has residual water, insufficient preheating, unresolved temperature difference, or compressed cycle time; 3. Determine whether the previous state insufficiency type can explain the energy consumption compensation method of the current process; 4. Compare the debt attribution strength of different debt source processes to the current process.

[0068] Debt attribution strength is calculated as follows:

[0069] in, The debt attribution strength of the j-th preceding process to the i-th subsequent process; The unconfirmed carbon footprint debt value transferred from the j-th preceding process to the i-th subsequent process is expressed in kgCO2e. This is a status indicator for the j-th preceding process. A value of 1 indicates that there is a status deficiency, and a value of 0 indicates that there is no status deficiency. This is a time sequence marker. A value of 1 indicates that the insufficient state of the j-th preceding process occurs before the energy consumption of the i-th subsequent process increases, while a value of 0 indicates that the time sequence is not satisfied.

[0070] The production process control system identifies the preceding process with the largest debt attribution strength greater than 0 as the priority controllable preceding process. If there are multiple preceding processes with debt attribution strength greater than 0, a sequence of controllable preceding processes is established according to the carbon footprint debt value to be confirmed from largest to smallest.

[0071] If there is a control parameter execution deviation in the current process, but there is no debt entry in the carbon footprint debt account pointing to the current process, it is determined to be current control deviation compensation; if there is a debt entry in the carbon footprint debt account pointing to the current process, and the type of insufficient status in the preceding process is consistent with the current energy consumption compensation method, it is determined to be debt repayment left over from the preceding process; if current control deviation compensation and debt repayment left over from the preceding process exist simultaneously, the production process control system shall establish current control deviation entries and preceding debt repayment entries respectively, and record the corresponding control parameter deviations and debt attribution coefficients in the subsequent debt attribution records respectively.

[0072] S330: Based on the preceding controllable node determined by S320, the production process control system reads the target parameters of the preceding controllable node in the planned process curve and the actual execution parameters in the process operation data, and determines the corresponding control parameter deviation.

[0073] Control parameter deviations include air volume deviation, preheating intensity deviation, heating slope deviation, heat preservation duration deviation, waste heat utilization ratio deviation, and process cycle time deviation. The data sources include planned process curves in the process formula library, control instruction records of the production process control system, batch carbon footprint control datasets generated by S130, and carbon footprint debt accounts formed by S230.

[0074] The control parameter deviation is calculated as follows:

[0075] in, Let k be the deviation of the k-th type of control parameter in the j-th preceding controllable node; k is the control parameter type number. This represents the actual execution value of the k-th type of control parameter in the j-th preceding controllable node; Let $\frac{j}{j}$ be the target value of the $k$ type of control parameter corresponding to the $j$-th preceding controllable node in the planned process curve.

[0076] If the sign of the control parameter deviation indicates that the actual executed value is lower than the target value, the subsequent adjustment direction is to supplement the parameter; if the sign of the control parameter deviation indicates that the actual executed value is higher than the target value, the subsequent adjustment direction is to compress the parameter or adjust its execution time.

[0077] The production process control system writes batch identifier, current process, preceding controllable nodes, carbon footprint debt value to be confirmed, debt attribution strength, control parameter deviation, state insufficiency type, energy consumption compensation method, quality risk performance, attribution conclusion, and adjustment priority into the debt attribution record. The debt attribution record is a cause identification record formed on the basis of the carbon footprint debt account. It is used to save the attribution results of the current process's energy consumption increase and serves as the direct input for the subsequent S410 to construct a set of candidate actions for dynamic energy consumption adjustment.

[0078] In one implementation, the preceding controllable node is the drying stage after molding. The planned process curve requires the air volume to reach the target value, but the actual air volume is lower than the target value for several consecutive control cycles. This results in the moisture content of the green body being higher than the allowable upper limit at the end of drying, and a temperature rise compensation is formed in the subsequent firing preparation stage. In this case, the production process control system writes the air volume deviation, residual moisture content, unconfirmed carbon footprint debt value, and temperature rise compensation into the debt attribution record.

[0079] Therefore, S410 can generate candidate actions for dynamic adjustment of air volume, cycle time, or preheating related energy consumption during the drying stage after molding, avoiding direct energy consumption reduction adjustment only for the current high-energy-consuming process.

[0080] S4 specifically includes the following sub-steps: S410, the production process control system reads the debt attribution record generated by S330, and constructs a set of candidate actions for dynamic energy consumption adjustment based on the preceding controllable nodes, control parameter deviations, state insufficiency types, energy consumption compensation methods, unconfirmed carbon footprint debt values, and quality risk performance.

[0081] The set of candidate actions for dynamic energy consumption adjustment refers to the set of alternative control actions that can correct the deviation of control parameters of the preceding controllable nodes and reduce the carbon footprint debt value. It includes process cycle adjustment actions, air volume distribution adjustment actions, preheating intensity adjustment actions, heating slope adjustment actions, heat preservation time adjustment actions, and waste heat utilization ratio adjustment actions.

[0082] Candidate actions are generated by mapping the state deficiency type and control parameter deviation in the debt attribution record: when the current state deficiency type is residual moisture and the control parameter deviation is insufficient air volume, candidate actions are generated to increase the drying air volume, extend the drying cycle time, or increase the proportion of waste heat preheating; when the current state deficiency type is insufficient preheating, candidate actions are generated to increase the preheating intensity, extend the preheating duration, or increase the proportion of waste heat utilization; when the current state deficiency type is not eliminated temperature difference, candidate actions are generated to adjust the air volume distribution, extend the temperature equalization cycle time, or reduce the subsequent temperature rise slope; when the current state deficiency type is cycle time compression, candidate actions are generated to restore the process cycle time, adjust the inlet interval, or correct the heat preservation time.

[0083] In one specific implementation, each action in the aforementioned set of candidate actions for dynamic energy consumption adjustment is a discrete step-size adjustment command.

[0084] The baseline step size for adjusting the air volume distribution is set to ±5% or ±10% of the rated air volume. The baseline step size for adjusting the cycle time of the process is set to ±5 min or ±10 min; The baseline step size for adjusting the preheating intensity and heating slope is set to ±0.5℃ / min or ±1℃ / min; the baseline step size for adjusting the waste heat recovery ratio is set to ±5%. The candidate action set is constructed by combining one or more of the above discrete step size actions to ensure that the underlying actuators (such as fan frequency converters, gas proportional valves, etc.) can accurately receive and respond to quantitative adjustment amplitudes.

[0085] The production process control system retrieves stored historical adjustment execution records and continuously updates the action selection priority using new adjustment execution records in subsequent S530 processes.

[0086] Reinforcement learning does not replace process rules, but rather updates the priority of action selection under the premise that candidate actions meet quality constraint boundaries. Its state is derived from debt attribution records, and its actions are candidate actions for dynamic energy consumption adjustment. Its feedback results include the reduction in carbon footprint debt value, the change in quality risk, and the change in the pressure on subsequent processes. The change in the pressure on subsequent processes refers to the change in the utilization of the remaining carbon footprint allowance, equipment load margin, and process adjustment margin of subsequent processes after the candidate action is executed, relative to before execution.

[0087] Feedback scores are calculated as follows:

[0088] in, The feedback score for the a-th candidate action; where a is the candidate action number; This represents the predicted reduction in carbon footprint debt value corresponding to the a-th candidate action, expressed in kgCO2e. Let be the predicted quality risk change corresponding to the a-th candidate action; This represents the predicted change in pressure on subsequent processes corresponding to the a-th candidate action. , , These are the weighting coefficients for the reduction of carbon footprint debt, changes in quality risk, and changes in subsequent burden.

[0089] It should be noted that, due to the predicted reduction in carbon footprint debt value... (Unit: kgCO2e) Predicted changes in quality risk (Dimensionless probability value) and predicted changes in pressure during subsequent processes The original dimensions of (dimensionless relative values) differ, and when substituted into the formula to calculate the feedback score... Previously, the production process control system used the extreme value standardization method to normalize each variable to the [0,1] interval.

[0090] The normalization calculation formula is:

[0091] in, X represents the dimensionless variable after normalization, and X represents the original variable to be processed. and These represent the maximum and minimum values ​​of the variable obtained statistically from the historical production database. In this embodiment, a specific set of values ​​for each weight coefficient is as follows: =0.5, =0.3, =0.2.

[0092] The resulting set of candidate actions for dynamic energy consumption adjustment is used in subsequent S420 to verify the quality constraint boundary and the carbon footprint acceptance capability of subsequent processes.

[0093] S420: The production process control system inputs the set of candidate actions for dynamic energy consumption adjustment obtained from S410 into the quality constraint boundary and the carbon footprint bearing capacity of subsequent processes for verification, and eliminates candidate actions that may lead to increased quality risks or transfer of carbon footprint debt, thus obtaining a set of executable adjustment actions.

[0094] Quality constraint boundaries refer to the quality and safety ranges determined by the production process control system based on the process formula library, historical qualified batch data, and the quality inspection system. These include the upper limit for drying crack risk, the upper limit for glaze defect risk, the allowable range for firing shrinkage deviation, the allowable range for body temperature difference, the allowable range for heating slope, and the allowable range for holding time. Quality risk prediction values ​​are derived from historical adjustment execution records, current batch quality feedback fields, and the allowable ranges in the planned process curve.

[0095] Quality constraints are determined by marking in the following manner:

[0096] in, The quality constraints for the a-th candidate action are defined by the marker. Let m be the predicted quality risk value for the a-th candidate action of the m-th class. This represents the upper limit of allowable quality risk for the m-th type; m is the quality risk type number; n is the total number of quality risk types. This indicates existence. If... The corresponding candidate action is eliminated.

[0097] Specifically, the predicted value of the m-th type of quality risk The probability of defect occurrence is calculated by constructing a multivariate logistic regression model based on the boundaries of historical qualified batches.

[0098] Taking the risk of drying cracks as an example, the system reads the deviation of the residual moisture content of the green body at the end of the current batch drying process. and the change in drying duration due to candidate action a The predicted risk value of drying cracks is calculated by substituting it into the following risk mapping equation. :

[0099] In the formula, e is the base of the natural logarithm (the natural constant). , , The regression coefficients, obtained by training based on historical repair and scrap records, are respectively set to values ​​of [values ​​to be filled in] in this embodiment. =-4.5, =2.1, =-1.3. When the calculated probability value... Risk exceeding the pre-set limit of the process If the value is set to 0.05, the action will be deemed to increase the quality risk and trigger the removal mechanism.

[0100] The carbon footprint carrying capacity of subsequent processes refers to the ability of subsequent processes to absorb changes in thermal state, cycle time, or waste heat utilization caused by preceding adjustment actions, without exceeding quality constraint boundaries, equipment load limits, and carbon footprint allowable ranges. Its data sources include the remaining carbon footprint allowable amount of subsequent processes, equipment load margin, allowable adjustment range of planned process curves, and historical adjustment results.

[0101] In one implementation, if a candidate action reduces the current carbon footprint by lowering the preheating intensity before glazing, but this would result in additional temperature compensation during the firing preparation stage, and the firing preparation stage is already close to the allowable range of carbon footprint, then the candidate action is eliminated due to excessive subsequent pressure. If another candidate action reduces the same carbon footprint debt value by increasing the proportion of waste heat utilization, without increasing the risk of drying cracks, glaze defects, or subsequent pressure, then this action enters the set of executable adjustment actions.

[0102] S430: The production process control system performs a comprehensive priority calculation on the set of executable adjustment actions obtained from S420, and selects the action with the highest comprehensive priority to generate a dynamic energy consumption adjustment command.

[0103] High-debt energy-saving actions refer to control actions that superficially reduce energy consumption in the current process, but lead to an increase in the carbon footprint debt value, quality risk, or rework risk of subsequent processes in historical adjustment execution records. The penalty value for high-debt energy-saving actions is used to reduce the overall priority of such actions. The overall priority is used to determine the execution order among multiple executable adjustment actions, taking into account the predicted reduction in carbon footprint debt value, the predicted change in quality risk, the predicted change in the pressure on subsequent processes, and the penalty value for high-debt energy-saving actions.

[0104] Overall priority is calculated as follows:

[0105] in, The overall priority of the a-th executable adjustment action; , , The meaning is the same as in S410; The penalty value for the high-debt energy-saving action corresponding to the a-th executable adjustment action; , , , These are the weighting coefficients for carbon footprint debt reduction, quality risk changes, subsequent burden changes, and penalties for high-debt energy-saving actions.

[0106] Similarly, in calculating the overall priority... When, in the formula , , and high debt energy saving action penalty value All of them were processed using the same extreme value standardization method as described above for dimensionless processing.

[0107] In this embodiment, a specific set of values ​​for the weighting coefficient of the overall priority is: =0.4, =0.3, =0.2, =0.1, and the penalty value The original base value is set in the range [0, 100] based on the frequency of secondary debt occurrence caused by this action in subsequent processes in historical statistics.

[0108] The production process control system incorporates the highest priority executable adjustment actions into the energy consumption dynamic adjustment instructions. These instructions include at least the execution process, execution time, adjustment target, adjustment range, duration, target debt reduction amount, quality constraint boundary, subsequent continuation restrictions, and prohibited action conditions. Prohibited action conditions are formed based on historical adjustment execution records. If a certain type of action, under the same or similar batch loading density, billet moisture content, and preceding controllable node conditions, results in no decrease in carbon footprint debt, increased quality risk, or an increase in carbon footprint debt in subsequent processes, then that type of action is included in the prohibited action conditions.

[0109] In one implementation, the batch loading density is close to 200 kg / m³. 3Furthermore, when the moisture content of the billet is high, historical records show that simply reducing the drying air volume increases the carbon footprint debt value during the firing preparation stage. Therefore, "reducing the drying air volume under the same loading density and moisture content" is written as a prohibited action condition. The resulting dynamic energy consumption adjustment command is then directly executed by the S510 and forms an adjustment execution record.

[0110] S5 specifically includes the following sub-steps: S510. Before the energy consumption dynamic adjustment command arrives at the execution time, the production process control system first verifies the execution conditions.

[0111] The execution conditions, including batch identifier, execution process, execution time, adjustment object, adjustment range, duration, quality constraint boundary, subsequent acceptance restrictions, and prohibited action conditions, are all consistent with those when the energy consumption dynamic adjustment instruction is generated in S430. If the current batch's billet moisture content, batch loading density, equipment load, quality risk performance, or subsequent process carbon footprint acceptance capacity exceeds the allowable range of the energy consumption dynamic adjustment instruction, the execution of the instruction is suspended, and the current batch status is fed back to S410 to S430 to generate a new energy consumption dynamic adjustment instruction.

[0112] If the execution conditions are met, the production process control system adjusts the corresponding process cycle time, air volume, preheating intensity, heating slope, heat preservation time or waste heat utilization ratio according to the energy consumption dynamic adjustment command, and simultaneously collects the actual control parameters, actual power consumption, actual gas consumption, actual waste heat utilization, process response status and quality feedback data after execution to form an adjustment execution record.

[0113] The adjustment execution record refers to the data record formed by the production process control system after executing the dynamic energy consumption adjustment command. It is used to save the adjustment object, adjustment range, execution time, actual energy consumption change, process response status and quality feedback results. It includes at least batch identifier, execution process, execution time, adjustment object, target adjustment range, actual adjustment range, duration, carbon footprint debt value before execution, actual electricity consumption after execution, actual gas consumption after execution, waste heat utilization after execution, process response status, quality feedback data and execution result mark.

[0114] In one implementation, if the energy consumption dynamic adjustment command requires the waste heat utilization ratio to be increased to the target value, but the actual waste heat supply is insufficient and only reaches 80% of the target value, then the adjustment execution record includes the target waste heat utilization ratio, the actual waste heat utilization ratio, the period of insufficient waste heat, the corresponding change in power consumption, and subsequent quality feedback, so that the subsequent S520 can distinguish between situations where the adjustment action is not fully executed and situations where the adjustment action is ineffective after execution.

[0115] S520: Based on the adjustment execution record formed in S510, the production process control system uses the same calculation method as S220 and S230 to recalculate the actual carbon footprint increment, process carbon footprint deviation, and carbon footprint debt value after adjustment, and compares them with the corresponding carbon footprint debt value in the carbon footprint debt account before adjustment.

[0116] The adjusted actual electricity consumption, actual gas consumption, and actual waste heat utilization will still be converted using the electricity carbon factor, gas carbon factor, and waste heat substitution carbon factor in S220; the adjusted carbon footprint debt value will still be registered according to the debt source process, subsequent undertaking process, insufficient status type, and debt attribution coefficient to ensure that the data before and after adjustment are comparable.

[0117] The production process control system calculates the reduction in carbon footprint debt after the execution of energy consumption dynamic adjustment commands as follows:

[0118] in, The measured reduction in carbon footprint debt after the execution of the u-th dynamic energy consumption adjustment command is expressed in kgCO2e; u is the command execution number. The carbon footprint debt value before the execution of the uth dynamic energy consumption adjustment instruction is expressed in kgCO2e. This represents the carbon footprint debt value after the execution of the u-th dynamic energy consumption adjustment command, expressed in kgCO2e. If... A value greater than 0 indicates that this adjustment has reduced the carbon footprint debt value; if A value less than or equal to 0 indicates that the carbon footprint debt value has not been reduced in this adjustment.

[0119] The production process control system further combines changes in quality risk and changes in pressure from subsequent processes to determine the effectiveness of adjustment actions.

[0120] in, This is a validity flag for the u-th dynamic energy consumption adjustment command. A value of 1 indicates a valid adjustment action, and a value of 0 indicates a restricted adjustment action. This represents the measured change in quality risk after the execution of the u-th dynamic energy consumption adjustment command. The measured change in pressure of subsequent processes after the execution of the u-th dynamic energy consumption adjustment command; This indicates a logical AND operation. This indicates a logical OR.

[0121] Effective adjustment actions refer to dynamic energy consumption adjustments that reduce carbon footprint debt without increasing quality risk or pressure on subsequent processes. Restrictive adjustment actions refer to dynamic energy consumption adjustments that do not reduce carbon footprint debt or increase quality risk or pressure on subsequent processes.

[0122] In one implementation, if the carbon footprint debt value decreases from 60 kg CO2e to 35 kg CO2e after a certain adjustment, and the risks of drying cracks, glaze defects, and subsequent bearing pressure do not increase, it is marked as an effective adjustment action; if another action decreases the carbon footprint debt value but increases the risk of firing shrinkage deviation, it is marked as a restrictive adjustment action.

[0123] S530, the production process control system feeds back the effective adjustment actions, limiting adjustment actions, updated carbon footprint debt accounts, adjustment execution records and corresponding batch quality feedback results obtained from S520 to the production process control model of subsequent batches, and updates the action selection priority under different batch states through reinforcement learning.

[0124] The state inputs for reinforcement learning include batch loading density, billet moisture content, preceding controllable nodes, control parameter deviation, state deficiency type, and carbon footprint debt value; the action inputs are the adjustment actions corresponding to the energy consumption dynamic adjustment commands; the feedback results include effectiveness markings, measured reduction in carbon footprint debt value, measured change in quality risk, and measured change in subsequent process pressure.

[0125] The reinforcement learning model employs a table-driven Q-learning algorithm. Its state input s is a vector composed of discretized state features. .

[0126] in, The batch loading density is 50 kg / m³ 3 The status codes corresponding to the interval being divided into 4 discrete intervals; The residual moisture content of the billet is classified into three levels: high, medium, and low, based on the degree of deviation from the upper limit of the process curve. This is the process identification number for the preceding controllable node; A direction marker for controlling parameter deviation (a value of 1 indicates too high, -1 indicates too low, and 0 indicates no deviation). The level code is the discretized value of the carbon footprint debt to be confirmed, with a step size of 20 kg CO2e.

[0127] The production process control system updates the action selection priority as follows:

[0128] in, The updated selection priority for the a-th candidate action in state s; is the selection priority before updating the a-th candidate action in state s; s is the state feature vector composed of batch loading density, billet moisture content, previous controllable nodes, control parameter deviation, state insufficiency type and carbon footprint debt value after discretization; a is the candidate action number. To learn step size; This is the execution feedback value of the u-th dynamic energy consumption adjustment command.

[0129] If the adjustment is a valid adjustment action, the execution feedback value is positive; if the adjustment is a limiting adjustment action, the execution feedback value is negative or 0.

[0130] During the initial cold start phase of the system, the initial action selection priority table (i.e., the Q table) contains all initial action selection priorities. All are uniformly initialized to 0.

[0131] The learning step size (learning rate) in the formula. The specific value range is 0.05. 0.2, in this embodiment, the specific value chosen is 0.2. =0.1.

[0132] Execution feedback value The quantitative given rule is: if the effectiveness label after this adjustment is... =1, then execute the feedback value If validity marker If the value is 0 and a quality defect is caused, then the feedback value is executed. =-50.

[0133] Through this adaptive iteration, the Q-values ​​(i.e. priorities) of the action control parameters in the Q table corresponding to low carbon footprint debt and high process safety will be continuously amplified, thereby achieving closed-loop dynamic optimization of the control parameters.

[0134] The reinforcement learning update results do not directly rewrite the entire planned process curve, but change the priority of selecting candidate actions in subsequent batches under the same or similar batch conditions; for control actions that are marked as restricted adjustment actions multiple times, the production process control system reduces their action selection priority and writes the prohibited action condition when the S430 prohibited action condition is met.

[0135] In one implementation, the batch loading density is close to 200 kg / m³. 3When the moisture content of the billet is high and the preceding controllable node is the drying stage after molding, if "increasing the drying air volume and increasing the proportion of waste heat utilization" is continuously marked as an effective adjustment action, then the selection priority of this action under the same conditions is increased; if "simply reducing the drying air volume" is continuously marked as a restrictive adjustment action, then its selection priority is reduced and a prohibited action condition is added. This allows the energy consumption control parameters of subsequent batches to be updated based on the changes in carbon footprint debt, forming a closed-loop process control for dynamic adjustment of energy consumption in the production of electrical porcelain insulators.

[0136] All the above formulas are performed using dimensionless numerical calculations; the relevant formulas are based on empirical models that approximate the real situation, obtained through extensive data collection and software simulation fitting. The preset parameters and thresholds involved in the formulas can be conventionally set and adjusted by those skilled in the art according to the physical constraints of the actual application scenario.

[0137] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0138] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0139] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback, characterized in that, Includes the following steps: S1. Establish batch process records with batch identifiers, collect process operation data and quality feedback data of electric porcelain insulator batches in drying after molding, preheating before glazing, firing preparation, firing holding, cooling recovery and process transfer, and generate batch carbon footprint control datasets. S2. Based on the batch carbon footprint control dataset, screen historical qualified batches, calculate the baseline carbon footprint, actual carbon footprint increment and process carbon footprint deviation for each process, and register the deviation portion transferred from insufficient previous state to subsequent process as carbon footprint debt value to be confirmed, forming a carbon footprint debt account. S3. Identify the attribution type of the current process energy consumption increase based on the carbon footprint debt account, determine the controllable nodes and control parameter deviations of the preceding process, and form a debt attribution record; S4. Generate a set of candidate actions for dynamic energy consumption adjustment based on debt attribution records. After verification by quality constraint boundaries and subsequent carbon footprint carrying capacity, obtain a set of executable adjustment actions and generate dynamic energy consumption adjustment instructions.

2. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 1, characterized in that, Also includes: S5. Execute the dynamic energy consumption adjustment command to form an adjustment execution record, recalculate the carbon footprint debt value and mark the effective adjustment action or the restriction adjustment action, and feed the updated carbon footprint debt account back to the subsequent batch production process control model.

3. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 1, characterized in that, S1 specifically includes: After the batch of electrical porcelain insulators enters the production process control system, a batch process record is established with the batch identifier as the main index and the process identifier and control time as auxiliary indexes. The product specifications, number of blanks, total weight of blanks, loading position, loading density, process flow sequence and planned process curve are written in. According to the batch process records, data on billet status, process execution, energy consumption, waste heat utilization, process waiting and quality inspection of each process are collected and compiled into process operation data according to the process control cycle.

4. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 3, characterized in that, Also includes: The batch process records, process operation data, and quality feedback data are aligned by batch identifier, process identifier, and collection time. Data on downtime, sensor failure, and non-production waiting are removed to generate a batch carbon footprint control dataset.

5. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 1, characterized in that, S2 specifically includes: Read product specifications, batch loading density, planned process curve, energy consumption field and quality feedback field from the batch carbon footprint control dataset, filter historical qualified batches that match the product specifications, loading density, process curve and quality results, and calculate the baseline carbon footprint of each process; The actual carbon footprint increment is calculated based on the actual power consumption, actual gas consumption, and actual waste heat utilization of each process in the current batch. The deviation of the process carbon footprint is obtained by comparing it with the benchmark carbon footprint. The deviation corresponding to the insufficient previous state is marked as the carbon footprint debt value to be confirmed. Register the carbon footprint debt value to be confirmed according to batch identification, debt source process, subsequent takeover process and control time to form a carbon footprint debt account.

6. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 1, characterized in that, S3 specifically includes: Read the actual carbon footprint increment, baseline carbon footprint, process carbon footprint deviation, and carbon footprint debt account; calculate the relative deviation rate of the current process's carbon footprint; and identify whether the increase in energy consumption in the current process is due to normal process heat absorption, current control deviation compensation, or repayment of debts left over from previous processes. When the process is not determined to be a normal process heat absorption, the preceding controllable nodes are determined based on the debt items pointing to the current process, the time of energy consumption increase, the response delay, the quality risk performance, and the carbon footprint debt value to be confirmed.

7. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 6, characterized in that, Also includes: Read the target parameters and actual execution parameters of the preceding controllable nodes, determine the control parameter deviation, and write the preceding controllable nodes, debt attribution strength, control parameter deviation, state insufficiency type, energy consumption compensation method, and attribution conclusion into the debt attribution record.

8. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 1, characterized in that, S4 specifically includes: Read the preceding controllable nodes, control parameter deviations, state insufficiency types, energy consumption compensation methods, and carbon footprint debt values ​​to be confirmed from the debt attribution records, map them to generate a set of candidate actions for dynamic energy consumption adjustment, and combine historical adjustment execution records and reinforcement learning to call the priority of candidate actions; The set of candidate actions for dynamic energy consumption adjustment is verified by inputting them into the quality constraint boundary and the carbon footprint bearing capacity of subsequent processes. Candidate actions that lead to increased quality risk or transfer of carbon footprint debt are eliminated to obtain the set of executable adjustment actions. Based on the predicted reduction in carbon footprint debt, the predicted change in quality risk, the predicted change in pressure on subsequent processes, and the penalty value for energy-saving actions with high debt, a comprehensive priority is calculated to generate dynamic energy consumption adjustment instructions.

9. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 2, characterized in that, S5 specifically includes: Before the energy consumption dynamic adjustment command arrives at the execution time, the batch identifier, execution process, adjustment object, adjustment range, quality constraint boundary and prohibited action conditions are verified. After the verification is passed, the command is executed, and actual control parameters, energy consumption, process response status and quality feedback data are collected to form an adjustment execution record. Based on the adjustment execution record, recalculate the adjusted carbon footprint debt value using the same calculation method as S2, compare the debt changes before and after adjustment, and mark effective adjustment actions or restricted adjustment actions.

10. The method for dynamic adjustment of energy consumption in the production of porcelain insulators based on carbon footprint feedback according to claim 9, characterized in that, Also includes: Effective adjustment actions, limiting adjustment actions, updated carbon footprint debt accounts, adjustment execution records, and quality feedback results are fed back to the subsequent batch production process control model, and action selection priorities are updated through reinforcement learning.

Images