Concrete pole production quality control and detection system

Abstract

The invention provides a concrete pole production quality control and detection system, which realizes accurate data acquisition, scientific abnormal value processing and dynamic iterative optimization, and comprises a parameter acquisition module for acquiring and preprocessing key process parameters; the quality detection module is used for judging and eliminating abnormal parameters in the preprocessed key process parameters, and calculating a production process qualification index in combination with a preset parameter weight; the intelligent decision regulation and control module is used for comparing the production process qualification index with a preset quality standard threshold value, determining the adjustment amount of each key process parameter according to a comparison result, and outputting a targeted production equipment parameter adjustment instruction; the execution feedback module is used for finishing key process parameter correction according to the parameter adjustment instruction and evaluating a parameter adjustment effect; if the adjustment effect reaches the standard, storing the corresponding adjustment parameter as an optimization template; if not, an evaluation result is fed back to trigger secondary adjustment and is synchronized to a parameter acquisition module, and closed-loop iteration is realized.

Description

Technical Field

[0001] This invention relates to the field of cement pole production technology, specifically to a cement pole production quality control and testing system. Background Technology

[0002] As a core supporting component for power transmission, new energy infrastructure, and urban and rural power distribution network construction, the quality stability and durability of cement poles directly affect the safety of power grid operation and the service life of infrastructure. With the advancement of 5G base station construction, new power system upgrades, and rural power grid consolidation and improvement projects, the market's performance requirements for cement poles are continuously upgrading. They not only need to meet the stringent requirements of high strength, corrosion resistance, and wind load resistance, but also set higher standards for product consistency, production efficiency, and green and low-carbon attributes. However, the cement pole manufacturing industry still faces many technical bottlenecks, which seriously restrict product quality upgrades and high-quality development of the industry. First, process parameter control is rudimentary, lacking automation and precision. Most small and medium-sized enterprises in the industry still rely on manual experience to regulate core processes, and the control of key parameters lacks standardized methods, leading to significant performance fluctuations within the same batch or even the same pole. Second, data processing and outlier identification capabilities are weak, resulting in lagging and inaccurate quality inspection. In traditional production, the collection of key process parameters relies on manual inspections or low-frequency sensor recordings, resulting in insufficient data integrity and a lack of effective preprocessing mechanisms. Simultaneously, quality inspection largely depends on offline sampling or manual visual inspection, which is not only inefficient but also fails to cover the entire production process, leading to a high rate of defective products. The defects were only discovered after entering the market, leading to increased rework costs and safety risks. Simultaneously, the production process lacks a complete closed loop from data collection, detection, decision-making, feedback to iteration. Adjustment effects cannot be verified in real time, and process parameters and optimization templates remain static for extended periods, making it difficult to adapt to dynamic factors such as raw material batch fluctuations, equipment aging, and changes in the production environment, resulting in poor product quality stability. In summary, existing cement pole production technology has significant shortcomings in precise control of process parameters, closed-loop detection throughout the entire process, and dynamic iterative optimization. Therefore, there is an urgent need for a quality closed-loop control and detection system that integrates precise data collection, scientific outlier handling, and dynamic iterative optimization. Summary of the Invention

[0003] In order to solve the technical problems mentioned in the background art, the purpose of this invention is to provide a cement pole production quality control and inspection system.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A quality control and inspection system for cement pole production, comprising: The parameter acquisition module collects and preprocesses key process parameters; The quality inspection module identifies and removes abnormal parameters from the key process parameters after preprocessing, and then calculates the production process qualification index by combining the preset parameter weights. The intelligent decision-making and control module compares the production process qualification index with the preset quality standard threshold, determines the adjustment amount of each key process parameter based on the comparison result, and outputs targeted production equipment parameter adjustment instructions. The execution feedback module completes the correction of key process parameters according to the parameter adjustment instructions and evaluates the effect of parameter adjustment. If the adjustment effect meets the standard, the corresponding adjusted parameters are stored as an optimization template. If the effect does not meet the standard, the evaluation result is fed back to trigger a second adjustment, and simultaneously synchronized to the parameter acquisition module to achieve closed-loop iteration.

[0005] Furthermore, the key process parameters include raw material ratio parameters, mixing process parameters, molding pressure parameters, curing temperature and humidity parameters, and finished product size parameters; the preprocessing includes deduplication, filling in missing values, format unification, physical boundary verification, and standardization.

[0006] Furthermore, the preprocessed key process parameters are grouped to obtain a subset of parameters corresponding to each category. Based on the sample size of each subset and the sample size classification criteria, a suitable normality test method is selected according to different sample types; raw material ratio parameters and finished product size parameters belong to small samples, and the appropriate method is... Shapiro-Wilk Inspection; the mixing process parameters, molding pressure parameters, and curing temperature and humidity parameters are within the range of medium-sized samples and are suitable. Kolmogorov-Smirnov Test: For parameters that conform to a normal distribution, use the Laida criterion to remove outliers; for parameters that do not conform to a normal distribution, use the interquartile range method to remove outliers.

[0007] Furthermore, the Raida criterion is expressed in the following formula:

[0008] in, Indicates the first k The parameters of the class parameter collected at time t; Indicates the first k Standardized mean of class parameters; Indicates the first k Standardized standard deviation of class parameters; If the parameter value is satisfied, it is marked as an outlier and removed. After removal, the valid data retained for each type of parameter forms a valid subset. The interquartile range method sorts and removes duplicates from the target parameter subset, then uses linear interpolation to calculate the first quartile, the third quartile, and the interquartile range. Subsequently, outlier boundaries are set according to parameter importance, and the sorted data is compared with the boundaries to identify outliers that exceed the range. These outliers are then verified in conjunction with production conditions and removed after confirmation. The remaining valid data for each type of parameter after removal forms a valid subset. For each type of valid subset, the ratio of the number of samples that meet the preset threshold of the process standard to the total number of valid samples is calculated to obtain the pass rate of that type of parameter. The formula for calculating the production process qualification index is:

[0009] in, Indicates parameter weights; Indicates the qualification index of the production process; Indicates the first k The pass rate of class parameters; k Indicates the first k Class parameters.

[0010] Furthermore, the production process qualification index is related to the preset quality standard threshold. To make a comparison, if If so, the production process is determined to be in a stable and qualified state; if If the process quality fails to meet the standard, the parameter adjustment process will be triggered. Calculate the first k The adjustment amount for the class parameter is calculated using the following formula:

[0011] in, Indicates the first k Adjustment amount for class parameters; Indicates the first k The sensitivity coefficient for adjusting class parameters; Indicates the first k Standard process values ​​for class parameters; Indicates the first k The average value of the class parameters; According to the k Adjustment amount of class parameters Generate targeted instructions for adjusting production equipment parameters and simultaneously generate structured traceability carriers; The structured traceability carrier is a 5-row, 4-column structured matrix. Each row corresponds to an adjustment record of a type of key process parameter, containing 4 core fields: parameter category, adjustment amount, corresponding equipment, and production batch.

[0012] Furthermore, the adjusted key process parameters are collected and preprocessed, outliers are removed, and the adjusted production process qualification index is calculated based on preset weights. ; The formula for calculating the effect evaluation value is:

[0013] in, Indicates the performance evaluation value; Evaluation values Compared with the preset optimization threshold The comparison triggers the following two types of decisions: like If the adjustment achieves the desired effect, then the parameter category and adjustment amount will be identified. and adjusting the sensitivity coefficient Store it as a reusable optimization template, and directly call the optimization template when similar quality fluctuations occur in the future; like Then the effect evaluation value E The adjusted key process parameters are synchronized with the intelligent decision-making and control module, triggering a secondary adjustment process. At the same time, a parameter monitoring priority update instruction is sent to the parameter acquisition module to increase the acquisition frequency of parameters whose adjustment effect has not met the standard.

[0014] Furthermore, the iteration mechanism of the optimization template is driven by both periodic triggering and event triggering. The triggering scenarios for the event are as follows: 1) Adjust the effect evaluation value after the template is called 3 times consecutively. ; 2) A batch of utility poles was found to have substandard quality indicators during random inspection or self-inspection, and the root cause of the problem was associated with a stored optimization template; 3) Changes in production conditions.

[0015] Furthermore, the optimization template iteration process is as follows: 1) Calculate the template effect decay rate using the following formula:

[0016] in, D Indicates the template effect decay rate; This represents the evaluation value of the adjustment effect when the template is first stored; This represents the average evaluation value of the call effectiveness over the past month; like If the template effect is significantly diminished, a full iteration is required; if Only local parameter adjustments are required; 2) Production feature update: Collect new production data after the triggered iteration and update the feature dimensions associated with the template; 3) Adjust the sensitivity coefficient according to the parameter weighting ratio of the new data; 4) Conduct trial production of the iterated new template to verify its control effect, call the new template to generate adjustment instructions, collect key process parameters of the trial production product, and calculate the adjusted pass index. ;like and If the verification passes, the model is considered valid; otherwise, it is rolled back to the production feature update stage to supplement missing features or adjust model parameters until the verification passes.

[0017] Compared with the prior art, the advantages of the present invention are as follows: 1. This invention collects key process parameters in real time through a high-frequency sensor network. At the same time, it adopts appropriate normality test methods for small and medium sample parameters, and uses the Raida criterion and interquartile range method to accurately remove outliers, effectively solving the quality fluctuation problem caused by rough parameter collection and data distortion in traditional production. 2. This invention constructs a closed-loop iterative mechanism throughout the entire process, with dual-drive updates to optimize templates, dynamically adapting to changes in production conditions; 3. This invention achieves full-chain traceability from raw material batches and process parameter adjustments to finished product quality through template version management and full-process data archiving, solving the pain points of traditional production where paper records are difficult to trace and accident responsibility cannot be accurately located; at the same time, through trial production verification and iterative rollback mechanisms, invalid templates are prevented from entering the production process, significantly reducing the risk of unqualified products flowing out. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the system workflow of the present invention; Figure 2 This is a schematic diagram of the intelligent decision-making and control module of the present invention; Figure 3 This is a schematic diagram of the execution feedback module of the present invention. Detailed Implementation

[0020] To achieve the above objectives, the present invention provides a quality control and inspection system for cement pole production. Please refer to the following technical solution. Figures 1 to 3The system includes the following modules: The parameter acquisition module collects and preprocesses key process parameters; The key process parameters include raw material ratio parameters, mixing process parameters, molding pressure parameters, curing temperature and humidity parameters, and finished product size parameters. During each batch of raw material preparation, the weight of each raw material is collected in real time by the electronic weighing scale in the batching silo, and the deviation between the preset ratio and the actual amount of raw material is recorded simultaneously. After each batch of preparation is completed, a raw material ratio collection form for that batch is automatically generated and stored in parallel with the batch number to obtain the raw material ratio parameters. After the mixing equipment is started, the speed sensor collects the rotation speed of the mixing shaft in real time and records it once every 2 seconds. The timer records the duration from the start to the stop of mixing. One minute before the end of mixing, the slump of the slurry is measured on-site by a slump meter. The mixing speed, mixing time, and slump data are bound to the corresponding mixing pot number to form a single pot mixing parameter record and obtain the mixing process parameters. After the centrifuge is started, pressure sensors are placed at both ends of the centrifuge mold to collect radial pressure in real time during the centrifugation process, recording it once every 5 seconds. The speed sensor collects the centrifugation speed synchronously, and the timer records the time from centrifugation start-up to shutdown. The data is associated with the number of the pole being formed to generate the forming parameter record for a single pole and obtain the forming pressure parameters. Before the curing kiln is started and heated, three sets of temperature and humidity sensors are evenly arranged inside the kiln. Throughout the entire cycle from heating up, maintaining constant temperature to cooling down, temperature and humidity data are automatically collected every 5 minutes. The data is synchronously linked to the batch number of the electric poles inside the kiln to form a batch curing environment parameter curve and obtain the curing temperature and humidity parameters. After demolding, the actual length of the entire pole is measured using a laser rangefinder. The diameter and wall thickness are measured at three sections (two ends and one middle) using a digital caliper. Two symmetrical points are measured at each section, and the average value is taken. The dimensional data is then linked to the finished product number of each pole, generating a finished product dimensional inspection record to obtain the finished product dimensional parameters. Sensor calibration and effectiveness verification are performed regularly. The calibration cycle for data acquisition equipment such as electronic weighing scales, mixing equipment, centrifuges, and curing kilns is set monthly. Calibration standards, data recording, and anomaly handling are recorded. If calibration fails, data acquisition is suspended and the equipment is replaced. Each sensor transmits raw data to the edge computing node via the RS485 bus Modbus RTU protocol. The edge computing node performs the parameter preprocessing operation, and the preprocessed data is uploaded to the local server of the intelligent decision-making and control module via industrial Ethernet. To eliminate redundancy, ensure data integrity, and resolve format conflicts, data cleaning is performed, including deduplication, missing value completion, and format standardization. Duplicate parameter values ​​are automatically identified and deleted using timestamps, production object identifiers, and parameter types as unique keys. In this embodiment, if the forming pressure of the same pole at the same time point is repeatedly collected twice, only the earliest valid value is retained to avoid redundant data interfering with subsequent calculations. For single-node missing values ​​caused by temporary sensor malfunctions, the average value from the same batch and process is used for completion. Then, the values ​​output by different sensors are uniformly converted to floating-point format and bound with timestamps and production object identifiers to ensure data structure consistency. Then, physical boundary verification is performed to initially screen invalid data, and invalid data that obviously exceeds the equipment range or process physical limits is removed in advance to reduce the deviation of subsequent statistical calculations; physical boundary thresholds are set for each type of key process parameter, and if the data exceeds the threshold range, it is marked as invalid data and removed; valid data is retained to form an initial screening valid dataset; Finally, parameters of different dimensions are mapped to a unified interval to eliminate the difference in dimensions. The min-max standardization method is used to map the initial screening effective data to the interval [0,1].

[0021] The quality inspection module identifies and removes abnormal parameters from the key process parameters after preprocessing, and then calculates the production process qualification index by combining the preset parameter weights. The preprocessed key process parameters were grouped into 5 categories to obtain a subset of parameters for each category. Based on the sample size of each subset and the sample size classification criteria, the industry standard for small samples is 3 to 50, and for medium samples it is 51 to 500. Appropriate normality tests are selected according to different sample types. Raw material ratio parameters and finished product size parameters belong to small samples, and appropriate normality tests are selected accordingly. Shapiro-Wilk The mixing process parameters, molding pressure parameters, and curing temperature and humidity parameters belong to a medium sample, and the appropriate normality test method is: Kolmogorov-Smirnov Test; determine whether it conforms to a normal distribution. For parameters that conform to a normal distribution, use the Laida criterion. For non-normal data, use the interquartile range method to remove outliers. For each subset of data, calculate the mean and standard deviation using the following formula:

[0022]

[0023] in, Indicates the first k Standardized mean of class parameters; Indicates the first k Standardized standard deviation of class parameters; This indicates the total number of time points at which this type of parameter was collected; Indicates the first k The parameters of the class parameter collected at time t; Indicates the time point at which this type of parameter was collected; The Raida criterion is used to identify outliers, and the formula is as follows:

[0024] If the parameter value is satisfied, it is marked as an outlier and removed. After removal, the valid data retained for each type of parameter forms a valid subset. Interquartile Range (IQR) is a core outlier removal technique for critical process parameters with non-normal distributions. It is suitable for parameters with non-normal characteristics such as material bridging, mold wear, and left or right skewness in the opening and closing of curing kiln doors due to fluctuations in production conditions. It does not rely on the data distribution pattern, but defines the reasonable range only through the quantile characteristics of the data itself, which is robust and not affected by extreme values. The target parameter subset is sorted and deduplicated, and then the first quartile, third quartile, and the difference between them (interquartile range) are calculated using linear interpolation. Then, outlier boundaries are set according to the importance of the parameters. The sorted data is compared with the boundaries to identify outliers that exceed the range. These outliers are verified in conjunction with production conditions and removed after confirmation. The remaining valid data for each type of parameter after removal form a valid subset. For each valid subset of data, the ratio of the number of samples meeting the process standard to the total number of valid samples is calculated to obtain the pass rate of that parameter. The process standard is a preset threshold; in this embodiment, the raw material ratio deviation is... Molding pressure fluctuation Deviation of stirring process parameters Maintenance temperature and humidity fluctuations and finished product size deviation If the conditions are met, the sample is considered a qualified sample; Based on preset parameter weights, the embedded processor of the quality inspection module calculates the production process qualification index. The calculation result is synchronized to the intelligent decision-making and control module via the OPC UA protocol. The formula is as follows:

[0025] in, Indicates parameter weights; This represents the production process qualification index, with a value range of [0,1]. The closer the value is to 1, the higher the quality of the process. Indicates the first k The pass rate of class parameters; k Indicates the first k Class parameters; The parameter weights are set based on the degree of influence of the process on the pole quality, satisfying a weight sum of 1. In this embodiment, the raw material ratio parameter... Stirring process parameters Molding pressure parameters Maintenance temperature and humidity parameters Finished product size parameters The raw material ratio parameters are the core foundation for pole strength and have the highest weight. The mixing process parameters, molding pressure parameters, and curing temperature and humidity parameters are key processes and have the next highest weight. The finished product size parameters are the final appearance and accuracy indicators and have a relatively low weight. The quantification and transformation of outlier boundary multipliers are based on preset parameter weights as the core quantification basis, combined with the interquartile range method as the basic multiplier, and the specific value matching the importance of the parameters is calculated through a linear mapping formula. The final boundary is then determined through production condition verification. The mapping rule based on parameter importance is: the higher the parameter weight, the higher the outlier boundary multiplier. m The smaller the value, the stricter the outlier detection. The formula for linear mapping is:

[0026] in, Represents the basic boundary multiplier; This represents the maximum weight value among all critical process parameters; represents the minimum multiplier threshold; m represents the outlier boundary multiplier.

[0027] The intelligent decision-making and control module compares the production process qualification index with the preset quality standard threshold, determines the adjustment amount of each key production parameter based on the comparison result, and outputs targeted parameter adjustment instructions for the production equipment. The intelligent decision-making and control module incorporates an industrial-grade PLC programmable logic controller and a local computing unit. Based on the production process qualification index combined with a preset quality standard threshold, a three-dimensional method is employed, including industry benchmark surveys, enterprise historical data statistics, and quality cost balance analysis. First, the average process qualification index of 30 leading domestic cement pole manufacturers is surveyed (0.82-0.88). Then, the qualification index of stable production batches from the pilot enterprise over the past year is statistically analyzed (0.84-0.86). Finally, through a cost-benefit model, the cost of reworking non-conforming products is balanced with the cost of achieving standard production to determine the quality standard threshold. The industry typically uses a specific quality standard threshold. That is, the production process qualification index The entire process is determined to be stable, from threshold comparison to adjustment calculation, then instruction generation, and finally record retention. like If the production process is deemed to be in a stable and qualified state, no parameter adjustment will be triggered; only a "Process Stability Record" will be generated and synchronized to the execution feedback module for filing. If the process quality fails to meet the standard, the parameter adjustment process is triggered, and the next step of adjustment calculation is initiated. For the five categories of key process parameters, based on the weight allocation and the current quality deviation, the adjustment amount for each category of parameters is calculated using the following formula:

[0028] in, Indicates the first k The adjustment amount for class parameters is as follows: positive values ​​adjust upwards towards the standard value, and negative values ​​adjust downwards. Indicates the first k The sensitivity coefficient of the class parameter adjustment is positively correlated with the weight; the higher the weight of the parameter, the higher the adjustment priority. Indicates the first k Standard process values ​​for class parameters; Indicates the first in the valid dataset k The average value of the class parameters, i.e., the first k The average level of the current actual performance of the class parameter; Adjusting the sensitivity coefficient requires appropriately increasing the slow response parameters. In this embodiment, the curing temperature and humidity parameters are increased by 20% to 0.24 based on a weight of 0.2, and the finished product size parameter is increased to 0.18. For the fast response parameters, they can be maintained or slightly reduced. In this embodiment, the raw material ratio parameter is reduced to 0.22, the stirring process parameter is reduced to 0.18, and the molding pressure parameter is reduced to 0.18. Adjustment amount based on the k-th type of parameter This generates targeted adjustment instructions for production equipment parameters: raw material proportioning parameters, outputting instructions to the batching system to adjust the material feeding amount, specifying the increase or decrease range of various raw materials; mixing process parameters, outputting instructions to the mixing equipment to adjust the mixing speed and mixing time, correcting the mixing parameters; molding pressure parameters, outputting instructions to the centrifugal molding equipment to adjust the speed and pressure, optimizing the centrifugal process; curing temperature and humidity parameters, outputting instructions to the curing kiln to adjust the curing environment; finished product size parameters, outputting instructions to the molding die to adjust the die precision, correcting dimensional deviations; parameter adjustment instructions must include the adjustment direction, adjustment range, and effective batch to ensure that the equipment can execute them directly. After generating production equipment parameter adjustment instructions, a structured traceability carrier is generated simultaneously to completely retain the key information of each adjustment. A 5-row, 4-column structured matrix is ​​generated, with each row corresponding to the adjustment record of a type of key process parameter, containing 4 core fields: The parameter categories are clearly marked as raw material ratio parameters, mixing process parameters, molding pressure parameters, curing temperature and humidity parameters, and finished product size parameters; Adjustment amount: Record the specific adjustment range of this parameter; The corresponding equipment records the production equipment that received the adjustment instructions; For each effective batch, record the production batch or pole number corresponding to the adjustment instruction to ensure that the adjustment action can be traced back to the specific production object.

[0029] The execution feedback module corrects key process parameters according to the parameter adjustment instructions and evaluates the effect of parameter adjustment. If the adjustment effect meets the standard, the corresponding adjusted parameters are stored as an optimization template. If the effect does not meet the standard, the evaluation result is fed back to trigger a second adjustment and is simultaneously synchronized to the parameter acquisition module to achieve closed-loop iteration. The instruction is issued to the corresponding production equipment, and the process parameters are corrected according to the adjustment amount: Based on the raw material proportioning parameters, an adjustment instruction for the feeding quantity is issued to the batching system, according to... Correct the raw material feed weight; issue adjustment instructions for mixing speed and mixing time to the mixing equipment based on the mixing process parameters, and proceed accordingly. Adjust the stirring speed and stirring time; for the molding pressure parameters, issue speed and pressure adjustment commands to the centrifugal molding equipment, and proceed accordingly. Adjust the rotation speed and pressure; issue temperature and humidity adjustment commands to the curing kiln according to the curing temperature and humidity parameters. Correct the constant temperature and humidity; for the finished product dimensional parameters, issue a mold precision adjustment command to the molding die, and proceed accordingly. Correct mold positioning deviation; Collect and preprocess the adjusted key process parameters, remove outliers, and calculate the adjusted production process qualification index based on preset weights. ; Comparison of production process qualification index before adjustment K Compared with the adjusted production process qualification index The formula for calculating the effect evaluation value is:

[0030] in, Indicates the effect evaluation value. The adjusted production process qualification index improves, resulting in a positive effect. If the production process qualification index does not improve or even decreases after the adjustment, the effect is negative. Preset optimization threshold The industry typically takes The effect evaluation value Compared with the preset optimization threshold The comparison triggers two types of decisions: like If the adjustment achieves the desired effect, then the parameter category and adjustment amount will be identified. and adjusting the sensitivity coefficient Store it as a reusable optimization template, and directly call the optimization template when similar quality fluctuations occur in the future; like Then the effect evaluation value E The adjusted key process parameters are synchronized with the intelligent decision-making and control module to trigger a secondary adjustment process. At the same time, a parameter monitoring priority update instruction is sent to the parameter acquisition module to increase the acquisition frequency of parameters that have not met the adjustment effect and enhance the granularity of subsequent parameters. To ensure that reusable optimized templates can continuously adapt to changes in production conditions, an optimized template iteration mechanism is established. Through a dual-drive approach of periodic triggering and event triggering, the template parameters are continuously corrected to ensure that they can still accurately match actual production needs in subsequent quality control. Automatically triggered at the end of each quarter, iterative processes are started by retrieving key process parameters and template call records from the past three months; or iterative processes can be initiated proactively when any of the following scenarios occur: 1) Adjust the effect evaluation value after the template is called 3 times consecutively. ; 2) A batch of utility poles was found to have substandard quality indicators during random inspection or self-inspection, and the root cause of the problem was associated with a stored optimization template; 3) Significant changes have occurred in production conditions; The optimized template iteration process is as follows: 1) Calculate the template effect decay rate using the following formula:

[0031] in, D Indicates the template effect decay rate; This represents the evaluation value of the adjustment effect when the template is first stored; This represents the average evaluation value of the call performance over the past month. like If the template effect is significantly diminished, a full iteration is required; if Only local parameter adjustments are required; 2) Production feature update: Collect new production data after triggering the iteration and update the feature dimensions associated with the template; if production conditions change, add the changed production conditions as associated features of the template; if equipment parameters drift, update the threshold range of equipment status parameters to ensure that the template matches the current production conditions. 3) Adjust the sensitivity coefficient according to the parameter weighting ratio of the new data; 4) Conduct trial production of the iterated new template to verify its control effect, call the new template to generate adjustment instructions, collect key process parameters of the trial production product, and calculate the adjusted pass index. ;like and If the verification is successful, the process is considered complete; otherwise, the process is reverted to the production feature update stage to supplement missing features or adjust model parameters until the verification is successful. During the trial production phase, 5-10 batches should be produced to ensure that the sample size is sufficient to cover variables such as equipment fluctuations and minor differences in raw materials. Template version management and archiving rules: Template version number naming rules adopt the format of V, iteration number and effective conditions to ensure rapid matching with production conditions; Historical templates are archived for at least 2 years, and the archived content includes the original key process parameters, iteration reasons and operating effect data to facilitate technical traceability and problem investigation; Template call priority is sorted by effective condition matching degree > version update time > effect evaluation value, and the latest template that fully matches the current production conditions is called first. 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 quality control and inspection system for cement pole production, characterized in that, include: The parameter acquisition module collects and preprocesses key process parameters; The quality inspection module identifies and removes abnormal parameters from the key process parameters after preprocessing, and then calculates the production process qualification index by combining the preset parameter weights. The intelligent decision-making and control module compares the production process qualification index with the preset quality standard threshold, determines the adjustment amount of each key process parameter based on the comparison result, and outputs targeted production equipment parameter adjustment instructions. The execution feedback module completes the correction of key process parameters according to the parameter adjustment instructions and evaluates the effect of parameter adjustment. If the adjustment effect meets the standard, the corresponding adjusted parameters are stored as an optimization template. If the effect does not meet the standard, the evaluation result is fed back to trigger a second adjustment, and simultaneously synchronized to the parameter acquisition module to achieve closed-loop iteration.

2. The system according to claim 1, characterized in that, The key process parameters include raw material ratio parameters, mixing process parameters, molding pressure parameters, curing temperature and humidity parameters, and finished product size parameters; the preprocessing includes deduplication, filling in missing values, format standardization, physical boundary verification, and standardization.

3. The system according to claim 2, characterized in that, The preprocessed key process parameters are grouped to obtain a subset of parameters that correspond one-to-one with each type of parameter. Based on the sample size of each subset, and combined with the sample size division criteria, select the appropriate normality test method according to different sample types. The raw material ratio parameters and finished product size parameters are for a small sample and are suitable for use. Shapiro-Wilk Inspection; the mixing process parameters, molding pressure parameters, and curing temperature and humidity parameters are within the range of medium-sized samples and are suitable. Kolmogorov-Smirnov Test: For parameters that conform to a normal distribution, use the Laida criterion to remove outliers; for parameters that do not conform to a normal distribution, use the interquartile range method to remove outliers.

4. The system according to claim 3, characterized in that, The Raida criterion is defined as follows: in, Indicates the first k The parameters of the class parameter collected at time t; Indicates the first k Standardized mean of class parameters; Indicates the first k Standardized standard deviation of class parameters; If the parameter value is satisfied, it is marked as an outlier and removed. After removal, the valid data retained for each type of parameter forms a valid subset. The interquartile range method sorts and removes duplicates from the target parameter subset, then uses linear interpolation to calculate the first quartile, the third quartile, and the interquartile range. Subsequently, outlier boundaries are set according to parameter importance, and the sorted data is compared with the boundaries to identify outliers that exceed the range. These outliers are then verified in conjunction with production conditions and removed after confirmation. The remaining valid data for each type of parameter after removal forms a valid subset. For each type of valid subset, the ratio of the number of samples that meet the preset threshold of the process standard to the total number of valid samples is calculated to obtain the pass rate of that type of parameter. The formula for calculating the production process qualification index is: in, Indicates parameter weights; Indicates the qualification index of the production process; Indicates the first k The pass rate of class parameters; k Indicates the first k Class parameters.

5. The system according to claim 4, characterized in that, The multiplier linear mapping of the outlier boundary is given by the following formula: in, Represents the basic boundary multiplier; This represents the maximum weight value among all critical process parameters; Indicates the minimum multiplier threshold; m Indicates the outlier boundary multiplier; The production process qualification index and the preset quality standard threshold To make a comparison, if If so, the production process is determined to be in a stable and qualified state; if If the process quality fails to meet the standard, the parameter adjustment process will be triggered. Calculate the first k The adjustment amount for the class parameter is calculated using the following formula: in, Indicates the first k Adjustment amount for class parameters; Indicates the first k The sensitivity coefficient for adjusting class parameters; Indicates the first k Standard process values ​​for class parameters; Indicates the first k The average value of the class parameters; According to the k Adjustment amount of class parameters Generate targeted instructions for adjusting production equipment parameters and simultaneously generate structured traceability carriers; The structured traceability carrier is a 5-row, 4-column structured matrix. Each row corresponds to an adjustment record of a type of key process parameter, containing 4 core fields: parameter category, adjustment amount, corresponding equipment, and production batch.

6. The system according to claim 5, characterized in that, Collect and preprocess the adjusted key process parameters, remove outliers, and calculate the adjusted production process qualification index based on preset weights. ; The formula for calculating the effect evaluation value is: in, Indicates the performance evaluation value; Evaluation values Compared with the preset optimization threshold The comparison triggers the following two types of decisions: like If the adjustment achieves the desired effect, then the parameter category and adjustment amount will be identified. and adjusting the sensitivity coefficient Store it as a reusable optimization template, and directly call the optimization template when similar quality fluctuations occur in the future; like Then the effect evaluation value E The adjusted key process parameters are synchronized with the intelligent decision-making and control module, triggering a secondary adjustment process. At the same time, a parameter monitoring priority update instruction is sent to the parameter acquisition module to increase the acquisition frequency of parameters whose adjustment effect has not met the standard.

7. The system according to claim 6, characterized in that, The iteration mechanism of the optimization template is driven by both periodic triggering and event triggering. The triggering scenarios for the event are as follows: 1) Adjust the effect evaluation value after the template is called 3 times consecutively. ; 2) A batch of utility poles was found to have substandard quality indicators during random inspection or self-inspection, and the root cause of the problem was associated with a stored optimization template; 3) Changes in production conditions.

8. The system according to claim 7, characterized in that, The optimization template iteration process is as follows: 1) Calculate the template effect decay rate using the following formula: middle, D Indicates the template effect decay rate; This represents the evaluation value of the adjustment effect when the template is first stored; This represents the average evaluation value of the call effectiveness over the past month; like If the template effect is significantly diminished, a full iteration is required; if Only local parameter adjustments are required; 2) Production feature update: Collect new production data after the triggered iteration and update the feature dimensions associated with the template; 3) Adjust the sensitivity coefficient according to the parameter weighting ratio of the new data; 4) The iterated new template is used in the trial production stage. The new template is called to generate adjustment instructions, key process parameters of the trial production product are collected, and the adjusted pass index is calculated. ;like and If the verification passes, the model is considered valid; otherwise, it is rolled back to the production feature update stage to supplement missing features or adjust model parameters until the verification passes.

Images