Patent Application
20250208612
This application claims priority to Chinese Patent Application No. 202510098636.2, filed on Jan. 22, 2025, the contents of which are hereby incorporated by reference to its entirety.
This disclosure relates to the field of quality management technology, and in particular to, methods, systems, and media for sampling product quality based on Industrial Internet of Things (IIoT).
In the complex production and/or assembly process, the production and assembly work involve a plurality of types, specifications, and materials of parts and components, and these parts and components in the overall product take on their different functions and have different requirements for the quality inspection standards. In consideration of this situation, differentiated sampling standards and strengths need to be implemented for different parts.
Therefore, there is a need to provide methods, systems, and media for sampling product quality based on Industrial Internet of Things (IIoT), to perform targeted quality inspection on products or workpieces of a production line, intelligently adjust the conditions and/or strength of the sampling, and improve the efficiency of the quality inspection.
One or more embodiments of the present disclosure provide a method for sampling product quality based on Industrial Internet of Things (IIoT). The method is executed by an IIoT management platform of a system for sampling product quality based the IIoT. The system further comprises an IIoT user platform, an IIoT service platform, an IIoT sensing network platform, an IIoT management platform, and an IIoT perception and control platform. The IIoT management platform is communicatively connected to the IIoT sensing network platform and the IIoT service platform; and the IIoT perception and control platform comprises production equipment for each production link in a production line, a sampling device deployed in the production line, and a quality inspection equipment, and is configured to realize data interaction with the IIoT management platform through the IIoT sensing network platform. The method comprises: acquiring model data through a data center, and acquiring, by the IIoT perception and control platform, production data and equipment data through the IIoT sensing network platform; determining a first preset condition, a first sampling parameter, a second preset condition, and a second sampling parameter based on the model data, the production data, and the equipment data; obtaining, by the IIoT perception and control platform, production feature data through the IIoT sensing network platform, in response to the production feature data satisfying the first preset condition, generating a first sampling inspection instruction based on the first sampling parameter and sending the first sampling inspection instruction to the IIoT perception and control platform by the IIoT sensing network platform to control the sampling device to sample and deliver a first sample to the quality inspection equipment based on the first sampling inspection instruction; obtaining a first quality inspection result from the quality inspection equipment by the IIoT sensing network platform; obtaining historical sampling inspection results and historical model data through the data center, in response to at least one of the historical sampling inspection results and the historical model data satisfying the second preset condition, generating a second sampling inspection instruction based on the second sampling parameter and sending the second sampling inspection instruction to the IIoT perception and control platform by the IIoT sensing network platform to control the sampling device to sample and deliver a second sample to the quality inspection equipment based on the second sampling inspection instruction; obtaining a second quality inspection result from the quality inspection equipment by the IIoT sensing network platform; and adjusting the first preset condition, the first sampling parameter, the second preset condition, and the second sampling parameter, based on the first quality inspection result and the second quality inspection result.
One or more embodiments of the present disclosure provide a system for sampling product quality based on IIoT. The system comprises an IIoT user platform, an IIoT service platform, an IIoT management platform, an IIoT sensing network platform, and an IIoT perception and control platform, and the IIoT management platform comprises a data center; wherein the IIoT management platform is communicatively connected to the IIoT sensing network platform and the IIoT service platform; the IIoT perception and control platform comprises production equipment for each production link in a production line, a sampling device deployed in the production line, and a quality inspection equipment, and is configured to realize data interaction with the IIoT management platform through the IIoT sensing network platform; the sampling device is configured to sample a workpiece as a sample and transmit the sample to the quality inspection equipment, and the quality inspection equipment is configured to perform a quality inspection on the sample; and the IIoT management platform is configured to the method for sampling product quality based on IIoT.
One or more embodiments of the present disclosure provide a non-transitory computer-readable storage medium, the storage medium storing computer instructions, and the computer executing the method described in the embodiment when the computer reads the computer instructions in the storage medium.
This disclosure will be further illustrated by way of exemplary embodiments, which will be described in detail by means of the accompanying drawings. These embodiments are not limiting, and in these embodiments, the same numbering denotes the same structure, wherein:
In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings, which are required to be used in the description of the embodiments, are briefly described below. The accompanying drawings do not represent the entirety of the embodiments.
It should be understood that the terms “system”, “device”, “unit”, and/or “module” as used herein refers to a way to distinguish between different components, elements, parts, sections, or assemblies at different levels. The words may be replaced by other expressions if other words accomplish the same purpose.
Unless the context clearly suggests an exception, the words “a”, “one”, “an”, and/or “the” do not refer specifically to the singular, but may also include the plural. In general, the terms “including” and “comprising” only suggest the inclusion of explicitly identified steps and elements that do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.
In the embodiments of the present disclosure, the order of the steps is all interchangeable and the steps may be omitted if not otherwise indicated, and other steps may be included in the process of operation.
In some embodiments, as shown in
The IIoT user platform 110 refers to a platform for interacting with a user. In some embodiments, the IIoT user platform is configured as a terminal device, etc. The terminal device includes a tablet, a cell phone, or the like.
The IIoT service platform 120 refers to a platform that provides information and services to the user. In some embodiments, the IIoT service platform is configured as a server. The IIoT service platform may interact with the IIoT user platform.
The IIoT management platform 130 refers to a platform that manages information and/or data related to the system 100 for sampling product quality based on IIoT. In some embodiments, the IIoT management platform is communicatively connected to an IIoT sensing network platform and the IIoT service platform. The IIoT management platform includes a processor and a storage device, etc. The storage device includes a data center.
The data center pre-stores information and/or data related to the system 100 for sampling product quality based on IIoT, for example, model data, etc. The data center includes a plurality of databases. The database refers to a database management system that supports high-concurrency access.
The processor includes Central Processing Units (CPUs), Application Specific Integrated Circuits (ASICs), Application Specific Instruction Processors (ASIPs), etc., or any combination thereof.
In some embodiments, the IIoT management platform is configured to perform the method for sampling product quality based on IIoT, for more contents, please refer to the related descriptions of
The IIoT sensing network platform 140 refers to a platform for integrated management of sensing information. In some embodiments, the IIoT sensing network platform is configured as a communication network or a gateway, etc. The IIoT sensing network platform may interact with the IIoT management platform as well as an IIoT perception and control platform.
The IIoT perception and control platform 150 refers to a functional platform for perception information generation and control information execution. In some embodiments, the IIoT perception and control platform includes production equipment for each production link on a production line, a sampling device deployed on the production line, a plurality of monitoring devices, and a quality inspection equipment, and the IIoT perception and control platform is configured to realize data interaction with the IIoT management platform through the IIoT sensing network platform.
The plurality of monitoring devices refer to devices for obtaining monitoring data. In some embodiments, the monitoring devices include a thermometer, a hygrometer, an air quality monitor, a manometer, etc. The monitoring devices are deployed at any feasible location on the production line.
The production equipment refers to equipment that perform production activities. The production activities include, for example, producing or assembling workpieces, etc. In some embodiments, a workpiece includes a single part or a component assembled from a plurality of parts, etc. The production equipment includes automated assembly machines, etc.
For the sake of illustration, the following is an example of a production activity as assembly.
The sampling device refers to a device that extracts a workpiece as a sample and transmits the sample to the quality inspection equipment. In some embodiments, the sampling device includes a robotic arm, a conveyor belt, or the like.
The quality inspection equipment refers to an equipment used to perform quality inspection on samples. In some embodiments, the quality inspection equipment includes a material testing machine, etc.
A detailed description of the foregoing may be found in the descriptions of
The system for sampling product quality based on IIoT can form a closed loop of information operation between various functional platforms, coordinate and run regularly, and realize the informatization and intelligence of sampling product quality.
In 210, model data is obtained through a data center, and production data and equipment data are collected through an IIoT perception and control platform and obtained through an IIoT sensing network platform.
More descriptions regarding the system for sampling product quality based on IIoT and the various platforms may be found in
The model data refers to data related to the model of the workpiece. In some embodiments, the model data includes a type and a model of the workpiece, etc.
The production data refers to data related to the operation during a production process of a production line. In some embodiments, the production data includes, for example, the production rate of the production line, etc. The production rate refers to a rate at which the production line assemblies to obtain a workpiece.
The equipment data refers to data related to a production equipment. In some embodiments, the equipment data includes at least one of an equipment type, a current operating power, a rated parameter, an equipment operating time, an equipment maintenance cycle, and equipment malfunction records, etc. The rated parameters include the maximum operating power and the minimum operating power of the production equipment. The equipment maintenance cycle and the equipment malfunction records, etc. are entered into the IIoT perception and control platform by technicians after maintenance and troubleshooting of the production equipment.
In 220, a first preset condition, a first sampling parameter, a second preset condition, and a second sampling parameter are determined based on the model data, the production data, and the equipment data.
The first preset condition refers to a condition used to determine whether or not to perform the first type of sampling. The first type of sampling refers to a sampling performed by the sampling device controlled by the management platform in response to the first preset condition being satisfied.
In some embodiments, the first preset condition includes the cumulative production time satisfying the sampling cycle and/or the cumulative production quantity reaching the sampling threshold, etc.
The sampling parameter refers to a parameter associated with the extracted sample. The sample refers to a workpiece used for quality inspection. In some embodiments, the sampling parameter includes, for example, the sampling quantity and the sampling interval, etc. The sampling quantity refers to a count of extracted samples during a single round of sampling. The sampling interval refers to a count of workpieces spaced between adjacent samples during a single round of sampling.
The first sampling parameter refers to a sampling parameter corresponding to the first type of sampling.
The second preset condition refers to a condition used to determine whether or not to perform the second type of sampling.
The second type of sampling refers to a sampling performed by a sampling device controlled by the management platform in response to the second preset condition being satisfied.
In some embodiments, the second preset condition includes at least one of the count of non-conforming items in the historical sampling inspection results reaching a quantity threshold, and/or the historical model data being different from the model data, etc.
In some embodiments, the second preset condition further includes a plurality of feasible conditions such as the monitoring data not meeting an evaluation parameter, etc. In response to one or more of a plurality of second preset conditions being satisfied, the management platform determines to perform a second type of sampling. More descriptions regarding the monitoring data not meeting the evaluation parameter may be found in
The second sampling parameter refers to a sampling parameter corresponding to the second type of sampling.
In some embodiments, the management platform is configured to query the first initial condition and the first initial parameter corresponding to the model data in the model table based on the model data, the production data, and the equipment data. The management platform is configured to adjust, based on the production data and the equipment data, the first initial condition and the first initial parameter according to the first adjustment direction to obtain the first preset condition and the first sampling parameter.
The model table is pre-set based on historical experience and includes a plurality of pieces of the model data and a first initial condition, a first initial parameter, a second initial condition, and a second initial parameter corresponding to each piece of the model data.
The first adjustment direction is pre-set based on historical experience. For example, the faster the production rate, the longer the equipment operation time, or the longer the equipment maintenance cycle, then the count of samples is adjusted to be greater and the sampling interval is adjusted to be smaller, the sampling cycle is adjusted to be shorter, and the sampling threshold and the count threshold are adjusted to be smaller.
In some embodiments, the management platform is configured to determine the second preset condition and the second sampling parameter similarly as described above for determining the first preset condition and the first sampling parameter.
In 230, production feature data is collected by the IIoT perception and control platform and obtained via the IIoT sensing network platform.
The production feature data refers to data related to the operation of the production line.
In some embodiments, the production feature data includes a cumulative production time and a cumulative production quantity of the production line, etc. The cumulative production time refers to a total duration of production on the production line. The cumulative production quantity refers to a total count of workpieces produced by the production line.
In 240, in response to the production feature data satisfying the first preset condition, a first sampling inspection instruction is generated based on the first sampling parameter, and a first quality inspection result is obtained from a quality inspection equipment via the IIoT sensing network platform.
The first sampling inspection instruction refers to an instruction for instructing the sampling device to perform the first type of sampling. In some embodiments, the first sampling inspection instruction includes a robotic arm instruction and a conveyor belt instruction. The robotic arm instruction is used to instruct the robotic arm to extract a sample. The conveyor belt instruction is used to instruct the conveyor belt to transport the sample to the quality inspection equipment.
The robotic arm instruction includes sampling interval and sampling quantity, etc. the conveyor belt instruction includes a transportation rate of the conveyor belt, etc. The transportation rate of the conveyor belt is pre-set based on historical experience.
In some embodiments, the management platform is configured to send the first sampling inspection instruction to the IIoT perception and control platform via the IIoT sensing network platform to control the sampling device to extract and deliver a first sample to the quality inspection equipment based on the first sampling inspection instruction.
The first sample refers to a sample extracted by the sampling device based on the first sampling inspection instruction.
More descriptions regarding the first sampling inspection instruction may be found in
The first quality inspection result refers to a result obtained by the quality inspection equipment performing quality inspection on the first sample. In some embodiments, the first quality inspection result includes a pass rate of the first sample, etc.
In some embodiments, the management platform is configured to obtain the first quality inspection result from the quality inspection equipment via the IIoT sensing network platform.
In 250, historical sampling inspection results and historical model data are obtained through a data center.
The historical sampling inspection results refer to quality inspection results from the historical sampling process. The historical sampling inspection results include historical first quality inspection results and historical second quality inspection results. More descriptions regarding second quality inspection results may be found in operation 260 and relevant descriptions thereof
The historical model data refers to model data from the historical sampling process.
In 260, in response to the historical sampling inspection results and/or the historical model data satisfying a second preset condition, a second sampling inspection instruction is generated based on the second sampling parameter, and the second quality inspection result is obtained from the quality inspection equipment via the IIoT sensing network platform.
The second sampling inspection instruction refers to an instruction for instructing the sampling device to perform the second type of sampling. In some embodiments, the second sampling inspection instruction contains the same type of data as the first sampling inspection instruction, and the process for determining the second sampling inspection instruction is similar to the process for determining the first sampling inspection instruction.
In some embodiments, the management platform is configured to send the second sampling inspection instruction to the IIoT perception and control platform via the IIoT sensing network platform to control the sampling device to sampling and deliver a second sample to the quality inspection equipment based on the second sampling inspection instruction.
The second sample refers to a sample extracted by the sampling device based on the second sampling inspection instruction.
More descriptions regarding the second sampling inspection instruction may be found in
The second quality inspection result refers to a result obtained by the quality inspection equipment performing the quality inspection on the second sample. In some embodiments, the second quality inspection result includes a pass rate of the second sample, etc.
In some embodiments, the management platform is configured to obtain the second quality inspection result from the quality inspection equipment via the IIoT sensing network platform.
In some embodiments, the management platform is configured to independently determine whether the first preset condition or the second preset condition is satisfied. In order to prevent the time interval of the multiple rounds of sampling from being too short, the management platform may determine whether the time interval between two adjacent rounds of sampling is greater than or equal to a time threshold when executing the multiple rounds of sampling. If the time interval is less than the time threshold, the management platform does not execute the sampling inspection with a later execution order in the two adjacent rounds of sampling inspections. The time threshold is positively correlated to the count of samples or the sampling interval.
In 270, the first preset condition, the first sampling parameter, and the second preset condition, and the second sampling parameter are adjusted based on the first quality inspection result and the second quality inspection result.
In some embodiments, the management platform is configured to adjust the first preset condition, the first sampling parameter, and the second preset condition, and the second sampling parameter according to the second adjustment direction based on the first quality inspection result and the second quality inspection result. For example, in response to the quality inspection result of a single round of sampling being lower than the first threshold or the quality inspection results of multiple rounds of sampling being lower than the second threshold, the management platform adjusts the first preset condition, the first sampling parameter, and the second preset condition, and the second sampling parameter according to the second adjustment direction. In response to the fact that the quality inspection result of the single round of sampling is above a third threshold or the quality inspection results of the multiple rounds of sampling are all above a fourth threshold, the management platform adjusts the first preset condition, the first sampling parameter, and the second preset condition, and the second sampling parameter according to a third adjustment direction. The quality inspection results include the first quality inspection result or the second quality inspection result. The four thresholds described above are pre-set based on historical experience, for example, a size relationship is the first threshold<the second threshold<the fourth threshold<the third threshold.
The second adjustment direction is pre-set based on historical experience. For example, in response to the quality inspection results of a single round of sampling being below the first threshold or the quality inspection results of multiple rounds of sampling being below the second threshold, the count of samples is adjusted to increase, the sampling interval is adjusted to decrease, the sampling period is adjusted to shorten, and the sampling threshold and the quantity threshold are adjusted to reduce.
The third adjustment direction is pre-set based on historical experience. For example, in response to the quality inspection results of a single round of sampling being higher than the third threshold or the quality inspection results of multiple rounds of sampling being higher than a fourth threshold, the count of samples is adjusted to reduce, the sampling interval is adjusted to increase, the sampling period is adjusted to lengthen, and the sampling threshold and the quantity threshold are adjusted to increase. The magnitude of the adjustment is negatively correlated to the size of the quality inspection results.
In some embodiments, the management platform may be also configured to adjust the thresholds described above after each adjustment is made according to the second adjustment direction or the third adjustment direction. The management platform is also configured to obtain the thresholds adjusted by the technician. The manner in which the thresholds described above are adjusted may be preset.
In some embodiments, the management platform may preset an upper limit of adjustment and a lower limit of adjustment for the first preset condition, the first sampling parameter, the second preset condition, and the second sampling parameter. If the magnitude of the adjustment by the management platform reaches the upper limit of adjustment or the lower limit of adjustment, the adjustment is stopped.
By obtaining a large amount of production-related data, then targeted sampling conditions and sampling parameters are generated. By determining whether different sampling conditions are met, the workpiece can be more comprehensively sampled, and based on the results of the quality inspection, the sampling conditions and sampling parameters can be adjusted intelligently, which can effectively improve the efficiency of the quality inspection and reduce the manpower and cost.
In some embodiments, the management platform obtains monitoring data through a plurality of monitoring devices of the IIoT perception and control platform. The second preset condition further includes that the monitoring data does not satisfy an evaluation parameter.
The monitoring data refers to data related to the production environment. In some embodiments, the monitoring data includes at least one of an ambient temperature, an ambient humidity, an air miscellaneous quality, an ambient air pressure, etc. In some embodiments, the monitoring data is acquired by the plurality of monitoring devices and uploaded to the IIoT perception and control platform by the monitoring devices.
The evaluation parameter refers to a value range used to determine whether or not to conduct sampling. The monitoring data not meeting the evaluation parameter refers to the monitoring data exceeding the value range of the evaluation parameter. In some embodiments, different kinds of the monitoring data correspond to different evaluation parameters. The evaluation parameter is pre-set according to actual needs.
In some embodiments, the management platform determines influencing factor data 360 for workpiece quality based on workpiece feature data 310, model data 320, equipment data 330, and a link process parameter 340 through the factor determination model 350 and determine an evaluation parameter 370 based on the influencing factor data 360.
More descriptions regarding model data and equipment data may be found in operation 210 and relevant descriptions thereof.
Workpiece feature data refers to data related to the features of the workpiece. In some embodiments, the workpiece feature data includes material features, shape features, manufacturing manners, etc.
The material features include the physical properties and chemical composition of the material of the workpiece, etc.
The shape features include the shape and size of the workpiece, etc.
The manufacturing manners include machining or forging and molding, etc.
The link process parameter refers to data related to a production link of a manufacturing manner. In some embodiments, the link process parameter includes a plurality of production links that realize the manufacturing manner and parameters associated with the production links. For example, the manufacturing manner is forging and molding, and the link process parameters include production links such as discharging, heating, an amount of the discharging, and a temperature of the heating.
In some embodiments, the management platform obtains historical workpiece feature data and historical link process parameters corresponding to the model data as the workpiece feature data and the link process parameters through a data center.
The influencing factor data refers to data related to factors that influence the workpiece quality. In some embodiments, the influencing factor data includes an impact factor and an impact score corresponding to the impact factor, etc.
In some embodiments, the impact factors include a plurality of factors, and each impact factor corresponds to one type of data in the monitoring data. As it is not necessary to obtain all kinds of the monitoring data in actual production, the kinds of the monitoring data corresponding to the impact factors may not be present in the currently obtained monitoring data.
Impact scores are used to characterize the degree of influence of the impact factors on the workpiece quality. In some embodiments, the impact scores are expressed by a numerical value, e.g., 0-1, etc. The larger the numerical value, the greater the degree of influence of the impact factors on the workpiece quality.
The factor determination model refers to a model used to determine the influencing factor data. In some embodiments, the factor determination model is a machine learning model, for example, Multilayer Perceptron (MLP) model, or other customized model structures, etc., or any combination thereof.
In some embodiments, the management platform trains the factor determination model based on the training dataset by a gradient descent manner. The training dataset includes training samples and training labels corresponding to the training samples.
The training samples include sample workpiece feature data, sample model data, sample equipment data, and sample link process parameters, and the training labels include the actual influencing factor data corresponding to the training samples.
In some embodiments, the factor determination model is obtained by training in the following manner: inputting a plurality of training samples with the training labels into the initial factor determination model, constructing a loss function based on the training labels and the prediction results of the initial factor determination model, iteratively updating parameters of the initial factor determination model based on the loss function, and when the loss function of the initial factor determination model satisfies a preset condition, obtaining the factor determination model. The preset condition includes the loss function converging, the count of iterations reaching a set value, etc.
In some embodiments, the count of training samples corresponding to each sample link process parameter satisfies a sample count condition.
In some embodiments, the sample count condition includes that the count of training samples corresponding to each sample link process parameter is not less than a corresponding sample count threshold. The sample count threshold corresponding to each sample link process threshold is pre-set based on the link process difficulty, and the greater the link process difficulty, the greater the sample count threshold. See the description below for description of the link process difficulty.
In some embodiments, the quality inspection results may also include a failure rate of the sample.
In some embodiments, the management platform determines historical influencing factor data as training labels based on a correlation between the historical monitoring data and the failure rate in the historical quality inspection results during the sample time period. The sample time period refers to a time period during which the training samples are collected.
In some embodiments, the management platform determines, based on the fluctuation of the historical monitoring data and the change in the failure rate in the historical sampling inspection results during the sample time period, the correlation between the historical monitoring data and the failure rate in the historical sampling inspection results via a correlation algorithm. The correlation algorithm includes SPSS, SAS, etc.
In some embodiments, in response to a correlation between a type of historical monitoring data and the failure rate in the historical quality inspection result, the management platform determines that the type of historical monitoring data as a historical impact factor, and the impact score of the historical impact factor is positively correlated with the correlation. The management platform determines a plurality of historical impact factors and impact scores by the manner described above, so as to obtain the historical influencing factor data, and determines the historical influencing factor data as training labels.
The large amount of historical data is used to train the factor determination model, which can improve the efficiency of the model training, and at the same time, the generalization of the factor determination model can be improved by training the model with the training samples corresponding to the different kinds of the sample link process parameter.
In some embodiments, the management platform determines the evaluation parameter based on the influencing factor data. For example, the management platform identifies impact factors with impact scores greater than an impact threshold as relevant impact factors based on the influencing factor data. The impact threshold is negatively correlated to the pass rate of the historical sampling inspection results.
In some embodiments, the management platform determines a span of a value range of the evaluation parameter based on an impact score corresponding to the relevant impact factor. For example, the span of the value range of the evaluation parameter is negatively correlated to the impact score corresponding to the relevant impact factor. The span of the range of values is represented by the difference between the maximum and minimum of the value range.
By determining whether the monitoring data satisfies the evaluation parameter, the content of the second preset condition can be enriched, thereby more comprehensively determining whether to conduct a sampling inspection. At the same time, utilizing the factor determination model to determine the influencing factor data, and then determining the evaluation parameter, it can determine a more accurate evaluation parameter.
In some embodiments, the management platform adjusts the evaluation parameter based on the second quality inspection result and the link process difficulty.
The link process difficulty is used to characterize the complexity of the production link. In some embodiments, the link process difficulty is represented by a numerical value, for example, the larger the numerical value, the higher the link process difficulty. More descriptions regarding the link process difficulty may be found in operation 410 and relevant descriptions thereof.
In some embodiments, the management platform adjusts the evaluation parameter based on the second quality inspection result and the link process difficulty in a plurality of ways. For example, in response to the second quality inspection result of a single round of sampling being higher than a third threshold or the second quality inspection result of a plurality of rounds of sampling all being higher than a fourth threshold, the management platform expands a value range of the evaluation parameter.
As another example, the span of the value range of the evaluation parameter is negatively correlated to the link process difficulty. More descriptions regarding the third threshold and the fourth threshold may be found in operation 220 and relevant descriptions thereof.
The evaluation parameter is adjusted based on the second quality inspection result and the link process difficulty, which can make the determined evaluation parameter more in line with the actual production situation.
In some embodiments, the management platform determines a monitoring parameter based on the influencing factor data and a monitoring device status and generates a monitoring instruction based on the monitoring parameter. The management platform sends the monitoring instruction to the IIoT perception and control platform to control a plurality of monitoring devices to conduct monitoring based on the monitoring parameter.
The monitoring device status refers to an operating status of the monitoring device. For example, opening or closing of the monitoring device, etc. The management platform collects the monitoring device status via the IIoT perception and control platform and obtains the monitoring device status via the IIoT sensing network platform.
In some embodiments, the monitoring parameter includes at least one of a monitoring device needing to be turned on, a monitoring frequency of the monitoring device, an amount of data to be collected, etc.
In some embodiments, the management platform identifies, based on the influencing factor data and the monitoring device status, the monitoring device corresponding to the impact factor as a monitoring device needing to be turned on, and determines the monitoring frequency and the amount of data to be collected based on the impact score of the impact factor. The monitoring frequency and the amount of data to be collected are positively correlated to the impact score.
It can determine the monitoring device needing to be turned on, the monitoring frequency of the monitoring device, and the amount of data to be collected based on the influencing factor data, which in turn can collect more monitoring data that has a more significant impact on the workpiece quality, improve the effectiveness of the monitoring data, and contribute to accurate determination of whether to carry out sampling.
In 410, a first inspection parameter corresponding to a first preset condition and a second inspection parameter corresponding to a second preset condition are determined.
More descriptions regarding the first preset condition and the second preset condition may be found in
The first inspection parameter refers to a parameter used by the quality inspection equipment to inspect the first sample. In some embodiments, the first inspection parameter includes a quality inspection program and a hardware and software allocation volume, etc.
The quality inspection program includes quality inspection content and a quality inspection algorithm. The quality inspection content includes detecting whether the appearance of the workpiece is worn or not, and whether the mechanical properties are up to standard. The quality inspection content is pre-set based on actual needs.
The quality inspection algorithm is an algorithm used by the quality inspection equipment to perform inspection. The quality inspection algorithm also includes the amount of data required for applying the quality inspection algorithm.
The hardware and software allocation volume refers to a hardware volume and a software volume allocated to the quality inspection program. The hardware volume includes, for example, a count of the quality inspection equipment and/or computing equipment. The software volume includes data transmission bandwidth, etc. Computing devices are configured to perform calculations for quality inspection algorithms.
In some embodiments, the management platform determines a workpiece susceptibility degree based on workpiece feature data and determines the first inspection parameter based on the workpiece susceptibility degree. More description regarding the workpiece feature data may be found in
Workpiece susceptibility degree is used to characterize the degree to which a workpiece is susceptible to damage.
In some embodiments, the management platform determines, based on the workpiece feature data, the reference susceptibility degree as a workpiece susceptibility degree by querying a reference susceptibility degree corresponding to the workpiece feature data in the susceptibility data table. The susceptibility data table is pre-set based on historical data and includes a plurality of workpiece feature data and a reference susceptibility degree corresponding to each workpiece feature data. The reference susceptibility degree is obtained by a technician performing a damage experiment on a workpiece corresponding to the workpiece feature data. The damage experiment includes a stress fatigue experiment, etc.
In some embodiments, the management platform queries, based on the workpiece susceptibility degree, a reference first inspection parameter corresponding to the workpiece susceptibility degree in the historical parameter table, and determines the reference first inspection parameter as the first inspection parameter. The historical parameter table is pre-set based on historical data and includes a plurality of workpiece susceptibility degrees and a reference first inspection parameter and a reference second inspection parameter corresponding to each workpiece susceptibility degree. The management platform selects a historical production process corresponding to the workpiece susceptibility degree in the historical data and records the first inspection parameter and the second inspection parameter actually used in the historical production process into the historical parameter table.
The second inspection parameter refers to a parameter used by the quality inspection equipment to inspect a second sample.
In some embodiments, the second inspection parameter is similar to the first inspection parameter. The manner for determining the second inspection parameter is similar to the manner for determining the first inspection parameter.
In some embodiments, the first inspection parameter and the second inspection parameter are positively correlated to a link process difficulty. More descriptions regarding the link process difficulty may be found in
In some embodiments, the management platform determines the link process difficulty based on a number of link processes. For example, the higher the number of link processes, the higher the link process difficulty.
The number of link processes refers to a count of production links that produce the workpiece. In some embodiments, the management platform obtains a count of historical link processes corresponding to model data through a data center, and uses the count of historical link processes as the number of link processes.
Based on the number of link processes, it can determine a more accurate link process difficulty, which in turn improves the accuracy of determining the first inspection parameter and second inspection parameter.
In 420, in response to the production feature data satisfying a first preset condition, the first quality inspection instruction is generated based on the first inspection parameter.
In some embodiments, the first sampling inspection instruction further includes the first quality inspection instruction.
The first quality inspection instruction refers to an instruction for controlling the quality inspection equipment to perform quality inspection on the first sample based on the first inspection parameter.
In some embodiments, the management platform sends the first quality inspection instruction to the IIoT perception and control platform via the IIoT sensing network platform to control the quality inspection equipment to perform quality inspection on the first sample based on the first quality inspection instruction.
In 430, in response to historical sampling inspection results and/or historical model data satisfying the second preset condition, the second quality inspection instruction is generated based on the second inspection parameter.
More descriptions regarding the historical sampling inspection results and the historical model data may be found in
The second sampling inspection instruction further includes the second quality inspection instruction.
The second quality inspection instruction refers to a control instruction for controlling the quality inspection equipment to perform quality inspection of the second sample based on the second inspection parameter.
In some embodiments, the management platform sends the second quality inspection instruction to the IIoT perception and control platform via the IIoT sensing network platform to control the quality inspection equipment to perform quality inspection on the second sample based on the second quality inspection instruction.
Based on the first quality inspection instruction and the second quality inspection instruction, the quality inspection equipment can be targeted to be controlled to perform the quality inspection on workpieces corresponding to different sampling conditions, which is conducive to the improvement of the efficiency and accuracy of quality inspection.
In some embodiments, the management platform determines a first inspection accuracy based on a first inspection parameter, and a second inspection accuracy based on a second inspection parameter.
The first inspection accuracy refers to an accuracy with which the quality inspection equipment performs quality inspection on a first sample.
In some embodiments, the management platform constructs a quality inspection feature vector based on the first inspection parameter, model data, and quality inspection equipment data, matches a reference quality inspection vector that satisfies the matching condition in the vector database, and determines a label corresponding to the reference quality inspection vector as the first inspection accuracy. The matching condition includes having the highest similarity to the quality inspection feature vector. The vector similarity is negatively correlated to a vector distance. The vector distance includes Euclidean distance, etc.
The vector database is pre-set based on historical data and includes multiple reference quality inspection vectors and labels corresponding to the reference quality inspection vectors. The management platform constructs the reference quality inspection vectors based on the historical first inspection parameter, historical model data, and the historical quality inspection equipment data in the historical data and uses the actual first inspection accuracy as a label of the reference quality inspection vector. The actual first inspection accuracy refers to a accuracy of the historical quality inspection result corresponding to the historical first inspection parameter, and the accuracy of the historical sampling inspection result is determined by manual examination of the quality inspection results. More descriptions regarding the model data may be found in operation 210 and relevant descriptions thereof.
The quality inspection equipment data includes the type of the quality inspection equipment, the equipment maintenance cycle, or the like. In some embodiments, the management platform obtains the quality inspection equipment data through a data center.
The second inspection accuracy refers to an accuracy with which the quality inspection equipment performs a quality inspection of a second sample. In some embodiments, the second inspection accuracy is similar to the first inspection accuracy, and the manner for determining the second inspection accuracy is similar to the manner for determining the first inspection accuracy.
In some embodiments, the management platform may determine first fluctuation data based on a first quality inspection result, and second fluctuation data based on a second quality inspection result.
The first fluctuation data refers to data used to characterize fluctuations in the first quality inspection result. In some embodiments, the first fluctuation data is represented by, for example, the variance of a plurality of first quality inspection results.
In some embodiments, the management platform calculates the first fluctuation data based on the plurality of first quality inspection results.
The second fluctuation data refers to data used to characterize fluctuations in the second quality inspection result. In some embodiments, the second fluctuation data is similar to the first fluctuation data. The management platform calculates the second fluctuation data based on the plurality of second quality inspection results.
Understandably, there may be a plurality of rounds of sampling during a production process, resulting in a plurality of quality inspection results.
In some embodiments, in response to the fact that the first fluctuation data, the second fluctuation data, the first inspection accuracy, and the second inspection accuracy all satisfy a correction condition, the management platform generates a plurality of candidate sampling parameters based on the first inspection accuracy and the second inspection accuracy.
The correction condition refers to a condition for determining whether to update the sampling parameter and inspection parameter. In some embodiments, the correction condition includes that the first fluctuation data and the second fluctuation data are both greater than a fluctuation threshold, and that the first inspection accuracy and the second inspection accuracy are both less than an accuracy threshold. In some embodiments, the fluctuation threshold is negatively correlated to the workpiece susceptibility degree and a link process difficulty. The accuracy threshold is positively correlated to the workpiece susceptibility degree and the link process difficulty.
More descriptions regarding the workpiece susceptibility degree and the link process difficulty may be found in operation 410 and relevant descriptions thereof.
The candidate sampling parameters are parameters to be determined in relation to the sampling process. In some embodiments, the candidate sampling parameters include a set of a first sampling parameter and a second sampling parameter, and the first inspection parameter and the second inspection parameter.
In some embodiments, the management platform determines the plurality of candidate sampling parameters based on the first inspection accuracy and the second inspection accuracy. For example, the first sampling parameter and the second sampling parameter are unchanged, and the management platform randomly adjusts the first inspection accuracy and the second inspection accuracy, inversely determines, via the second vector database, the first inspection parameter and the second inspection parameter corresponding to the adjusted first inspection accuracy and the adjusted second inspection accuracy, and determines the obtained first inspection parameter and second inspection parameter, and the unchanged first sampling parameter and the unchanged second sampling parameter as a candidate sampling parameter.
As another example, the management platform inversely determines, via the second vector database, a first inspection parameter and a second inspection parameter corresponding to the adjusted first inspection accuracy and the adjusted second inspection accuracy, and adjusts, based on the obtained first inspection parameter and the obtained second inspection parameter, the first sampling parameter and the second sampling parameter. The management platform adjusts the sampling parameters based on the adjustment direction of the inspection parameter.
In some embodiments, the adjustment direction of the inspection parameter is to increase the hardware and software allocation volume, and the sampling parameter is correspondingly adjusted to shorten the sampling interval and increase the count of samples.
In some embodiments, the management platform may preset an upper limit of adjustment and a lower limit of adjustment for the first inspection parameter and the second inspection parameter. If the adjustment by the management platform reaches the upper limit of adjustment or the lower limit of adjustment, the adjustment is stopped.
In some embodiments, the management platform further determines, based on the first quality inspection result and the first fluctuation data, a first adjustment amplitude; determines, based on the second quality inspection result and the second fluctuation data, a second adjustment amplitude; adjusts, based on the first adjustment amplitude and the second adjustment amplitude, the first inspection accuracy and the second inspection accuracy.
The first adjustment amplitude refers to an amplitude for adjusting the first inspection accuracy. The second adjustment amplitude refers to an amplitude for adjusting the second inspection accuracy.
In some embodiments, the first adjustment amplitude and the second adjustment amplitude are positively correlated to the corresponding fluctuation data and negatively correlated to the corresponding quality inspection results.
Based on the fluctuation data of the quality inspection results and the historical sampling inspection results, it is possible to determine a more reasonable adjustment amplitude, which in turn allows for precise adjustments to the inspection accuracy.
In some embodiments, the management platform determines a quality inspection coverage 540 of a plurality of candidate sampling parameters 510 based on the plurality of candidate sampling parameters 510 and production process data 520 via a coverage model 530.
In some embodiments, the production process data includes a first preset condition, a second preset condition, a first quality inspection result, a second quality inspection result, a historical first quality inspection result, a historical second quality inspection result, workpiece feature data, model data, equipment data, etc. More description is provided in
The quality inspection coverage is used to characterize the quality inspection accuracy of the overall quality inspection results of a workpiece after sampling based on the candidate sampling parameters.
The coverage model refers to a model that is used to determine the quality inspection coverage. In some embodiments, the coverage model is machine learning model, e.g., Recurrent Neural Network (RNN) or other customized model structures, etc.
In some embodiments, the management platform may train the coverage model based on a large count of quality inspection training samples and corresponding quality inspection labels by a gradient descent manner. The quality inspection training samples include sample sampling parameters and sample production process data, the quality inspection labels include the actual quality inspection coverage corresponding to the quality inspection training samples.
In some embodiments, the management platform determines the quality inspection training samples and their quality inspection labels based on the historical data. For example, the management platform uses historical sampling parameters and historical production process data from the historical data as the quality inspection training samples and historical quality inspection coverage as the quality inspection labels.
In some embodiments, the management platform controls, in a single historical time period of the historical production process, the quality inspection equipment to re-inspect each workpiece corresponding to the quality inspection training sample to obtain an overall quality inspection result, compares the overall quality inspection result with the historical quality inspection result of the historical production process in the same historical time period, calculates the difference between the pass rate corresponding to the overall quality inspection results and the pass rate corresponding to the historical sampling inspection results, and determines the actual quality inspection coverage corresponding to the quality inspection training samples in the historical time period based on the difference. The smaller the difference, the higher the actual quality inspection coverage corresponding to the quality inspection training sample. The management platform takes the mean value of the quality inspection coverages as the actual quality inspection coverage corresponding to the quality inspection training sample by obtaining the quality inspection coverages corresponding to a plurality of historical time periods.
In some embodiments, the process of training the coverage model is similar to the process of training a factor determination model. For more description about training the coverage model, please see the related description of
In some embodiments, the management platform determines a composite sampling inspection score for the candidate sampling parameter based on the quality inspection coverage.
The composite sampling inspection score refers to a composite score for a candidate sampling parameter. The higher the composite sampling inspection score, the higher the efficiency or accuracy of sampling based on the candidate sampling parameter.
In some embodiments, the management platform determines a reference composite score corresponding to the quality inspection coverage as the composite sampling inspection score for the candidate sampling parameter based on the quality inspection coverage by querying the scoring table. The scoring table is pre-set based on historical data and includes a plurality of quality inspection coverages and a reference composite score corresponding to each quality inspection coverage. The reference composite scores are manually labeled by a technician based on actual sampling.
In some embodiments, the management platform performs a weighted summation of the quality inspection coverage, the single-piece quality inspection time, and the count of candidate samples in the candidate sampling parameter and takes the summation result as the composite sampling inspection score. The weights of the data in the weighted summation are pre-set based on historical experience. The management platform obtains the individual piece quality inspection time corresponding to the model data through the data center.
In some embodiments, the management platform updates the first sampling parameter, the first inspection parameter, the second sampling parameter, and the second inspection parameter based on the quality inspection coverage, the composite sampling inspection score, and the plurality of candidate sampling parameters. For example, the management platform uses the candidate sampling parameter with the highest composite sampling inspection score and the quality inspection coverage greater than the first coverage threshold as the update sampling parameter and uses each parameter in the update sampling parameter as a new inspection parameter and a new sampling parameter. The first coverage threshold is pre-set based on historical experience.
By determining a plurality of candidate sampling parameters and comprehensively evaluating the candidate sampling parameters based on their composite sampling inspection scores and quality inspection coverages, it can determine the candidate sampling parameter that best ensures the quality of sampling and update the existing sampling parameters and inspection parameters, making the quality inspection results more accurate.
In some embodiments, in response to the quality inspection coverage of the plurality of candidate sampling parameters satisfying a filtering condition, the management platform screens the plurality of candidate sampling parameters to obtain a target sampling parameter and adjusts the first preset condition and the second preset condition until the target sampling parameter satisfies an update condition.
The filtering condition refers to a condition used to filter the target sampling parameter. In some embodiments, the filtering condition includes that the quality inspection coverages corresponding to the plurality of candidate sampling parameters are greater than a first coverage threshold.
In some embodiments, in response to the quality inspection coverages corresponding to the plurality of candidate sampling parameters satisfying the filtering condition, the management platform eliminates the candidate sampling parameters having a quality inspection coverage less than a second coverage threshold and uses the remaining candidate sampling parameters as the target sampling parameter. The first coverage threshold is greater than the second coverage threshold.
In some embodiments, the first coverage threshold and the second coverage threshold are positively correlated to the workpiece cost and the link process difficulty. The workpiece cost is positively correlated to workpiece susceptibility degree.
In some embodiments, the management platform adjusts the first preset condition and the second preset condition by the manner of adjusting the first preset condition and the second preset condition in
By updating the first preset condition and the second preset condition, it is possible to carry out a more reasonable sampling in the subsequent sampling of the workpiece, which is conducive to reducing the workload of quality inspection and reducing the redundant data generated by quality inspection.
Some embodiments of the present disclosure further provide a non-transitory computer-readable storage medium, the storage medium storing computer instructions, and when the computer reads the computer instructions in the storage medium, the computer executes the method described in any of the above embodiments.
In addition, certain features, structures, or features of one or more embodiments of the present disclosure may be suitably combined.
Some embodiments use numbers describing the count of components or attributes, and it should be understood that such numbers used in the description of the embodiments are modified in some examples by the modifiers “about”, “approximately”, or “substantially”. Unless otherwise noted, the terms “about,” “approximately,” or “substantially” indicate that a ±20% variation in the stated count is allowed. Correspondingly, in some embodiments, the numerical parameters used in the disclosure and claims are approximations, which can change depending on the desired features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified count of valid digits and employ general place-keeping methods. While the numerical domains and parameters used to confirm the breadth of their ranges in some embodiments of this disclosure are approximations, in specific embodiments, such values are set to be as precise as practicable.
In the event of any inconsistency or conflict between the descriptions, definitions, and/or use of terms in the materials referenced in this specification and those described in this specification, the descriptions, definitions, and/or use of terms in this specification shall prevail.
| Number | Date | Country | Kind |
|---|---|---|---|
| 202510098636.2 | Jan 2025 | CN | national |
