Smart factory production control method and system based on Internet of Things

Abstract

The invention discloses a smart factory production control method and system based on the Internet of Things, and relates to the technical field of smart factory production control, key quality parameters including humidity, temperature, vibration frequency and flow velocity in the production process are collected in real time through an Internet of Things sensor, and changes of various parameters in the production process can be accurately monitored. The real-time data acquisition mechanism solves the problem of quality monitoring lag in a traditional system, and provides accurate and timely production data support. The system automatically predicts the variation trend of production quality according to data collected in real time by establishing a dynamic quality model, and automatically adjusts production parameters including humidity and flow velocity. By means of automatic linkage regulation and control, errors caused by human intervention can be avoided, and the stability and consistency of the production process are improved. The system can simultaneously consider the linkage effect of a plurality of production parameters, including the relationship between humidity and flow velocity, and ensures the optimal balance of each parameter by adjusting the production parameters in real time.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent factory production control, in particular to an intelligent factory production control method and system based on the Internet of Things. BACKGROUND

[0002] Under the broad background of industrial automation and intelligent manufacturing, the Internet of Things technology is widely used to improve production efficiency, quality control, and equipment management. With the continuous development of intelligent manufacturing, more and more manufacturing enterprises begin to introduce Internet of Things systems to monitor and optimize the production process in real time. The intelligent factory production control system based on the Internet of Things improves the accuracy and stability of production through real-time data collection, intelligent analysis, and automatic adjustment. In this system, key quality parameters such as temperature, humidity, pressure, and speed in the production process are monitored in real time by sensors and fed back to the central control system to form a closed-loop adjustment to ensure the consistency of product quality.

[0003] Although modern manufacturing has gradually adopted Internet of Things technology to improve production efficiency, the current quality control method still has some significant problems. Traditional quality control mostly relies on manual intervention or batch monitoring, often ignoring the dynamic changes of real-time data in the production process, resulting in large fluctuations in production quality. The mutual influence between temperature and humidity, if not adjusted and optimized in real time, may lead to unstable product quality, and even cause waste.

[0004] Traditional quality control systems are difficult to make rapid adjustments in the case of real-time data changes. Many production lines still rely on manual judgment or adjustments based on historical experience, lacking real-time automated feedback mechanisms. Such phenomena lead to frequent quality fluctuations in production, when external environmental changes cannot be timely reflected in the control system, which will adversely affect the quality of the final product, thereby increasing the waste rate, production cost, and even affecting brand reputation and customer satisfaction. SUMMARY

[0005] To overcome the deficiencies of the prior art, the present application provides an intelligent factory production control method and system based on the Internet of Things, which solves the problems mentioned in the background art.

[0006] To achieve the above purpose, the present application is implemented by the following technical scheme: an intelligent factory production control system based on the Internet of Things, comprising a quality parameter acquisition module, a data preprocessing and transmission module, a multi-parameter dynamic modeling module, a real-time linkage control module, an effect feedback and adjustment module

[0007] and an optimization learning and model updating module;

[0008] The quality parameter acquisition module collects data in real time from the production process, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and forms a dataset W.

[0009] The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB.

[0010] The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q.

[0011] The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control.

[0012] The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect.

[0013] The optimization learning and model update module automatically adjusts the dynamic quality model based on feedback data of the control effect, and updates the parameter weights and control proportional coefficient Kp, integral coefficient Ki and derivative coefficient Kd using machine learning algorithms.

[0014] Preferably, the quality parameter acquisition module includes a data acquisition unit and a rate calculation unit;

[0015] The data acquisition unit collects the humidity Hm of the production material through a microwave humidity sensor, the ambient temperature Te through a temperature sensor, the vibration frequency Fv of the equipment through a vibration velocity sensor, and the fluid velocity Vf through an ultrasonic flow meter.

[0016] The rate calculation unit calculates the humidity change rate ΔHm and the flow rate change rate ΔVf based on the collected production material humidity Hm and fluid flow rate Vf; and fits the data collected by the data acquisition unit and the data collected by the rate calculation unit to obtain the dataset W.

[0017] The humidity change rate ΔHm is obtained using the following formula:

[0018]

[0019] In the formula, Hm(t) represents the humidity of the production material at time t, Hm(t-1) represents the humidity of the production material at time t-1, and Δt represents the sampling time interval;

[0020] The rate of change of flow velocity ΔVf is obtained by the following formula:

[0021]

[0022] In the formula, Vf(t) represents the fluid velocity Vf at time t, Vf(t-1) represents the fluid velocity Vf at time t-1, and Δt represents the sampling time interval.

[0023] Preferably, the data preprocessing and transmission module includes a data denoising unit and a data standardization and packetization unit;

[0024] The data denoising unit uses a filter to denoise the dataset W, eliminate high-frequency noise, and obtain the denoised data set WZ.

[0025] The data standardization and packaging unit uses Z-Score standardization to standardize the data denoising set WZ, transforming data of different dimensions to the same scale, obtaining the data standard set WB, and packaging it, which is then transmitted to the central processing unit via the network.

[0026] The dataset standard WB is obtained using the following formula:

[0027]

[0028] In the formula, WBd represents the data value of the d-th data in the standard dataset WB, WZd represents the data value of the d-th data in the denoised dataset WZ, μd represents the mean of the data value of the d-th data, and σd represents the mean and standard deviation of the data value of the d-th data.

[0029] Preferably, the multi-parameter dynamic modeling module includes a multi-parameter feature extraction unit and a dynamic quality model construction and scoring unit;

[0030] The parameter feature extraction unit extracts features from the data standard set WB and historical data, including humidity change rate ΔHm, average temperature Tavg, vibration frequency standard deviation σFv, and flow velocity change rate ΔVf, forming a feature set F; and analyzes the relationship between different features using the Pearson correlation coefficient.

[0031] The average temperature Tavg is obtained by the ratio of the sum of the ambient temperatures Te over all time periods to the value of the time period.

[0032] The standard deviation of the vibration frequency σFv is obtained by the following formula:

[0033]

[0034] In the formula, n represents the total number of vibration frequency Fv records, μFv represents the mean value of vibration frequency Fv, and Fva represents the vibration frequency value of the a-th record.

[0035] The formula for the Pearson correlation coefficient is:

[0036]

[0037] In the formula, rFxy represents the correlation coefficient of any feature xi and yi in feature set F, μx and μy represent the mean values ​​of features x and y respectively, m represents the number of data points in the time window, and xi and yi represent any feature in feature set F;

[0038] The dynamic quality model construction and scoring unit constructs a dynamic quality model based on the feature set F using a multivariate regression model, analyzes the impact of each feature on quality, and calculates the current quality score Q.

[0039] The quality score Q is obtained using the following formula:

[0040]

[0041] In the formula, Q(t) represents the quality score at time t, N1 represents the number of data, including the humidity of production materials Hm, ambient temperature Te, equipment vibration frequency Fv, and fluid velocity Vf, N2 represents the number of features, including the humidity change rate ΔHm, average temperature Tavg, vibration frequency standard deviation σFv, and velocity change rate ΔVf, fia(t) represents the value of the ia-th data at time t, gja(t) represents the interaction correlation between the ja-th features at time t, αia represents the preset weight value of the ia-th data, βja represents the preset weight value of the interaction correlation between the ja-th features, and C represents the constant term.

[0042] Preferably, the real-time linkage control module includes a quality scoring feedback processing unit and a linkage control and feedback control unit;

[0043] The quality scoring feedback processing unit receives the quality score Q and compares it with the preset scoring threshold TQ to determine the current production status and obtain the scoring deviation ΔQ.

[0044] The current production status is obtained by matching in the following way:

[0045] When the quality score Q ≥ the score threshold TQ, it indicates that the current production status is normal and no parameters need to be adjusted.

[0046] When the quality score Q < the score threshold TQ, it indicates that the current production status is abnormal. Calculate the score deviation ΔQ and adjust the parameters accordingly.

[0047] The scoring deviation ΔQ is obtained using the following formula:

[0048] ΔQ = TQ - Q;

[0049] Based on the scoring deviation ΔQ, adjust the production parameters, including fluid flow rate Vf and production material humidity Hm, and calculate the flow rate adjustment amount ΦVf and humidity adjustment amount ΦHm.

[0050] The flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm are obtained by the following formulas:

[0051] ΦVf=γv*ΔQ;

[0052] ΦHm=γh*ΔQ;

[0053] In the formula, γv and γh represent the adjustment coefficients for fluid velocity Vf and production material humidity Hm, respectively.

[0054] Preferably, the linkage control and feedback control unit adjusts the fluid flow rate Vf and the production material humidity Hm during the production process according to the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm, calculates and obtains the adjusted fluid flow rate VfN and the adjusted production material humidity HmN, and re-introduces them into the production process for production control.

[0055] The adjusted fluid velocity VfN and the adjusted production material humidity HmN are obtained using the following formulas:

[0056] VfN=Vf(t)+ΦVf;

[0057] HmN=Hm(t)+ΦHm;

[0058] In the formula, Vf(t) represents the fluid velocity Vf at time t, and Hm(t) represents the humidity Hm of the production material at time t;

[0059] During the adjustment process, the adjustment of fluid flow rate Vf and production material humidity Hm is limited by using a PID controller to avoid over-adjustment or under-adjustment and ensure the stability of the adjustment process.

[0060] The limiting formula is:

[0061]

[0062] In the formula, u(t) represents the adjustment control signal, ΔQ(t) represents the scoring deviation at time t, Kp represents the proportional coefficient, Ki represents the integral coefficient, and Kd represents the derivative coefficient. Let dt represent the integral term, dt represent the time period, and d represent the derivative. This represents the differential term.

[0063] Preferably, the effect feedback and adjustment module includes a quality monitoring and feedback acquisition unit and a control effect evaluation and feedback unit;

[0064] The quality monitoring and feedback acquisition unit is responsible for real-time monitoring of the adjusted production quality, recording the current quality score Q, and collecting the impact of the adjusted fluid flow rate VfN and the adjusted production material humidity HmN on the production quality.

[0065] The regulation effect evaluation and feedback unit analyzes the impact of the adjusted production parameters on quality based on real-time monitoring data and historical data, evaluates the regulation effect, and obtains the real-time deviation ΔQ(t).

[0066] The real-time deviation ΔQ(t) is obtained by the difference between the real-time quality score Q(t) and the score threshold TQ:

[0067] The real-time deviation ΔQ(t) is obtained by the following formula:

[0068] ΔQ(t) = TQ - Q(t);

[0069] The effects of fluid velocity Vf and production material humidity Hm on real-time quality score Q(t) were evaluated using machine learning algorithms.

[0070] The formula for the influence is:

[0071] ΔQ(t)=αV*ΦVf+αH*ΦHm;

[0072] In the formula, ΔQ(t) represents the real-time deviation, ΦVf represents the flow rate adjustment amount, ΦHm represents the humidity adjustment amount, and αV and αH represent the sensitivity coefficients of the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm to the production quality, respectively.

[0073] The calculation results are used to analyze the contribution of flow rate and humidity changes to quality and provide a basis for adjustment strategies; the current control effect is compared with historical data to evaluate the effectiveness of the adjustment strategy and generate optimization suggestions based on the results; if the quality score Q(t) reaches the expected target, the optimization suggestions will be used for the next production control; if the deviation is large, the coefficient or control method needs to be adjusted.

[0074] Preferably, the optimization learning and model includes a feedback data analysis and feature extraction unit and a model update and optimization unit;

[0075] The feedback data analysis and feature extraction unit evaluates the impact of the adjusted parameters on production quality based on the acquired real-time deviation ΔQ(t), flow rate adjustment amount ΦVf, and humidity adjustment amount ΦHm.

[0076] The real-time quality score Q(t), the adjusted fluid velocity VfN, and the adjusted production material humidity HmN are extracted from the feedback data to form a data set D, and the relationship between the PID controller adjustment parameters and the quality score Q is evaluated.

[0077] Preferably, the model update and optimization unit uses a machine learning algorithm to train a dynamic quality model based on data set D. After training, the machine learning model automatically adjusts the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd, making the system's adjustment effect more precise. The optimization objective of the controller parameters is to minimize the quality deviation ΔQ, and to update the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller, and recalculate the feedback signal to adjust the actual production parameters.

[0078] After updating the controller parameters, the performance of quality control is evaluated by verifying the effect of the new model in actual production; if the control effect does not meet expectations, the learning algorithm or model structure is adjusted until satisfactory results are obtained.

[0079] The evaluation methods include verifying the difference between model predictions and actual production quality through simulation and actual production data;

[0080] If the evaluation results show that the control accuracy has improved, the updated model will continue to be used for regulation.

[0081] A smart factory production control method based on the Internet of Things includes the following steps:

[0082] Step 1: The quality parameter acquisition module collects data in real time from the production process using IoT sensors, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and then forms a dataset W.

[0083] Step 2: The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB;

[0084] Step 3: The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q;

[0085] Step 4: The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control.

[0086] Step 5: The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect.

[0087] Step Six: Optimize the learning and model update module. Based on the feedback data of the regulation effect, use machine learning algorithms to automatically adjust the dynamic quality model and update the parameter weights and control proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd.

[0088] This invention provides a smart factory production control method and system based on the Internet of Things, which has the following beneficial effects:

[0089] (1) During system operation, key quality parameters, including humidity, temperature, vibration frequency, and flow rate, are collected in real time through IoT sensors, enabling precise monitoring of changes in various parameters during production. This real-time data acquisition mechanism solves the problem of lagging quality monitoring in traditional systems, providing accurate and timely production data support and a reliable basis for subsequent control. By establishing a dynamic quality model, the system automatically predicts the trend of production quality changes based on the real-time collected data and automatically adjusts production parameters, including humidity and flow rate. This automatic linkage control avoids errors caused by human intervention, improves the stability and consistency of the production process, and reduces the impact of human factors on quality control.

[0090] This system can simultaneously consider the interrelationship of multiple production parameters, including the relationship between humidity and flow rate, and ensure optimal balance among these parameters by adjusting them in real time. This multi-parameter linkage control avoids the drawbacks of traditional quality control methods that neglect the interrelationships between parameters, thus optimizing the quality stability of the production process.

[0091] (2) The quality parameter acquisition module collects key quality parameters in real time during the production process using multiple advanced sensors. This multi-sensor combination can comprehensively and accurately monitor various factors that may affect product quality during the production process, thereby improving the accuracy and completeness of the data. The rate calculation unit calculates the humidity change rate ΔHm and the flow rate change rate ΔVf based on the real-time collected humidity and flow rate data, providing immediate feedback for dynamic adjustment of production parameters. By calculating the change rate, the system can more sensitively capture minute changes that may cause quality fluctuations during the production process, thereby making timely adjustments and reducing the impact of quality fluctuations.

[0092] (3) Key features such as humidity change rate, average temperature, vibration frequency standard deviation, and flow velocity change rate are extracted from the data standard set and historical data through a multi-parameter feature extraction unit to construct a feature set F. This method comprehensively captures various parameters in the production process, especially the changing trends of humidity and flow velocity, providing a rich source of information for the subsequent construction of a dynamic quality model, making quality control more in-depth and accurate. In addition, the feature extraction process also includes Pearson correlation coefficient analysis, which can identify the correlation between different features, providing a scientific basis for the establishment of an accurate quality model.

[0093] (4) Through the quality monitoring and feedback acquisition unit, the system can monitor the adjusted production quality in real time, obtain the current quality score Q and the impact of the adjusted production parameters on quality. This process can comprehensively reflect the effect of production adjustments and provide data support for subsequent optimization. Compared with traditional quality control methods, this real-time monitoring and feedback mechanism significantly improves the quality control capability in the production process, enabling the system to respond to quality fluctuations at any time during the production process, ensuring the consistency and stability of product quality.

[0094] By using machine learning algorithms, the system can automatically assess the impact of flow rate and humidity adjustments on the production quality score Q(t). Based on the real-time deviation ΔQ(t) and changes in flow rate and humidity adjustments, the system can identify which parameters have a significant impact on product quality, thereby enabling precise optimization and adjustment. Attached Figure Description

[0095] Figure 1 This is a schematic diagram of the block diagram of a smart factory production control system based on the Internet of Things according to the present invention.

[0096] Figure 2 This is a schematic diagram illustrating the steps of a smart factory production control method based on the Internet of Things according to the present invention. Detailed Implementation

[0097] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0098] Example 1

[0099] This invention provides a smart factory production control method and system based on the Internet of Things (IoT). Please refer to [link / reference]. Figure 1 It includes a quality parameter acquisition module, a data preprocessing and transmission module, a multi-parameter dynamic modeling module, a real-time linkage control module, and an effect feedback and adjustment module.

[0100] And the optimization learning and model update module;

[0101] The quality parameter acquisition module collects data in real time from the production process, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and forms a dataset W.

[0102] The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB.

[0103] The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q.

[0104] The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control.

[0105] The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect.

[0106] The optimization learning and model update module automatically adjusts the dynamic quality model based on feedback data of the regulation effect using machine learning algorithms.

[0107] In this embodiment, key quality parameters during the production process, including humidity, temperature, vibration frequency, and flow rate, are collected in real time using IoT sensors, enabling precise monitoring of changes in multiple parameters. This real-time data acquisition mechanism solves the problem of lagging quality monitoring in traditional systems, providing accurate and timely production data support and a reliable basis for subsequent adjustments. The system establishes a dynamic quality model, automatically predicts the trend of production quality changes based on the real-time collected data, and automatically adjusts production parameters, including humidity and flow rate. This automatic linkage control avoids errors caused by human intervention, improves the stability and consistency of the production process, and reduces the impact of human factors on quality control.

[0108] This system can simultaneously consider the interrelationship of multiple production parameters, including the relationship between humidity and flow rate, and ensure optimal balance among these parameters by adjusting them in real time. This multi-parameter linkage control avoids the drawbacks of traditional quality control methods that neglect the interrelationships between parameters, thus optimizing the quality stability of the production process.

[0109] By automatically adjusting production parameters and dynamically optimizing the control model, the system can significantly reduce the scrap rate caused by parameter fluctuations. Real-time monitoring and control help maintain the production process in optimal operating condition, thereby improving the consistency and stability of product quality and ensuring high-quality standards for each batch. Through machine learning optimization and model updates, the quality control model can be continuously adjusted and optimized based on feedback from the control effects, gradually improving control accuracy. As data accumulates and learning deepens, the system's control capabilities will gradually improve, enabling it to better cope with complex and dynamic changes in the production process.

[0110] Example 2

[0111] This embodiment is an explanation based on Embodiment 1. Please refer to it. Figure 1Specifically: the quality parameter acquisition module includes a data acquisition unit and a rate calculation unit;

[0112] The data acquisition unit collects the humidity Hm of the production material through a microwave humidity sensor, the ambient temperature Te through a temperature sensor, the vibration frequency Fv of the equipment through a vibration velocity sensor, and the fluid velocity Vf through an ultrasonic flow meter.

[0113] The rate calculation unit calculates the humidity change rate ΔHm and the flow rate change rate ΔVf based on the collected production material humidity Hm and fluid flow rate Vf; and fits the data collected by the data acquisition unit and the data collected by the rate calculation unit to obtain the dataset W.

[0114] The humidity change rate ΔHm is obtained using the following formula:

[0115]

[0116] In the formula, Hm(t) represents the humidity of the production material at time t, Hm(t-1) represents the humidity of the production material at time t-1, and Δt represents the sampling time interval;

[0117] The rate of change of flow velocity ΔVf is obtained by the following formula:

[0118]

[0119] In the formula, Vf(t) represents the fluid velocity Vf at time t, Vf(t-1) represents the fluid velocity Vf at time t-1, and Δt represents the sampling time interval.

[0120] The data preprocessing and transmission module includes a data denoising unit and a data standardization and packetization unit;

[0121] The data denoising unit uses a filter to denoise the dataset W, eliminate high-frequency noise, and obtain the denoised data set WZ.

[0122] The data standardization and packaging unit uses Z-Score standardization to standardize the data denoising set WZ, transforming data of different dimensions to the same scale, obtaining the data standard set WB, and packaging it, which is then transmitted to the central processing unit via the network.

[0123] The dataset standard WB is obtained using the following formula:

[0124]

[0125] In the formula, WBd represents the data value of the d-th data in the standard dataset WB, WZd represents the data value of the d-th data in the denoised dataset WZ, μd represents the mean of the data value of the d-th data, and σd represents the mean and standard deviation of the data value of the d-th data.

[0126] In this embodiment, the quality parameter acquisition module collects key quality parameters in real time during the production process using multiple advanced sensors. This multi-sensor combination enables comprehensive and accurate monitoring of various factors that may affect product quality during production, thereby improving the accuracy and completeness of the data. The rate calculation unit calculates the humidity change rate ΔHm and flow rate change rate ΔVf based on the real-time collected humidity and flow rate data, providing immediate feedback for dynamically adjusting production parameters. By calculating the change rates, the system can more sensitively detect minute changes during production that may lead to quality fluctuations, allowing for timely adjustments and reducing the impact of these fluctuations.

[0127] The data denoising unit uses filters to denoise the dataset, effectively eliminating high-frequency noise during the acquisition process and ensuring data quality. Filtering removes unnecessary noise, ensuring that the acquired data more accurately reflects the actual fluctuations in quality parameters during production. This is crucial for ensuring the accuracy of subsequent quality model establishment and control. The data standardization and packaging unit employs the Z-Score standardization method to convert data of different dimensions to the same scale, thereby eliminating inconsistencies caused by differences in dimensions between different data. Standardization allows sensor data from different sources to be analyzed and compared on the same scale, ensuring data comparability and thus improving the accuracy of multi-parameter dynamic modeling and quality control.

[0128] The standardized dataset is effectively packaged and transmitted over the network to the central processing unit, enabling rapid data transmission and efficient processing. This transmission method reduces information delivery time, allowing the system to respond in real time to changes occurring during production and quickly implement control measures, thus improving the responsiveness and real-time performance of production control.

[0129] Through real-time data acquisition, computation, noise reduction, and standardization, the system can acquire high-quality data promptly and perform real-time analysis, thereby continuously optimizing quality control strategies. Each step of data processing and optimization makes the system more adaptable, enabling it to cope with different changes in the production process and maintain the consistency and stability of product quality.

[0130] Example 3

[0131] This embodiment is an explanation based on Embodiment 2. Please refer to it. Figure 1 Specifically: the multi-parameter dynamic modeling module includes a multi-parameter feature extraction unit and a dynamic quality model construction and scoring unit;

[0132] The parameter feature extraction unit extracts features from the data standard set WB and historical data, including humidity change rate ΔHm, average temperature Tavg, vibration frequency standard deviation σFv, and flow velocity change rate ΔVf, forming a feature set F; and analyzes the relationship between different features using the Pearson correlation coefficient.

[0133] The average temperature Tavg is obtained by the ratio of the sum of the ambient temperatures Te over all time periods to the value of the time period.

[0134] The standard deviation of the vibration frequency σFv is obtained by the following formula:

[0135]

[0136] In the formula, n represents the total number of vibration frequency Fv records, μFv represents the mean value of vibration frequency Fv, and Fva represents the vibration frequency value of the a-th record.

[0137] The formula for the Pearson correlation coefficient is:

[0138]

[0139] In the formula, rFxy represents the correlation coefficient of any feature xi and yi in feature set F, μx and μy represent the mean values ​​of features x and y respectively, m represents the number of data points in the time window, and xi and yi represent any feature in feature set F;

[0140] The dynamic quality model construction and scoring unit constructs a dynamic quality model based on the feature set F using a multivariate regression model, analyzes the impact of each feature on quality, and calculates the current quality score Q.

[0141] The quality score Q is obtained using the following formula:

[0142]

[0143] In the formula, Q(t) represents the quality score at time t, N1 represents the number of data, N2 represents the number of features, fia(t) represents the value of the ia-th data at time t, gja(t) represents the interaction correlation between the ja-th features at time t, αia represents the preset weight value of the ia-th data, βja represents the preset weight value of the interaction correlation between the ja-th features, and C represents the constant term.

[0144] The real-time linkage control module includes a quality scoring feedback processing unit and a linkage control and feedback control unit.

[0145] The quality scoring feedback processing unit receives the quality score Q and compares it with the preset scoring threshold TQ to determine the current production status and obtain the scoring deviation ΔQ.

[0146] The current production status is obtained by matching in the following way:

[0147] When the quality score Q ≥ the score threshold TQ, it indicates that the current production status is normal and no parameters need to be adjusted.

[0148] When the quality score Q < the score threshold TQ, it indicates that the current production status is abnormal. Calculate the score deviation ΔQ and adjust the parameters accordingly.

[0149] The scoring deviation ΔQ is obtained using the following formula:

[0150] ΔQ = TQ - Q;

[0151] Based on the scoring deviation ΔQ, adjust the production parameters, including fluid flow rate Vf and production material humidity Hm, and calculate the flow rate adjustment amount ΦVf and humidity adjustment amount ΦHm.

[0152] The flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm are obtained by the following formulas:

[0153] ΦVf=γv*ΔQ;

[0154] ΦHm=γh*ΔQ;

[0155] In the formula, γv and γh represent the adjustment coefficients for fluid velocity Vf and production material humidity Hm, respectively.

[0156] The linkage control and feedback control unit adjusts the fluid flow rate Vf and the production material humidity Hm during the production process according to the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm, calculates and obtains the adjusted fluid flow rate VfN and the adjusted production material humidity HmN, and re-introduces them into the production process for production control.

[0157] The adjusted fluid velocity VfN and the adjusted production material humidity HmN are obtained using the following formulas:

[0158] VfN=Vf(t)+ΦVf;

[0159] HmN=Hm(t)+ΦHm;

[0160] In the formula, Vf(t) represents the fluid velocity Vf at time t, and Hm(t) represents the humidity Hm of the production material at time t;

[0161] During the adjustment process, the adjustment process of fluid flow rate Vf and production material humidity Hm is limited by using a PID controller;

[0162] The limiting formula is:

[0163]

[0164] In the formula, u(t) represents the adjustment control signal, ΔQ(t) represents the scoring deviation at time t, Kp represents the proportional coefficient, Ki represents the integral coefficient, and Kd represents the derivative coefficient. Let dt represent the integral term, dt represent the time period, and d represent the derivative. This represents the differential term.

[0165] In this embodiment, a multi-parameter feature extraction unit extracts key features such as humidity change rate, average temperature, vibration frequency standard deviation, and flow rate change rate from the data standard set and historical data to construct a feature set F. This method comprehensively captures various parameters in the production process, especially the changing trends of humidity and flow rate, providing a rich source of information for the subsequent construction of a dynamic quality model, making quality control more in-depth and accurate. Furthermore, the feature extraction process includes Pearson correlation coefficient analysis, which can identify the correlation between different features, providing a scientific basis for establishing an accurate quality model. Based on the extracted feature set F, the system uses a multivariate regression model to construct a dynamic quality model. This model can not only evaluate the impact of each feature on quality but also calculate the real-time quality score Q. This process accurately predicts the quality change trend in the production process through multivariate regression analysis and judges the current quality status of the production process based on the quality score output by the model. This data-driven dynamic modeling method greatly improves the accuracy of quality prediction, enabling the prediction of potential quality problems in the early stages of production, thereby allowing for the implementation of effective preventative measures.

[0166] Through the quality scoring feedback processing unit, the system can compare the real-time calculated quality score Q with a preset scoring threshold TQ to determine whether the production status is normal. When the quality score is lower than the threshold, the system automatically calculates the scoring deviation ΔQ and adjusts key production parameters based on the deviation value. This feedback mechanism enables the system to respond to anomalies in the production process in real time, automatically adjusting parameters, avoiding the lag and inconsistencies caused by manual operation, and ensuring the high efficiency and stability of the production process.

[0167] The real-time linkage control module optimizes the production process by adjusting fluid flow rate and the humidity of production materials. The flow rate and humidity adjustments are dynamically calculated based on the scoring deviation ΔQ, ensuring immediate adjustment of the production environment and quality control. The use of a PID controller ensures smooth and precise adjustments to flow rate and humidity, avoiding oscillations and instability caused by over-adjustment. This automated adjustment process precisely maintains product quality consistency and reduces fluctuations in the production process. Through an automated quality scoring and feedback mechanism, the system can detect quality deviations in real time during production and adjust production parameters promptly. This precise real-time adjustment reduces scrap and rework caused by quality fluctuations, significantly improving production efficiency. Furthermore, because the system ensures the production process is always in an optimal state, it can stably produce products that meet quality requirements, reducing scrap rates and non-conforming product rates.

[0168] Example 4

[0169] This embodiment is an explanation based on Embodiment 3. Please refer to it. Figure 1 Specifically: the effect feedback and adjustment module includes a quality monitoring and feedback acquisition unit and a control effect evaluation and feedback unit;

[0170] The quality monitoring and feedback acquisition unit is responsible for real-time monitoring of the adjusted production quality, recording the current quality score Q, and collecting the impact of the adjusted fluid flow rate VfN and the adjusted production material humidity HmN on the production quality.

[0171] The regulation effect evaluation and feedback unit analyzes the impact of the adjusted production parameters on quality based on real-time monitoring data and historical data, evaluates the regulation effect, and obtains the real-time deviation ΔQ(t).

[0172] The real-time deviation ΔQ(t) is obtained by the difference between the real-time quality score Q(t) and the score threshold TQ:

[0173] The real-time deviation ΔQ(t) is obtained by the following formula:

[0174] ΔQ(t) = TQ - Q(t);

[0175] The effects of fluid velocity Vf and production material humidity Hm on real-time quality score Q(t) were evaluated using machine learning algorithms.

[0176] The formula for the influence is:

[0177] ΔQ(t)=αV*ΦVf+αH*ΦHm;

[0178] In the formula, ΔQ(t) represents the real-time deviation, ΦVf represents the flow rate adjustment amount, ΦHm represents the humidity adjustment amount, and αV and αH represent the sensitivity coefficients of the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm to the production quality, respectively.

[0179] The optimization learning and model includes a feedback data analysis and feature extraction unit and a model update and optimization unit;

[0180] The feedback data analysis and feature extraction unit evaluates the impact of the adjusted parameters on production quality based on the acquired real-time deviation ΔQ(t), flow rate adjustment amount ΦVf, and humidity adjustment amount ΦHm.

[0181] The real-time quality score Q(t), the adjusted fluid velocity VfN, and the adjusted production material humidity HmN are extracted from the feedback data to form a data set D, and the relationship between the PID controller adjustment parameters and the quality score Q is evaluated.

[0182] The model update and optimization unit uses a machine learning algorithm to train a dynamic quality model based on data set D. After training, the machine learning model automatically adjusts the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd; updates the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller; and recalculates the feedback signal to adjust the actual production parameters.

[0183] In this embodiment, through the quality monitoring and feedback acquisition unit, the system can monitor the adjusted production quality in real time, obtain the current quality score Q and the impact of the adjusted production parameters on quality. This process can comprehensively reflect the effect of production adjustments and provide data support for subsequent optimization. Compared with traditional quality control methods, this real-time monitoring and feedback mechanism significantly improves the quality control capability in the production process, enabling the system to respond to quality fluctuations at any time during production, ensuring the consistency and stability of product quality.

[0184] By using machine learning algorithms, the system can automatically assess the impact of flow rate and humidity adjustments on the production quality score Q(t). Based on the real-time deviation ΔQ(t) and changes in flow rate and humidity adjustments, the system can identify which parameters have a significant impact on product quality, thus enabling precise optimization and adjustment. This automated assessment and adjustment process significantly improves the system's intelligence level, reduces human intervention, and ensures the efficiency and accuracy of quality control.

[0185] The feedback data analysis and feature extraction unit can extract real-time quality scores, adjusted flow rates, and humidity parameters from the feedback data, forming a dataset D to provide training data for the machine learning model. By evaluating the relationship between the PID controller's adjustment parameters and the quality score, the system can continuously optimize its control strategy and update the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd. This mechanism enables the system to have self-learning and adaptive capabilities, allowing it to continuously improve control precision and optimize production quality as production continues.

[0186] In the model update and optimization unit, machine learning training based on data set D automatically adjusts the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller. This automatic optimization of parameters enables the PID controller to respond more accurately to real-time changes in production when adjusting flow rate and humidity, reducing over-adjustment and instability during the adjustment process. This process not only improves the stability of the production process but also enhances the consistency of product quality, making the entire production process more efficient and robust.

[0187] The system can automatically update and optimize its quality control model based on real-time feedback data and machine learning algorithms, improving the intelligence level of each stage of the production process. This continuous optimization capability ensures that the system can adapt to different production environments and needs, avoiding the limitations of manual intervention and making the production process more flexible, efficient, and controllable.

[0188] This system enhances the transparency of the production process through real-time monitoring, feedback adjustment, and automatic optimization, ensuring precise control over every step. This makes every step in the production process traceable and adjustable, providing more refined control methods for production management and enhancing the factory's controllability and ability to respond to emergencies.

[0189] Example 5

[0190] A smart factory production control method based on the Internet of Things (IoT) is described in the following reference: Figure 2 Specifically, it includes the following steps:

[0191] Step 1: The quality parameter acquisition module collects data in real time from the production process using IoT sensors, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and then forms a dataset W.

[0192] Step 2: The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB;

[0193] Step 3: The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q;

[0194] Step 4: The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control.

[0195] Step 5: The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect.

[0196] Step Six: Optimize the learning and model update module. Based on the feedback data of the regulation effect, use machine learning algorithms to automatically adjust the dynamic quality model.

[0197] In this embodiment, by fully implementing steps one through six, the system can effectively reduce the scrap rate caused by parameter fluctuations during production, improving the consistency and stability of product quality. The combined effects of intelligent quality prediction, automatic adjustment, real-time feedback, and model optimization make quality control during production more precise and stable, thereby improving production efficiency and quality levels. This system reduces human intervention and the probability of operational errors through automated quality control and optimization learning modules. Through the system's intelligent analysis and adjustment, each link in the production process can automatically respond to quality changes, ensuring stable production operation. Compared to traditional methods relying on manual adjustments, this reduces human error and improves production efficiency and quality reliability.

[0198] The entire control process, through a closed-loop feedback mechanism and intelligent adjustments, makes the production process more controllable. Adjustments at each stage can be traced and optimized, and every detail of the production process can be monitored in real time. This transparency and controllability provide data support for production management, making the production process more efficient and precise, while also helping to resolve unexpected problems during production.

[0199] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart factory production control system based on the Internet of Things, characterized in that: It includes a quality parameter acquisition module, a data preprocessing and transmission module, a multi-parameter dynamic modeling module, a real-time linkage control module, an effect feedback and adjustment module, and an optimization learning and model update module; The quality parameter acquisition module collects data in real time from the production process, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and forms a dataset W. The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB. The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q. The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control. The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect. The optimization learning and model update module automatically adjusts the dynamic quality model based on feedback data of the regulation effect using machine learning algorithms.

2. The smart factory production control system based on the Internet of Things according to claim 1, characterized in that: The quality parameter acquisition module includes a data acquisition unit and a rate calculation unit; The data acquisition unit collects the humidity Hm of the production material through a microwave humidity sensor, the ambient temperature Te through a temperature sensor, the vibration frequency Fv of the equipment through a vibration velocity sensor, and the fluid velocity Vf through an ultrasonic flow meter. The rate calculation unit calculates the humidity change rate ΔHm and the flow rate change rate ΔVf based on the collected production material humidity Hm and fluid flow rate Vf; and fits the data collected by the data acquisition unit and the data collected by the rate calculation unit to obtain the dataset W. The humidity change rate ΔHm is obtained using the following formula: In the formula, Hm(t) represents the humidity of the production material at time t, Hm(t-1) represents the humidity of the production material at time t-1, and Δt represents the sampling time interval; The rate of change of flow velocity ΔVf is obtained by the following formula: In the formula, Vf(t) represents the fluid velocity Vf at time t, Vf(t-1) represents the fluid velocity Vf at time t-1, and Δt represents the sampling time interval.

3. The smart factory production control system based on the Internet of Things according to claim 1, characterized in that: The data preprocessing and transmission module includes a data denoising unit and a data standardization and packetization unit; The data denoising unit uses a filter to denoise the dataset W, eliminate high-frequency noise, and obtain the denoised data set WZ. The data standardization and packaging unit uses Z-Score standardization to standardize the data denoising set WZ, transforming data of different dimensions to the same scale, obtaining the data standard set WB, and packaging it, which is then transmitted to the central processing unit via the network. The dataset standard WB is obtained using the following formula: In the formula, WBd represents the data value of the d-th data in the standard dataset WB, WZd represents the data value of the d-th data in the denoised dataset WZ, μd represents the mean of the data value of the d-th data, and σd represents the mean and standard deviation of the data value of the d-th data.

4. The smart factory production control system based on the Internet of Things according to claim 1, characterized in that: The multi-parameter dynamic modeling module includes a multi-parameter feature extraction unit and a dynamic quality model construction and scoring unit; The parameter feature extraction unit extracts features from the data standard set WB and historical data, including humidity change rate ΔHm, average temperature Tavg, vibration frequency standard deviation σFv, and flow velocity change rate ΔVf, forming a feature set F; and analyzes the relationship between different features using the Pearson correlation coefficient. The average temperature Tavg is obtained by the ratio of the sum of the ambient temperatures Te over all time periods to the value of the time period. The standard deviation of the vibration frequency σFv is obtained by the following formula: In the formula, n represents the total number of vibration frequency Fv records, μFv represents the mean value of vibration frequency Fv, and Fva represents the vibration frequency value of the a-th record. The formula for the Pearson correlation coefficient is: In the formula, rFxy represents the correlation coefficient of any feature xi and yi in feature set F, μx and μy represent the mean values ​​of features x and y respectively, m represents the number of data points in the time window, and xi and yi represent any feature in feature set F; The dynamic quality model construction and scoring unit constructs a dynamic quality model based on the feature set F using a multivariate regression model, analyzes the impact of each feature on quality, and calculates the current quality score Q. The quality score Q is obtained using the following formula: In the formula, Q(t) represents the quality score at time t, N1 represents the number of data, N2 represents the number of features, fia(t) represents the value of the ia-th data at time t, gja(t) represents the interaction correlation between the ja-th features at time t, αia represents the preset weight value of the ia-th data, βja represents the preset weight value of the interaction correlation between the ja-th features, and C represents the constant term.

5. The smart factory production control system based on the Internet of Things according to claim 1, characterized in that: The real-time linkage control module includes a quality scoring feedback processing unit and a linkage control and feedback control unit. The quality scoring feedback processing unit receives the quality score Q and compares it with the preset scoring threshold TQ to determine the current production status and obtain the scoring deviation ΔQ. The current production status is obtained by matching in the following way: When the quality score Q ≥ the score threshold TQ, it indicates that the current production status is normal and no parameters need to be adjusted. When the quality score Q < the score threshold TQ, it indicates that the current production status is abnormal. Calculate the score deviation ΔQ and adjust the parameters accordingly. The scoring deviation ΔQ is obtained using the following formula: ΔQ = TQ - Q; Based on the scoring deviation ΔQ, adjust the production parameters, including fluid flow rate Vf and production material humidity Hm, and calculate the flow rate adjustment amount ΦVf and humidity adjustment amount ΦHm. The flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm are obtained by the following formulas: ΦVf=γv*ΔQ; ΦHm=γh*ΔQ; In the formula, γv and γh represent the adjustment coefficients for fluid velocity Vf and production material humidity Hm, respectively.

6. The smart factory production control system based on the Internet of Things according to claim 5, characterized in that: The linkage control and feedback control unit adjusts the fluid flow rate Vf and the production material humidity Hm during the production process according to the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm, calculates and obtains the adjusted fluid flow rate VfN and the adjusted production material humidity HmN, and re-introduces them into the production process for production control. The adjusted fluid velocity VfN and the adjusted production material humidity HmN are obtained using the following formulas: VfN=Vf(t)+ΦVf; HmN=Hm(t)+ΦHm; In the formula, Vf(t) represents the fluid velocity Vf at time t, and Hm(t) represents the humidity Hm of the production material at time t; During the adjustment process, the adjustment process of fluid flow rate Vf and production material humidity Hm is limited by using a PID controller; The limiting formula is: In the formula, u(t) represents the adjustment control signal, ΔQ(t) represents the scoring deviation at time t, Kp represents the proportional coefficient, Ki represents the integral coefficient, and Kd represents the derivative coefficient. Let dt represent the integral term, dt represent the time period, and d represent the derivative. This represents the differential term.

7. A smart factory production control system based on the Internet of Things according to claim 6, characterized in that: The effect feedback and adjustment module includes a quality monitoring and feedback acquisition unit and a control effect evaluation and feedback unit; The quality monitoring and feedback acquisition unit is responsible for real-time monitoring of the adjusted production quality, recording the current quality score Q, and collecting the impact of the adjusted fluid flow rate VfN and the adjusted production material humidity HmN on the production quality. The regulation effect evaluation and feedback unit analyzes the impact of the adjusted production parameters on quality based on real-time monitoring data and historical data, evaluates the regulation effect, and obtains the real-time deviation ΔQ(t). The real-time deviation ΔQ(t) is obtained by the difference between the real-time quality score Q(t) and the score threshold TQ: The real-time deviation ΔQ(t) is obtained by the following formula: ΔQ(t) = TQ - Q(t); The effects of fluid velocity Vf and production material humidity Hm on real-time quality score Q(t) were evaluated using machine learning algorithms. The formula for the influence is: ΔQ(t)=αV*ΦVf+αH*ΦHm; In the formula, ΔQ(t) represents the real-time deviation, ΦVf represents the flow rate adjustment amount, ΦHm represents the humidity adjustment amount, and αV and αH represent the sensitivity coefficients of the flow rate adjustment amount ΦVf and the humidity adjustment amount ΦHm to the production quality, respectively.

8. A smart factory production control system based on the Internet of Things according to claim 7, characterized in that: The optimization learning and model update module includes a feedback data analysis and feature extraction unit and a model update and optimization unit; The feedback data analysis and feature extraction unit evaluates the impact of the adjusted parameters on production quality based on the acquired real-time deviation ΔQ(t), flow rate adjustment amount ΦVf, and humidity adjustment amount ΦHm. The real-time quality score Q(t), the adjusted fluid velocity VfN, and the adjusted production material humidity HmN are extracted from the feedback data to form a data set D, and the relationship between the PID controller adjustment parameters and the quality score Q is evaluated.

9. A smart factory production control system based on the Internet of Things according to claim 8, characterized in that: The model update and optimization unit uses a machine learning algorithm to train a dynamic quality model based on data set D. After training, the machine learning model automatically adjusts the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd; updates the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd of the PID controller; and recalculates the feedback signal to adjust the actual production parameters.

10. A smart factory production control method based on the Internet of Things (IoT), applied to the smart factory production control system based on the IoT as described in any one of claims 1 to 9, characterized in that: Includes the following steps: Step 1: The quality parameter acquisition module collects data in real time from the production process using IoT sensors, including the humidity Hm of the production materials, the ambient temperature Te, the equipment vibration frequency Fv, the fluid velocity Vf, the humidity change rate ΔHm, and the velocity change rate ΔVf, and then forms a dataset W. Step 2: The data preprocessing and transmission module performs noise reduction and standardization on the dataset W to obtain the data standard set WB; Step 3: The multi-parameter dynamic modeling module establishes a dynamic quality model based on the data standard set WB and historical data, predicts the trend of growth quality changes, and obtains the quality score Q; Step 4: The real-time linkage control module automatically adjusts production parameters based on the quality score Q, and regulates control variables through feedback control. Step 5: The effect feedback and adjustment module monitors the adjusted production quality in real time, records the impact of the adjusted production parameters on quality, and combines the feedback results with historical data to evaluate the control effect. Step Six: Optimize the learning and model update module. Based on the feedback data of the regulation effect, use machine learning algorithms to automatically adjust the dynamic quality model.

Images