Industrial control system for intelligent valve production

Abstract

The invention discloses an industrial control system for intelligent valve production, and belongs to the field of valve production industrial control systems, and the system comprises a multi-dimensional sensing module which is used for collecting the assembling angle, the assembling depth and the assembling height data in the valve assembling process in real time; the central control module is in communication connection with the multi-dimensional sensing module and is used for comparing the real-time assembly data with a deviation value of a preset process standard and generating an assembly qualification judgment result; the traceability analysis module responds to the assembly disqualification judgment result, and traces at least one process link defect causing assembly disqualification based on the distribution characteristics of the deviation values and the time sequence relevance; and the adjustment execution mechanism is controlled by the central control module and dynamically corrects the operation parameters of the assembly equipment according to the deviation value. According to the method, whether assembly is qualified or not can be analyzed in a multi-aspect and high-precision mode from the assembly angle, the assembly depth and the assembly height, and dynamic correction is conducted in time when the assembly is unqualified.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of valve production industrial control system, and particularly relates to an industrial control system for intelligent valve production. BACKGROUND

[0002] The industrial control system for valve production (industrial control system) is a computer integrated system for monitoring, managing and automatically controlling the whole process of valve production. It realizes precise control of production equipment, process parameters, quality detection and other links through the cooperative work of software and hardware, so as to improve production efficiency, guarantee product quality and reduce operating costs.

[0003] The existing patent 202411396471.9, a valve production measurement and control system based on artificial intelligence, checks the prepared production materials before the production of each component of the valve. If the prepared production materials are consistent with the standard production materials used for the production of the corresponding components in the valve production plan, the production is allowed; otherwise, the production materials need to be adjusted. The components obtained from each process of the production process are scanned and inspected for their qualification. If qualified, the components are allowed to enter the next process; if not, the components are transported to the recycling department. The installation position and connection sequence of the components in the valve assembly process are monitored online, and the valves with assembly errors are transported to the valve secondary assembly storage point, thereby comprehensively realizing valve production quality control.

[0004] However, the existing industrial control system for valve production usually focuses on the valve casting link, mechanical processing link and pressure testing link, and does not pay enough attention to the valve assembly link. It cannot analyze whether the assembly is qualified from the assembly angle, assembly depth and assembly height, and cannot dynamically correct the assembly when it is not qualified, which has certain defects. Therefore, the technical personnel in the field provide an industrial control system for intelligent valve production to solve the problems raised in the above background technology. SUMMARY

[0005] The purpose of the present application is to provide an industrial control system for intelligent valve production, which can analyze whether the assembly is qualified from the assembly angle, assembly depth and assembly height with high precision, and dynamically correct the assembly when it is not qualified, so as to solve the problems raised in the above background technology.

[0006] To achieve the above purpose, the present application provides the following technical solutions:

[0007] An industrial control system for intelligent valve production, comprising:

[0008] A multi-dimensional sensing module for real-time acquisition of assembly angle, assembly depth and assembly height data in the valve assembly process;

[0009] A central control module, in communication connection with the multi-dimensional sensing module, built-in process standard database and dynamic analysis algorithm, for comparing deviation value of real-time assembly data with preset process standard, and generating assembly qualification judgment result;

[0010] A traceability analysis module, in response to assembly unqualified judgment result, based on distribution characteristics of deviation value and time sequence correlation, tracing at least one process link defect causing assembly unqualification;

[0011] An adjustment execution mechanism, controlled by the central control module, dynamically correcting operation parameters of the assembly equipment according to the deviation value to realize self-adaptive calibration.

[0012] As a further scheme of the present application, the multi-dimensional sensing module comprises:

[0013] A high-precision angle encoder, installed on a rotating shaft of an assembly fixture, for measuring assembly angle of a valve body and a valve cover;

[0014] A laser ranging array, deployed on a reference plane of an assembly station, for real-time detection of three-dimensional coordinates of valve rod insertion depth;

[0015] A pressure-sensitive height sensor, integrated in an assembly press-fitting mechanism, for feeding back real-time data of press-fitting height of a sealing ring.

[0016] As a still further scheme of the present application, the traceability analysis module performs the following operations:

[0017] (a1), multivariate correlation analysis is performed on deviation values of assembly angle, depth and height to identify dominant factors of abnormal parameters;

[0018] (b1), a Bayesian probability model is constructed based on historical process data to calculate fault reason probability weight of assembly equipment failure, material size out-of-tolerance and clamping positioning deviation;

[0019] (c1), a visual traceability report is outputted, with key process parameter abnormal nodes and confidence evaluation results being marked.

[0020] As a still further scheme of the present application, the adjustment execution mechanism comprises:

[0021] An angle compensation device driven by a servo motor, for dynamically adjusting rotating angle of the assembly fixture within ±5°;

[0022] A hydraulic closed-loop control press-fitting module, adjusting valve rod press-in depth with 0.01mm precision;

[0023] A pneumatic floating positioning platform, automatically compensating planeness error of an assembly reference plane according to feedback of the height sensor.

[0024] As a further scheme of the present application: further comprising:

[0025] A process knowledge graph database is configured to store a valve model-assembly parameter mapping relationship, a typical failure mode library and an expert correction strategy.

[0026] When it is detected that the assembly parameter exceeds the tolerance range for three consecutive times, the knowledge graph retrieval is automatically triggered and an optimized process scheme is pushed.

[0027] As a further scheme of the present application: the central control module is integrated with an edge computing unit, which is configured to perform the following in real time:

[0028] (1) Fourier transform analysis of assembly angle data to detect periodic vibration interference of the assembly fixture;

[0029] (2) Kalman filter processing of assembly depth time series data to eliminate the influence of measurement noise on the determination result.

[0030] As a further scheme of the present application: further comprising:

[0031] A multi-spectral visual detection module is configured to perform defect scanning on the valve sealing surface after assembly to generate a visual detection result.

[0032] The visual detection result is stored in association with the assembly process data to form a full life cycle quality traceability chain.

[0033] As a further scheme of the present application: the central control module is configured with an adaptive learning function, which is configured to optimize the process standard by the following ways:

[0034] Collecting the actual assembly parameter distribution characteristics of qualified products;

[0035] Using a clustering algorithm to dynamically update the process parameter confidence interval;

[0036] Triggering a process standard revision warning when the parameter distribution center deviates more than 2σ from the standard deviation.

[0037] As a further scheme of the present application: the multi-dimensional sensing module can also be configured in an intermittent acquisition mode, and the data acquisition mode of the multi-dimensional sensing module includes the following three modes:

[0038] (a2) real-time acquisition mode;

[0039] (b2) intermittent acquisition mode;

[0040] (c2) real-time acquisition mode + intermittent acquisition mode.

[0041] As a further scheme of the present application: the intermittent acquisition mode triggers data acquisition under at least one of the following conditions:

[0042] (a3), when the assembly device enters a critical work station;

[0043] (b3), when an assembly pressure threshold change is detected to exceed 10%;

[0044] (c3), periodically collected according to a preset time interval.

[0045] Compared with the prior art, the beneficial effects of the present application are:

[0046] 1. The present application can analyze whether the assembly is qualified in multiple aspects of assembly angle, assembly depth and assembly height with high precision, and timely dynamic correction when the assembly is unqualified, and through the multi-dimensional data real-time monitoring, dynamic calibration and intelligent learning function, the full-process closed-loop control of valve assembly is realized, and the production automation level and product quality consistency are significantly improved.

[0047] 2. The functions of each module of the present application are complementary, from data acquisition (such as high-precision sensors), dynamic judgment (edge computing), defect tracing (Bayesian model) to process optimization (knowledge graph and adaptive learning), which comprehensively improves the assembly quality and efficiency, and significantly improves the production automation level and product quality consistency.

[0048] 3. When switching to produce other types of valves, the present application configures the corresponding data acquisition mode for the multi-dimensional sensing module according to the preset data acquisition rule, so as to balance performance and cost while ensuring production precision.

[0049] 4. The adaptive learning function of the present application can realize dynamic iteration of process parameters, reduce the frequency of manual intervention, and realize self-evolution of process standards through machine learning, adapt to dynamic changes of the production line, and thus comprehensively improve the assembly quality and efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 is a structural block diagram of an industrial control system for intelligent valve production. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0052] As mentioned in the background technology of this application, research has found that the existing industrial control systems for valve production usually focus on the valve casting, machining and pressure testing links, and do not pay enough attention to the valve assembly link. They are unable to perform high-precision analysis of whether the assembly is qualified from multiple aspects such as assembly angle, assembly depth and assembly height, and are unable to make timely dynamic corrections when the assembly is unqualified. There are certain defects.

[0053] In order to solve the above-mentioned defects, the present application discloses an industrial control system for intelligent valve production, which can perform high-precision analysis of whether the assembly is qualified from multiple aspects such as assembly angle, assembly depth and assembly height, and make timely dynamic corrections when the assembly is unqualified.

[0054] The following will describe in detail how the solution of this application solves the above technical problems with reference to the accompanying drawings.

[0055] See also Figure 1 In an embodiment of the present invention, an industrial control system for intelligent valve production includes: a multi-dimensional sensing module for real-time acquisition of assembly angle, assembly depth, and assembly height data during valve assembly; a central control module, communicatively connected to the multi-dimensional sensing module and equipped with a built-in process standard database and dynamic analysis algorithm, for comparing deviations between real-time assembly data and preset process standards and generating an assembly acceptance determination result; a traceability analysis module, responsive to an assembly failure determination result, tracing back at least one process defect that led to the failure based on the distribution characteristics and time series correlation of the deviations; and an adjustment actuator, controlled by the central control module, for dynamically modifying the operating parameters of the assembly equipment based on the deviations to achieve adaptive calibration. The system uses the multi-dimensional sensing module to acquire real-time valve assembly angle, depth, and height data. The central control module compares the real-time data with preset process standards to determine assembly acceptance. If the assembly is deemed unacceptable, the traceability analysis module traces the process defect based on the deviation characteristics and adjusts the actuator to dynamically modify the assembly parameters, achieving closed-loop adaptive calibration. The system covers the complete chain of data acquisition, analysis, execution, and traceability, ensuring controllable assembly precision and quality.

[0056] In this embodiment, the multi-dimensional sensing module includes a high-precision angle encoder mounted on the assembly fixture's rotating axis to measure the assembly angle between the valve body and bonnet, ensuring an angular deviation of ≤±0.5°; a laser ranging array deployed on the assembly station's reference plane to measure the three-dimensional coordinates of the valve stem insertion depth in real time, with a resolution of 0.005mm; and a pressure-sensitive height sensor integrated into the assembly press mechanism to provide real-time data on the seal ring's press-fit height, with an accuracy of ±0.01mm. These sensors work together to provide multi-dimensional data support for the assembly process.

[0057] In this embodiment, the traceability analysis module performs the following operations: (a1) performs multivariate correlation analysis on the deviation values ​​of assembly angle, depth, and height to identify the dominant factors of abnormal parameters (such as angle deviation causing depth error); (b1) constructs a Bayesian probability model based on historical process data to calculate the probability weights of the failure causes of assembly equipment failure, material size deviation, and fixture positioning offset (such as equipment failure weight 60% and material deviation 30%), and quantify the source of defects; (c1) outputs a visual traceability report, marking the abnormal nodes of key process parameters (such as "fixture positioning offset") and the confidence level (such as 95%) evaluation results to support rapid maintenance decisions.

[0058] Multivariate correlation analysis is a statistical method used to study the degree of interrelationships between multiple variables (such as assembly angle, depth, and height) and their combined impact on system output. In this application, this analysis aims to identify the interactions between parameter deviations during the assembly process, thereby pinpointing the dominant factors leading to assembly failures.

[0059] The specific process of identifying the dominant factors of abnormal parameters through multivariate correlation analysis is as follows:

[0060] Step 1: Data Preprocessing

[0061] Data normalization: Standardize parameters of different dimensions, such as angle (unit: °), depth (unit: mm), and height (unit: mm) (e.g., Z-score normalization) to eliminate the impact of dimensional differences on the analysis;

[0062] Timing alignment: Align real-time data collected by multiple sensors based on the timestamps of the assembly process to ensure temporal consistency in analysis.

[0063] Step 2: Correlation calculation

[0064] Pearson correlation coefficient: Calculates the pairwise correlation coefficient of the deviation values ​​of each parameter (range: -1 to 1), for example:

[0065] The correlation coefficient between angle deviation and depth deviation is 0.85 (strong positive correlation);

[0066] The correlation coefficient between height deviation and angle deviation is -0.2 (weak negative correlation).

[0067] Covariance matrix: Construct a multi-parameter covariance matrix to analyze the coordinated change trends between parameters.

[0068] Step 3: Identification of dominant factors

[0069] Contribution ranking: Sort by the absolute value of the correlation coefficient to select the variables (such as angle deviation) that have the greatest impact on the target parameter (such as depth deviation);

[0070] Causal verification: Combined with the Bayesian probability model, the fault contribution weight of each parameter is calculated (for example, the contribution of angle deviation is 70%, and the contribution of material deviation is 25%).

[0071] Visual mapping: Mark the dominant parameter in the traceability report (such as "angle deviation is the main cause of depth deviation") and provide a confidence level (such as 95%).

[0072] Application scenario examples:

[0073] Scenario 1: The sealing ring is not pressed high enough

[0074] Correlation analysis found that the height deviation was highly correlated with the pressure sensor data (correlation coefficient 0.78);

[0075] The dominant factor is the positioning error of the pneumatic floating platform (weight 65%), which triggers the automatic calibration of the platform.

[0076] Scenario 2: Valve stem insertion depth fluctuations

[0077] The correlation coefficient between angle and depth is 0.9, but the probability of material size exceeding the tolerance is only 15%;

[0078] The dominant factor is fixture vibration (weight 70%), and the edge computing unit starts the vibration suppression algorithm.

[0079] In this embodiment, before building the Bayesian probability model, the following data of the past N assembly tasks are extracted from the database:

[0080] Fault cause label (equipment failure, material out of tolerance, clamping offset);

[0081] Corresponding assembly parameter deviation values ​​(angle deviation Δθ, depth deviation Δd, height deviation Δh).

[0082] The specific process of calculating the probability weight of the fault cause based on the Bayesian probability model is as follows:

[0083] Step 1: Define the prior probability

[0084] The prior probability of each fault cause is calculated based on historical data:

[0085]

[0086] Among them, A i ∈{equipment failure (E), material tolerance (M), clamping offset (F)}.

[0087] Step 2: Calculate the conditional probability (likelihood probability)

[0088] Assuming that the deviation of each parameter follows normal distribution, calculate the iThe probability of observing the deviation B = (Δθ, Δd, Δh) is:

[0089]

[0090] Among them, μ ij and σ ij Fault cause A i The mean and standard deviation of the lower parameter j (angle, depth, height).

[0091] Step 3: Calculate the posterior probability

[0092] According to Bayes' theorem, calculate the fault cause A when the deviation B is observed. i The posterior probability of :

[0093]

[0094] Step 4: Output probability weights

[0095] The fault cause with the largest posterior probability is selected as the dominant factor, and the probability weight of each cause is output.

[0096] Model application examples

[0097] Scenario: During an assembly, an angle deviation of Δθ = 0.6°, a depth deviation of Δd = 0.03 mm, and a height deviation of Δh = 0.02 mm were detected.

[0098] Calculation process:

[0099] (1) Prior probability: In historical data, equipment failure (E) accounts for 50%, material tolerance (M) accounts for 30%, and fixture offset (F) accounts for 20%;

[0100] (2) Conditional probability:

[0101] Assume that under equipment failure, the mean angle deviation μ Eθ =0.5°, standard deviation σ Eθ =0.2°;

[0102] Calculate P(B|E) = P(Δθ|E)·P(Δd|E)·P(Δh|E);

[0103] Similarly, calculate P(B|M) and P(B|F).

[0104] (3) Posterior probability:

[0105]

[0106] Final output: Equipment failure probability 68.9%, material out-of-tolerance 25.6%, and clamping offset 5.5%.

[0107] In this embodiment, the adjustment actuator includes a servo-motor-driven angle compensation device for dynamically adjusting the assembly fixture's rotation angle within a ±5° range, with a response time of ≤50ms; a hydraulic closed-loop press-fit module for adjusting the valve stem's insertion depth with a precision of 0.01mm, within a pressure control range of 0-10MPa; and a pneumatic floating positioning platform that automatically compensates for the flatness error of the assembly reference surface (with an accuracy of ±0.02mm) based on feedback from a height sensor. Through integrated electromechanical and hydraulic control, real-time correction of assembly parameters is achieved. Specifically, the ±5° angle compensation, 0.01mm depth adjustment, and automatic flatness error compensation improve the assembly qualification rate by over 30%.

[0108] In this embodiment, the application also includes: a process knowledge graph database that stores valve model-assembly parameter mappings (such as the standard value of 45° for ball valve angles), a typical failure mode library (such as 80% of seal ring press-fits), and expert correction strategies (such as adjusting the press-fit pressure to 8MPa); when it is detected that the assembly parameters are out of tolerance three times in a row (such as depth deviation ≥ 0.03mm), the knowledge graph retrieval is automatically triggered and an optimized process solution (such as replacing the fixture locating pin) is pushed. This setting reduces the cost of manual intervention through knowledge-driven optimization.

[0109] In this embodiment, the edge computing unit is integrated into the central control module to perform in real time: (1) Fourier transform analysis of assembly angle data to detect periodic vibration interference of the assembly fixture (such as 10Hz abnormal vibration) to avoid angle measurement distortion; (2) Kalman filter processing of assembly depth time series data to eliminate the influence of measurement noise (such as ±0.002mm random noise) on the judgment result and improve the reliability of depth data. The edge computing unit is embedded in the central control module to achieve low-latency (<5ms) data processing and ensure the real-time nature of the judgment result. Among them, Fourier transform is a mathematical tool that converts time domain signals (signals that change over time) into frequency domain signals (distribution of different frequency components). Through this conversion, the periodic features implicit in the signal, such as vibration frequency, noise components, etc., can be identified. Kalman filter is an optimal estimation algorithm that estimates the optimal solution of the system state in real time by fusing the predicted value of the system dynamic model and the sensor observation value. Its core advantage is that it can effectively handle noise interference, especially suitable for time series data with random noise.

[0110] In this embodiment, the present application further includes: a multispectral visual inspection module, which performs defect scanning on the valve sealing surface after assembly is completed to generate visual inspection results. More specifically, multispectral imaging technology (wavelength range 400-1000nm) is used to detect defects such as scratches and cracks on the sealing surface (resolution 0.02mm); the visual inspection results are associated with the assembly process data and stored to form a full life cycle quality traceability chain. For example, the visual inspection results (such as "sealing surface scratch length 2mm") are associated with the assembly process data (such as angle deviation 0.3°) to generate a unique traceability code (such as a QR code) to support full life cycle quality traceability.

[0111] In this embodiment, an adaptive learning function is configured within the central control module to optimize process standards through the following methods: collecting actual assembly parameter distribution characteristics of qualified products (e.g., an angle mean of 45.2° and a standard deviation of 0.1°); dynamically updating process parameter confidence intervals using a clustering algorithm (e.g., updating the angle tolerance to ±0.6°); and triggering a process standard revision alert when the center of the parameter distribution shifts by more than 2σ of the standard deviation (e.g., a mean shift to 45.6°). This setting uses machine learning to enable the self-evolution of process standards, adapting to dynamic changes in the production line. The adaptive learning function enables dynamic iteration of process parameters, reducing the frequency of manual intervention. Clustering algorithms are unsupervised machine learning methods that aim to divide objects in a dataset into groups (called "clusters"), ensuring that data points within the same cluster have high similarity (e.g., consistent distribution patterns of assembly angle, depth, and height); and that data points between different clusters are significantly different (e.g., separating clusters of normal process parameters from clusters of abnormal parameters). Common algorithms include K-means and hierarchical clustering.

[0112] In this embodiment, the multi-dimensional sensing module can also be configured as an intermittent acquisition mode, and the data acquisition modes of the multi-dimensional sensing module include the following three: (a2), real-time acquisition mode, suitable for high-precision scenarios (such as ball valve assembly), data sampling rate ≥ 1kHz; (b2), intermittent acquisition mode, in low-cost scenarios (such as gate valve rough assembly), the amount of data is reduced by triggering conditions (such as pressure changes exceeding 10%); (c2), real-time acquisition mode + intermittent acquisition mode, key parameters (such as angle) are collected in real time, and secondary parameters (such as height) are collected intermittently, balancing performance and cost. When switching to the production of other types of valves, this application configures the corresponding data acquisition mode for the multi-dimensional sensing module according to the preset data acquisition rules, so as to balance performance and cost while ensuring that the production accuracy meets the standards.

[0113] In this embodiment, the intermittent data collection mode triggers data collection under at least one of the following conditions: (a3) ​​when the assembly equipment enters a critical workstation, for example, when the assembly equipment enters the valve stem press-fitting workstation, thereby avoiding redundant data; (b3) when a change in the assembly pressure threshold exceeds 10% (e.g., from 5 MPa to 5.5 MPa); and (c3) when data is collected periodically at a preset interval, for example, every 2 seconds, which is applicable to the steady-state assembly phase. This intelligent triggering strategy ensures that critical data is not missed during the intermittent data collection mode.

[0114] In order to further illustrate the present invention, the industrial control system for intelligent valve production provided by the present invention is described in detail below with reference to embodiments.

[0115] Example 1: High-precision ball valve assembly scenario

[0116] Workflow:

[0117] Data collection: Real-time collection of the assembly angle between the valve body and the valve cover (tolerance ±0.5°) and the valve stem pressing depth (tolerance ±0.02mm);

[0118] Dynamic judgment: The central control module compares real-time data with process standards and triggers adjustment instructions if the deviation exceeds the limit;

[0119] Adaptive calibration: The hydraulic module adjusts the press-fitting pressure according to the deviation value, and the pneumatic platform compensates for the reference surface error;

[0120] Quality traceability: After multispectral visual inspection of the sealing surface, the assembly data is associated to generate a unique quality traceability code.

[0121] Example 2: Low-cost gate valve intermittent data collection mode

[0122] Workflow:

[0123] Trigger acquisition: When the assembly pressure threshold changes by more than 10%, angle and depth data acquisition is initiated;

[0124] Defect tracing: Compensate for collection intervals by interpolating historical data and constructing a Bayesian model to analyze the cause of the fault;

[0125] Process optimization: The knowledge graph database pushes expert correction strategies that adapt to current parameters.

[0126] The application can analyze whether the assembly is qualified in multiple aspects of assembly angle, assembly depth and assembly height with high precision, and dynamically correct in time when the assembly is unqualified, and through the multi-dimensional data real-time monitoring, dynamic calibration and intelligent learning function, realize the full-process closed-loop control of valve assembly, and further significantly improve the production automation level and product quality consistency. The functions of the modules of the application are complementary, from data acquisition (such as high-precision sensors), dynamic judgment (edge computing), defect tracing (Bayesian model) to process optimization (knowledge graph and adaptive learning), which comprehensively improves the assembly quality and efficiency, and significantly improves the production automation level and product quality consistency. When switching to produce other types of valves, the application configures corresponding data acquisition modes for multi-dimensional sensing modules according to the preset data acquisition rules, so as to balance performance and cost while ensuring production precision to meet standards. The adaptive learning function of the application can realize dynamic iteration of process parameters, reduce the frequency of manual intervention, and realize self-evolution of process standards through machine learning, adapt to dynamic changes of production lines, and thus comprehensively improve the assembly quality and efficiency.

[0127] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

[0128] The above is only the preferred embodiment of the present application, but the protection scope of the present application is not limited to this. Any skilled person in the art can make equivalent substitutions or changes to the technical solutions and inventive concepts of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. An industrial control system for intelligent valve production, characterized in that: include: Multi-dimensional sensing module, used to collect real-time data on assembly angle, assembly depth, and assembly height during valve assembly; A central control module, which is in communication with the multi-dimensional sensing module and has a built-in process standard database and a dynamic analysis algorithm, is used to compare the deviation value of the real-time assembly data with the preset process standard and generate an assembly qualification determination result; The traceability analysis module, in response to the assembly failure determination result, traces back to at least one process defect that caused the assembly failure based on the distribution characteristics of the deviation value and the time series correlation; The adjustment actuator is controlled by the central control module and dynamically corrects the operating parameters of the assembly equipment according to the deviation value to achieve adaptive calibration.

2. The industrial control system for intelligent valve production according to claim 1, characterized in that: The multi-dimensional sensing module includes: High-precision angle encoder, installed on the rotating axis of the assembly fixture, used to measure the assembly angle of the valve body and valve cover; A laser ranging array is deployed on the reference plane of the assembly station to detect the three-dimensional coordinates of the valve stem insertion depth in real time; The pressure-sensitive height sensor is integrated into the assembly press mechanism to provide real-time feedback on the seal ring press height.

3. The industrial control system for intelligent valve production according to claim 2, characterized in that: The traceability analysis module performs the following operations: (a1) Perform multivariate correlation analysis on the deviation values ​​of assembly angle, depth, and height to identify the dominant factors of abnormal parameters; (b1) Building a Bayesian probability model based on historical process data to calculate the probability weights of the failure causes of assembly equipment failure, material size deviation, and clamping positioning deviation; (c1) Output a visual traceability report, marking abnormal nodes of key process parameters and confidence assessment results.

4. The industrial control system for intelligent valve production according to claim 3, characterized in that: The adjustment execution mechanism includes: Servo motor driven angle compensation device for dynamically adjusting the rotation angle of the assembly fixture within a range of ±5°; Hydraulic closed-loop control press-fit module adjusts the valve stem pressing depth with an accuracy of 0.01mm; The pneumatic floating positioning platform automatically compensates for the flatness error of the assembly reference surface based on feedback from the height sensor.

5. The industrial control system for intelligent valve production according to claim 4, characterized in that: Also includes: Process knowledge graph database, which stores valve model-assembly parameter mapping, typical failure mode library and expert correction strategy; When it is detected that the assembly parameters are out of tolerance for three consecutive times, the knowledge graph retrieval is automatically triggered and the optimized process plan is pushed.

6. The industrial control system for intelligent valve production according to claim 5, characterized in that: The central control module integrates an edge computing unit to perform real-time: (1) Fourier transform analysis of assembly angle data to detect periodic vibration interference of assembly fixtures; (2) Kalman filter processing is performed on the assembled depth time series data to eliminate the influence of measurement noise on the judgment results.

7. The industrial control system for intelligent valve production according to claim 6, characterized in that: Further including: Multispectral visual inspection module, which scans the valve sealing surface for defects after assembly and generates visual inspection results; The visual inspection results are stored in association with the assembly process data to form a full life cycle quality traceability chain.

8. The industrial control system for intelligent valve production according to claim 7, characterized in that: The central control module is equipped with an adaptive learning function to optimize process standards in the following ways: Collect the actual assembly parameter distribution characteristics of qualified products; Clustering algorithm is used to dynamically update the confidence interval of process parameters; When the center deviation of the parameter distribution exceeds the standard deviation 2σ, a process standard revision warning is triggered.

9. The industrial control system for intelligent valve production according to claim 8, characterized in that: The multi-dimensional sensing module can also be configured in an intermittent acquisition mode, and the data acquisition modes of the multi-dimensional sensing module include the following three: (a2), real-time acquisition mode; (b2), intermittent acquisition mode; (c2), real-time acquisition mode + intermittent acquisition mode.

10. The industrial control system for intelligent valve production according to claim 9, characterized in that: The intermittent acquisition mode triggers data acquisition under at least one of the following conditions: (a3) When the assembly equipment enters the key station; (b3), detecting that the assembly pressure threshold changes by more than 10%; (c3) Periodic collection according to preset time intervals.

Images