Method for monitoring spiral case ring welding process quality

Abstract

This invention discloses a method for monitoring the welding process quality of a volute housing seat ring, comprising deploying a digital dial indicator and configuring a wireless data acquisition module to establish a deformation data acquisition and transmission link; acquiring deformation data in parallel through multiple channels of the link and performing preprocessing; constructing a deformation monitoring model based on the preprocessed data, calculating the risk rate of a single measuring point, and then calculating the joint risk rate of multiple measuring points; and synchronizing the data to multiple terminals and performing interactive adaptation. This invention replaces manual reading with a high-precision digital dial indicator and wireless automatic acquisition, achieving objective and accurate acquisition and real-time early warning of deformation data; by constructing a deformation monitoring model and a quantitative model of the joint risk rate of multiple measuring points, it accurately characterizes the deformation correlation of multiple measuring points, realizing quantitative prediction of the overall structural deformation risk and proactive process guidance; and through multi-terminal data synchronization, it achieves traceability of the entire welding cycle data of the volute housing seat ring and multi-scenario adapted closed-loop monitoring, while being compatible with existing processes, thus improving the level of quality control.

Description

Technical Field

[0001] This invention belongs to the technical field of welding quality monitoring methods, specifically relating to a method for monitoring the welding process quality of a volute housing seat ring. Background Technology

[0002] The volute housing seat ring is a key component of core equipment such as hydropower generator sets and large pressure vessels, and its welding quality directly affects the installation accuracy, operational stability, and service life of the equipment. During the welding process between the volute housing and the seat ring, thermal deformation is the core factor affecting the installation accuracy of the flange face; therefore, it is necessary to monitor the flange face offset throughout the entire process with high precision.

[0003] Currently, the industry commonly uses mechanical dial indicators for visual monitoring. While this method meets basic monitoring needs, it has significant shortcomings in the context of intelligent construction development: The monitoring process relies on manual reading, which is prone to misjudgment due to visual fatigue and subjective differences, leading to insufficient quality control accuracy; manual data recording is inefficient, failing to achieve real-time synchronization and continuous storage of welding process data, creating a "data black box" that hinders quality traceability and problem review; the lack of a real-time early warning mechanism means that when deformation approaches or exceeds a threshold, it cannot promptly alert operators to intervene, easily leading to welding defects; the absence of deformation trend prediction and process guidance functions makes it impossible to provide welders with targeted parameter adjustment suggestions, hindering proactive control of the welding process; and the data is scattered and lacks a standardized archiving format, resulting in a lack of systematic data support for subsequent quality analysis and process optimization.

[0004] While some welding monitoring devices exist in the current technology, they are not specifically adapted for the large size, high precision, and complex working conditions of volute seat ring welding. Furthermore, they generally suffer from poor compatibility with existing welding processes, high operational complexity, and insufficient level of automation. Therefore, developing a volute seat ring welding quality monitoring solution that balances accuracy, real-time performance, and traceability to replace traditional mechanical visual monitoring methods has become an urgent need to meet the requirements of intelligent construction. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring the welding process quality of the volute housing ring, which solves the problem that existing welding quality monitoring relies on manual mechanical visual inspection.

[0006] The technical solution adopted in this invention is: a method for monitoring the welding process quality of the volute housing ring, comprising the following steps: Step 1: Deploy a digital display dial indicator and configure a wireless data acquisition module to establish a deformation data acquisition and transmission link; Step 2: The multi-channel parallel acquisition deformation data obtained in Step 1 is sequentially subjected to time alignment, outlier removal, smoothing and noise reduction and normalization to obtain preprocessed data; Step 3: Construct a deformation monitoring model based on the preprocessed data obtained in Step 2. Calculate the risk rate of a single measurement point by using the residual between the model fitting value and the measured value. Then, use the Frank Copula function to calculate the joint risk rate of multiple measurement points. Step 4: Synchronize the deformation data and risk rate from Step 3 to multiple terminals and perform interactive adaptation.

[0007] The invention is further characterized in that, Step 1 is as follows: Based on the distribution requirements of the welding measurement points of the volute seat ring, multiple monitoring points are deployed on the flange face and the welding deformation sensitive area. Each monitoring point is equipped with a digital display dial indicator with a resolution of 0.01mm as a monitoring terminal. Then, a wireless data acquisition module is configured for each digital display dial indicator to convert the digital signal output by the digital display dial indicator into transmittable data, and a wireless data transmission link is established using Bluetooth.

[0008] Step 2 specifically includes the following steps: Step 2.1: Set the data sampling frequency to 2000Hz, the measurement range to ±5V~±20V, and the filter parameter to be adjustable from 0 to 100 with a default value of 62. Step 2.2: Acquire welding deformation data of each measuring point in parallel through the link from Step 1 and record the acquisition timestamp to form the raw data stream; Step 2.3: Perform timing alignment, outlier removal, smoothing and noise reduction, and normalization on the original data stream in sequence.

[0009] Step 2.3 specifically includes the following steps: Step 2.3.1: Using the data collection timestamp as a reference, linear interpolation is used to align the deformation data with the process parameter data in time sequence, so that the welding deformation at the same moment corresponds one-to-one with the welding process parameter. Step 2.3.2: Use the 3σ criterion to remove outliers from the time-aligned data. When the sampled values ​​within the sliding window satisfy | x i - m |>3 s When an outlier is identified, it is removed. m This represents the mean of the data within the sliding window. s The standard deviation of the data within the sliding window. x i For the first in the window i Each sample value; the mean of adjacent valid data is used to complete the elimination position; Step 2.3.3: Use the moving average filtering algorithm to perform noise reduction on the data after removing outliers; Step 2.3.4: Use the Min-Max normalization method to map all the smoothed and denoised deformation data and process parameter data to the [0,1] interval to obtain the normalized deformation data and process parameter data.

[0010] Step 3 specifically includes the following steps: Step 3.1: Using the normalized process parameter data from Step 2 as input and the normalized deformation data from Step 2 as output, construct the following deformation monitoring model:

[0011] In the formula, For the first j The first measuring point i The fitted value of deformation at each sampling time; Q = I × U / v This refers to the welding line energy, i.e., the heat input. I For welding current, U For welding voltage, v For welding speed, T Temperature is a variable during the welding process; T s For temperature-related variables; g = i / 100 is the aging factor, which characterizes the aging deformation effect of welding heat accumulation; a jr , b js , c j1 , c j2 , d j The model fitting coefficients are obtained using the least squares method. Step 3.2: Based on the residual between the fitted value and the measured value of the deformation monitoring model obtained in Step 3.1, the real-time deformation risk rate of a single measuring point is calculated using the normal distribution probability integral method. :

[0012] In the formula, d ji For the first j The first measuring point i The measured value of deformation at each sampling time. s j For the first j The residual standard deviation of the deformation monitoring model at each measuring point is calculated using the following formula:

[0013] In the formula, n This represents the total number of sampling times for a single measuring point. Step 3.3: Based on the residual sequence of each measuring point output in Step 3.2, calculate the overall welding deformation risk rate of the volute housing ring, which specifically includes the following steps: Step 3.3.1: Using the residual sequences of each measurement point obtained in Step 3.2 as input, select the Gamma distribution as the marginal probability distribution function of the residual sequences of each measurement point, and use the maximum likelihood estimation method to solve for the shape parameters of the Gamma distribution. α j and scale parameters β j , obtained the j Marginal probability distribution function of measurement points :

[0014] In the formula, For gamma function, e It is a natural constant. t For integration variables; Step 3.3.2: Substitute the marginal probability distribution sequence of all measurement points into the Frank Copula function, and use a nonlinear fitting method combined with maximum likelihood estimation to solve for the correlation parameter of the Copula function. i A joint distribution model of multiple measurement points was obtained. :

[0015] In the formula, m This represents the total number of measurement points. Step 3.3.3: Calculate the combined risk rate of overall welding deformation of the volute bearing ring at each sampling time using the following formula. :

[0016] In the formula, X This represents the deviation between the current deformation state and the theoretical normal state of the corresponding measuring point.

[0017] Step 3.1, the process of solving the model fitting coefficients, specifically includes the following steps: Step 3.1.1, calculate the first... j The first measuring point i Measured values ​​at each sampling time d ji with fitted value The sum of the remaining squares S :

[0018] Step 3.1.2: Calculate the partial derivatives of each fitted coefficient using the least squares method and set them equal to 0 to construct the normal equation system:

[0019] Step 3.1.3: Solve the normal equation system to obtain the optimal solution for each fitting coefficient.

[0020] In step 4, the measured value of deformation will be... d ji Fitted values ​​of deformation obtained in step 3 Single measurement point risk rate Multi-point joint risk rate The data is simultaneously pushed to the monitoring screen, on-site tablets or mobile terminals, and the back-end management system via wireless network.

[0021] Step 4, the interactive adaptation, includes real-time collection of the communication status and power information of the wireless data acquisition module and display on the terminal system. When the communication status is interrupted or the signal is abnormal, the wireless data acquisition module is triggered to start local data storage. After the communication is restored, the locally stored data is automatically synchronized to the background management system. Different operating permissions are configured for on-site operators, administrators, and supervisors on the terminal system.

[0022] The beneficial effects of this invention are as follows: The volute housing ring welding process quality monitoring method of this invention replaces manual reading with a high-precision digital display dial gauge and wireless automatic acquisition, realizing objective and accurate acquisition and real-time early warning of deformation data; by constructing a deformation monitoring model and a multi-measurement point joint risk rate quantification model based on the Frank Copula function, the deformation correlation of multiple measurement points is accurately characterized, realizing quantitative prediction of overall structural deformation risk and proactive process guidance; through multi-terminal data synchronization, the traceability of the entire volute housing ring welding cycle data and multi-scenario adaptable closed-loop monitoring are realized, and it is compatible with existing processes, thus improving the level of quality control. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the architecture of the volute housing seat ring welding process quality monitoring method of the present invention. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0025] Example 1 This invention provides a method for monitoring the welding process quality of volute housing seat rings, applicable to thermal deformation monitoring, quality traceability, and process optimization in the welding process of volute housings and seat rings in fields such as hydropower and energy equipment. By upgrading a high-precision digital dial indicator as the monitoring terminal and building a wireless data acquisition and transmission module, a terminal system is constructed that includes three core functions: real-time monitoring, electronic archiving, and trend prediction. This achieves accurate acquisition of welding deformation data (resolution 0.01mm, sampling frequency 2000Hz), real-time presentation and out-of-tolerance warnings, automatic generation of full-cycle electronic archives, deformation trend prediction, and intelligent guidance of process parameters. It solves the problems of subjective judgment, inefficient processes, data black boxes, and lack of intelligent features inherent in traditional mechanical visual monitoring. It is compatible with existing welding processes, reduces implementation costs, provides traceable data, and supports multi-terminal and multi-role adaptation, providing data support for stable control and process optimization of volute housing seat ring welding quality. Figure 1 As shown, it includes the following steps: 1. System architecture setup and device adaptation; 2. Multi-channel data acquisition and preprocessing; 3. Based on the preprocessed historical dataset, a monitoring model for welding deformation of the volute seat ring, a single-point risk rate quantification model, and a multi-point joint risk rate model are constructed to provide core algorithm support for real-time monitoring, early warning, and prediction. 4. Terminal presentation and interaction adaptation.

[0026] Through the above methods, the volute housing seat ring welding process quality monitoring method of the present invention can achieve the following: 1. Solve the problem of subjective misjudgment in traditional mechanical visual monitoring: Replace manual reading with a high-precision digital dial indicator (0.01mm resolution) and automatic data acquisition to ensure the objectivity and accuracy of deformation data. The acquisition frequency reaches 2000Hz, meeting the needs of high-frequency monitoring.

[0027] 2. Improve monitoring efficiency and real-time performance: Data is wirelessly transmitted to the terminal system in real time, and the dynamic curves intuitively show the deformation trend. When the deviation exceeds the limit, an audible and visual alarm is triggered immediately, avoiding quality defects caused by delayed intervention.

[0028] 3. Achieve full-cycle data traceability: Automatically generate standardized electronic archives, support multi-dimensional retrieval, historical playback and data export, provide complete data support for quality analysis and problem review, and fill the gap of the "data black box".

[0029] 4. Provide proactive process optimization support: Predict and warn of out-of-tolerance risks by forecasting deformation trends, and push targeted process adjustment suggestions based on the knowledge base to achieve proactive control of the welding process.

[0030] 5. Compatible with existing processes and reduced implementation costs: The original welding process is not changed, only the monitoring equipment and data transmission module are upgraded, resulting in low learning costs for operators and minimal resistance to implementation.

[0031] 6. Capable of emergency handling and multi-scenario adaptation: Supports local data storage in case of communication failure, multi-terminal and multi-role permission configuration, and adapts to different scenario requirements such as on-site operation, back-end management, and supervision verification.

[0032] 7. Provides standardized data output and report generation functions: Supports exporting raw data in CSV / Excel format and generating standardized reports in PDF format to meet the requirements of quality acceptance and process archiving.

[0033] Example 2 This invention provides a method for monitoring the welding process quality of the volute housing seat ring. Based on Example 1, step 1 preferably includes the following steps: Step 1.1 Equipment Upgrade Configuration: Select a high-precision digital display dial indicator with a resolution of 0.01mm as the monitoring terminal. Based on the distribution requirements of the welding measurement points of the volute seat ring, deploy multi-channel monitoring equipment to cover key flange surfaces and welding deformation sensitive areas.

[0034] Step 1.2, Data Transmission Module Setup: Construct a "hardware-software" data transmission bridge. The digital signal from the high-precision digital display dial indicator is converted into transmittable data through a wireless data acquisition module. Bluetooth transmission is used to achieve real-time synchronization between the data and the terminal system.

[0035] Step 1.3, Existing process compatibility: Keep the original volute seat ring welding process unchanged, only replace the monitoring equipment and data recording method to reduce the learning cost for on-site operators and the resistance to implementation of the solution.

[0036] Example 3 This invention provides a method for monitoring the welding process quality of the volute housing ring. Based on Example 1, step 2 preferably includes the following steps: Step 2.1, Parameter Configuration: Set the data sampling frequency to 2000Hz, configure the range to ±5V~±20V according to the welding deformation requirements, and set the filter parameter to 0~100 adjustable (default 62) to ensure the high frequency and accuracy of data acquisition.

[0037] Step 2.2, Real-time Data Acquisition: The welding deformation data (physical quantity unit mm) of each measuring point is acquired in parallel through a multi-channel synchronous acquisition module, and the acquisition timestamp is recorded to form the raw data stream.

[0038] Step 2.3, Data Preprocessing: The collected raw data is denoised and deduplicated to remove abnormal interference signals and ensure that the data truly reflects the welding deformation process.

[0039] 1) Time alignment: Based on the acquisition timestamp, linear interpolation is used to align the deformation data with the process parameter data to ensure a one-to-one correspondence between the deformation amount and the process parameter at the same moment. The interpolation formula is as follows:

[0040] In the formula, The interpolated data is at time t. t 0、 t 1 are respectively t Sampling times adjacent to each other before and after time point, , They are respectively t 0、 t Measured data at time 1.

[0041] 2) Outlier removal: using 3 s The criteria remove outliers from the original data and eliminate jumps in data caused by equipment vibration, electromagnetic interference, and welding spatter. The specific formula is as follows:

[0042] In the formula, m This represents the mean of the data within the sliding window. s The standard deviation of the data within the sliding window. n This represents the number of sampling points within the sliding window. x i For the first in the window i One sample value; when the sample value satisfies | x i - m |>3 s If an outlier is detected, it is removed and replaced with the mean of the adjacent valid data.

[0043] 3) Smoothing and Denoising: A moving average filtering algorithm is used to denoise the data after outlier removal, further weakening random interference. The formula is as follows:

[0044] In the formula, For the first k Filtered data from each sampling point m The length of the sliding window. For the first kj Valid data from each sampling point.

[0045] 4) Data Normalization: The Min-Max normalization method is used to standardize the filtered deformation data and process parameter data, mapping all data to the [0,1] interval to eliminate the influence of different units on subsequent algorithm modeling. The formula is as follows:

[0046] In the formula, x norm For the normalized data, x The original data, x max , x min These are the maximum and minimum values ​​of the corresponding parameter sequences, respectively.

[0047] Example 4 This invention provides a method for monitoring the welding process quality of the volute housing ring. Based on Example 1, step 3 preferably includes the following steps: Step 3.1, Construction of the volute housing seat ring welding deformation monitoring model: Using welding process parameters as input and deformation at measuring points as output, a statistical model for deformation monitoring considering welding heat input and aging accumulation is constructed to quantify the mapping relationship between process parameters and deformation. The model expression is as follows:

[0048] In the formula, For the first j The first measuring point i The fitted value of deformation at each sampling time, in mm; Q = I × U / v This refers to the welding line energy, i.e., the heat input. I The welding current (A) is used. U The welding voltage is (V). v Welding speed (mm / s); T Temperature is a variable during the welding process; r Subscripts for the order of the welding heat input polynomial; T s For temperature-related variables, where, s The index of the temperature parameter is the serial number. s When =1, T 1 represents the interlayer temperature (°C); when s When =2, T 2 represents the preheating temperature (°C); g = i / 100 is the aging factor, which characterizes the aging deformation effect of welding heat accumulation; a jr , b js , c j1 ,c j2 , d j The model fitting coefficients are obtained using the least squares method. The solution process is as follows: 1) Calculate the measured value d ji with fitted value The sum of the remaining squares S :

[0049] 2) Based on the least squares principle, take the partial derivatives of each fitted coefficient and set them equal to 0 to construct a normal system of equations:

[0050] 3) Solve the normal equation system to obtain the optimal solution for each fitting coefficient, completing the first step. j Deformation monitoring model construction for one measuring point; model construction for the remaining measuring points is completed using the same method.

[0051] Step 3.2, Quantification of Real-Time Deformation Risk Rate at a Single Measurement Point: Based on the residual between the fitted value and the measured value of the deformation monitoring model, the real-time deformation risk rate at a single measurement point is calculated using the normal distribution probability integral method, thus achieving a quantitative characterization of deformation risk. The formula is as follows:

[0052] In the formula: For the first j The first measuring point i The deformation risk rate at each sampling time, with a value range of [0,1]. The larger the value, the higher the risk of deformation deviating from the normal state. For the first j The first measuring point i The fitted value of deformation at each sampling time, in mm; d ji For the first j The first measuring point i Measured values ​​of deformation at each sampling time, in mm; s j For the first j The residual standard deviation of the deformation monitoring model at each measuring point is calculated using the following formula:

[0053] in, n The total number of sampling times within a complete modeling cycle for a single measurement point.

[0054] Step 3.3, Quantification of Joint Deformation Risk Rate at Multiple Measurement Points: Based on the Frank Copula function, the deformation correlation and synergy among multiple measurement points are characterized, and the joint risk rate of overall welding deformation of the volute housing ring is calculated. This solves the problem that single-measurement-point early warning cannot reflect the overall deformation risk of the structure. The specific steps are as follows: 1) Marginal distribution fitting: The Gamma distribution is selected as the marginal probability distribution function of the residual sequence at each measurement point. The sequence and corresponding residuals calculated in step 3.2 are used as known conditions to fit the Gamma distribution parameters. α j , β j The distribution function expression is:

[0055] In the formula, For gamma function, α j For shape parameters, β j Let be the scaling parameter; the parameter is solved using the maximum likelihood estimation method to obtain the th . j Marginal probability distribution function of measurement points .

[0056] 2) Construction of the Copula joint distribution function: Substitute the marginal distribution sequences of all measurement points into the Frank Copula function to construct a multi-measurement point joint distribution model. The function expression is as follows:

[0057] In the formula, m The total number of measurement points. i The correlation parameter of the Copula function is solved by combining nonlinear fitting with maximum likelihood estimation to obtain the specific expression of the joint distribution function.

[0058] 3) Overall Risk Rate Calculation: For welding deformation of the volute seat ring, any abnormal deformation at a single measuring point is considered an abnormality in the overall structural condition. Therefore, the formula for calculating the overall deformation risk rate is:

[0059] In the formula, For the first i The risk rate of overall deformation of the volute seat ring welding at each sampling time is denoted as [0,1]; where, X This represents the deviation between the current deformation state and the theoretical normal state of the corresponding measuring point.

[0060] Example 5 This invention provides a method for monitoring the welding process quality of the volute housing ring. Based on Example 1, step 4 preferably includes the following steps: Step 4.1 Multi-terminal adaptation: Supports data synchronization across multiple terminals, including monitoring screens (real-time display of global data), tablets / mobile phones (on-site operation and receiving notifications), and back-end management systems (data storage and analysis).

[0061] Step 4.2, Status Monitoring and Emergency Handling: Real-time feedback on device communication status (online / weak signal / offline) and power information. When communication is interrupted or signal is abnormal, local data storage is automatically triggered to avoid data loss.

[0062] Step 4.3, Multi-role permission configuration: Supports multi-role permission division for on-site operators (operation execution, receiving guidance), administrators (parameter configuration, system maintenance), and supervisors (read-only viewing, quality verification), adapting to different use cases.

[0063] Example 6 This invention provides a method for monitoring the welding process quality of the volute housing ring. By upgrading monitoring equipment, establishing a wireless data transmission link, and developing a multi-module intelligent terminal system, it achieves intelligent monitoring of the entire welding process of the volute housing ring. The specific implementation process is as follows: 1. System architecture setup and device adaptation 1.1 Deployment of Monitoring Equipment A high-precision digital dial indicator with a resolution of 0.01 mm was selected as the monitoring terminal. Based on the structural characteristics and deformation-sensitive areas of the volute seat ring, four monitoring points were deployed around the key flange face and weld, sharing one acquisition module to ensure that there were no blind spots in the monitoring range.

[0064] 1.2 Data Transmission Module Configuration The acquisition module is equipped with a wireless data acquisition module (as a "hardware-software" data bridge), which supports the Bluetooth wireless transmission protocol to convert the digital signal of the digital display dial gauge into standardized data and transmit it to the central database with a transmission delay of ≤5ms.

[0065] 1.3 Process compatibility adaptation Keep the original volute seat ring welding process parameters (such as welding current 180~280A, voltage 23~28.4V, speed 18cm / min~35cm / min, etc.) unchanged, only replace the traditional mechanical dial indicator with a high-precision digital display device, and the operators continue to use the original welding operation process.

[0066] 2. Multi-channel data acquisition and preprocessing 2.1 Initialization of Acquisition Parameters The acquisition parameters are set through the background management system. The sampling frequency is 2000Hz, the measurement range is configured as ±5V~±20V according to the deformation requirements of the measurement point, and the filter parameter is set to 62 to ensure the anti-interference ability and accuracy of data acquisition.

[0067] 2.2 Real-time data acquisition During the welding process, the multi-channel acquisition module collects deformation data of each measuring point in parallel and records the acquisition timestamps accurate to the millisecond level. The raw data includes AD raw values, physical quantities (mm), acquisition time, equipment number, measuring point number and other information.

[0068] 2.3 Data Preprocessing The terminal system performs noise reduction processing on the collected raw data, removes outliers caused by equipment vibration and electromagnetic interference, retains valid deformation data, and synchronously transmits it to the central database for storage.

[0069] 3. Construct a monitoring model for welding deformation of the volute housing ring, a single-point risk rate quantification model, and a multi-point joint risk rate model. 3.1 Real-time monitoring and out-of-tolerance early warning of welding deformation Threshold configuration: According to the welding process requirements of the volute seat ring, the early warning threshold (e.g., 2.8mm) and alarm threshold (e.g., 3.0mm) of each measuring point are set through the terminal system and marked on the monitoring screen with yellow and red curves respectively.

[0070] Real-time display: The monitoring screen synchronously displays the deformation changes of each measuring point (measuring point 1 to measuring point 4, etc.) in the form of dynamic curves, and updates the current value, maximum value, minimum value and rate of change (such as 0.12mm / min) in real time.

[0071] Out-of-tolerance alarm: When the deformation of measuring point 4 reaches 2.91mm (exceeding the warning threshold of 2.8mm), the corresponding curve will flash brightly; when it reaches 3.12mm (exceeding the alarm threshold of 3.0mm), an audible and visual alarm will be triggered, and an alarm notification will pop up on the on-site tablet terminal, indicating the alarm location (equipment A-12 / tunnel south entrance K22+300), the current deformation, and suggested measures (immediately verify the situation, suspend construction in the relevant area, and increase the measurement frequency).

[0072] 3.2 Generation of electronic archives for the entire welding process Data storage: The system automatically stores data of the entire welding process, including deformation data, process parameters (welding current, voltage, speed, temperature), equipment status, alarm events, etc., and names them in the format of "file ID + workpiece number + date" (e.g., file ID: W20231026008, workpiece number: JX-1023, date: 2023-10-26).

[0073] Search and playback: Supports one-click search based on metadata such as project number, workpiece ID / KKS code, date, operator, etc. Search results can be played back at 1x / 2x / 5x speed to display the entire welding process data, and simultaneously display deformation curves and process parameter changes.

[0074] Report generation: Supports exporting raw data in CSV / Excel format (single file data size is approximately 4.7MB), or generating standardized PDF reports with one click. The reports include deformation curve change graphs, process parameter summary tables, alarm event records, quality assessment results, and other content.

[0075] Trend prediction: Based on machine learning algorithms, inputting characteristic parameters such as deformation rate and acceleration within the current 5 minutes, predicting the trend of deformation change within the next 1 hour; when it is predicted that weld #B3-2 will have insufficient penetration depth leading to deformation exceeding the tolerance in 3 minutes, an early warning notification will be issued in advance.

[0076] Process guidance: The system fits the backend welding knowledge base (including historical optimization data for working conditions such as 12mm thickness of Q345B steel plate and gas shielded welding), and pushes process adjustment suggestions for abnormal deformation, such as "reduce the welding current from 180A to 165A and slow down the welding speed by 5%", "increase the welding speed by 10% and reduce the heat input", "increase the preheating temperature to 120℃ and reduce thermal stress", etc. Operators can select "adopt suggestion" or "ignore suggestion" through tablet terminals. After adoption, the system can synchronously push parameter adjustment instructions to the welding equipment.

[0077] 4. Terminal presentation and interaction adaptation 4.1 Multi-terminal data synchronization The monitoring screen displays global data in real time (deformation trends of all measuring points, equipment status, and alarm summary); on-site operators receive alarm notifications and process guidance suggestions via tablet / mobile terminals and provide feedback on operation results; the back-end administrator configures collection parameters, maintains the welding knowledge base, and manages user permissions through the management system; and supervisors view welding data and quality reports in read-only mode to conduct quality checks.

[0078] 4.2 Equipment Status Monitoring The terminal system provides real-time feedback on the communication status (online / weak signal / offline) and power information of each acquisition module (e.g., module MOD-001 has 85% power and normal connection; module MOD-003 has 18% power and weak signal). When communication is interrupted, the acquisition module automatically starts local data storage and synchronizes to the central database after the connection is restored to avoid data loss.

[0079] 4.3 Access Control By assigning roles and permissions to user accounts, operators can only view data from their assigned workstation and receive and execute process recommendations; administrators have permissions for parameter configuration, knowledge base maintenance, and user management; supervisors can only view data and reports and have no operational permissions, thus ensuring system data security and operational standards.

[0080] Through the above implementation methods, this invention achieves quality monitoring of the volute housing ring welding process, solving the problems of subjective misjudgment, low efficiency, and lack of data traceability in traditional mechanical monitoring. Through four core capabilities—precise data collection, real-time early warning, data archiving, and intelligent guidance—it provides full-cycle assurance for the welding quality of the volute housing ring, adapting to the high-quality, high-efficiency, and traceable requirements of intelligent construction for the welding process.

Claims

1. A method for monitoring the welding process quality of the volute housing seat ring, characterized in that, Includes the following steps: Step 1: Deploy a digital display dial indicator and configure a wireless data acquisition module to establish a deformation data acquisition and transmission link; Step 2: The multi-channel parallel acquisition deformation data obtained in Step 1 is sequentially subjected to time alignment, outlier removal, smoothing and noise reduction and normalization to obtain preprocessed data; Step 3: Construct a deformation monitoring model based on the preprocessed data obtained in Step 2. Calculate the risk rate of a single measurement point by using the residual between the model fitting value and the measured value. Then, use the Frank Copula function to calculate the joint risk rate of multiple measurement points. Step 4: Synchronize the deformation data and risk rate from Step 3 to multiple terminals and perform interactive adaptation.

2. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 1, characterized in that, Step 1 specifically involves: based on the distribution requirements of welding measurement points on the volute seat ring, deploying multiple monitoring points on the flange face and welding deformation sensitive areas, with each monitoring point configured with a digital dial indicator with a resolution of 0.01mm as a monitoring terminal; then configuring a wireless data acquisition module for each digital dial indicator to convert the digital signal output by the digital dial indicator into transmittable data, and establishing a wireless data transmission link using Bluetooth.

3. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 1, characterized in that, Step 2 specifically includes the following steps: Step 2.1: Set the data sampling frequency to 2000Hz, the measurement range to ±5V~±20V, and the filter parameter to be adjustable from 0 to 100 with a default value of 62. Step 2.2: Acquire welding deformation data of each measuring point in parallel through the link from Step 1 and record the acquisition timestamp to form the raw data stream; Step 2.3: Perform timing alignment, outlier removal, smoothing and noise reduction, and normalization on the original data stream in sequence.

4. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 3, characterized in that, Step 2.3 specifically includes the following steps: Step 2.3.1: Using the data collection timestamp as a reference, linear interpolation is used to align the deformation data with the process parameter data in time sequence, so that the welding deformation at the same moment corresponds one-to-one with the welding process parameter. Step 2.3.2: Use the 3σ criterion to remove outliers from the time-aligned data. When the sampled values ​​within the sliding window satisfy | x i - μ |>3 σ When an outlier is identified, it is removed. μ This represents the mean of the data within the sliding window. σ The standard deviation of the data within the sliding window. x i For the first in the window i Each sample value; the mean of adjacent valid data is used to complete the elimination position; Step 2.3.3: Use the moving average filtering algorithm to perform noise reduction on the data after removing outliers; Step 2.3.4: Use the Min-Max normalization method to map all the smoothed and denoised deformation data and process parameter data to the [0,1] interval to obtain the normalized deformation data and process parameter data.

5. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 1, characterized in that, Step 3 specifically includes the following steps: Step 3.1: Using the normalized process parameter data from Step 2 as input and the normalized deformation data from Step 2 as output, construct the following deformation monitoring model: In the formula, For the first j The first measuring point i The fitted value of deformation at each sampling time; Q = I × U / v This refers to the welding line energy, i.e., the heat input. I For welding current, U For welding voltage, v For welding speed, T Temperature is a variable during the welding process; T s For temperature-related variables; ζ = i / 100 is the aging factor, which characterizes the aging deformation effect of welding heat accumulation; a jr , b js , c j1 , c j2 , d j The model fitting coefficients are obtained using the least squares method. Step 3.2: Based on the residual between the fitted value and the measured value of the deformation monitoring model obtained in Step 3.1, the real-time deformation risk rate of a single measuring point is calculated using the normal distribution probability integral method. : In the formula, δ ji For the first j The first measuring point i The measured value of deformation at each sampling time. σ j For the first j The residual standard deviation of the deformation monitoring model at each measuring point is calculated using the following formula: In the formula, n This represents the total number of sampling times for a single measuring point. Step 3.3: Based on the residual sequence of each measuring point output in Step 3.2, calculate the overall welding deformation risk rate of the volute housing ring, which specifically includes the following steps: Step 3.3.1: Using the residual sequences of each measurement point obtained in Step 3.2 as input, select the Gamma distribution as the marginal probability distribution function of the residual sequences of each measurement point, and use the maximum likelihood estimation method to solve for the shape parameters of the Gamma distribution. α j and scale parameters β j , obtained the j Marginal probability distribution function of measurement points : In the formula, For gamma function, e It is a natural constant. t For integration variables; Step 3.3.2: Substitute the marginal probability distribution sequence of all measurement points into the Frank Copula function, and use a nonlinear fitting method combined with maximum likelihood estimation to solve for the correlation parameter of the Copula function. θ A joint distribution model of multiple measurement points was obtained. : In the formula, m This represents the total number of measurement points. Step 3.3.3: Calculate the combined risk rate of overall welding deformation of the volute bearing ring at each sampling time using the following formula. : In the formula, X This represents the deviation between the current deformation state and the theoretical normal state of the corresponding measuring point.

6. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 5, characterized in that, The process of solving the model fitting coefficients in step 3.1 specifically includes the following steps: Step 3.1.1, calculate the first... j The first measuring point i Measured values ​​at each sampling time δ ji with fitted value The sum of the remaining squares S : Step 3.1.2: Calculate the partial derivatives of each fitted coefficient using the least squares method and set them equal to 0 to construct the normal equation system: Step 3.1.3: Solve the normal equation system to obtain the optimal solution for each fitting coefficient.

7. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 1, characterized in that, In step 4, the measured value of deformation will be... δ ji Fitted values ​​of deformation obtained in step 3 Single measurement point risk rate Multi-point joint risk rate The data is simultaneously pushed to the monitoring screen, on-site tablets or mobile terminals, and the back-end management system via wireless network.

8. The method for monitoring the welding process quality of the volute housing seat ring as described in claim 1, characterized in that, The interactive adaptation in step 4 includes real-time acquisition of the communication status and power information of the wireless data acquisition module and displaying it on the terminal system. When the communication status is interrupted or the signal is abnormal, the wireless data acquisition module is triggered to start local data storage. After the communication is restored, the locally stored data is automatically synchronized to the background management system. Different operating permissions are configured for on-site operators, administrators, and supervisors in the terminal system.

Images