A Joint Online Monitoring Method for Serial Machining Manufacturing Systems

Abstract

This invention discloses a joint online monitoring method for tandem machining manufacturing systems, aiming to achieve simultaneous monitoring and intelligent early warning of equipment operating status and product quality. This method collects tool status data and workpiece dimensional and quality data at each machining stage, mines the linear relationship between tool degradation and workpiece dimensions, and constructs a system proportional risk function for the tool and a multivariate generalized likelihood ratio (MGLR) control chart for workpiece quality. Historical data is used to estimate model parameters, and based on a set system risk rate threshold and control limits from Monte Carlo simulation, the monitoring and diagnosis of tool status and workpiece quality are achieved. This method enables online joint monitoring of tool status and workpiece quality in tandem machining systems, improving equipment reliability and product consistency, reducing failure and rework rates, and ultimately achieving intelligent optimization and efficient management of the manufacturing process.

Description

Technical Field

[0001] This invention relates to the field of computer integration and automated manufacturing technology, and in particular to a joint online monitoring method for serial machining manufacturing systems. Background Technology

[0002] In the context of rapid industrial development, manufacturers use sensors and advanced quality inspection devices to monitor equipment status and product quality in real time, providing timely warnings of potential problems in the production process. However, due to the close relationship between equipment status and product quality, relying solely on monitoring either equipment status or product quality is insufficient to comprehensively guarantee equipment reliability and product quality stability. Machining systems, as a crucial component of manufacturing systems, typically consist of multiple CNC machine tools, machining centers, and auxiliary equipment connected in series. Their machining accuracy, operating status, and coordination directly impact product quality and production efficiency. As a key component in the machining process, the condition of cutting tools has a particularly significant impact on the final workpiece quality. While random factors such as uneven material hardness and differences in operator skill can exacerbate workpiece quality fluctuations, tool condition remains a critical factor affecting workpiece quality. During machining, cutting tools gradually wear down due to cutting forces, friction, and material properties, directly affecting machining accuracy. For example, worn cutting edges lead to increased cutting forces, which in turn cause workpiece dimensional errors or excessive surface roughness. Currently, monitoring in tandem machining manufacturing systems primarily focuses on workpiece quality monitoring. This involves collecting workpiece dimensional data at each stage and setting specification lines to determine workpiece quality compliance. Tool status is assessed based on tool usage to determine replacement needs, with very little simultaneous quantitative monitoring of tool wear. However, this approach has several drawbacks: First, ignoring tool status data may lead to undetected potential tool malfunctions, resulting in sudden downtime, resource waste, and even safety hazards. Second, this monitoring method is lagging; quality issues are often only discovered after machining is complete, failing to provide early warnings of workpiece quality problems and causing significant rework and resource waste. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a joint online monitoring method for tandem machining manufacturing systems. By mining historical data, a linear model between tool degradation and workpiece dimensions is constructed to determine the normality of tool status and workpiece quality in real time. When an anomaly is detected, the system issues an early warning and provides diagnostic decision support, achieving comprehensive monitoring and optimized management. The specific implementation path includes: First, collecting tool degradation and workpiece size data at each machining stage and establishing a linear relationship between them based on this data; second, constructing a system proportional risk function for the tool and a verification statistic for workpiece quality; then, estimating model parameters using historical data to determine appropriate thresholds and control limits to ensure the system can effectively monitor potential risks; finally, real-time monitoring of whether the system proportional risk function value exceeds a preset threshold or whether the workpiece quality verification statistic exceeds a control limit. If any limit is exceeded, the system immediately triggers an alarm and provides diagnostic decisions for the maintenance phase. By comprehensively monitoring the tool status and workpiece quality in a tandem machining system, the reliability of the equipment and the stability of the product can be effectively improved, and real-time optimization of the production process can be provided. Through real-time monitoring and early warning mechanisms, the system can intervene in the early stages of problems, avoiding sudden failures and quality fluctuations, thereby improving production efficiency, reducing rework rates, and ultimately optimizing the stability and efficiency of the entire production process.

[0004] The objective of this invention is achieved through the following technical solution:

[0005] A joint online monitoring method for serial machining manufacturing systems includes:

[0006] S1. Collect historical data and perform preprocessing: When the working state of the serial machining manufacturing system is stable, collect historical data of tool status in each stage through sensors, and collect historical data of workpiece size in each stage using measuring equipment; perform preprocessing on all data, including removing missing values, removing outliers, and normalization; historical data of tool status includes historical data of tool degradation and historical data of tool failure.

[0007] S2. Construct a linear model between tool degradation and workpiece dimensions;

[0008] S3. Model parameter estimation: Use experimental design to estimate the coefficients in the linear model; use the maximum likelihood method to estimate the distribution parameters of tool failure time based on historical tool failure data;

[0009] S4. Construct a joint monitoring model: First, use the tool degradation at each stage to construct the trend function of the workpiece size at each stage, solve the difference between the workpiece size and the trend function to form an error vector, and construct a multivariate generalized likelihood ratio (MGLR) control chart; Second, use the tool degradation at each stage as a covariate to construct the proportional risk function of the tool at each stage, and sum the proportional risk functions of each stage to obtain the proportional risk function of the serial machining manufacturing system.

[0010] S5. Simulate the control limits of the multivariate generalized likelihood ratio (MGLR) control chart using the Monte Carlo method; set the threshold of the proportional risk function for the serial machining manufacturing system based on engineering experience;

[0011] S6. Online Joint Monitoring and Diagnosis: Collect tool status data and workpiece size data at each stage online, calculate the multivariate generalized likelihood ratio (MGLR) statistic and the system proportional risk function value; if the multivariate generalized likelihood ratio (MGLR) statistic exceeds the control limit or the proportional risk function value exceeds the threshold, a maintenance stage is required, and the tool at the corresponding stage is prompted to be replaced.

[0012] The present invention also provides a joint online monitoring device for serial machining manufacturing systems, comprising:

[0013] The historical data storage module is used to store historical data of tool status and workpiece dimensions at each stage when the working state of the serial machining manufacturing system is stable.

[0014] The joint monitoring, storage, and computing module is used to clean historical data and construct a linear model between tool degradation and workpiece size, as well as a joint monitoring model of the serial machining manufacturing system. The linear model coefficients and the distribution parameters of tool failure time are obtained through model parameter estimation.

[0015] The real-time data storage module is used to collect tool status data and workpiece size data at each stage in real time through sensors and measuring devices.

[0016] The real-time joint monitoring module is used to solve the real-time multivariate generalized likelihood ratio (MGLR) statistic and the system proportional hazards function value, and to make judgments with the control limits and thresholds respectively.

[0017] The alarm feedback module is used to provide alarm feedback based on the judgment results of the real-time joint monitoring module. If the result is that the system proportional risk function value exceeds the preset threshold, or the multivariate generalized likelihood ratio (MGLR) statistic value exceeds the control limit, an alarm will be triggered, providing diagnostic decisions for the maintenance phase.

[0018] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the joint online monitoring method for a serial machining manufacturing system.

[0019] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the joint online monitoring method for a serial machining manufacturing system.

[0020] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0021] 1. Comprehensive Real-Time Monitoring for Enhanced System Reliability: This invention establishes a linear relationship between tool degradation and workpiece dimensions at each stage by simultaneously collecting and processing such data. By constructing a system proportional risk function and MGLR control chart, it enables real-time comprehensive monitoring of potential tool wear and workpiece quality fluctuations during manufacturing, thereby promptly identifying potential faults and significantly improving equipment reliability and production process stability.

[0022] 2. Targeted early warning to avoid sudden downtime and quality accidents: Utilizing MGLR statistics and the system proportional risk function, the system quantitatively monitors workpiece quality fluctuations and tool risk rates. When the MGLR statistics exceed control limits or the system risk rate exceeds a preset threshold, the system immediately triggers an alarm and provides diagnostic decision support, ensuring targeted intervention in the early stages of problems and avoiding downtime, rework, and resource waste caused by sudden tool failures or workpiece defects.

[0023] 3. Data preprocessing ensures monitoring accuracy: Preprocessing historical data (including removing missing and outliers, and normalization) ensures the accuracy and reliability of the data, making the constructed linear model and monitoring indicators more realistically reflect the on-site situation, thereby enhancing the accuracy and effectiveness of the early warning system.

[0024] 4. Statistical Modeling and Simulation Techniques for Optimizing Monitoring Indicators: Experimental design and maximum likelihood estimation are applied to solve for model parameters, and the Monte Carlo method is used to simulate control limits. Accurate estimation of model parameters and reasonable control limit settings ensure that the monitoring model is more sensitive to tool risk rates and workpiece quality fluctuations, enabling it to detect abnormal changes earlier and thus optimize the production process and management decisions.

[0025] 5. Joint Monitoring Strategy, Balancing Tool Status and Workpiece Quality: A joint online monitoring solution is adopted, focusing on both tool risk rate and real-time monitoring of workpiece quality fluctuations at each stage. This avoids the shortcomings of single monitoring methods. On the one hand, when workpiece quality is stable but tool risk is high, early warnings can be issued to prevent production interruptions caused by tool failure. On the other hand, when tool status is good but workpiece quality fluctuates significantly, timely measures can be taken to prevent defective products, effectively reducing rework rates and production waste, and improving overall system production efficiency and economic benefits.

[0026] In summary, by jointly monitoring tool status and workpiece quality, and utilizing precise data preprocessing, statistical model building, and real-time online early warning, this invention can effectively identify and intervene in risks in the early stages of production, ensuring reliable equipment operation and product quality, thereby bringing higher reliability, stability, and production efficiency to serial machining manufacturing systems. Attached Figure Description

[0027] Figure 1 Schematic diagram of the working process of the joint online monitoring device for a tandem machining system;

[0028] Figure 2 This is a schematic diagram of the scroll machining process;

[0029] Figure 3 Schematic diagram simulating tool degradation values ​​at different stages;

[0030] Figure 4 A schematic diagram simulating workpiece dimensional values ​​at different stages;

[0031] Figure 5 A schematic diagram showing the error variance values ​​at different stages;

[0032] Figure 6 A combined monitoring graph showing alarms triggered by workpiece quality fluctuations;

[0033] Figure 7 A joint monitoring chart showing alarms triggered by high risk rates of cutting tools;

[0034] Figure 8 This is a joint monitoring chart showing simultaneous alarms. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0036] like Figure 1 As shown, this embodiment provides a joint online monitoring device for a serial machining manufacturing system, comprising:

[0037] The historical data storage module is used to store historical data of tool status and workpiece dimensions at each stage when the machining manufacturing system is in a stable working state.

[0038] The joint monitoring, storage, and computing module is used to clean historical data and construct a linear model between tool degradation and workpiece size, as well as a joint monitoring model of the serial machining manufacturing system. The linear model coefficients and the distribution parameters of tool failure time are obtained through model parameter estimation.

[0039] The real-time data storage module is used to collect tool status data and workpiece size data at each stage in real time through sensors and measuring devices.

[0040] The real-time joint monitoring module is used to solve the real-time multivariate generalized likelihood ratio (MGLR) statistic and the system proportional hazards function value, and to make judgments with the control limits and thresholds respectively.

[0041] The alarm feedback module is used to provide alarm feedback based on the judgment results of the real-time joint monitoring module. If the result is that the system proportional risk function value exceeds the preset threshold or the MGLR statistic value exceeds the control limit, an alarm will be triggered, providing diagnostic decisions for the maintenance phase.

[0042] Specifically, the workflow of the aforementioned joint online monitoring device is as follows:

[0043] (1) Historical Data Collection and Preprocessing: Acquire m sets of historical data on tool status and workpiece dimensions under stable working conditions. The failure time of the m sets in the kth stage is represented as... The tool degradation data for the m groups in stage k are represented as follows: Workpiece dimension data is represented as in, This represents the tool failure time in stage k. and These represent the tool degradation data and workpiece dimension data for the t-th group, respectively. This represents the i-th degradation index in the k-th stage at time t. This represents the j-th workpiece position in the k-th stage at time t, with a total of N stages, p workpiece positions, and q degradation indices. Records with many missing values ​​are deleted; K-means is used to identify and remove outliers in the tool degradation and workpiece dimension data, and the data is normalized. The historical tool degradation data and historical workpiece dimension data are arranged by time to form a historical data table, as shown in Table 1.

[0044] Table 1 Historical Data on Tool Degradation and Workpiece Dimensions

[0045]

[0046] (2) Constructing a linear model between tool degradation and workpiece size: Using a generalized linear model, the relationship between tool degradation and the i-th workpiece size in the k-th stage is constructed as follows:

[0047]

[0048] here It is a scalar value. This represents the coefficient vector affected by tool degradation. The coefficient vector represents the coefficients affected by random factors. This represents the coefficient matrix that is affected by the interaction between tool degradation and random factors. Represents a vector of random factors and

[0049] (3) Constructing a joint monitoring model: First, a monitoring model for workpiece quality fluctuations in a serial machining manufacturing system is constructed. It is assumed that the error value of the q workpiece dimensions in the k-th stage is... and here Represents the covariance of the system under control. This represents the covariance of the system under runaway conditions. Assume... Let represent the trend function of the i-th workpiece size in the k-th stage at time t, and

[0050]

[0051] here yes The observed value at time t. Then the trend function for the dimensions of the q workpieces in the k-th stage is: And there are

[0052]

[0053] According to equation (1), the trend function Further derivation as

[0054]

[0055] The n sets of workpiece size vectors collected from observations With trend function vector The difference is used to obtain the n groups of error values ​​for the k-th stage.

[0056]

[0057] Consider about The log-likelihood function of the covariance is

[0058]

[0059] Assume the error vector at time τ in the k-th stage A drift occurred, meaning the covariance changed from It became The MGLR statistic for stage k at time t is constructed using the MGLR test.

[0060]

[0061] here and Furthermore, the MGLR statistic for the k-th stage at time t is obtained as follows:

[0062]

[0063] make Available Approximate, further,

[0064]

[0065] In fact, It can be approximated as Detailed derivation is omitted here. The MGLR statistic for the k-th stage at time t is:

[0066]

[0067] Based on this, the MGLR statistic of the serial machining manufacturing system at time t is:

[0068]

[0069] Next, based on the tool degradation data The proportional risk function of the tool in stage k at time t is constructed as follows:

[0070]

[0071] Here λ k and η k These represent the distribution parameters of the failure time in stage k, respectively. Let represent the influence coefficient of tool degradation in stage k on the tool risk rate. Based on this, the proportional risk function of the tandem machining manufacturing system at time t is:

[0072]

[0073] (4) Joint monitoring model parameter estimation: Based on historical data of tool degradation and workpiece size, experimental design is used to estimate the coefficients in the generalized nonlinear model. Based on historical tool failure data F kFor k = 1, ..., N, use maximum likelihood estimation to estimate the distribution parameter λ of the tool failure time. k ,η k ,,k=1,…,N.

[0074] (5) Simulation of control limits and threshold setting of proportional hazard function value for MGLR control chart: Given a Type I error, the control limit L is simulated using the Monte Carlo method. Q Based on engineering experience, set the system proportional risk function threshold L. D .

[0075] (6) Joint online monitoring and diagnosis: Based on the real-time collection of tool degradation and workpiece size data at each stage, the proportional risk function H of the serial machining manufacturing system is calculated respectively. t With MGLR statistic R t If the proportional risk function value of the serial machining manufacturing system exceeds the threshold, i.e., H t ≥L D Or the MGLR statistic value exceeds the control limit, i.e., R t ≥L Q An alarm signal will be issued if the joint monitoring model triggers an alarm. When an alarm occurs, it means that the cutting tool in the serial machining manufacturing system is operating under high risk or the workpiece dimensional quality is unstable. At this point, it is necessary to diagnose the corresponding stage and replace the cutting tool for that stage. The diagnosis for the maintenance stage is given by the following formula.

[0076]

[0077] here and

[0078] Furthermore, a simulation example is used to verify the applicability and effectiveness of the combined online monitoring method of the present invention. The simulation case takes the four-stage tandem compressor scroll machining process as the object, covering several key machining operations, mainly including milling, turning, drilling, and grooving. The machining process is shown below. Figure 2 Different tool degradation indices are involved in each machining stage, including tool cutting force, tool temperature, spindle vibration, and motor current. Furthermore, the workpiece dimensions at each machining stage include three key locations, which directly affect the performance and quality of the vortex-machined parts. We first simulate tool degradation data using a gamma process, and then simulate workpiece dimension data based on the linear relationship between tool degradation and workpiece dimensions. Simultaneously, to verify the superiority of the joint monitoring scheme, it is compared with existing schemes that only monitor tool condition and only monitor workpiece quality. The main simulation process is as follows:

[0079] (1) Instance parameter settings

[0080] Consider a machining process with a monitoring field of view T = 100 hours. Assume that the i-th degradation index of the tool in the k-th stage follows a scale parameter of... and shape parameters are The gamma process, namely Assume the error value in the k-th stage Controlled covariance make

[0081]

[0082] here,

[0083]

[0084] The specific parameter values ​​are shown in Table 2.

[0085] Table 2 Parameter Value Settings for Each Stage

[0086]

[0087] (2) Simulation Result Analysis

[0088] Based on the parameter values ​​in Table 2, Figure 3 The simulation showed tool degradation values ​​at different stages, with the degradation value increasing over time, leading to a continuous increase in the failure risk rate. Meanwhile, Figure 4 The simulation showed workpiece size values ​​at different stages, with the red line representing the trend of workpiece size values, which gradually deviated from the trend value as time increased. Figure 5 The data provides the variance values ​​for errors at different stages. It can be seen that the variance values ​​for each stage of error change relatively little within a certain time period, but increase rapidly after a certain time. This indicates that workpiece quality fluctuations are small in the early stages of processing, but gradually increase in the later stages as the tool wears down. Therefore, timely monitoring of changes in workpiece quality fluctuations during the process is crucial for early warning of workpiece quality problems. Figure 6-8The table shows different monitoring results for simultaneously monitoring the tool system risk rate and workpiece size fluctuation, representing alarms caused by workpiece quality fluctuation, tool system risk rate, and both. According to the different alarm results in Table 3, workpiece quality fluctuation caused a joint monitoring alarm at hour 52, indicating the need to replace the tool in stage 4; high tool risk rate caused a joint monitoring alarm at hour 56, indicating the need to replace the tool in stage 2; and both caused a joint monitoring alarm at hour 56, indicating the need to replace the tool in stage 4. To compare the superior performance of the proposed monitoring scheme compared to existing schemes, the average alarm time will be compared with schemes that only monitor workpiece quality and only monitor tool status. Table 4 represents the average alarm time of different monitoring schemes under different drift conditions, where small drift, medium drift, and large drift are represented by values ​​of 0.1, 0.5, and 1, respectively. Drift type 1, type 2, and type 3 represent the number of workpiece size drifts in stage k. Small drift corresponds to drift type 1, type 2, and type 3 respectively.

[0089]

[0090] Mid-drift corresponds to drift type 1, type 2 and type 3 respectively.

[0091]

[0092] Large drift corresponds to drift type 1, while types 2 and 3 are respectively...

[0093]

[0094] As shown in Table 4, under different drift amplitudes and types, the average alarm time of the proposed joint monitoring scheme is shorter than that of the schemes that only monitor workpiece quality and the schemes that only monitor tool status. This indicates that the joint monitoring scheme can detect high-risk rates of workpiece quality fluctuations and tool failures at each stage earlier, ensuring the stability of workpiece quality and the reliability of the tool in the system. Furthermore, as the drift amplitude increases, the average alarm time of both the proposed joint monitoring scheme and the scheme that only monitors workpiece quality decreases continuously, while the average alarm time of the scheme that only monitors tool status remains unchanged. Under different drift types, both the proposed joint monitoring scheme and the scheme that only monitors workpiece quality are sensitive to the number of drifts in workpiece dimensions; that is, the average alarm time for type 3 is shorter than that for type 2, and the average alarm time for type 2 is shorter than that for type 1. In summary, the proposed joint monitoring scheme can promptly detect high-risk tool operating states and workpiece size fluctuations at each stage, ensuring the stability and reliability of the cascaded vortex machining process.

[0095] Table 3 Different types of alarms in the joint monitoring scheme

[0096]

[0097] Table 4. Average alarm time of different monitoring schemes under different drift conditions.

[0098]

[0099] This invention is not limited to the embodiments described above. The above description of specific embodiments is intended to illustrate and explain the technical solutions of this invention. The specific embodiments described above are merely illustrative and not restrictive. Without departing from the spirit and scope of the claims, those skilled in the art can make many specific modifications based on the teachings of this invention, and these modifications all fall within the scope of protection of this invention.

Claims

1. A joint online monitoring method for serial machining manufacturing systems, characterized in that, include: S1. Collect historical data and perform preprocessing: When the working state of the serial machining manufacturing system is stable, collect historical data of tool status in each stage through sensors, and collect historical data of workpiece size in each stage using measuring equipment; perform preprocessing on all data, including removing missing values, removing outliers, and normalization. Tool condition history data includes tool degradation history data and tool failure history data; S2. Construct a linear model between tool degradation and workpiece dimensions; S3. Model parameter estimation: Use experimental design to estimate the coefficients in the linear model; use the maximum likelihood method to estimate the distribution parameters of tool failure time based on historical tool failure data; S4. Construct a joint monitoring model: First, use the tool degradation at each stage to construct the trend function of the workpiece size at each stage, solve the difference between the workpiece size and the trend function to form an error vector, and construct a multivariate generalized likelihood ratio (MGLR) control chart; Second, use the tool degradation at each stage as a covariate to construct the proportional risk function of the tool at each stage, and sum the proportional risk functions of each stage to obtain the proportional risk function of the serial machining manufacturing system. S5. Simulate the control limits of the multivariate generalized likelihood ratio (MGLR) control chart using the Monte Carlo method; set the threshold of the proportional risk function for the serial machining manufacturing system based on engineering experience; S6. Online Joint Monitoring and Diagnosis: Collect tool status data and workpiece size data at each stage online, calculate the multivariate generalized likelihood ratio (MGLR) statistic and the system proportional risk function value; if the multivariate generalized likelihood ratio (MGLR) statistic exceeds the control limit or the proportional risk function value exceeds the threshold, a maintenance stage is required, and the tool at the corresponding stage is prompted to be replaced.

2. A joint online monitoring device for serial machining manufacturing systems, characterized in that, include: The historical data storage module is used to store historical data of tool status and workpiece dimensions at each stage when the working state of the serial machining manufacturing system is stable. The joint monitoring, storage, and computing module is used to clean historical data and construct a linear model between tool degradation and workpiece size, as well as a joint monitoring model of the serial machining manufacturing system. The linear model coefficients and the distribution parameters of tool failure time are obtained through model parameter estimation. The real-time data storage module is used to collect tool status data and workpiece size data at each stage in real time through sensors and measuring devices. The real-time joint monitoring module is used to solve the real-time multivariate generalized likelihood ratio (MGLR) statistic and the system proportional hazards function value, and to make judgments with the control limits and thresholds respectively. The alarm feedback module is used to provide alarm feedback based on the judgment results of the real-time joint monitoring module. If the result is that the system proportional risk function value exceeds the preset threshold, or the multivariate generalized likelihood ratio (MGLR) statistic value exceeds the control limit, an alarm will be triggered, providing diagnostic decisions for the maintenance phase.

3. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the joint online monitoring method for serial machining manufacturing systems as described in claim 1.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the joint online monitoring method for serial machining manufacturing systems as described in claim 1.

Images