Intelligent pressure forming equipment state sensing method fusing multi-source sensor signals

Abstract

The invention belongs to the field of intelligent manufacturing, and particularly relates to a pressure forming equipment state intelligent sensing method fusing multi-source sensor signals, which comprises the following steps: S1, setting data acquisition equipment; s2, a data acquisition device acquires monitoring data of each sensor and sends the monitoring data to a data processing center; s3, the data processing center preprocesses the received monitoring data and extracts feature data; s4, the data processing center calculates the characteristic data through a distributed Kalman filtering algorithm to generate monitoring indexes of the equipment state, and the monitoring indexes comprise local monitoring indexes and global monitoring indexes; s5, the data processing center analyzes the running state of the equipment according to the monitoring indexes and the monitoring data of the sensors; and S6, the data processing center sends an analysis result to an equipment management system, and the equipment management system generates a control instruction according to the analysis result and stores and displays the control instruction. The problems that pressure forming equipment is poor in product forming quality and low in efficiency are solved.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent manufacturing, and in particular relates to an intelligent perception method for the state of pressure forming equipment by fusing multi-source sensor signals. Background Art

[0002] Pressure forming equipment has an irreplaceable position in the process of modern manufacturing. Pressure forming equipment is used to accurately control pressure to achieve high-precision forming, meet the standards of the precision field, optimize material performance and ensure the consistency of batch product quality. With the booming development of intelligent manufacturing, pressure forming equipment operation status monitoring faces challenges and opportunities. The traditional way of relying on a single sensor to monitor individual parameters has limitations in data collection and can only provide specific parameter information, which is difficult to adapt to its complex working conditions. Single monitoring may miss early warning signals, resulting in hidden accumulation of faults until the problem becomes serious. In this case, not only will the maintenance cost increase, but it may also cause production line shutdowns, disrupt production rhythms, and affect production efficiency; in the field of precision manufacturing, product quality consistency is crucial. A single sensor may not ensure that all key process parameters are within the optimal range, thereby affecting the quality and performance of the final product; modern intelligent manufacturing should have real-time monitoring, intelligent analysis and rapid response. The amount of information provided by a single sensor is limited, which is not enough to support complex decision-making processes, nor can it meet the needs of smart factories for automation and optimization. Due to the interweaving of multiple factors such as equipment mechanical stability, mold working conditions, and power system coordination, single monitoring seriously affects product molding quality and efficiency, which does not meet the demands of intelligent manufacturing. Summary of the invention

[0003] The present invention provides a method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals. By fusing multi-sensor data, all-round monitoring is achieved to solve the problems of poor product forming quality and low efficiency of pressure forming equipment.

[0004] The basic solution provided by the present invention is a method for intelligently sensing the state of a pressure forming equipment by fusing multi-source sensor signals, which specifically includes the following steps: S1: Data acquisition equipment is installed on the equipment according to production requirements, and the data acquisition equipment includes a force sensor, a pressure sensor, a temperature sensor, a velocity sensor, an acceleration sensor, a displacement sensor and a proximity sensor; S2: The data acquisition device collects the monitoring data of each sensor and sends the monitoring data to the data processing center; S3: The data processing center pre-processes the received monitoring data and extracts feature data; S4: The data processing center calculates the characteristic data through a distributed Kalman filter algorithm to generate monitoring indicators of equipment status, wherein the monitoring indicators include local monitoring indicators and global monitoring indicators; S4-1) Inputting the characteristic data of each detection data into the corresponding local filter respectively, obtaining the local state prediction and local covariance prediction of each monitoring data, and each local filter updates and outputs the corresponding local state prediction and local covariance prediction as the local monitoring index along with the monitoring data; S4-2) performing weighted average on the latest output local state prediction and local covariance prediction to obtain a global state estimate as a global monitoring indicator output of the equipment state, and returning the global state estimate to S4-1) to update the parameters of each local filter; S5: The data processing center analyzes the equipment operation status according to the monitoring indicators and the monitoring data of each sensor to obtain the analysis results; S6: The data processing center sends the analysis results to the equipment management system. The equipment management system generates control instructions based on the analysis results and sends them to the equipment actuator, and stores and displays the analysis results.

[0005] The working principle and advantages of the present invention are as follows: by integrating the monitoring signals of pressure forming equipment from force sensors, pressure sensors, temperature sensors, speed sensors, acceleration sensors, displacement sensors and proximity sensors, monitoring indicators are generated, and the status of the monitoring indicators is analyzed through the system management system, the analysis results are displayed and stored, and production is guided. This method can not only integrate multiple types of sensor signals to keenly capture subtle changes in equipment operation for accurate monitoring, change the previous disadvantages of single sensor monitoring, reduce the risk of sudden failures, and avoid downtime caused by accumulated failures that disrupt production rhythm and affect production efficiency; and can rely on accurate control of equipment status to ensure product quality stability, reduce defective and scrap products, and meet the demand for high-quality production; at the same time, each monitoring data can monitor the equipment status in real time and assist in issuing control instructions to avoid invalid operation and help optimize process parameters to improve production efficiency; each monitoring data and monitoring indicators are collected and analyzed for a long time, which is helpful to draw a "health curve", optimize equipment maintenance strategies, realize the transformation from passive maintenance to preventive maintenance, and extend equipment life; at the same time, it provides underlying data support for intelligent manufacturing management and control upgrades, achieves remote monitoring and intelligent scheduling, and enhances the intelligence level and coordination of the system; and multi-source sensors are redundant backups for each other, which enhances the adaptability and reliability of the system and ensures the continuous and stable operation of equipment in complex environments; The monitoring indicators of the equipment are generated by the distributed Kalman filter algorithm, which can integrate the data from multiple sensors, integrate the information advantages of each sensor, reduce the measurement error and noise interference of a single sensor, and thus more accurately estimate the state of the equipment. In addition, the Kalman filter has a certain fault tolerance. When one of the sensors is abnormal, the data of other sensors can still be used for state estimation, ensuring the normal output of monitoring indicators and improving the reliability of the pressure forming equipment. Compared with the centralized filtering method, the distributed Kalman filter performs local filtering calculations on each sensor independently, reducing the burden of data transmission and processing, improving the real-time and response speed of the system, and can obtain the equipment status information in a timely and accurate manner during the high-speed operation of the pressure forming equipment to provide support for real-time control. At the same time, the distributed Kalman filter only needs to perform calculations at local nodes, and only necessary information needs to be exchanged between nodes, which greatly reduces the amount of communication and can reduce the communication cost. In addition, the distributed Kalman filter algorithm can increase the sensors of the pressure forming equipment as needed, improving the scalability of the equipment. The distributed Kalman filter can provide more accurate feedback information for the control of the pressure forming equipment, so that the control system can adjust the operating parameters of the equipment in time according to the actual state, and improve the forming quality and production efficiency.

[0006] Preferably, in step S4-1, the expressions for state prediction and covariance prediction of each monitoring data are as follows: The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the force sensor are:

[0007] In the formula, is the local state prediction of the monitoring data of the force sensor at time point k1, is the local state prediction of the monitoring data of the force sensor at time point k1-1, is the local covariance prediction of the monitoring data of the force sensor at time point k1, is the local covariance prediction of the monitoring data of the force sensor at time point k1-1, is the process noise covariance matrix of the monitoring data of the force sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the pressure sensor are:

[0008] In the formula, is the local state prediction of the monitoring data of the pressure sensor at time point k2, is the local state prediction of the monitoring data of the pressure sensor at time point k2-1, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2-1, is the process noise covariance matrix of the monitoring data of the pressure sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the temperature sensor are:

[0009] In the formula, is the local state prediction of the temperature sensor monitoring data at time point k3, is the local state prediction of the temperature sensor monitoring data at time point k3-1, is the local covariance prediction of the temperature sensor monitoring data at time point k3, is the local covariance prediction of the temperature sensor monitoring data at time point k3-1, is the process noise covariance matrix of the temperature sensor monitoring data, is the state transfer matrix of the monitoring data of the temperature sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the speed sensor are:

[0010] In the formula, is the local state prediction of the speed sensor monitoring data at time point k4, is the local state prediction of the speed sensor monitoring data at time point k4-1, is the local covariance prediction of the speed sensor monitoring data at time point k4, is the local covariance prediction of the speed sensor monitoring data at time point k4-1, is the process noise covariance matrix of the speed sensor monitoring data, is the state transfer matrix of the monitoring data of the speed sensor, is the control input matrix; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the acceleration sensor are:

[0011] In the formula, is the local state prediction of the acceleration sensor monitoring data at time point k5, is the local state prediction of the acceleration sensor monitoring data at time point k5-1, is the local covariance prediction of the acceleration sensor monitoring data at time point k5, is the local covariance prediction of the acceleration sensor monitoring data at time point k5-1, is the noise covariance matrix of the monitoring data of the acceleration sensor in the process; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the displacement sensor are:

[0012] In the formula, is the local state prediction of the displacement sensor monitoring data at time point k5, is the local state prediction of the displacement sensor monitoring data at time point k6-1, is the local covariance prediction of the monitoring data of the displacement sensor at time point k6, is the local covariance prediction of the displacement sensor monitoring data at time point k6-1, is the noise covariance matrix of the displacement sensor monitoring data in the process, is the state transfer matrix of the monitoring data of the displacement sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the proximity sensor are:

[0013] In the formula, is the local state prediction of the monitoring data of the proximity sensor at time point k7, is the local state prediction of the monitoring data of the proximity sensor at time point k7-1, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7-1, is the process noise covariance matrix of the monitoring data of the proximity sensor, It is the state transfer matrix of the monitoring data of the proximity displacement sensor. The local state prediction and local covariance prediction of the monitoring data of multiple sensors are respectively performed to accurately grasp the working conditions of each sensor monitoring object and facilitate the subsequent data fusion.

[0014] Preferably, in step S4-2, the global state estimation includes global state prediction and global covariance prediction, and the global state estimation expression is as follows:

[0015] In the formula, is the global state prediction, is the global covariance prediction, is the state estimate of the latest output of each local filter, is the estimated covariance of the latest output of each local filter, is the weight of each local filter, is the number of local filters.

[0016] By outputting the weighted average comprehensive data as monitoring indicators, the equipment management system can make more accurate judgments and suggestions, effectively improving the performance of the measurement system and making it more robust and efficient when facing complex and changeable actual application scenarios.

[0017] Preferably, in step S1, the force sensor is used to monitor the force applied by the mold during stamping; the pressure sensor is used to monitor the pressure in the hydraulic system pipeline; the temperature sensor is used to monitor the temperature of the workpiece; the speed sensor is used to monitor the speed of the mold movement; the acceleration sensor is used to monitor the vibration frequency and amplitude of the mold; the displacement sensor is used to monitor the working stroke of the mold; and the proximity sensor is used to monitor the position of the moving parts. The pressure forming equipment is monitored by multiple sensors to facilitate the operation status of the equipment at any time. The force applied by the mold is monitored by the force sensor to ensure that the workpiece reaches the required shape and size. The system pressure is ensured to be within the specified range by the pressure sensor. The mold temperature is monitored by the temperature sensor to obtain the temperature distribution of the workpiece, which is convenient for grasping the molding state and helps to make a comprehensive judgment based on the temperature and other information. The speed sensor is used to accurately control the speed, optimize the stamping process, make the contact time and force between the mold and the workpiece meet the process requirements, improve production efficiency and product consistency, monitor the vibration of the mold by the acceleration sensor, judge whether the equipment is working normally, and locate the source of vibration, and monitor the impact force of the mold to avoid damage to the mold. The displacement sensor is used to accurately measure the working stroke to control the molding pressure and the deformation degree of the workpiece. The proximity sensor is used to monitor whether the moving parts are beyond the safety range to prevent collisions. In addition, the monitoring indicators integrate multi-source data to achieve multi-dimensional information integration, which can evaluate equipment from a system level and comprehensively present the equipment operation status, overcoming the defects of analyzing sensors one by one only to obtain local, single-dimensional information and easily ignore parameter correlations. Data fusion technology is used to reduce the impact of measurement noise, and data can be verified and supplemented with each other to greatly improve work efficiency, providing strong support for equipment management and production optimization.

[0018] Preferably, in step S2, the data acquisition device performs signal amplification, noise filtering, and linearization calibration on each monitoring data in turn, and then sends it to the data processing center through the field bus. By amplifying the signal, it is ensured that the signal will not be submerged or lost in subsequent processing and transmission, and the resolution and accuracy of the final measurement result can be improved. By filtering the signal noise, the anti-interference ability of the signal can be enhanced, the influence of external factors on the measurement result can be reduced, and a cleaner signal can be provided. By linearizing and calibrating the data, the measured value is more accurate, and the readings between different sensors are more consistent, which is convenient for comparison and integration. By processing the data, the data quality and the overall performance of the system can be significantly improved, and the communication efficiency can be optimized; and the field bus is used for data transmission, and the signal transmission is relatively stable, meeting the requirements of intelligent production.

[0019] Preferably, in step S5, the analysis results include monitoring indicator performance analysis and equipment operation performance analysis, wherein the monitoring indicator performance analysis includes the change trend of each monitoring indicator and the correlation between each local monitoring indicator, and the equipment operation performance analysis includes equipment production efficiency, equipment accuracy, equipment health status and remaining service life, equipment fault type and location. By analyzing the change trend and correlation of each monitoring indicator to analyze the equipment operation status, the monitoring indicators integrate multi-source data, and by analyzing the trend of the monitoring indicators, the equipment operation status can be presented more comprehensively. By analyzing the correlation of the monitoring indicators, the operation status of each component of the equipment can be better grasped to grasp the overall operation information of the equipment; and through the monitoring data of multi-source sensors, the fault can be located more accurately, the analysis results can avoid being misled by noise, and the equipment status can be quickly judged as a whole, maintenance decisions can be optimized, potential faults can be discovered in advance, maintenance can be reasonably arranged, downtime can be reduced, work efficiency can be greatly improved, and strong support can be provided for equipment management and production optimization.

[0020] Further preferably, the analysis results also include equipment development forecast analysis, and the historical data of each monitoring indicator is fitted through a mathematical model to obtain the equipment development forecast result, and the corresponding monitoring indicator threshold is set. When the equipment development forecast result exceeds the monitoring indicator threshold, the equipment management system will issue a warning signal. By predicting the monitoring indicators, when the equipment pre-occupancy forecast result exceeds the set threshold, an alarm is issued to facilitate maintenance personnel to have sufficient time to perform equipment inspections, troubleshooting, and necessary maintenance operations, effectively avoiding sudden equipment failures, greatly reducing equipment downtime, and ensuring the continuity and stability of production.

[0021] Further preferably, the variation trend of each monitoring indicator is analyzed by a data trend analysis model, and the expression of the data trend analysis model is as follows:

[0022] In the formula, is the value of each monitoring indicator at time t, is the lag operator , p is the autoregressive order, d is the difference order, q is the sliding average order, and are the autoregressive and moving average coefficients of the model, It is a white noise sequence. By analyzing the changing trend of the monitoring indicators through the analysis model, the changing trend of each monitoring indicator can be accurately obtained, which is convenient for real-time evaluation of the equipment operation status. By observing the changes of each indicator at different stages, it can be accurately judged whether the equipment is operating normally.

[0023] Preferably, in step S6, the display content of the equipment management system includes each monitoring indicator and the monitoring data of each sensor, and the monitoring indicators and the monitoring data of each sensor displayed by the equipment management system are displayed in a table or a chart. The monitoring indicators and monitoring data are displayed in real time, and the staff can intuitively observe the changes in the monitoring data and monitoring indicators, fully understand the equipment operation status, and respond in time.

[0024] Preferably, in step S6, the equipment management system generates production suggestions according to the evaluation results. Production suggestions are automatically generated according to the evaluation results obtained from the monitoring data of the multi-source sensors to guide production. Suggestions based on precise data analysis are usually more accurate and reliable than human judgment, helping employees make more informed choices and reducing the risk of wrong decisions. Automatically generated production suggestions reduce the time and energy required for manual analysis and decision-making, which not only reduces labor intensity, but also enables staff to focus on higher-value tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of the present invention; Figure 2 It is a schematic diagram of field bus transmission of the present invention; Figure 3 A flowchart of the global monitoring index calculation of the present invention; Figure 4 A comparison chart of each monitoring data after data fusion and before data fusion of the present invention; Figure 5 It is a monitoring index curve diagram after fusion of the present invention; Figure 6 It is the mean square error diagram of the monitoring indicators after fusion of the present invention; Figure 7 A histogram of the sensor monitoring data of the present invention; Figure 8 This is a box plot comparison diagram of the monitoring indicators of the present invention after fusion with the original monitoring data. DETAILED DESCRIPTION

[0026] The following is further described in detail through specific implementation methods: The specific implementation process is as follows: Figures 1 to 8 , a method for intelligently sensing the state of a pressure forming equipment by fusing multi-source sensor signals, specifically comprising the following steps: S1: Data acquisition equipment is installed on the equipment according to production requirements, and the data acquisition equipment includes a force sensor, a pressure sensor, a temperature sensor, a velocity sensor, an acceleration sensor, a displacement sensor and a proximity sensor; In step S1, the force sensor is used to monitor the force applied when the mold is stamping; the pressure sensor is used to monitor the pressure in the hydraulic system pipeline; the temperature sensor is used to monitor the workpiece temperature; the speed sensor is used to monitor the mold movement speed; the acceleration sensor is used to monitor the vibration frequency and amplitude of the mold; the displacement sensor is used to monitor the working stroke of the mold; and the proximity sensor is used to monitor the position of the moving parts.

[0027] In this embodiment, the force sensor is installed at the connection between the stamping die and the press to monitor the force applied by the die during stamping in real time. The shape of automobile parts is complex and the forming force is strictly required. The force sensor ensures that each part is accurately subjected to force, guarantees the forming quality, and avoids defects caused by insufficient or excessive pressure. The pressure sensor is arranged at the key nodes of the hydraulic system pipeline to monitor the pressure changes of the system. When the pressure is abnormal, such as exceeding the normal range, it indicates that the hydraulic pump is faulty or the pipeline is blocked. Timely alarm can prevent equipment damage, ensure stable pressure supply, and maintain production continuity. Temperature sensors are installed in both the hot pressing mold and the heating furnace. The temperature sensor on the mold monitors the forming temperature of the workpiece, and the heating temperature in the heating furnace is controlled to ensure that the workpiece is formed at a suitable temperature to avoid temperature problems affecting material properties and product quality. The speed sensor is installed on the drive shaft of the press slide to monitor the movement speed of the slide. By accurately controlling the speed, the stamping process is optimized, so that the contact time and force between the mold and the workpiece meet the process requirements, and the production efficiency and product consistency are improved. Acceleration sensors are installed at key points of the mold and transmission system to monitor vibration. The mold will vibrate due to impact force during stamping and wear or imbalance of transmission parts. Acceleration sensors monitor in real time, determine the operating status of the equipment, locate the vibration source, and avoid mold damage and equipment failure. Displacement sensors are installed on the active parts and sliders of the mold to monitor the motion trajectory and working stroke. Through precise measurement, they control the molding pressure and the degree of deformation of the workpiece to ensure the dimensional accuracy of parts. When the displacement is abnormal, it indicates mold wear or installation problems. Proximity sensors are used in the feeding mechanism to monitor whether the material is delivered on time and accurately to prevent the loss or dislocation of materials from affecting production. At the same time, they are installed at the end of the mold stroke to prevent the mold from damaging the equipment due to excessive stroke, ensuring the safety of equipment and operators.

[0028] S2: The data acquisition device collects the monitoring data of each sensor and sends the monitoring data to the data processing center; In step S2, the data acquisition device amplifies, filters noise, and linearizes each monitoring data in turn through a signal conditioning circuit, and then sends it to a data processing center through a field bus. The field bus uses a CAN bus or HART, and the CAN bus is preferably used in this embodiment.

[0029] S3: The data processing center pre-processes the received monitoring data and extracts feature data; The preprocessing includes data conversion and removal of outliers.

[0030] S4: The data processing center calculates the characteristic data through a distributed Kalman filter algorithm to generate monitoring indicators of equipment status, wherein the monitoring indicators include local monitoring indicators and global monitoring indicators; S4-1) Inputting the characteristic data of each detection data into the corresponding local filter respectively, obtaining the local state prediction and local covariance prediction of each monitoring data, and each local filter updates and outputs the corresponding local state prediction and local covariance prediction as the local monitoring index along with the monitoring data; In step S4-1, the expressions for state prediction and covariance prediction of each monitoring data are as follows: The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the force sensor are:

[0031] In the formula, is the local state prediction of the monitoring data of the force sensor at time point k1, is the local state prediction of the monitoring data of the force sensor at time point k1-1, is the local covariance prediction of the monitoring data of the force sensor at time point k1, is the local covariance prediction of the monitoring data of the force sensor at time point k1-1, is the process noise covariance matrix of the monitoring data of the force sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the pressure sensor are:

[0032] In the formula, is the local state prediction of the monitoring data of the pressure sensor at time point k2, is the local state prediction of the monitoring data of the pressure sensor at time point k2-1, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2-1, is the process noise covariance matrix of the monitoring data of the pressure sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the temperature sensor are:

[0033] In the formula, is the local state prediction of the temperature sensor monitoring data at time point k3, is the local state prediction of the temperature sensor monitoring data at time point k3-1, is the local covariance prediction of the temperature sensor monitoring data at time point k3, is the local covariance prediction of the temperature sensor monitoring data at time point k3-1, is the process noise covariance matrix of the temperature sensor monitoring data, is the state transfer matrix of the monitoring data of the temperature sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the speed sensor are:

[0034] In the formula, is the local state prediction of the speed sensor monitoring data at time point k4, is the local state prediction of the speed sensor monitoring data at time point k4-1, is the local covariance prediction of the speed sensor monitoring data at time point k4, is the local covariance prediction of the speed sensor monitoring data at time point k4-1, is the process noise covariance matrix of the speed sensor monitoring data, is the state transfer matrix of the monitoring data of the speed sensor, is the control input matrix; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the acceleration sensor are:

[0035] In the formula, is the local state prediction of the acceleration sensor monitoring data at time point k5, is the local state prediction of the acceleration sensor monitoring data at time point k5-1, is the local covariance prediction of the acceleration sensor monitoring data at time point k5, is the local covariance prediction of the acceleration sensor monitoring data at time point k5-1, is the noise covariance matrix of the monitoring data of the acceleration sensor in the process; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the displacement sensor are:

[0036] In the formula, is the local state prediction of the displacement sensor monitoring data at time point k5, is the local state prediction of the displacement sensor monitoring data at time point k6-1, is the local covariance prediction of the monitoring data of the displacement sensor at time point k6, is the local covariance prediction of the displacement sensor monitoring data at time point k6-1, is the noise covariance matrix of the displacement sensor monitoring data in the process, is the state transfer matrix of the monitoring data of the displacement sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the proximity sensor are:

[0037] In the formula, is the local state prediction of the monitoring data of the proximity sensor at time point k7, is the local state prediction of the monitoring data of the proximity sensor at time point k7-1, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7-1, is the process noise covariance matrix of the monitoring data of the proximity sensor, is the state transfer matrix of the monitoring data of the proximity displacement sensor.

[0038] In step S4-1, the local state prediction and local covariance prediction expressions output by each local filter according to the Kalman filter update output along with the monitoring data are as follows: The local state prediction and local covariance prediction expressions of the force sensor monitoring data update output are:

[0039] In the formula, is the Kalman gain, Update the output observation matrix for the monitoring data of the force sensor, Update the measurement noise covariance matrix output by the monitoring data of the force sensor, is the monitoring data of the force sensor, I is the unit matrix, is the estimated state value after the force sensor monitoring data is updated, is the estimated covariance after the monitoring data of the force sensor is updated; The local state prediction and local covariance prediction expressions of the pressure sensor monitoring data update output are:

[0040] In the formula, is the Kalman gain, Update the output observation matrix for the monitoring data of the pressure sensor, Update the measurement noise covariance matrix output by the pressure sensor monitoring data, is the monitoring data of the pressure sensor, I is the unit matrix, is the estimated value of the state after the monitoring data of the pressure sensor is updated, is the estimated covariance after the monitoring data of the pressure sensor is updated; The local state prediction and local covariance prediction expressions of the temperature sensor monitoring data update output are:

[0041] In the formula, is the Kalman gain, Update the output observation matrix for the temperature sensor’s monitoring data, Update the measurement noise covariance matrix output by the temperature sensor monitoring data, is the monitoring data of the temperature sensor, I is the unit matrix, is the estimated value of the state after the temperature sensor monitoring data is updated, is the estimated covariance after the monitoring data of the temperature sensor is updated; The local state prediction and local covariance prediction expressions of the speed sensor monitoring data update output are:

[0042] In the formula, is the Kalman gain, Update the output observation matrix for the speed sensor monitoring data, Update the measurement noise covariance matrix output by the speed sensor monitoring data, is the monitoring data of the speed sensor, I is the unit matrix, is the state estimation value after the speed sensor monitoring data is updated, is the estimated covariance after the monitoring data of the speed sensor is updated; The local state prediction and local covariance prediction expressions of the monitoring data update output of the acceleration sensor are:

[0043] In the formula, is the Kalman gain, Update the output observation matrix for the monitoring data of the acceleration sensor, Update the measurement noise covariance matrix output by the monitoring data of the acceleration sensor, is the monitoring data of the acceleration sensor, I is the unit matrix, is the estimated value of the state after the monitoring data of the acceleration sensor is updated, is the estimated covariance after the monitoring data of the acceleration sensor is updated; The local state prediction and local covariance prediction expressions of the displacement sensor's monitoring data update output are:

[0044] In the formula, is the Kalman gain, Update the output observation matrix for the monitoring data of the displacement sensor, Update the measurement noise covariance matrix output by the displacement sensor monitoring data, is the monitoring data of the displacement sensor, I is the unit matrix, is the state estimation value after the displacement sensor monitoring data is updated, is the estimated covariance after updating the monitoring data of the displacement sensor; The local state prediction and local covariance prediction expressions of the monitoring data update output of the proximity sensor are:

[0045] In the formula, is the Kalman gain, Update the output observation matrix for the monitoring data of the proximity sensor, Update the output measurement noise covariance matrix for the monitoring data of the proximity sensor, is the monitoring data of the proximity sensor, I is the unit matrix, is the estimated value of the state after the monitoring data of the proximity sensor is updated, is the estimated covariance after updating the monitoring data of the proximity sensor.

[0046] S4-2) performing weighted average on the latest output local state prediction and local covariance prediction to obtain a global state estimate as a global monitoring indicator output of the equipment state, and returning the global state estimate to S4-1) to update the parameters of each local filter; In step S4-2, the global state estimation includes global state prediction and global covariance prediction. The global state estimation expression is as follows:

[0047] In the formula, is the global state prediction, is the global covariance prediction, is the state estimate of the latest output of each local filter, is the estimated covariance of the latest output of each local filter, is the weight of each local filter, is the number of local filters.

[0048] The state estimation value and estimated covariance of the latest output of each local filter are fused by weighted averaging to obtain the global monitoring index.

[0049] like Figure 4 As shown in the figure, it is a comparison chart of the monitoring data before and after data fusion. It can be seen that Figure 4 The monitoring data before and after data fusion were compared. Before fusion, the data had large fluctuations, high discreteness, and obvious interference; after fusion, the data curve is smoother, with fewer discrete points, and can integrate multi-source information, reduce the impact of noise, be more stable and reliable, and fully and accurately reflect the operating status of the equipment, which is conducive to early detection of fault hazards; Figure 5 As shown in the figure, it is a curve chart of monitoring indicators after fusion. It can be seen that the curve of monitoring indicator data after fusion is stable and has small fluctuations. The stability of display data is improved, and the law of equipment operation status can be intuitively presented. Abnormal changes in the curve can indicate potential equipment failures. Figure 6 This is the mean square error diagram of the monitoring indicators after fusion. It can be seen that the mean square error of the monitoring indicators after data fusion is lower than the monitoring data before fusion. This shows that data fusion improves the equipment status monitoring and prediction accuracy, optimizes the model performance, reduces data noise and uncertainty, and enhances the system reliability and effectiveness.

[0050] Figure 8 This is a box plot comparison chart of the monitoring indicators after fusion and the original monitoring data. Before fusion, the monitoring data of each sensor had large dispersion, large fluctuations, strong interference, weak median representativeness and many outliers. After fusion, the dispersion of monitoring indicators is significantly reduced, the data is more concentrated, the median can better represent the typical value of equipment operation, the number of outliers is reduced, and the risk of misjudgment can be reduced.

[0051] S5: The data processing center analyzes the equipment operation status according to the monitoring indicators and the monitoring data of each sensor to obtain the analysis results; The analysis results include monitoring indicator performance analysis and equipment operation performance analysis. The monitoring indicator performance analysis includes the changing trend of each monitoring indicator and the correlation between each local monitoring indicator. The equipment operation performance analysis includes equipment production efficiency, equipment accuracy, equipment health status and remaining service life, equipment fault type and location.

[0052] The changing trends of each monitoring indicator are analyzed by a data trend analysis model, and the expression of the data trend analysis model is as follows:

[0053] In the formula, is the value of each monitoring indicator at time t, is the lag operator , p is the autoregressive order, d is the difference order, q is the sliding average order, and are the autoregressive and moving average coefficients of the model, is a white noise sequence; The data processing center analyzes the changing trends of the monitoring indicators in the startup phase, stable operation phase, and stop phase of the equipment operation process. According to the normal operation of the equipment: in the startup phase, the various components of the equipment start to operate, the energy is gradually transferred, and the global monitoring indicators will quickly rise to the normal range. In the stable operation phase, the various systems of the equipment work in coordination, and the global indicators should remain relatively stable with a small fluctuation range. In the stop phase, the various components of the equipment gradually stop working, the energy consumption is reduced, and the global indicators gradually drop to the initial value. By judging the changing trends of the monitoring indicators in these three stages, if the monitoring indicators of each stage do not meet the response of the normal operation phase of the equipment, such as the monitoring indicators showing a continuous upward or downward trend, this is often an important signal of changes in equipment performance. For example, when the temperature-related part of the monitoring indicators continues to rise, it may mean that there is a problem with the equipment's heat dissipation system, such as a failure of the cooling fan, a blockage of the cooling channel, etc. If not handled in time, it may cause the equipment to overheat and be damaged. In this way, the equipment's operating status can be fully evaluated.

[0054] The correlation between the local monitoring indicators is judged by a correlation judgment method, and the expression of the correlation judgment method is:

[0055] In the formula, and are the sample values ​​of two local monitoring indicators for correlation analysis (such as the local monitoring indicators of the force sensor and the pressure sensor), and are the sample means of the two local monitoring indicators for correlation analysis, and n is the number of samples; The expression of the Granger causality test is:

[0056] In the formula, and are time series data representing two different local monitoring indicators (e.g. It can be a sequence of local temperature monitoring indicators of the equipment. Can be the power sequence of the device); and are the autoregressive coefficients, which respectively indicate the influence of the observations of the y and x sequences themselves at the past i moments on the current moment; and is the cross regression coefficient, It indicates the influence of the observation value of the past i moments of the x sequence on the current moment of the y sequence. It indicates the influence of the observations of the past i moments of the y series on the current moment of the x series; p, q, m, n are the lag orders, which are used to determine how many past moments of observations are considered in the regression model; and They are the error terms in the two regression equations, representing the parts that the model cannot explain.

[0057] The operation of each part of the equipment is interrelated, and the analysis of the correlation between each local monitoring index can determine the operation of the equipment. For example, when the pressure forming equipment is operating normally, the change in the force applied by the mold and the change in the pressure of the hydraulic system should show a positive correlation. Once this correlation is abnormal, such as the pressure decreases when the force increases, it indicates that the hydraulic system or force transmission components of the equipment may have a fault, such as hydraulic pipeline leakage, seal damage, etc.

[0058] In the equipment operation performance analysis, the following model is used to evaluate production efficiency:

[0059] Where n is the number of decision-making units (DMUs), representing different production units, equipment or departments, etc. Specifically, it is understood as pressure forming equipment in different time periods or different batches of production processes. By analyzing different decision-making units, the efficiency differences of equipment under different operating conditions can be compared; m is the number of input indicators, Indicates DMU No. The values ​​of the input indicators may include the input of raw materials, equipment operation time, energy consumption, etc. These indicators reflect the resource input of the equipment in the production process; s is the number of output indicators, Indicates The value of the rth output indicator of a DMU can be set as product output, product qualification rate, etc., which is used to measure the production results of the equipment; is the efficiency value, indicating the The relative efficiency of a DMU, ​​ranging from 0 to 1. This indicates that the DMU is on the production frontier and is relatively effective; It is a weight variable. When evaluating the efficiency of pressure forming equipment, different decision units are linearly combined. Its value is determined in the optimization process. It reflects the relative importance of pressure forming equipment in different operating states (different decision units) in constructing the production frontier, which helps to analyze the impact of different production conditions on equipment efficiency.

[0060] The accuracy of the equipment is predicted by the following model:

[0061] ,

[0062] In the formula, is the characteristic vector of each monitoring index. In the pressure forming equipment, It is composed of multiple sensor monitoring data, such as , It can be the mold working stroke data measured by the displacement sensor, It can be the mold movement speed data obtained by the speed sensor, etc. d is the dimension of the feature. These multi-source data comprehensively reflect the factors affecting the accuracy of the equipment operation status; is the corresponding equipment performance index value, which is used to measure the accuracy of the products actually produced by the equipment; is the weight vector of each monitoring indicator, , its components Determine the corresponding monitoring indicators The degree of influence on the equipment accuracy prediction results, express The transpose of ; b is the bias term; is the kernel function mapping, which maps the input data to the high-dimensional feature space, so that the data that may be linearly inseparable in the low-dimensional space becomes linearly separable; C is the penalty parameter. The larger the C is, the greater the penalty of the model for the error, and the more it tends to reduce the prediction error, but it may cause overfitting and poor adaptability to new data; the smaller the C is, the simpler the model is, but it may be underfitting and fail to accurately capture the relationship between equipment operation data and accuracy; and It is a relaxation variable, which allows some sample points not to meet the constraint conditions. That is, considering the noise and abnormal data that may appear during the operation of the pressure forming equipment, the model is made more robust by setting relaxation variables. It is the parameter of the insensitive loss function. Reasonable setting of the value can control the tolerance of the model to prediction errors and improve the stability of the prediction.

[0063] The model is used to analyze the health status and remaining service life of equipment: First, extract data features through the following model:

[0064] In the formula, x is each monitoring indicator; K is the convolution kernel; is the result of the convolution operation; M and n are the sizes of the pooling windows; Then, the time series of data is processed through the following model, and the health index threshold is set for early warning:

[0065] In the formula, is the input gate; is the sigmoid activation function; is the weight matrix of the input gate, which is used to transform the hidden state of the previous moment and the current input Linear combinations; is the bias term of the input gate;

[0066] In the formula, For the Gate of Oblivion; is the weight matrix of the forget gate; is the bias term of the forget gate;

[0067] In the formula, It is the output gate, which determines the cell state at the current moment. How much information can be output to the hidden state? , is the weight matrix of the output gate, is the bias term of the output gate.

[0068]

[0069] In the formula, is the cell state, storing long-term memory information, is the hyperbolic tangent activation function, which maps the input to values ​​between -1 and 1. is the weight matrix for cell state update, is the bias term for cell state update.

[0070]

[0071] In the formula, It is a hidden state, which contains both the short-term memory information of the current moment and the long-term memory information transmitted from the cell state.

[0072] The equipment failure type and location are analyzed through the following models:

[0073] In the formula, is the predicted value of the monitoring indicator; x is the operating time or the number of times the equipment is operated; are the coefficients of the polynomial, which are solved by the least square method to determine the shape of the polynomial function. Different combinations of coefficients will give the polynomial curve different shapes and trends; The following steps are then used to predict the health status and remaining service life of the equipment: A. Original data sequence Generate cumulative sequence , In the formula, , is the monitoring data sequence before fusion; The new sequence generated by accumulation is used to establish the differential equation of grey prediction.

[0074] B. Establishing differential equations , In the formula, and are the parameters solved by the least squares method, Reflecting the development trend, is a constant term; C. The prediction formula is , and then restored to , In the formula, and They are the predicted values ​​of the cumulative sequence and the monitoring data sequence, which are used to predict future monitoring indicators.

[0075] S6: The data processing center sends the analysis results to the equipment management system, which generates control instructions based on the analysis results and sends them to the equipment actuator, and stores and displays the analysis results. The equipment management system includes a control system, a display device, and data storage.

[0076] In step S6, the display content of the equipment management system includes each monitoring indicator and the monitoring data of each sensor. The monitoring indicators and the monitoring data of each sensor displayed by the equipment management system are displayed in tables or charts. By displaying the historical data and real-time values ​​of the monitoring indicators in a table, it is convenient for the staff to fully grasp the equipment operation information. By displaying the monitoring indicators and monitoring data in the histogram or line chart in the chart, the changing trend of the data can be clearly and intuitively seen, which is convenient for the staff to respond in time. In this implementation, if Figure 7 As shown in the figure, the use of histograms to display monitoring data can intuitively present the data distribution characteristics, help understand the parameter fluctuation range under normal operating conditions of the equipment, and also facilitate the detection of abnormal data through abnormal histogram shapes and the detection of hidden dangers of equipment failure.

[0077] Further preferably, the data processing center is provided with an indicator alarm threshold for each monitoring indicator. When the monitoring indicator reaches the indicator alarm threshold, the equipment management system will highlight the monitoring indicator that has reached the indicator alarm threshold in red and issue an indicator abnormality alarm. The data processing center has a data alarm threshold for each sensor monitoring data. When the monitoring data reaches the data alarm threshold, the equipment management system will highlight the monitoring indicator that has reached the data alarm threshold in yellow and issue a data abnormality alarm. The data processing center sets alarm thresholds for monitoring indicators and detection data, and the equipment management system highlights and issues an alarm, so that management personnel can grasp the equipment operation status in a timely and accurate manner.

[0078] The analysis results also include equipment development forecast analysis. The historical data of each monitoring indicator is fitted through a mathematical model to obtain the equipment development forecast result, and the corresponding monitoring indicator threshold is set. When the equipment development forecast result exceeds the monitoring indicator threshold, the equipment management system will issue a warning signal. In this embodiment, linear fitting is used to predict the equipment development forecast analysis. Through continuous tracking and analysis of the monitoring indicators, once an abnormal trend is found, even if the equipment has not yet shown obvious fault symptoms, the system can issue a warning signal in advance, which allows maintenance personnel to have sufficient time to perform equipment inspections, troubleshooting and necessary maintenance operations, effectively avoiding sudden equipment failures, greatly reducing equipment downtime, and ensuring the continuity and stability of production.

[0079] It also includes the equipment management system generating production recommendations based on the evaluation results.

[0080] The production suggestions include equipment performance optimization suggestions, equipment maintenance suggestions, and production resource allocation suggestions; The equipment performance optimization suggestions include adjusting process parameters and optimizing equipment operation rhythm; the specific adjustment of process parameters is that if the equipment operation performance analysis shows that the product molding accuracy is slightly deviated, the equipment management system can suggest fine-tuning the mold stamping speed and pressure parameters. If the current mold stamping speed is too fast, resulting in minor flaws on the edge of the workpiece, it is recommended to reduce the speed by 5%-10% and increase the stamping pressure by 5%-10% accordingly to make the mold and workpiece contact more fully and improve product accuracy. The specific optimization of equipment operation rhythm is that when monitoring finds that the equipment production efficiency does not meet expectations and each component is operating normally, the equipment operation rhythm can be adjusted according to production tasks and equipment characteristics. If the equipment has a short idle time during the stable operation stage, it is recommended to reasonably shorten the equipment start and stop interval to reduce unnecessary waiting and increase output per unit time.

[0081] The equipment maintenance recommendations include preventive maintenance arrangements and parts replacement plans.

[0082] The preventive maintenance arrangement is specifically based on the assessment of the health status and remaining service life of the equipment. If the wear monitoring index of a key component of the equipment (such as a hydraulic pump) is close to the warning threshold, although it has not yet affected the normal operation of the equipment, in order to avoid sudden failures, it is recommended to inspect and maintain the component during the next production interval, such as replacing seals and replenishing hydraulic oil.

[0083] The parts replacement plan is specifically recommended when the equipment operation performance analysis indicates that the wear of a certain part (such as a mold) affects product quality and the repair cost is high. If the displacement sensor monitoring data of the mold shows that the mold working stroke has deviated, resulting in a decrease in product dimensional accuracy, it is recommended to replace the mold immediately after completing the current emergency production task to ensure the stability of subsequent product quality.

[0084] The production resource allocation suggestions include material supply adjustment and human resource deployment; The material supply adjustment is specifically that when the proximity sensor detects that the material delivery of the feeding mechanism is not timely, which affects the production efficiency, the system can suggest optimizing the material supply frequency and quantity according to the equipment operation speed and production requirements. For example, increase the quantity of each feeding, or shorten the feeding interval time, to ensure the timely supply of materials and avoid the equipment stopping and waiting due to lack of materials.

[0085] Specifically, if the equipment management system finds that equipment failures occur frequently during a certain period and require more manual intervention for repairs, it can recommend temporary deployment of technicians to the production line to strengthen equipment inspection and maintenance, ensure the normal operation of the equipment, and reduce the impact of downtime on production progress.

[0086] Generate a production proposal by following these steps: 1. Coding: Encode the key parameters of the pressure forming equipment (such as mold temperature, stamping pressure, stamping speed, etc.) and convert them into chromosome forms that can be processed by the genetic algorithm. Generally, binary coding or real number coding is used. For example, the mold temperature range [100,300]℃ is encoded with real numbers, and each temperature value directly corresponds to a gene in the chromosome. If binary coding is used, the temperature value is converted into a binary string, such as a temperature of 200℃, which is assumed to be represented by an 8-bit binary code as "11001000". These encoded chromosomes are combined to form the initial population, representing different parameter combination schemes for pressure forming equipment.

[0087] 2. Initialize the population: According to the actual operating range and experience of the pressure forming equipment, a certain number of initial chromosomes are randomly generated to form the initial population. For example, when producing a specific automotive part, based on past production experience, the stamping pressure is generally 5-10MPa, the stamping speed is 0.1-0.5m / s, and the mold temperature is 150-250℃. Within this range, 50 sets of parameter combinations are randomly generated as the initial population. These initial parameter combinations are like different "experimental plans", laying the foundation for the subsequent genetic algorithm to find the optimal parameters.

[0088] 3. Fitness function: Determine a fitness function to evaluate the adaptability of each chromosome (i.e., parameter combination) to the production of pressure forming equipment. The fitness function can be designed according to the production objectives, such as improving product qualification rate, reducing energy consumption, and improving production efficiency. Taking improving product qualification rate as an example, each set of parameters is applied to the simulated production of pressure forming equipment (which can be done by establishing a production model or actual trial production) to obtain the product qualification rate. The higher the product qualification rate, the higher the corresponding chromosome fitness value. If a set of parameters (chromosomes) makes the product qualification rate reach 95% in simulated production, while the other set only reaches 80%, the former has a higher fitness value, indicating that the parameter combination is more conducive to production.

[0089] 4. Selection operation: According to the fitness value, roulette selection, tournament selection and other methods are used to select better chromosomes from the current population, so that they have more chances to be inherited to the next generation. For example, using the roulette selection method, the probability of each chromosome being selected is proportional to its fitness value. The chromosome with high fitness occupies a large "area" on the roulette wheel, and the probability of being selected is also high. In this way, after the selection operation, the parameter combination with high fitness is more likely to be passed on to the next generation of population, so that the population as a whole develops in a better direction.

[0090] 5. Crossover operation: Perform a crossover operation on the selected chromosomes to simulate the gene exchange process in biological genetics. Randomly select two chromosomes from the population, and exchange some genes at specific positions of the chromosomes according to a certain crossover probability (such as 0.8) to generate new offspring chromosomes. For example, there are two chromosomes A "11001000" (representing a set of parameters) and B "00110111". If the crossover point is at the 4th position, new chromosomes A' "11000111" and B' "00111000" are generated after crossover. These new chromosomes represent new combinations of pressure forming equipment parameters, which provide the possibility of finding a better solution.

[0091] 6. Mutation operation: Perform mutation operation on chromosomes with a certain mutation probability (such as 0.01) to simulate the gene mutation phenomenon in biological genetics. Randomly change the values ​​of certain genes in the chromosome to increase the diversity of the population and prevent the algorithm from falling into the local optimal solution. For binary-coded chromosomes, mutation is to change the 0 on the gene position to 1, or 1 to 0. For example, if the chromosome "11001000" mutates at the third position, it becomes "11101000". Mutation operation allows the genetic algorithm to explore a wider area in the search space, and it is possible to find better parameter combinations.

[0092] 7. Termination conditions: Check whether the termination conditions are met, such as reaching the predetermined number of iterations, fitness values ​​converging to a certain range, etc. If not, return to the fitness calculation step and continue with selection, crossover, mutation and other operations; if satisfied, output the chromosome with the highest fitness in the current population, and the parameter combination corresponding to the chromosome is the optimal parameter combination of the pressure forming equipment optimized by the genetic algorithm. For example, set the number of iterations to 100 times. After the genetic algorithm runs to 100 iterations, output the mold temperature, stamping pressure, stamping speed and other parameters corresponding to the chromosome with the highest fitness at this time. These parameters can be applied to the actual production of pressure forming equipment to improve production efficiency and product quality.

[0093] In terms of production guidance, the equipment management system is integrated with MES to automatically adjust process parameters; the operation process and precautions are displayed on the operator terminal, and guidance is provided to on-site operators to ensure that production suggestions are accurately implemented and to ensure efficient and stable production operations.

[0094] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals, characterized in that: The specific steps include: S1: Data acquisition equipment is installed on the equipment according to production requirements, and the data acquisition equipment includes a force sensor, a pressure sensor, a temperature sensor, a velocity sensor, an acceleration sensor, a displacement sensor and a proximity sensor; S2: The data acquisition device collects the monitoring data of each sensor and sends the monitoring data to the data processing center; S3: The data processing center pre-processes the received monitoring data and extracts feature data; S4: The data processing center calculates the characteristic data through a distributed Kalman filter algorithm to generate monitoring indicators of equipment status, wherein the monitoring indicators include local monitoring indicators and global monitoring indicators; S4-1) Inputting the characteristic data of each detection data into the corresponding local filter respectively, obtaining the local state prediction and local covariance prediction of each monitoring data, and each local filter updates and outputs the corresponding local state prediction and local covariance prediction as the local monitoring index along with the monitoring data; S4-2) performing weighted average on the latest output local state prediction and local covariance prediction to obtain a global state estimate as a global monitoring indicator output of the equipment state, and returning the global state estimate to S4-1) to update the parameters of each local filter; S5: The data processing center analyzes the equipment operation status according to the monitoring indicators and the monitoring data of each sensor to obtain the analysis results; S6: The data processing center sends the analysis results to the equipment management system. The equipment management system generates control instructions based on the analysis results and sends them to the equipment actuator, and stores and displays the analysis results.

2. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1 is characterized in that: In step S4-1, the expressions for state prediction and covariance prediction of each monitoring data are as follows: The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the force sensor are: In the formula, is the local state prediction of the monitoring data of the force sensor at time point k1, is the local state prediction of the monitoring data of the force sensor at time point k1-1, is the local covariance prediction of the monitoring data of the force sensor at time point k1, is the local covariance prediction of the monitoring data of the force sensor at time point k1-1, is the process noise covariance matrix of the monitoring data of the force sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the pressure sensor are: In the formula, is the local state prediction of the monitoring data of the pressure sensor at time point k2, is the local state prediction of the monitoring data of the pressure sensor at time point k2-1, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2, is the local covariance prediction of the monitoring data of the pressure sensor at time point k2-1, is the process noise covariance matrix of the monitoring data of the pressure sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the temperature sensor are: In the formula, is the local state prediction of the temperature sensor monitoring data at time point k3, is the local state prediction of the temperature sensor monitoring data at time point k3-1, is the local covariance prediction of the temperature sensor monitoring data at time point k3, is the local covariance prediction of the temperature sensor monitoring data at time point k3-1, is the process noise covariance matrix of the temperature sensor monitoring data, is the state transfer matrix of the monitoring data of the temperature sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the speed sensor are: In the formula, is the local state prediction of the speed sensor monitoring data at time point k4, is the local state prediction of the speed sensor monitoring data at time point k4-1, is the local covariance prediction of the speed sensor monitoring data at time point k4, is the local covariance prediction of the speed sensor monitoring data at time point k4-1, is the process noise covariance matrix of the speed sensor monitoring data, is the state transfer matrix of the monitoring data of the speed sensor, is the control input matrix; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the acceleration sensor are: In the formula, is the local state prediction of the acceleration sensor monitoring data at time point k5, is the local state prediction of the acceleration sensor monitoring data at time point k5-1, is the local covariance prediction of the acceleration sensor monitoring data at time point k5, is the local covariance prediction of the acceleration sensor monitoring data at time point k5-1, is the noise covariance matrix of the monitoring data of the acceleration sensor in the process; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the displacement sensor are: In the formula, is the local state prediction of the displacement sensor monitoring data at time point k5, is the local state prediction of the displacement sensor monitoring data at time point k6-1, is the local covariance prediction of the monitoring data of the displacement sensor at time point k6, is the local covariance prediction of the displacement sensor monitoring data at time point k6-1, is the noise covariance matrix of the displacement sensor monitoring data in the process, is the state transfer matrix of the monitoring data of the displacement sensor; The local state prediction and local covariance prediction expressions corresponding to the monitoring data of the proximity sensor are: In the formula, is the local state prediction of the monitoring data of the proximity sensor at time point k7, is the local state prediction of the monitoring data of the proximity sensor at time point k7-1, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7, is the local covariance prediction of the monitoring data of the proximity sensor at time point k7-1, is the process noise covariance matrix of the monitoring data of the proximity sensor, is the state transfer matrix of the monitoring data of the proximity displacement sensor.

3. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1 is characterized in that: In step S4-2, the global state estimation includes global state prediction and global covariance prediction. The global state estimation expression is as follows: In the formula, is the global state prediction, is the global covariance prediction, is the state estimate of the latest output of each local filter, is the estimated covariance of the latest output of each local filter, is the weight of each local filter, is the number of local filters.

4. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1 is characterized in that: In step S1, the force sensor is used to monitor the force applied when the mold is stamping; the pressure sensor is used to monitor the pressure in the hydraulic system pipeline; the temperature sensor is used to monitor the workpiece temperature; the speed sensor is used to monitor the mold movement speed; the acceleration sensor is used to monitor the vibration frequency and amplitude of the mold; the displacement sensor is used to monitor the working stroke of the mold; and the proximity sensor is used to monitor the position of the moving parts.

5. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1 is characterized in that: In step S2, the data acquisition device performs signal amplification, noise filtering, and linearization calibration on each monitoring data in turn, and then sends it to the data processing center through the field bus.

6. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1, characterized in that: In step S5, the analysis results include monitoring indicator performance analysis and equipment operation performance analysis. The monitoring indicator performance analysis includes the changing trend of each monitoring indicator and the correlation between each local monitoring indicator. The equipment operation performance analysis includes equipment production efficiency, equipment accuracy, equipment health status and remaining service life, equipment fault type and location.

7. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 6 is characterized in that: The analysis results also include equipment development forecast analysis. The historical data of each monitoring indicator is fitted through a mathematical model to obtain the equipment development forecast result, and corresponding monitoring indicator thresholds are set respectively. When the equipment development forecast result exceeds the monitoring indicator threshold, the equipment management system will issue a warning signal.

8. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 6 is characterized in that: The changing trends of each monitoring indicator are analyzed by a data trend analysis model, and the expression of the data trend analysis model is as follows: In the formula, is the value of each monitoring indicator at time t, is the lag operator , p is the autoregressive order, d is the difference order, q is the sliding average order, and are the autoregressive and moving average coefficients of the model, is a white noise sequence.

9. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1, characterized in that: In step S6, the display content of the equipment management system includes various monitoring indicators and monitoring data of various sensors. The monitoring indicators and monitoring data of various sensors displayed by the equipment management system are displayed in tables or charts.

10. The method for intelligently sensing the state of pressure forming equipment by fusing multi-source sensor signals according to claim 1, characterized in that: In step S6, the equipment management system also generates production suggestions according to the evaluation results.

Images