A pneumatic actuator remote monitoring and intelligent diagnosis system based on internet of things

Abstract

This invention discloses a remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things (IoT), belonging to the field of pneumatic actuator monitoring and diagnostic technology. It includes a sensor monitoring module containing pressure sensors, temperature sensors, displacement sensors, vibration sensors, and a data acquisition card; a data processing module employing a 32-bit ARM processor, supporting 4G / 5G and Ethernet communication, and with built-in 2GB storage; an IoT communication module including an NB-IoT unit and an edge gateway, supporting multi-protocol conversion and AES-128 encryption; a remote monitoring platform based on a B / S architecture, including real-time dashboards, historical queries, and equipment maps; and an intelligent diagnostic module identifying eight types of faults. This invention enables remote monitoring and intelligent diagnostics of pneumatic actuators, improving fault response speed and diagnostic accuracy, supporting predictive maintenance and adaptive adjustment, reducing operation and maintenance costs, extending equipment lifespan, and improving operating efficiency.

Description

Technical Field

[0001] This invention relates to the field of pneumatic actuator monitoring and diagnostic technology, and in particular to a remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things. Background Technology

[0002] Pneumatic actuators, with their advantages of simple structure, low cost, and strong environmental adaptability, are widely used in industrial automation, intelligent manufacturing, logistics and warehousing, and other fields. They are core components for realizing functions such as valve opening and closing, robotic arm movement, and assembly line conveying. With the development of industrial Internet of Things (IIoT) technology, the traditional "offline monitoring and manual diagnosis" operation and maintenance mode of pneumatic actuators has gradually exposed obvious shortcomings. Currently, most pneumatic actuators lack real-time monitoring capabilities and can only obtain their operating status through regular manual inspections. The inspection cycle is usually 1-7 days, making it difficult to detect sudden faults in time. For example, slow leakage caused by aging actuator seals may initially be small and difficult to detect visually during manual inspections. By the time the fault expands to affect production, it has already caused a large waste of air resources and production stoppage. Especially in industries with continuous production, such as chemical and automotive manufacturing, the loss from a single failure downtime can reach tens of thousands of yuan.

[0003] The few existing pneumatic actuators with monitoring functions also suffer from problems such as limited monitoring dimensions and weak diagnostic capabilities. These devices often only collect single parameters such as pressure or temperature, failing to comprehensively reflect the actuator's operating status. For example, monitoring only pressure cannot identify jamming faults caused by mechanical wear, and monitoring only temperature cannot determine internal corrosion caused by compressed air contamination. Furthermore, fault diagnosis often relies on simple threshold judgments, such as assuming a fault occurs when the pressure falls below a set value, failing to distinguish between different causes such as "fluctuations in air source pressure" and "air leakage in the actuator itself," resulting in a false alarm rate exceeding 20%. This not only increases the workload of maintenance personnel but may also delay the handling of genuine faults due to misjudgments. In addition, data transmission is mostly done via wired methods, resulting in high deployment costs, poor flexibility, difficulty in adapting to distributed actuator clusters, and a lack of remote access capabilities, requiring maintenance personnel to physically visit the site to view data, leading to low efficiency.

[0004] The lack of predictive maintenance capabilities further exacerbates the operational challenges. Traditional maintenance employs a "periodic replacement" strategy, replacing vulnerable parts at fixed intervals regardless of the actuator's actual condition. This "over-maintenance" not only increases spare parts costs but may also cause additional damage to the actuator during disassembly and reassembly. Furthermore, some actuators may fail prematurely within their designated cycle due to individual differences, leading to "under-maintenance" and subsequent malfunctions. Simultaneously, actuator operating conditions are susceptible to factors such as gas flow rate and ambient temperature. Fixed operating parameters are difficult to adapt to dynamic conditions. For example, gas viscosity increases at low temperatures; using the same operating pressure as at room temperature will result in sluggish actuator action and reduced production efficiency. These problems collectively lead to high maintenance costs, slow fault response, and low operating efficiency for pneumatic actuators, failing to meet the modern industrial demands for "high reliability, high intelligence, and low-cost maintenance." Therefore, a comprehensive solution integrating multi-dimensional monitoring, intelligent diagnostics, remote control, and predictive maintenance is urgently needed. Summary of the Invention

[0005] This invention proposes an Internet of Things-based remote monitoring and intelligent diagnostic system for pneumatic actuators to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things, comprising the following modules:

[0007] The sensing and monitoring module is equipped with a diffused silicon pressure sensor, a PT100 temperature sensor, a magnetostrictive displacement sensor, a piezoelectric vibration sensor, and an integrated data acquisition card to collect all dimensions of the pneumatic actuator's operating parameters.

[0008] The data processing module uses an STM32H743 microprocessor with a main frequency of 400MHz and is equipped with an FPU floating-point arithmetic unit. The data processing module performs 5th-order Butterworth filtering, wavelet transform threshold denoising, and time-domain feature extraction on the sensor data. The data processing module has a built-in 8GBeMMC storage unit and supports 4G, 5G and Gigabit Ethernet communication.

[0009] The IoT communication module includes an NB-IoT wireless transmission unit and an edge computing gateway. The NB-IoT wireless transmission unit operates at 800 / 900MHz and has a transmission distance of 3-10km. The edge computing gateway uses an ARM Cortex-A53 quad-core processor and supports MODBUSRTU / TCP, PROFINET, and EtherCAT protocol conversion. The IoT communication module uses the AES-128 algorithm for data encryption, with a key update cycle of 24 hours.

[0010] The remote monitoring platform is based on a B / S architecture, with an Nginx web server and a MySQL cluster database. The platform includes real-time data dashboards, historical curve queries, and electronic maps of device status. The remote monitoring platform supports access from both PC and mobile devices.

[0011] The intelligent diagnostic module incorporates a hybrid diagnostic model that integrates an expert system and a random forest algorithm. It identifies eight typical faults, including three levels for air leakage, four levels for jamming, and two levels for abnormal pressure. The intelligent diagnostic module has a diagnostic response time of ≤5s, a fault location accuracy of ≤0.5m, and supports automatic labeling of fault samples and iterative updates of the model.

[0012] Furthermore, it also includes a predictive maintenance module, which uses formulas... Calculate the actuator reliability; where, Let be the reliability at time t; Failure rate; t is runtime; when When the value is ≤0.85, the system will automatically generate a maintenance reminder. The reminder methods include platform pop-ups, SMS, and APP push notifications. Platform pop-ups will continue to be displayed until the user confirms, and APP push notifications will implement a 3-retry mechanism. The maintenance reminder content includes a list of parts to be replaced, operation steps, and spare parts models.

[0013] Furthermore, it also includes an adaptive adjustment module, which uses a formula... Calculate the optimal working pressure; where To achieve optimal work pressure; Where is the flow coefficient; Q is the gas flow rate; T is the temperature coefficient; T is the operating temperature. Environmental pressure coefficient; The ambient atmospheric pressure is used; the adaptive adjustment module has an adjustment accuracy of ≤±0.02MPa and a response time of ≤100ms. The adjustment actuator adopts a proportional pressure valve with a nominal diameter of 6-20mm.

[0014] Furthermore, the sensing and monitoring module is also equipped with a laser oil mist sensor and a capacitive humidity sensor; the laser oil mist sensor and the capacitive humidity sensor are used to monitor the oil mist concentration and relative humidity in the compressed air. When the oil mist concentration is ≥5ppm and this state lasts for more than 10 seconds, or the relative humidity is ≥80% and this state lasts for more than 30 seconds, the system triggers a pollution warning. The warning adopts a three-level warning mechanism, and different warning levels correspond to different handling strategies.

[0015] Furthermore, the data processing module employs an improved Kalman filter algorithm for dynamic noise reduction; the filter window size of the improved Kalman filter algorithm is set to 8-32 sampling points, and the window size is automatically switched according to the noise intensity; the initial value of the covariance matrix of the state equation of the algorithm is adjusted within the range of 0.1-1.0, and the measurement noise variance is adaptively updated within the range of 0.01-0.1.

[0016] Furthermore, the IoT communication module supports LoRaWAN edge node self-organizing network, with the self-organizing network operating frequency bands being 433 / 868 / 915MHz; the maximum number of nodes in the edge node self-organizing network is 100-500, and the number of nodes can be expanded through gateway cascading; the self-organizing network adopts a time division multiple access mechanism, with a time slot length of 10-100ms.

[0017] Furthermore, the remote monitoring platform has a built-in energy consumption analysis unit, which analyzes energy consumption using formulas. Calculate energy consumption per unit distance traveled; where, Energy consumption per unit distance; For real-time pressure; t represents real-time traffic; t represents trip time. The total travel time is calculated; the energy consumption analysis unit generates daily / monthly energy consumption reports, and also supports comparison with historical data for the same period. When the data deviation exceeds 10%, an alert is triggered.

[0018] Furthermore, the intelligent diagnostic module adopts an improved BP neural network model; the input layer of the improved BP neural network model contains 16 feature parameters; the hidden layer of the model has 3 layers, each with 24-32 neurons, and the activation function is ReLU; the output layer of the model corresponds to 8 types of faults, and the activation function is Softmax; the model training iterations are 5000-10000 times, and the learning rate is adaptively adjusted within the range of 0.001-0.01.

[0019] Furthermore, the predictive maintenance module also includes a remaining lifetime assessment unit, which assesses remaining lifetime using a formula... Calculate the remaining lifetime; where, Remaining lifespan; Rated lifespan; The failure rate function varies with time, and a Weibull distribution model is adopted. Where m is the shape parameter, The remaining life assessment error is ≤5%, which is obtained by comparing it with the actual life of the actuator.

[0020] Furthermore, the adaptive adjustment module is also equipped with a pressure compensation unit; when the ambient temperature changes within 10 minutes... Or gas source pressure fluctuation within 5 minutes At that time, the pressure compensation unit automatically activates the compensation mechanism; the compensation amount is determined by the formula. Calculation, in the formula The pressure compensation unit corrects the output pressure in real time through a proportional valve.

[0021] Compared with existing technologies, the beneficial effects of this invention are:

[0022] In terms of comprehensiveness and real-time monitoring, the system constructs a multi-dimensional monitoring network through multiple types of sensors. Compared with traditional single-parameter monitoring, it can simultaneously collect key parameters such as pressure, temperature, displacement, vibration, and air quality, accurately capturing subtle changes in actuator operation, such as early air leakage and slight jamming, which are easily overlooked potential faults, achieving "early detection and early warning." At the same time, the IoT communication module supports wireless transmission and edge networking, enabling centralized monitoring of distributed actuators without wiring. Data is transmitted to a remote platform in real time with a fast refresh rate. Maintenance personnel can grasp the equipment status without on-site inspections, completely solving the problems of "long cycles and high missed detection rates" of traditional manual inspections, and significantly improving fault response speed.

[0023] In terms of diagnostic accuracy and intelligence, the system breaks through the limitations of traditional threshold judgment. It adopts a hybrid diagnostic model that integrates expert systems and machine learning, which can accurately identify and classify multiple typical faults, as well as distinguish the causes of faults, significantly reducing the false alarm rate. The diagnostic response is rapid, and the fault location is accurate. Maintenance personnel can quickly find the fault point and formulate a repair plan based on the system prompts, avoiding the inefficiency of "blind troubleshooting" in traditional diagnostics. At the same time, the system supports automatic labeling of fault samples and model iteration. As the usage time increases, the diagnostic capabilities are continuously optimized, adapting to the individual differences of different brands and models of actuators, solving the shortcomings of "poor universality and easy misjudgment" in traditional diagnostics, and improving the reliability of equipment operation.

[0024] In terms of operational economy and foresight, the predictive maintenance module achieves "on-demand maintenance" through reliability calculation and remaining life assessment. This avoids the excessive waste of traditional periodic maintenance and prevents failures caused by insufficient maintenance, significantly reducing spare parts costs and maintenance workload. When equipment reliability drops to a threshold, the system automatically generates maintenance reminders and pushes information through multiple channels to ensure timely and efficient maintenance. The adaptive adjustment module can calculate and adjust the optimal working pressure in real time based on dynamic operating conditions such as flow rate and temperature, ensuring that actuators maintain efficient operation in different environments. This avoids inefficiency or damage caused by fixed parameters, improving equipment operating efficiency and lifespan.

[0025] In terms of ease of management and flexibility, the remote monitoring platform is based on a B / S architecture, supporting multi-terminal access. Maintenance personnel can view real-time data, query historical curves, and manage equipment status anytime, anywhere, achieving "remote control and centralized scheduling," especially suitable for the management needs of large-scale actuator clusters. The platform's built-in energy consumption analysis unit can accurately calculate energy consumption per unit stroke, generate energy consumption reports, and compare historical data, helping enterprises identify energy-saving opportunities and reduce operating costs. Overall, the system upgrades pneumatic actuators from "passive maintenance" to "proactive early warning," and from "manual management" to "intelligent control," fully meeting the modern industrial demands for high reliability, high intelligence, and low-cost equipment operation and maintenance, demonstrating significant economic and practical value. Attached Figure Description

[0026] Figure 1 This is a schematic block diagram of the IoT-based remote monitoring and intelligent diagnostic system for pneumatic actuators proposed in this invention.

[0027] Figure 2 A comparison chart of fault development stages and diagnostic response time;

[0028] Figure 3 This is a diagram showing the relationship between actuator runtime, reliability, and maintenance reminders. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0031] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0032] Reference Figures 1 to 3 A remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things (IoT) includes the following modules:

[0033] The sensing and monitoring module is used to collect all-dimensional operating parameters of the pneumatic actuator. It is equipped with a diffused silicon pressure sensor, a PT100 temperature sensor, a magnetostrictive displacement sensor, a piezoelectric vibration sensor, and an integrated data acquisition card. The diffused silicon pressure sensor has a measurement range of 0-1.6 MPa, an accuracy of ±0.5%FS, and a response time ≤5ms. The PT100 temperature sensor has a measurement range of -40℃ to 85℃, an accuracy of ±0.3℃, and a sampling interval adjustable from 1 to 10s. The magnetostrictive displacement sensor has a measurement range of 0-100mm, a resolution of 0.01mm, and a linearity error ≤0.02%FS. The piezoelectric vibration sensor has a measurement frequency range of 10-1000Hz, a sensitivity of 100mV / g±5%, and a range of ±50g. The integrated data acquisition card has a sampling frequency continuously adjustable from 1-10kHz and an A / D conversion accuracy of 16 bits.

[0034] The data processing module uses an STM32H743 microprocessor with a main frequency of 400MHz and an FPU floating-point unit. The module performs 5th-order Butterworth filtering, wavelet transform threshold denoising, and time-domain feature extraction on the sensor data. The cutoff frequency of the Butterworth filter can be adjusted within the range of 50-500Hz, and the threshold range for wavelet transform threshold denoising is 0.01-0.1V. Time-domain feature extraction includes 12 types of features such as peak value, mean, and variance. The module has a built-in 8GBeMMC storage unit with a continuous storage time of ≥720h. It supports 4G, 5G, and Gigabit Ethernet communication, with a 4G communication rate of 150Mbps and a 5G communication rate of 1.2Gbps. The data buffer latency is ≤100ms.

[0035] The IoT communication module includes an NB-IoT wireless transmission unit and an edge computing gateway. The NB-IoT wireless transmission unit operates at 800 / 900MHz, with a transmission distance of 3-10km, a sleep current ≤5μA, and a wake-up time ≤200ms. The edge computing gateway uses an ARM Cortex-A53 quad-core processor with a clock speed of 1.2GHz, supporting MODBUSRTU / TCP, PROFINET, and EtherCAT protocol conversion with a protocol conversion latency ≤10ms. Data encryption in the IoT communication module uses the AES-128 algorithm, with a key update cycle of 2... The system supports 4 hours of operation and also supports data interruption and resumption. The remote monitoring platform is based on a B / S architecture, using Nginx as the web server and a MySQL cluster as the database. The platform includes real-time data dashboards, historical curve queries, and electronic maps of device status. The real-time data dashboard refreshes at 0.5-1 seconds, the storage period for historical curve queries can be configured from 1 to 5 years, and the positioning accuracy of the electronic map of device status is ±5 meters. The remote monitoring platform supports access from both PC and mobile devices. The PC version is compatible with Windows and Linux systems, and the mobile version is compatible with Android and iOS systems. The platform can support ≥500 concurrent users.

[0036] The intelligent diagnostic module incorporates a hybrid diagnostic model that integrates an expert system and a random forest algorithm. It can identify eight typical faults, including air leakage, jamming, and abnormal pressure. Air leakage faults are classified into three levels, jamming faults into four levels, and abnormal pressure faults into two levels. The intelligent diagnostic module has a diagnostic response time of ≤5s and a fault location accuracy of ≤0.5m. It also supports automatic labeling of fault samples and iterative updates of the model, with the model iteration cycle adjustable within the range of 7-30 days.

[0037] This invention also includes a predictive maintenance module, which uses a formula... Calculate actuator reliability; in the formula, The reliability at time t is dimensionless and ranges from 0 to 1. Failure rate, in units of 1 / h, obtained by fitting historical failure data, with a value range of 0.0001-0.01; t is the running time, in units of hours; when When the value is ≤0.85, the system will automatically generate a maintenance reminder. The reminder methods include platform pop-ups, SMS, and APP push notifications. The platform pop-up will continue to be displayed until the user confirms. The SMS sending delay is ≤30 seconds. The APP push notification will implement a 3-retry mechanism. The maintenance reminder content includes a list of parts to be replaced, operation steps, and spare parts models.

[0038] This invention also includes an adaptive adjustment module, which uses a formula... Calculate the optimal working pressure; in the formula, The optimal working pressure is expressed in MPa, and its range is 0.4-1.2 MPa. is the flow coefficient, with a value ranging from 0.02 to 0.05 MPa·min / L, and the specific value is related to the actuator diameter; Q is the gas flow rate, in L / min, with a measurement range of 0-500 L / min; The value is the temperature coefficient, ranging from 0.001 to 0.003 MPa / ℃, with the specific value depending on the type of gas; T is the operating temperature, in ℃. This is the environmental pressure coefficient, ranging from 0.002 to 0.005, and is dimensionless. The ambient atmospheric pressure is measured in MPa, with a measurement range of 0.09-0.11 MPa. The adaptive adjustment module has an adjustment accuracy of ≤±0.02 MPa and a response time of ≤100 ms. The adjustment actuator uses a proportional pressure valve with a nominal diameter of 6-20 mm.

[0039] In this invention, the sensing and monitoring module is further equipped with a laser oil mist sensor and a capacitive humidity sensor. The laser oil mist sensor has a detection range of 0-50ppm, an accuracy of ±0.1ppm, and a response time of ≤1s. The capacitive humidity sensor has a measurement range of 0-100%RH and an accuracy of ±2%RH. The laser oil mist sensor and the capacitive humidity sensor are used to monitor the oil mist concentration and relative humidity in compressed air. When the oil mist concentration is ≥5ppm and this state lasts for more than 10s, or the relative humidity is ≥80% and this state lasts for more than 30s, the system triggers a pollution warning. The warning adopts a three-level warning mechanism, with different warning levels corresponding to different handling strategies.

[0040] In this invention, the data processing module employs an improved Kalman filter algorithm for dynamic noise reduction. The filter window size of the improved Kalman filter algorithm is set to 8-32 sampling points, and the window size can be automatically switched according to the noise intensity. The initial value of the covariance matrix of the algorithm's state equation can be adjusted within the range of 0.1-1.0, and the measurement noise variance is adaptively updated within the range of 0.01-0.1. After processing by the improved Kalman filter algorithm, the data signal-to-noise ratio is improved by 20-40dB, and it can effectively eliminate pulse interference with an amplitude ≥ 3 times the root mean square error and high-frequency noise with a frequency > 500Hz.

[0041] In this invention, the IoT communication module supports LoRaWAN edge node self-organizing networks, with the self-organizing network operating in the 433 / 868 / 915MHz frequency band. The maximum number of nodes in the edge node self-organizing network is 100-500, and the number of nodes can be expanded through gateway cascading. The self-organizing network uses a Time Division Multiple Access (TDMA) mechanism to avoid signal collisions, with a time slot length of 10-100ms. At a receiver sensitivity of -85dBm, the data transmission packet loss rate is ≤0.1%. The IoT communication module supports node sleep / wake-up scheduling, and the scheduling period can be adjusted within the range of 10s-1h.

[0042] In this invention, the remote monitoring platform has a built-in energy consumption analysis unit, which analyzes energy consumption using formulas. Calculate energy consumption per unit distance; in the formula, Energy consumption per unit stroke, expressed in J / mm; This is real-time pressure, in Pa. Real-time traffic, unit: megabytes (m). 3 / s; t represents the travel time, in seconds; The total travel distance is in mm; the energy consumption analysis unit generates daily / monthly energy consumption reports with a statistical error of ≤2%, and also supports comparison with historical data for the same period. An alert is triggered when the data deviation exceeds 10%.

[0043] In this invention, the intelligent diagnostic module employs an improved BP neural network model. The input layer of the improved BP neural network model contains 16 feature parameters, covering pressure fluctuation values, temperature change rates, etc. The model has three hidden layers, each with 24-32 neurons, and uses ReLU as the activation function. The output layer corresponds to eight types of faults, and uses Softmax as the activation function. The model undergoes 5000-10000 training iterations, with the learning rate adaptively adjusted within the range of 0.001-0.01. With a test sample size ≥1000 groups, the model's diagnostic accuracy is ≥95%.

[0044] In this invention, the predictive maintenance module further includes a remaining useful life assessment unit, which assesses remaining useful life using a formula. Calculate remaining lifetime; in the formula, Remaining lifetime, in hours (h); Rated life, in hours, with a range of 5000-10000 hours. The specific value varies depending on the actuator model. The failure rate function varies with time, with units of 1 / h, and adopts the Weibull distribution model. Where m is a shape parameter, with a value ranging from 1.2 to 3.0. The value is a scale parameter, ranging from 1000 to 5000 hours; the remaining life assessment error is ≤5%, which is obtained by comparing it with the actual life of the actuator.

[0045] In this invention, the adaptive adjustment module is further equipped with a pressure compensation unit; when the ambient temperature changes within 10 minutes... Or the amount of gas source pressure fluctuation within 5 minutes At that time, the pressure compensation unit automatically activates the compensation mechanism; the compensation amount is determined by the formula. Calculation, in the formula The pressure compensation unit corrects the output pressure in real time through a proportional valve to ensure that the actuator output force stability is ≥98%, that is, the output force fluctuation is ≤±2%.

[0046] The following two examples further illustrate the specific implementation of this system:

[0047] Example 1: Monitoring and Diagnostic System for Pneumatic Actuators of Valves in the Chemical Industry (Application Scenario: Valve control for chemical reactors, 30 actuators distributed across a 2000㎡ workshop)

[0048] I. System Deployment and Module Detailed Implementation

[0049] 1. Deployment and parameter configuration of sensor monitoring modules

[0050] Each pneumatic actuator is equipped with one sensing and monitoring unit. A diffused silicon pressure sensor is installed at the actuator's air inlet and fixed via an M12 threaded interface. It collects real-time air source pressure and internal cavity pressure, with a measurement range of 0-1.6MPa, an accuracy of ±0.5%FS, and a response time of 4ms, ensuring the capture of detailed pressure fluctuations. A PT100 temperature sensor is embedded in the actuator cylinder surface and bonded with thermally conductive adhesive. It measures from -40℃ to 85℃, with an accuracy of ±0.3℃ and a sampling interval of 5s, adapting to the diurnal temperature variations in chemical workshops. A magnetostrictive displacement sensor is fixed next to the actuator piston rod via a bracket. The actuator features a 0-100mm travel range, 0.01mm resolution, and 0.015%FS linearity error, providing real-time monitoring of valve displacement. A piezoelectric vibration sensor, mounted on the actuator housing, measures at a frequency of 10-1000Hz with a sensitivity of 100mV / g±5%, capturing abnormal mechanical vibrations. A laser oil mist sensor and a capacitive humidity sensor are connected in series in the air supply line. The laser oil mist sensor has a detection range of 0-50ppm, an accuracy of ±0.1ppm, and a response time of 0.8s. The capacitive humidity sensor has a measurement range of 0-100%RH and an accuracy of ±2%RH, monitoring compressed air quality. An integrated data acquisition card is fixed to the nearest power distribution box to the actuator, with a sampling frequency of 5kHz and a 16-bit A / D conversion accuracy. Shielded cables connect the sensors to avoid electromagnetic interference in the chemical plant environment.

[0051] 2. Implementation of Data Processing and IoT Communication Module

[0052] The data processing module uses an STM32H743 microprocessor with a main frequency of 400MHz and an FPU floating-point unit, installed in the workshop control cabinet. It performs a 5th-order Butterworth filter on the acquired data, with a cutoff frequency set to 200Hz to filter high-frequency mechanical noise in the workshop; a wavelet transform threshold denoising threshold is set to 0.05V to retain effective signal characteristics; time-domain feature extraction includes 12 types of features such as peak value, mean, and variance, generating one set of feature data every 100ms. The data processing module has a built-in 8GBeMMC storage unit with a continuous storage time of 750 hours, supports 4G and Gigabit Ethernet communication, with a 4G communication rate of 150Mbps for remote data transmission, and a Gigabit Ethernet connection to the local server in the workshop for dual data backup. An improved Kalman filter algorithm was used for dynamic noise reduction. The filter window size was set to 16 sampling points, the initial value of the state equation covariance matrix was 0.5, and the measurement noise variance was 0.05 and updated adaptively. After processing, the signal-to-noise ratio of the data was improved by 35dB, and pulse interference (such as the impact of starting and stopping workshop equipment) and high-frequency noise above 500Hz with a deviation of more than 3 times the root mean square were effectively eliminated.

[0053] The IoT communication module includes an NB-IoT wireless transmission unit and an edge computing gateway. The NB-IoT unit is installed next to the sensing and monitoring unit, operating at 900MHz, with a transmission distance of 8km, a sleep current of 4μA, and a wake-up time of 180ms, suitable for large-scale coverage in the workshop. The edge computing gateway is deployed in the workshop's central control room, using an ARM Cortex-A53 quad-core processor with a main frequency of 1.2GHz, supporting MODBUSRTU / TCP, PROFINET, and EtherCAT protocol conversion with a protocol conversion latency of 8ms, achieving protocol compatibility with actuators from different brands. The gateway supports LoRaWAN edge node self-organizing networking, operating at 433MHz, with a maximum of 300 nodes, using a TDMA mechanism to avoid signal collisions, a time slot length of 50ms, and a data transmission packet loss rate of 0.08% at a receiver sensitivity of -85dBm. The node sleep / wake-up scheduling cycle is set to 30s to balance energy consumption and real-time performance. Data encryption uses the AES-128 algorithm, the key is automatically updated at 3:00 AM every day, and it supports data interruption resumption. When the network connection is restored after an interruption, it automatically retransmits the cached data during the interruption.

[0054] 3. Implementation of remote monitoring and intelligent diagnostic module

[0055] The remote monitoring platform is based on a B / S architecture, using Nginx as the web server and a MySQL cluster as the database, deployed on an enterprise cloud server. The real-time data dashboard refreshes every 0.8 seconds, displaying parameters such as pressure, temperature, displacement, and vibration for 30 actuators, with color-coded status (green for normal, yellow for warning, and red for fault). Historical curve queries have a 3-year storage period and support searching by actuator number and time range; curves can be zoomed in for detailed viewing. An electronic map marks the location of the 30 actuators with a positioning accuracy of ±5m; clicking the icon displays device details. The platform supports access from PCs (Windows / Linux) and mobile devices (Android / iOS), with a concurrent user capacity of 500, meeting the needs of multi-person collaborative management by the operations and maintenance team. A built-in energy consumption analysis unit is included, using formulas... Calculate energy consumption per unit distance, where Real-time pressure (Pa). Real-time traffic (m) 3 / s), where t is the travel time (s). Generate a daily energy consumption report for the total travel (mm), with a statistical error of 1.5%. An alert is triggered when the energy consumption deviates from the historical average by more than 10%.

[0056] The intelligent diagnostic module incorporates a hybrid diagnostic model that integrates an expert system and a random forest algorithm. The input layer contains 16 feature parameters, including pressure fluctuations, temperature change rates, and vibration peak values. The hidden layer consists of three layers, each with 28 neurons, using ReLU activation. The output layer corresponds to eight fault types, including air leakage, jamming, and abnormal pressure. Air leakage is categorized into three levels: slight (≤5% flow loss), moderate (5%-15%), and severe (≥15%). Jamming is categorized into four levels: slight (≤5% travel delay), moderate (5%-15%), severe (15%-30%), and extremely severe (≥30%). Abnormal pressure is categorized into two levels: mild (±10% deviation) and severe (±20% deviation). The model underwent 8000 training iterations with a learning rate of 0.005 that adaptively adjusts. With 1200 test samples, the diagnostic accuracy reached 96%, the diagnostic response time was 4 seconds, and the fault location accuracy was 0.4 meters. This module supports automatic labeling of fault samples and weekly iterative updates to the model to adapt to the characteristic changes caused by actuator aging.

[0057] 4. Implementation of Predictive Maintenance and Adaptive Adjustment Module

[0058] The predictive maintenance module uses formulas Calculate actuator reliability. Let t be the reliability (0-1). Let be the failure rate (1 / h, fitted to historical failure data to obtain 0.0005), and t be the running time (h). When the actuator runs continuously for 3000h, If the value falls below the 0.85 threshold, the system automatically generates a maintenance reminder, notifying maintenance personnel via a platform pop-up (displayed continuously until confirmation), SMS (with a 25-second delay), and APP push (retrying 3 times). The maintenance content includes a list of seal replacement parts (model O-ring 30×5), operating procedures (shutting off the air supply → disassembling the cylinder → replacing the seals → airtightness test), and spare parts inventory information. The remaining life assessment unit uses a formula... calculate, The rated lifespan is 8000 hours. Using the Weibull distribution model When t=3000h, , The evaluation error is 4%.

[0059] The adaptive adjustment module uses formulas Calculate the optimal working pressure. (Actuator diameter 15mm), Q is the gas flow rate (measured 100L / min). (The gas is compressed air), T is the operating temperature (measured at 25℃). , ,but (Actual value: 0.7 MPa, adapted to the actuator's rated pressure), adjustment accuracy: ±0.015 MPa, response time: 90 ms, the actuator uses a proportional pressure valve (15 mm diameter). The pressure compensation unit monitors ambient temperature and air source pressure; if the temperature drops by 6°C within 10 minutes... Or, the gas source pressure fluctuation within 5 minutes is 0.12 MPa. When this happens, a compensation mechanism will be activated, and the compensation amount will be... The proportional valve corrects the output pressure in real time, ensuring that the actuator output force is stable at 98.5% (fluctuation 1.5%).

[0060] Table 1: Comparison of the Management Effects of Pneumatic Actuators in Chemical Workshops between Example 1 System and Traditional Operation and Maintenance Methods

[0061] Evaluation indicators Traditional operation and maintenance methods This system Improvement effect Fault discovery time Average 48 hours Average 5 minutes Efficiency increased by 576 times False alarm rate 22% 3% Significantly reduced Maintenance costs (per year) 150,000 yuan 80,000 yuan Reduced by 46.7% Actuator lifespan 5000 hours on average 8000 hours on average Increase by 60% Energy consumption (per unit distance traveled) 120J / mm 95J / mm Reduced by 20.8%

[0062] Table 1 clearly demonstrates the significant advantages of this system in a chemical plant setting. Traditional maintenance relies on manual inspections, with an average fault detection time of 48 hours, often leading to escalation of problems. This system, through real-time monitoring and rapid diagnosis, can detect faults within 5 minutes, significantly shortening response time and reducing production losses. The false alarm rate has decreased from 22% to 3% because the system uses a hybrid diagnostic model to accurately distinguish the causes of faults, avoiding the blindness of traditional threshold-based judgments and reducing the ineffective workload of maintenance personnel. Annual maintenance costs have decreased from 150,000 yuan to 80,000 yuan, thanks to predictive maintenance that allows for on-demand replacement of spare parts, avoiding over-maintenance waste. Simultaneously, actuator lifespan has increased by 60%, reducing equipment replacement investment. Energy consumption has decreased by 20.8% because the adaptive adjustment module optimizes pressure parameters based on operating conditions, avoiding energy waste and meeting the energy-saving and consumption-reducing requirements of the chemical industry. Overall, this verifies the system's value in high reliability and low-cost maintenance.

[0063] Example 2: Monitoring and Diagnostic System for Pneumatic Actuators of Robotic Arms in Automobile Manufacturing Production Line (Application Scenario: Automobile welding production line, 6 robotic arms, each containing 4 pneumatic actuators)

[0064] I. System Adaptation and Detailed Implementation

[0065] 1. Sensor monitoring and communication module adaptation

[0066] Each robotic arm's four pneumatic actuators (responsible for gripper opening and closing, and joint rotation) are equipped with one set of sensing units: diffused silicon pressure sensors with a measurement range of 0-1.2MPa (suitable for the low-pressure requirements of robotic arms), an accuracy of ±0.5%FS, and a response time of 3ms; PT100 temperature sensors with a sampling interval of 2s, suitable for the high-temperature environment of welding workshops (up to 60℃); magnetostrictive displacement sensors with a travel range of 0-80mm (for the short-travel requirements of robotic arms) and a resolution of 0.01mm; and piezoelectric vibration sensors with a measurement frequency of 10-800Hz (capturing high-frequency vibrations of the robotic arm). The laser oil mist sensor has a detection threshold of 4ppm (due to the high risk of oil mist contamination in welding workshops), and the capacitive humidity sensor has a threshold of 75%RH, triggering an early warning within 10 seconds when the threshold is exceeded. The IoT communication module uses a LoRaWAN self-organizing network, operating at 868MHz, with a maximum of 24 nodes (6 robotic arms × 4 actuators), a time slot length of 30ms, and a data transmission packet loss rate of 0.05%, suitable for densely deployed production lines.

[0067] 2. Intelligent Diagnosis and Predictive Maintenance Optimization

[0068] The intelligent diagnostic module for robotic arm actuators adds a new fault type, "gripper loosening" (9 categories in total), with refined fault classification (5 levels of jamming to meet the high-precision movement requirements of robotic arms). The model input layer adds a "displacement synchronization" feature parameter (a key indicator of multi-actuator coordination in robotic arms). With 1500 test samples, the diagnostic accuracy is 97%, the diagnostic response time is 3 seconds, and the fault location accuracy is 0.3 meters. Predictive maintenance module. A value of 0.0008 indicates that the robotic arm actuator operates frequently and has a high failure rate. ,when hour, The system sends a maintenance reminder 24 hours in advance, which includes a suggested shutdown window for the robotic arm (1 hour lunch break) and a spare parts preparation list (guide sleeve, sealing ring).

[0069] 3. Adaptive adjustment and platform function compatibility

[0070] Adaptive adjustment module The value is 0.04 MPa・min / L (10 mm diameter of robotic arm actuator). (The temperature in the welding workshop fluctuates greatly.) The value range is 0.5-1.0 MPa, and the adjustment response time is 80ms to ensure precise and synchronized robotic arm movements. The remote monitoring platform has added a "production line cycle time correlation analysis" function, which links the actuator status with the production line cycle time. When the actuator movement delay exceeds 50ms, the production line speed is automatically adjusted to avoid product accumulation. Energy consumption reports are compiled by robotic arm number and support correlation analysis with production output to identify low-output, high-consumption equipment.

[0071] Table 2: Comparison of the Management Effects of Pneumatic Actuators in Automobile Assembly Lines between Example 2 System and Traditional Operation and Maintenance Methods

[0072] Evaluation indicators Traditional operation and maintenance methods This system Improvement effect Actuator motion accuracy ±0.5mm ±0.1mm Increase by 80% Number of production line downtime Average 12 times / month Average 2 times / month Reduced by 83.3% Number of maintenance personnel 3 people / class 1 person / class Reduced by 66.7% Fault repair time Average 90 minutes Average 20 minutes Shortened by 77.8% Product defect rate 1.5% 0.3% Reduce by 80%

[0073] Table 2 highlights the core value of this system in automotive manufacturing assembly line scenarios. Traditional maintenance methods achieve an actuator motion accuracy of ±0.5mm, which is insufficient to meet the high-precision positioning requirements of automotive welding fixtures, resulting in a product defect rate of 1.5%. This system, through precise adjustment and monitoring, improves motion accuracy to ±0.1mm, reducing the product defect rate to 0.3%, significantly improving product quality. Assembly line downtime is reduced from 12 times / month to 2 times, as the system provides early warnings of potential faults, avoiding sudden downtime. Fault repair time is shortened by 77.8%, as precise fault location allows maintenance personnel to quickly troubleshoot, reducing production interruption losses. Maintenance personnel are reduced from 3 per shift to 1 per shift, as the system enables centralized remote management, eliminating the need for manual on-site inspections and reducing labor costs. Simultaneously, the system's synchronization with the assembly line cycle ensures production continuity, further improving manufacturing efficiency.

[0074] Reference Figure 2 This figure highlights the rapid response and high accuracy advantages of the intelligent diagnostic module in this system. Traditional threshold diagnosis relies on manually setting fixed thresholds, requiring 120 minutes to accumulate sufficient deviation to trigger an alarm for early minor faults, and has low accuracy. Even when the fault progresses to a later stage, the response time still requires 30 minutes, easily missing the optimal repair opportunity. This system adopts a hybrid model that integrates expert systems and random forest algorithms, which can identify early faults from multi-parameter correlation analysis. For example, by comprehensively judging minor air leaks through pressure fluctuation variance and vibration peak value changes, the response time is only 3.5 seconds, with an accuracy of 92%. As the fault worsens, the model can call on more fault sample features, shortening the response time to 2.5 seconds and improving the accuracy to 98%.

[0075] Reference Figure 3 This diagram illustrates the forward-looking advantages of the predictive maintenance module in this system. Traditional periodic maintenance replaces spare parts at fixed intervals (e.g., 5000 hours), failing to consider individual actuator differences. For example, after 3000 hours of operation, the reliability may have dropped to 0.7, and failure to maintain it at this stage can easily lead to malfunctions. Conversely, some actuators may still have a reliability of 0.6 after 5000 hours due to good operating conditions, yet they are still forcibly replaced, resulting in waste. This system uses a reliability formula... With remaining lifetime formula The system dynamically assesses equipment status: when reliability drops to 0.85 after 3000 hours of operation, a maintenance reminder is triggered, with a remaining lifespan of 4500 hours, allowing maintenance personnel to plan downtime for repairs in advance; when reliability reaches 0.75 after 5000 hours of operation, with a remaining lifespan of 2500 hours, an emergency reminder is issued. This "on-demand maintenance" model avoids the cost waste of traditional "over-maintenance" and prevents the failure risks of "under-maintenance," significantly improving the economic efficiency of operation and maintenance.

[0076] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things, characterized in that, Includes the following modules: The sensing and monitoring module is equipped with a diffused silicon pressure sensor, a PT100 temperature sensor, a magnetostrictive displacement sensor, a piezoelectric vibration sensor, and an integrated data acquisition card to collect all dimensions of the pneumatic actuator's operating parameters. The data processing module uses an STM32H743 microprocessor with a main frequency of 400MHz and is equipped with an FPU floating-point arithmetic unit. The data processing module performs 5th-order Butterworth filtering, wavelet transform threshold denoising, and time-domain feature extraction on the sensor data. The data processing module has a built-in 8GBeMMC storage unit and supports 4G, 5G and Gigabit Ethernet communication. The Internet of Things (IoT) communication module includes an NB-IoT wireless transmission unit and an edge computing gateway; the NB-IoT wireless transmission unit operates in the 800 / 900MHz frequency band and has a transmission distance of 3-10km. The edge computing gateway uses an ARM Cortex-A53 quad-core processor and supports MODBUSRTU / TCP, PROFINET, and EtherCAT protocol conversion; the IoT communication module uses the AES-128 algorithm for data encryption, with a key update cycle of 24 hours. The remote monitoring platform is based on a B / S architecture, with an Nginx web server and a MySQL cluster database. The platform includes real-time data dashboards, historical curve queries, and electronic maps of device status. The remote monitoring platform supports access from both PC and mobile devices. The intelligent diagnostic module incorporates a hybrid diagnostic model that integrates an expert system and a random forest algorithm. It identifies eight typical faults, including three levels for air leakage, four levels for jamming, and two levels for abnormal pressure. The intelligent diagnostic module has a diagnostic response time of ≤5s, a fault location accuracy of ≤0.5m, and supports automatic labeling of fault samples and iterative model updates. It also includes a predictive maintenance module, which uses formulas Calculate the actuator reliability; where, Let be the reliability at time t; t represents the failure rate; t represents the running time; when When the value is ≤0.85, the system will automatically generate a maintenance reminder. The reminder methods include platform pop-ups, SMS, and APP push notifications. Platform pop-ups will continue to be displayed until the user confirms, and APP push notifications will implement a 3-retry mechanism. The maintenance reminder content includes a list of parts to be replaced, operation steps, and spare parts models. It also includes an adaptive adjustment module, which uses a formula Calculate the optimal working pressure; where To achieve optimal work pressure; Where is the flow coefficient; Q is the gas flow rate; T is the temperature coefficient; T is the operating temperature. Environmental pressure coefficient; The ambient atmospheric pressure is used; the adaptive adjustment module has an adjustment accuracy of ≤±0.02MPa and a response time of ≤100ms. The adjustment actuator adopts a proportional pressure valve with a nominal diameter of 6-20mm. The predictive maintenance module also includes a remaining lifetime assessment unit, which assesses remaining lifetime using a formula. Calculate the remaining lifetime; where, Remaining lifespan; Rated lifespan; The failure rate function varies with time, and a Weibull distribution model is adopted. Where m is the shape parameter, The remaining life assessment error is ≤5%, which is obtained by comparing it with the actual life of the actuator. The adaptive adjustment module is also equipped with a pressure compensation unit; when the ambient temperature changes within 10 minutes... Or gas source pressure fluctuation within 5 minutes At that time, the pressure compensation unit automatically activates the compensation mechanism; the compensation amount is determined by the formula. Calculation, in the formula The pressure compensation unit corrects the output pressure in real time through a proportional valve.

2. The remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things as described in claim 1, characterized in that, The sensing and monitoring module is also equipped with a laser oil mist sensor and a capacitive humidity sensor. The laser oil mist sensor and the capacitive humidity sensor are used to monitor the oil mist concentration and relative humidity in the compressed air. When the oil mist concentration is ≥5ppm and the state lasts for more than 10 seconds, or the relative humidity is ≥80% and the state lasts for more than 30 seconds, the system triggers a pollution warning. The warning adopts a three-level warning mechanism, and different warning levels correspond to different handling strategies.

3. The remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things according to claim 1, characterized in that, The data processing module uses an improved Kalman filter algorithm for dynamic noise reduction. The filter window size of the improved Kalman filter algorithm is set to 8-32 sampling points, and the window size is automatically switched according to the noise intensity. The initial value of the covariance matrix of the state equation of the algorithm is adjusted in the range of 0.1-1.0, and the measurement noise variance is adaptively updated in the range of 0.01-0.

1.

4. The remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things according to claim 1, characterized in that, The IoT communication module supports LoRaWAN edge node self-organizing network, with the self-organizing network operating frequency bands of 433 / 868 / 915MHz; the maximum number of nodes in the edge node self-organizing network is 100-500, and the number of nodes can be expanded through gateway cascading; the self-organizing network adopts a time division multiple access mechanism, with a time slot length of 10-100ms.

5. The remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things according to claim 1, characterized in that, The remote monitoring platform has a built-in energy consumption analysis unit, which analyzes energy consumption using formulas. Calculate energy consumption per unit distance traveled; where, Energy consumption per unit distance; For real-time pressure; 'a' represents real-time traffic; 'a' represents trip time. The total travel time is calculated; the energy consumption analysis unit generates daily / monthly energy consumption reports, and also supports comparison with historical data for the same period. When the data deviation exceeds 10%, an alert is triggered.

6. The remote monitoring and intelligent diagnostic system for pneumatic actuators based on the Internet of Things according to claim 1, characterized in that, The intelligent diagnostic module adopts an improved BP neural network model; the input layer of the improved BP neural network model contains 16 feature parameters; the hidden layer of the model has 3 layers, each with 24-32 neurons, and the activation function is ReLU; the output layer of the model corresponds to 8 types of faults, and the activation function is Softmax. The model is trained 5,000-10,000 times, and the learning rate is adaptively adjusted within the range of 0.001-0.01.

Images