A remote operation and maintenance oil particle counting and classification system

Abstract

This invention discloses a remote oil particle counting and classification system, relating to the field of industrial equipment condition monitoring technology. The system includes: a working condition data acquisition module, employing industrial-grade high-precision sensors installed on key parts of the equipment's hydraulic system and power output to collect hydraulic pressure and operating load parameters in real time. These parameters are then transmitted to a remote terminal via a remote transmission module. This invention achieves dynamic adaptive adjustment of the classification threshold by deeply integrating oil particle detection with the equipment's real-time working conditions, greatly improving detection accuracy. The working condition data acquisition module and the oil particle detection module work synchronously, capturing core parameters and raw particle data under different operating conditions of the equipment in real time. A remote machine learning modeling module trains a model adapted to different working conditions based on this data, and a dynamic threshold control module generates precise threshold commands based on the model.

Description

Technical Field

[0001] This invention relates to the field of industrial equipment condition monitoring technology, specifically to a remote oil particle counting and classification system for operation and maintenance. Background Technology

[0002] In the field of industrial equipment condition monitoring technology, hydraulic equipment is widely used as a key piece of equipment in various industrial scenarios. Its operating status is directly related to the stability and efficiency of the entire production process. As the medium for transmitting power and signals in hydraulic equipment, the cleanliness of oil is closely related to the wear condition of internal components. Oil particle detection technology is the core means to judge the wear condition of hydraulic equipment and assess the cleanliness of oil. By analyzing the characteristics of particles in the oil, such as the number, size, and material, it can promptly detect potential faults in the equipment and provide an important basis for equipment operation and maintenance decisions.

[0003] Existing oil particle counting and classification systems generally employ a fixed threshold classification mode, which pre-sets the criteria for judging particle material and size grades. Regardless of the equipment's operating conditions, detection and classification are performed according to a uniform threshold. However, the generation mechanism, material composition, and size distribution of particles in hydraulic equipment vary significantly under different operating conditions. During high-load operation, increased wear of equipment components generates a large number of tiny metal particles, requiring improved sensitivity in identifying such particles. Under low-load conditions, the proportion of non-metallic impurities (such as dust and oil degradation products) increases, necessitating optimization of the classification accuracy of non-metallic particles. The fixed threshold mode cannot be dynamically adjusted according to the real-time operating conditions of the equipment, leading to the problem of missing tiny wear particles under high-load conditions and misclassifying non-metallic particles as metallic particles under low-load conditions, seriously affecting the accuracy of equipment fault warnings. Furthermore, most existing remote operation and maintenance technologies only achieve remote acquisition and display of oil particle data, failing to deeply link remote data processing with particle classification threshold adjustment. They cannot dynamically adjust detection standards according to the real-time operating conditions of the equipment, making it difficult to meet the needs of refined operation and maintenance of industrial equipment. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a remote oil particle counting and classification system. This system achieves dynamic adaptive adjustment of the classification threshold by deeply integrating oil particle detection with real-time equipment operating conditions and combining machine learning algorithms. The system consists of a field end and a remote end. The field end includes an operating condition data acquisition module, an oil particle detection module, and a field control module, responsible for real-time acquisition of equipment operating parameters and raw oil particle data, and executing commands issued by the remote end. The remote end includes a remote transmission module, a machine learning modeling module, and a dynamic threshold adjustment module, responsible for data transmission, model training, and threshold adjustment command generation. The operating condition data acquisition module employs high-performance... The system features high-precision industrial-grade sensors that simultaneously collect core operating parameters such as hydraulic pressure and operating load. The oil particle detection module integrates a laser-based shading detection component and a magnetic response material recognition component, enabling accurate detection of particle quantity, size, and material. The remote transmission module employs a wired + wireless dual-mode design to ensure stable and secure data transmission. The machine learning modeling module trains an adaptive model based on historical correlation data and outputs dynamic judgment thresholds. The dynamic threshold adjustment module generates control commands based on real-time equipment operating conditions, enabling dynamic adjustment of thresholds. This solves the problems of poor adaptability and high false positive / false negative rates associated with fixed threshold modes, significantly improving the accuracy of oil particle detection and the efficiency of remote operation and maintenance.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a remote oil particle counting and classification system, the system comprising: Operating data acquisition module: Industrial-grade high-precision sensors are installed in key parts of the equipment's hydraulic system and power output to collect hydraulic pressure and operating load parameters in real time, and upload the parameters to a remote terminal through a remote transmission module; Oil Particle Detection Module: Deployed at key sampling points in the hydraulic system of the equipment, it integrates a laser shading detection component and a magnetic response material identification component. It detects the number, size range, and material properties of particles through light scattering and magnetic response signals, generates raw data, and uploads it synchronously to a remote terminal. Remote transmission module: It is compatible with wired + wireless dual-mode communication, establishes a two-way data channel between the field end and the remote end, ensures transmission security through encryption and verification mechanisms, and follows industrial control standard communication protocols to realize field data uploading and remote command issuance; Machine learning modeling module: Based on historical working conditions and oil particle detection data, a training dataset is built, and a correlation model is trained through machine learning algorithms. The output is a dynamic judgment threshold for particle material and size level that is adapted to the current working conditions. It also has data storage and retrieval functions. Dynamic threshold control module: After receiving real-time operating parameters, it calls the associated model to determine the equipment operating status and generate targeted threshold control instructions. Under high load, it tightens the metal particle detection threshold and under low load, it optimizes the non-metal particle threshold. Field control module: An industrial-grade PLC controller or embedded control module is installed in the equipment's electrical control box. After receiving remote control commands, it parses and decodes them, drives the oil particle detection module to update the threshold and execute detection operations, and feeds back the results to the remote end to form a closed-loop control.

[0006] Furthermore, the oil particle detection module is deployed at the oil tank outlet or key oil circuit sampling point of the equipment's hydraulic system. It adopts an integrated design, internally comprising a laser shading detection component and a magnetic response material identification component. The laser shading detection component emits a laser beam of stable wavelength through the oil sample. Utilizing the changes in light scattering signals generated by particles blocking the laser, a size comprehensive judgment algorithm is used to count the number of particles and determine their preliminary size range. The magnetic response material identification component generates a stable magnetic field through a built-in electromagnetic induction coil. Based on the differences in magnetic response signals generated by ferrous, non-ferrous, and non-metallic particles in the magnetic field, a particle material judgment algorithm is used to calculate the particle material judgment coefficient, distinguishing the material characteristics of the three types of particles. Finally, raw detection data containing particle number, size range, and material properties is generated. The module uses a sealed sampling structure to prevent oil leakage and a quick-plug interface for easy installation and maintenance. The raw detection data is transmitted via cable to the field control module and then synchronously uploaded to the remote terminal via a remote transmission module.

[0007] Furthermore, the oil particle detection module uses a size comprehensive judgment algorithm to count the number of particles and determine the initial size range. The algorithm formula is as follows: ,in, It is the first Precise size calculation value for each particle. It is the first The laser blocking intensity corresponding to each particle It is the reference value of laser transmittance intensity when there are no particles. It is a size correction factor that compensates for the influence of oil refractive index and temperature on the test results.

[0008] Furthermore, the oil particle detection module uses a particle material determination algorithm to calculate the particle material determination coefficient to distinguish the material characteristics of the three types of particles. The algorithm formula is as follows: ,in, It is the first The material determination coefficient of each particle is used to distinguish between ferrous / non-ferrous metals / non-metals; It is a signal fusion weight that balances the influence of light scattering and magnetic response signals. It is the first The light scattering signal intensity of each particle It is the reference value of the light scattering signal of standard particles. It is the first The amplitude of the magnetic response signal of each particle. It is the reference value for the magnetic response signal of standard iron particles.

[0009] Furthermore, the remote transmission module adopts a wired + wireless dual-mode adaptive design. The wired communication is based on an industrial Ethernet architecture, achieving a stable connection between the field end and the remote end through shielded network cables. The wireless communication uses an industrial-grade wireless module, supporting long-distance data transmission and possessing anti-electromagnetic interference capabilities. It also integrates data encryption and verification components to encrypt the uploaded operating parameters and raw particle detection data, and adds a unique verification code to the issued control commands. The communication protocol strictly follows industrial control standards to ensure compatibility between different devices and modules, establishing a bidirectional data channel between the field end and the remote end. On the one hand, it enables real-time uploading of field data to the remote end, and on the other hand, it enables the accurate issuance of control commands from the remote end to the field control module.

[0010] Furthermore, the machine learning modeling module constructs a standardized training dataset based on historical operating condition data accumulated by similar equipment under different operating scenarios, as well as oil particle detection data under corresponding operating conditions. The dataset contains samples of the correlation between different operating conditions and particle material and size distribution. The operating condition-particle characteristic correlation model algorithm is used for model training, and the algorithm parameters are optimized through multiple iterations of training on the dataset. The input of the model is the operating condition parameter vector uploaded in real time from the field, and the output is the dynamic judgment threshold of particle material and size level for the current operating condition. At the same time, the module has a historical data storage function, supporting long-term storage and fast retrieval of training datasets and model running data.

[0011] Furthermore, the machine learning modeling module employs a working condition-particle characteristic correlation model algorithm for model training, the algorithm formula of which is: ,in, It is the first Material type Dynamic determination thresholds for each size level, It is the first Threshold weights for material types It is the first Class 1 working condition parameters and the first Material type Correlation coefficients for each size level It is the first Material type Threshold correction factor for each size level, It is the first under the current working conditions Material type Historical frequency of particle occurrence at each size grade It is the first under all working conditions Material type Average frequency of occurrence of particles in each size class These are the normalized operating parameters.

[0012] Furthermore, the dynamic threshold control module receives current operating parameters uploaded from the field terminal in real time, determines the operating status of the equipment through the built-in operating condition recognition program, and then automatically calls the operating condition-particle characteristic association model trained by the machine learning modeling module. It inputs the current operating parameters to calculate the appropriate dynamic threshold and optimizes the initial threshold using a dynamic threshold correction algorithm. For high-load operating conditions, it generates control instructions to tighten the detection thresholds for ferrous and non-ferrous metal particles, lowering the size judgment limit for both types of metal particles. For low-load operating conditions, it generates control instructions to optimize the non-metallic particle threshold, defining the size judgment range for non-metallic particles and reducing the misjudgment of dust-like non-metallic particles as metal particles. After the instructions are generated, they are encoded in a standardized format and sent to the field control module through the downlink channel of the remote transmission module.

[0013] Furthermore, the dynamic threshold control module employs a dynamic threshold correction algorithm to optimize the initial threshold, the algorithm formula of which is: ,in, It was the first one that was finally issued to the site. Material type Thresholds after adjustment at each size level, It is the threshold control sensitivity coefficient. It is the normalized value of the current operating load of the equipment. It is the load threshold, the dividing point that distinguishes between high and low loads. It is the normalized value of the equipment's maximum load. It is a symbolic function. When the threshold is set to 1, the threshold is tightened. When the threshold is -1, optimize the threshold.

[0014] Compared with existing technologies, this remote operation and maintenance oil particle counting and classification system has the following advantages: I. This invention achieves dynamic adaptive adjustment of classification thresholds by deeply integrating oil particle detection with real-time equipment operating conditions, greatly improving detection accuracy. The operating condition data acquisition module and the oil particle detection module work synchronously, capturing core parameters and raw particle data under different operating conditions of the equipment in real time. The remote machine learning modeling module trains a model adapted to different operating conditions based on this data. The dynamic threshold adjustment module generates precise threshold instructions based on the model. After the on-site control module executes the instructions to update the threshold, the detection module operates according to the new standard. Under high load, it can improve the sensitivity of identifying tiny metal particles, and under low load, it can optimize the classification accuracy of non-metallic particles. It effectively solves the problem of poor adaptability of fixed threshold mode, significantly reduces the false positive rate and false negative rate, and provides more accurate data support for equipment fault early warning.

[0015] Second, this invention adopts a wired + wireless dual-mode design through a remote transmission module, which can be flexibly selected according to the industrial field communication environment. The wired communication is based on an industrial Ethernet architecture, and the wireless communication uses an industrial-grade wireless module to ensure the stability of data transmission. At the same time, it integrates data encryption and verification components to encrypt uploaded data and add a unique verification code to the issued instructions, ensuring the security and integrity of data transmission. The communication protocol strictly follows industrial control standards, ensuring compatibility between different devices and modules.

[0016] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 This is a block diagram of a remote operation and maintenance oil particle counting and classification system; Figure 2 This is a flowchart of a remote operation and maintenance oil particle counting and classification system. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0020] Example 1 The operational data acquisition module utilizes industrial-grade pressure sensors, load sensors, and speed sensors, which are respectively installed via vibration-damping brackets at key pressure nodes in the hydraulic system of the hydraulic excavator, such as the hydraulic pump outlet and main relief valve; power output components, such as the engine output; and equipment operation status monitoring points. All sensors are vibration-resistant, adaptable to high and low temperature industrial environments, and have an IP65 protection rating, resisting dust and rain corrosion during outdoor operations. They establish a stable signal connection with the field control module through standardized interfaces and operate synchronously with the oil particle detection module according to a preset sequence, capturing core operational parameters such as equipment hydraulic pressure, operating load, start / stop frequency, and operating duration in real time. Preprocessing the raw data by normalizing operating parameters effectively eliminates the interference of different parameter magnitudes on subsequent modeling, ensuring data format uniformity. The oil particle detection module is deployed at the oil outlet of the hydraulic system tank, adopting an IP67 sealed sampling structure. It connects to the oil circuit via a quick-plug interface, facilitating daily installation and maintenance while preventing oil leakage. The module integrates a laser shading detection component and a magnetic response material identification component. Through a comprehensive particle material and size determination algorithm, it accurately captures the light scattering signal, magnetic response signal, and laser shading intensity of the particles. Specifically, the size determination algorithm is used to count the number of particles and determine the preliminary size range. The algorithm formula is as follows: ,in, It is the first Precise size calculation value for each particle. It is the first The laser blocking intensity corresponding to each particle It is the reference value of laser transmittance intensity when there are no particles. This is a size correction factor that compensates for the influence of oil refractive index and temperature on the detection results. A particle material determination algorithm is used to calculate the material determination coefficient of the particles, thus distinguishing the material characteristics of the three types of particles. The algorithm formula is as follows: ,in, It is the first The material determination coefficient of each particle is used to distinguish between ferrous / non-ferrous metals / non-metals; It is a signal fusion weight that balances the influence of light scattering and magnetic response signals. It is the first The light scattering signal intensity of each particle It is the reference value of the light scattering signal of standard particles. It is the first The amplitude of the magnetic response signal of each particle. It serves as a reference value for the magnetic response signal of standard ferrous particles, efficiently enabling particle quantity counting, size range determination, and differentiation of material properties among ferrous, non-ferrous metallic, and non-metallic particles. It generates raw detection data containing the core characteristics of the particles and transmits this data stably to the field control module via shielded cables. The field control module uses an industrial-grade PLC controller, fixedly installed inside the equipment's electrical control box, and possesses vibration resistance and adaptability to high and low temperature industrial environments. Figure 1 As shown, the preprocessed parameters from the working condition data acquisition module and the raw detection data from the oil particle detection module are received synchronously through the input interface, and a wireless connection is established with the remote transmission module through the communication interface.

[0021] The remote transmission module uses an industrial-grade wireless module, which is fixed to a location with good signal on the top of the excavator cab using a waterproof mounting bracket, avoiding obstruction from the cab and the influence of mechanical vibration. It establishes a stable wireless communication connection with the industrial computer at the remote end. The module adopts wireless communication to precisely adapt to the excavator's outdoor operation scenario. It integrates data encryption and verification components, and performs symmetrical encryption processing on the uploaded pre-processed parameters of working conditions and the original particle detection data to prevent theft or tampering during data transmission. It adds a unique check code to the issued control commands to ensure the integrity of the commands. At the same time, it monitors the data transmission quality in real time. When the signal weakens or the delay increases, it automatically adjusts the transmission power and frequency to ensure the security, stability and real-time performance of the transmission. The communication protocol strictly follows industrial control standards and is compatible with remote industrial computers and field PLC controllers. It establishes a two-way data channel between the field end and the remote end, realizing the real-time uploading of field data to the remote end and the precise issuance of remote commands to the field control module.

[0022] The machine learning modeling module and the dynamic threshold control module are jointly deployed on a remote industrial computer. This industrial computer is equipped with a high-performance processor and a large-capacity hard drive, enabling long-term storage and rapid retrieval of massive historical data from multiple excavators. The machine learning modeling module collects historical operating condition data accumulated by this model of hydraulic excavator under different operating scenarios, such as high-load digging, low-load transport, and idle start-stop, as well as corresponding oil particle detection data under these operating conditions, through a remote transmission channel. After being categorized and organized according to operating condition type, a standardized training dataset is constructed. The dataset contains samples of the correlation between different operating conditions and particle material and size distribution. The dataset is iteratively trained multiple times using an operating condition-particle characteristic correlation model algorithm, the formula of which is: ,in, It is the first Material type Dynamic determination thresholds for each size level, It is the first Threshold weights for material types It is the first Class 1 working condition parameters and the first Material type Correlation coefficients for each size level It is the first Material type Threshold correction factor for each size level, It is the first under the current working conditions Material type Historical frequency of particle occurrence at each size grade It is the first under all working conditions Material type By continuously optimizing the key parameters of the algorithm based on the average occurrence frequency of particles at each size level, the model's adaptability and generalization performance under different working conditions are significantly improved. Ultimately, an association model that can accurately output the dynamic judgment threshold for particle classification adapted to the current working condition is formed. The dynamic threshold control module establishes a real-time signal connection with the machine learning modeling module through an internal bus. It has integrated functions of data reception, working condition analysis, and instruction generation. It can quickly call the trained association model based on the working condition data uploaded on site to ensure the timeliness of threshold control.

[0023] When a hydraulic excavator enters a high-load excavation state during mining operations, the working condition data acquisition module captures various working condition parameters such as hydraulic pressure and operating load in real time. After preprocessing by a working condition parameter normalization algorithm, the data is stably uploaded to a remote industrial computer via a remote transmission module. Upon receiving the real-time working condition parameters, the dynamic threshold control module quickly determines that the equipment is in a high-load operating state using its built-in working condition recognition program. It automatically calls the working condition-particle characteristic correlation model trained by the machine learning modeling module to obtain a preliminary dynamic threshold. This preliminary threshold is then optimized using a dynamic threshold correction algorithm, the formula of which is: ,in, It was the first one that was finally issued to the site. Material type Thresholds after adjustment at each size level, It is the threshold control sensitivity coefficient. It is the normalized value of the current operating load of the equipment. It is the load threshold, the dividing point that distinguishes between high and low loads. It is the normalized value of the equipment's maximum load. It is a symbolic function. When the threshold is set to 1, the threshold is tightened. The system uses a threshold value of -1 to optimize the detection threshold and generates control commands to tighten the detection thresholds for ferrous and non-ferrous metal particles. This effectively lowers the size judgment limit for both types of metal particles, significantly improves the sensitivity of identifying minute wear particles, and promptly captures early wear signals of equipment components. When the excavator completes its digging operation and switches to a low-load transport state, the dynamic threshold control module uses the same process to determine the working condition and generates control commands to optimize the non-metallic particle threshold. This rationally defines the size judgment range for non-metallic particles, significantly reducing the misjudgment of dust-like non-metallic particles as metal particles and ensuring the accuracy of detection results. The control commands are encoded in a standardized format and accurately transmitted to the field control module via the downlink channel of the remote transmission module. The field control module receives these commands. Upon receiving the command, the built-in command parsing program is immediately activated to decode and verify its validity. This quickly drives the oil particle detection module to update the classification threshold parameters, controlling the detection module to efficiently perform particle counting and classification operations according to the new threshold. After the detection is completed, the on-site control module transmits the final detection results in real time to the machine learning modeling module and dynamic threshold adjustment module at the remote end via the remote transmission module, forming a complete closed-loop control. Maintenance personnel can view the equipment status, particle detection results, and threshold adjustment records in real time through the visual interface of the remote industrial computer. When the number of metal particles detected exceeds the preset warning value, the system automatically sends fault warning information via both SMS and platform pop-up windows, reminding maintenance personnel to arrange equipment maintenance in a timely manner.

[0024] Example 2 Based on the workshop production layout, industrial hydraulic press models, and stamping production requirements, the selection and deployment planning of each system module were completed. The working condition data acquisition module uses an industrial-grade pressure sensor with an accuracy of 0.5 and a load sensor with a range of 0-200kN. The remote transmission module adopts an industrial Ethernet architecture. The field control module uses an embedded control module that supports multiple interfaces. The remote end is configured with a cluster architecture server. At the same time, the historical working condition data of the hydraulic press for the past 3 years, including stamping, pressure holding, and return, are collected, including hydraulic pressure curves, operating load records, and corresponding oil particle detection reports, to provide a basic dataset for subsequent model training.

[0025] As planned, the pressure sensor of the working condition data acquisition module is installed at the oil circuit node of the main hydraulic cylinder inlet using a fixed bracket with shock-absorbing pads. The load sensor is integrated into the load-bearing part of the hydraulic press workbench. The sensor wiring uses double-shielded cables protected by galvanized pipes, laid along the workshop cable tray to the electrical control box, effectively resisting electromagnetic interference generated by welding equipment and motors in the workshop. A signal connection is established with the field control module through a standard RS485 interface. The oil particle detection module is deployed at the sampling point of the main hydraulic cylinder return oil circuit, using an IP65 sealed sampling connector to connect with the oil circuit. The quick-plug design facilitates periodic calibration. The module is connected to the field control module through a shielded cable. The field control module is fixedly installed in the hydraulic press electrical control cabinet, realizing signal linkage with the hydraulic press main control system. After power-on debugging is completed, as... Figure 2 As shown, the raw data collected by the sensor is preprocessed by normalizing the operating parameters to eliminate the differences in the magnitude of different parameters. At the same time, the working status of the laser shading method component and the magnetic response component of the oil particle detection module is verified to ensure that the particle characteristic data can be accurately output through the comprehensive judgment algorithm of particle material and size.

[0026] The remote transmission module is connected to the workshop's industrial Ethernet switch via a shielded network cable, with the other end connected to the remote server cluster. The network cable is properly grounded and laid in a separate trench from the power cables to avoid interference. A machine learning modeling module and a dynamic threshold control module are deployed on the remote server, and communication protocol adaptation is completed to ensure bidirectional data interaction with the hydraulic press's field control module. Based on previously prepared historical data, the machine learning modeling module constructs a training dataset including three working conditions: high-load stamping, low-load holding, and no-load return. Iterative training is performed using a working condition-particle characteristic correlation model algorithm, focusing on optimizing the correlation coefficient of metal particle detection under stamping conditions, generating a dedicated correlation model adapted to this hydraulic press model. The dynamic threshold control module establishes a real-time signal connection with the modeling module, completing the matching and debugging of the command encoding format with the field control module.

[0027] The system was started and entered the trial operation phase. The detection cycle was set in sync with the production rhythm of the hydraulic press. When the hydraulic press performed the stamping operation, the working condition data acquisition module uploaded hydraulic pressure and load data in real time. The remote end monitored the data transmission quality through the transmission reliability verification algorithm. The dynamic threshold control module called the correlation model to generate the initial threshold, which was then optimized by the threshold dynamic correction algorithm and sent to the site. During the trial operation, the threshold adjustment effect, detection error and command response time under different working conditions were recorded. At the same time, the algorithm weight parameters were fine-tuned according to the detection results to improve the detection accuracy.

[0028] After the system is officially put into use, it is integrated into the workshop production process. When the hydraulic press starts and receives the stamping command to enter the high-load condition, the working condition data acquisition module captures parameters such as the main hydraulic cylinder pressure and load in real time. After normalization and preprocessing of the working condition parameters, the data is transmitted to the remote server via industrial Ethernet. After receiving the data, the dynamic threshold control module completes the working condition identification within 1 second, calls the correlation model to obtain the initial threshold, and then tightens the metal particle detection threshold through the threshold dynamic correction algorithm. The instruction is encoded and sent to the field control module. The field control module quickly parses the instruction and drives the oil particle detection module to update the threshold. The detection is performed according to the new standard, focusing on identifying tiny metal particles generated by punch wear. When the hydraulic press completes the stamping and switches to the pressure holding and low-load condition, the system judges the working condition according to the same logic and generates an instruction to optimize the non-metallic particle threshold, reducing the misjudgment of dust and hydraulic oil degradation products in the oil. The detection data is fed back to the remote server in real time.

[0029] A remote server enables centralized monitoring of hydraulic presses through a visualization platform. Maintenance personnel can filter and view the operating curves, particle detection results, and threshold adjustment records for each piece of equipment by production line. The platform automatically calculates the particle change trend for each individual piece of equipment. When the number of metal particles on a hydraulic press exceeds the warning value three consecutive times during the stamping process, the system immediately triggers a three-level warning: a platform pop-up notification, a notification sent to the maintenance group, and an audible and visual alarm from the on-site electrical control box. Maintenance personnel can review the historical data of the equipment through the platform to formulate maintenance plans. After maintenance, the system re-executes the detection. If the data returns to normal, the warning is automatically lifted; if it remains abnormal, the warning is escalated and the equipment's production access is locked to prevent the fault from escalating. Incremental training of the model is performed monthly via a remote terminal to ensure the system adapts to changes in operating conditions caused by the aging of the hydraulic presses, extending the system's lifespan.

[0030] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A remote operation and maintenance oil particle counting and classification system, characterized in that, The system includes the following components: Operating data acquisition module: Industrial-grade high-precision sensors are installed in key parts of the equipment's hydraulic system and power output to collect hydraulic pressure and operating load parameters in real time, and upload the parameters to a remote terminal through a remote transmission module; Oil Particle Detection Module: Deployed at key sampling points in the hydraulic system of the equipment, it integrates a laser shading detection component and a magnetic response material identification component. It detects the number, size range, and material properties of particles through light scattering and magnetic response signals, generates raw data, and uploads it synchronously to a remote terminal. Remote transmission module: It is compatible with wired + wireless dual-mode communication, establishes a two-way data channel between the field end and the remote end, ensures transmission security through encryption and verification mechanisms, and follows industrial control standard communication protocols to realize field data uploading and remote command issuance; Machine learning modeling module: Based on historical working conditions and oil particle detection data, a training dataset is built, and a correlation model is trained through machine learning algorithms. The output is a dynamic judgment threshold for particle material and size level that is adapted to the current working conditions. It also has data storage and retrieval functions. Dynamic threshold control module: After receiving real-time operating parameters, it calls the associated model to determine the equipment operating status and generate targeted threshold control instructions. Under high load, it tightens the metal particle detection threshold and under low load, it optimizes the non-metal particle threshold. Field control module: An industrial-grade PLC controller or embedded control module is installed in the equipment's electrical control box. After receiving remote control commands, it parses and decodes them, drives the oil particle detection module to update the threshold and execute detection operations, and feeds back the results to the remote end to form a closed-loop control.

2. The remote operation and maintenance oil particle counting and classification system according to claim 1, characterized in that, The oil particle detection module is deployed at the oil tank outlet or key oil circuit sampling point of the equipment's hydraulic system. It adopts an integrated design, internally comprising a laser shading detection component and a magnetic response material identification component. The laser shading detection component emits a stable wavelength laser beam through the oil sample. Utilizing the changes in light scattering signals generated by particles blocking the laser, a size comprehensive judgment algorithm is used to count the number of particles and determine their preliminary size range. The magnetic response material identification component generates a stable magnetic field through a built-in electromagnetic induction coil. Based on the differences in magnetic response signals generated by ferrous, non-ferrous, and non-metallic particles in the magnetic field, a particle material judgment algorithm is used to calculate the particle material judgment coefficient, distinguishing the material characteristics of the three types of particles. Finally, raw detection data containing particle number, size range, and material properties is generated. The module uses a sealed sampling structure to prevent oil leakage and employs a quick-plug interface for connection to the oil circuit, facilitating installation and maintenance. The raw detection data is transmitted via cable to the field control module and then synchronously uploaded to the remote terminal via a remote transmission module.

3. The remote operation and maintenance oil particle counting and classification system according to claim 2, characterized in that, The oil particle detection module uses a size comprehensive judgment algorithm to count the number of particles and determine the initial size range. The algorithm formula is as follows: ,in, It is the first Precise size calculation value for each particle. It is the first The laser blocking intensity corresponding to each particle It is the reference value of laser transmittance intensity when there are no particles. It is a size correction factor that compensates for the influence of oil refractive index and temperature on the test results.

4. The remote operation and maintenance oil particle counting and classification system according to claim 2, characterized in that, The oil particle detection module uses a particle material determination algorithm to calculate the material determination coefficient of the particles, thereby distinguishing the material characteristics of the three types of particles. The algorithm formula is as follows: ,in, It is the first The material determination coefficient of each particle is used to distinguish between ferrous / non-ferrous metals / non-metals; It is a signal fusion weight that balances the influence of light scattering and magnetic response signals. It is the first The light scattering signal intensity of each particle It is the reference value of the light scattering signal of standard particles. It is the first The amplitude of the magnetic response signal of each particle. It is the reference value for the magnetic response signal of standard iron particles.

5. The remote operation and maintenance oil particle counting and classification system according to claim 1, characterized in that, The remote transmission module adopts a wired + wireless dual-mode adaptive design. The wired communication is based on an industrial Ethernet architecture, and a stable connection between the field end and the remote end is achieved through shielded network cables. The wireless communication uses an industrial-grade wireless module, which supports long-distance data transmission and has anti-electromagnetic interference capabilities. It also integrates data encryption and verification components to encrypt the uploaded operating parameters and raw particle detection data, and adds a unique verification code to the issued control commands. The communication protocol strictly follows industrial control standards to ensure compatibility between different devices and modules, and establishes a two-way data channel between the field end and the remote end. On the one hand, it realizes the real-time uploading of field data to the remote end, and on the other hand, it realizes the accurate issuance of control commands from the remote end to the field control module.

6. The remote operation and maintenance oil particle counting and classification system according to claim 1, characterized in that, The machine learning modeling module constructs a standardized training dataset based on historical operating condition data accumulated by the same type of equipment under different operating scenarios, as well as oil particle detection data under the corresponding operating conditions. The dataset contains samples of the correlation between different operating conditions and particle material and size distribution. The model is trained using an operating condition-particle characteristic correlation model algorithm, and the algorithm parameters are optimized through multiple iterations of training on the dataset. The input of the model is the operating condition parameter vector uploaded in real time from the field, and the output is the dynamic judgment threshold for particle material and size level for the current operating condition. At the same time, the module has a historical data storage function, supporting long-term storage and fast retrieval of training datasets and model running data.

7. The remote operation and maintenance oil particle counting and classification system according to claim 6, characterized in that, The machine learning modeling module uses the working condition-particle characteristic correlation model algorithm for model training, and the algorithm formula is as follows: ,in, It is the first Material type Dynamic determination thresholds for each size level, It is the first Threshold weights for material types It is the first Class 1 working condition parameters and the first Material type Correlation coefficients for each size level It is the first Material type Threshold correction factor for each size level, This is the first under the current working conditions. Material type Historical frequency of particle occurrence at each size grade It is the first under all working conditions Material type Average frequency of occurrence of particles in each size class These are the normalized operating parameters.

8. The remote operation and maintenance oil particle counting and classification system according to claim 1, characterized in that, The dynamic threshold control module receives the current operating condition parameters uploaded by the field terminal in real time, determines the operating status of the equipment through the built-in operating condition recognition program, and then automatically calls the operating condition-particle characteristic association model trained by the machine learning modeling module. It inputs the current operating condition parameters to calculate the appropriate dynamic threshold and uses a threshold dynamic correction algorithm to optimize the initial threshold. For high load operating conditions, it generates control instructions to tighten the detection thresholds of ferrous and non-ferrous metal particles, and lowers the size judgment lower limit of the two types of metal particles. For low-load operation, control instructions are generated to optimize the threshold of non-metallic particles, define the size judgment range of non-metallic particles, and reduce the situation where dust-like non-metallic particles are misjudged as metallic particles. After the instructions are generated, they are encoded in a standardized format and sent to the field control module through the downlink channel of the remote transmission module.

9. A remote oil particle counting and classification system for operation and maintenance according to claim 8, characterized in that, The dynamic threshold control module uses a dynamic threshold correction algorithm to optimize the initial threshold. The algorithm formula is as follows: ,in, It was the first one that was finally issued to the site. Material type Thresholds after adjustment at each size level, It is the threshold control sensitivity coefficient. It is the normalized value of the current operating load of the equipment. It is the load threshold, the dividing point that distinguishes between high and low loads. It is the normalized value of the equipment's maximum load. It is a symbolic function. When the threshold is set to 1, the threshold is tightened. When the threshold is -1, optimize the threshold.

Images