Intelligent operation and maintenance system of crusher

Abstract

The invention relates to the technical field of mechanical intelligent control, in particular to an intelligent operation and maintenance system of a crusher, which is characterized in that a data acquisition layer acquires key parameters during working of the crusher in real time; the data modeling layer establishes a behavior model and a rule model of the crusher according to the key parameters; the fault analysis layer performs state monitoring and coupling analysis according to the behavior model established by the data modeling layer, performs fault diagnosis on the crusher in combination with a rule model, and predicts a maintenance scheme for the fault; and the execution layer adjusts the crusher according to the maintenance scheme obtained by the fault analysis layer. By constructing the high-fidelity digital twin model of the crusher, real-time mapping and fault prediction of the equipment state are realized, and real-time state monitoring, fault early warning, performance optimization and predictive maintenance of the eccentric roller crusher are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of mechanical intelligent control, and particularly relates to an intelligent operation and maintenance system of a crusher. BACKGROUND

[0002] As a high-efficiency and energy-saving crushing device, the eccentric roller crusher is widely used in the mining, cement and metallurgical industries due to its compact structure, strong processing capacity and low energy consumption. However, in actual production, the traditional eccentric roller crusher usually adopts a mechanical adjustment mode, and the adjustment amount depends on the experience of the staff. Therefore, it is difficult to timely and effectively predict the faults of the eccentric roller crusher. For serious faults such as bearing failure and main shaft fracture, maintenance can only be carried out after the faults occur, resulting in a long unexpected downtime. The energy consumption of the crusher under different working conditions is significantly different. The traditional control method cannot adjust the operating parameters in real time according to the changes in ore hardness and feed size, resulting in high energy consumption. The digital twin technology, as a full-dimensional mapping of a physical entity in a virtual space, provides an innovative solution to the above problems. By integrating physical models, sensor data and operation history, the digital twin technology constructs a digital twin body that is updated synchronously with the physical entity, and realizes functions such as state monitoring, fault diagnosis and performance prediction. In the field of mining machinery, the digital twin technology has been applied to crushing equipment. However, there is no complete technical solution for the digital twin intelligent operation and maintenance system of the eccentric roller crusher. SUMMARY

[0003] Therefore, the present application aims to provide an intelligent operation and maintenance system of a crusher. By constructing a high-fidelity digital twin model of the crusher, the real-time mapping and fault prediction of the equipment state are realized, a complete intelligent operation analysis scheme is formed, and the real-time state monitoring, fault warning, performance optimization and predictive maintenance of the eccentric roller crusher are realized, effectively solving the technical problems of response lag, dependence on manual experience and inability to predict hidden faults in the traditional crusher operation and maintenance process.

[0004] To achieve the above-mentioned purpose, the technical scheme of the present application is as follows: An intelligent operation and maintenance system of a crusher, comprising: a data acquisition layer, which acquires key parameters of the crusher in real time; a data modeling layer, which establishes a behavior model and a rule model of the crusher according to the key parameters; a fault analysis layer, which performs state monitoring and coupling analysis according to the behavior model established by the data modeling layer, performs fault diagnosis on the crusher in combination with the rule model, and predicts a maintenance scheme for the fault; and an execution layer, which adjusts the crusher according to the maintenance scheme obtained by the fault analysis layer.

[0005] Further, the data acquisition layer comprises: a three-axis accelerometer for monitoring the vibration of the crusher during operation; a temperature sensor for real-time monitoring of the working temperature of the crusher; an industrial camera for periodically monitoring the discharge of the crusher; an ultrasonic thickness gauge for periodically measuring the thickness of the lower fixed jaw plate of the crusher; and a current sensor for real-time monitoring of the motor current of the crusher.

[0006] Further, in the data modeling layer: a geometric model of the crusher is constructed by three-dimensional modeling software; a physical model of the crusher is simulated based on the analysis of the physical field distribution of the geometric model; the operation rules of the crusher are learned based on the physical model through an LSTM neural network to construct a behavior model of the crusher; and the device operation specifications and maintenance knowledge of the crusher are coded based on the behavior model to obtain a rule model.

[0007] Further, the physical model comprises a force field model, a temperature field model, and a wear prediction model, wherein: the force field model is established by finite element analysis of the geometric model; the temperature field model is established for predicting the temperature rise of the key components of the crusher by simulating different ore hardness and material properties; and the wear prediction model is obtained for correcting the thickness of the liner based on the Archard wear theory and the periodically measured thickness of the crusher.

[0008] Further, the fault analysis layer comprises: a state monitoring module that receives data from the data acquisition layer, statistically analyzes key parameters in combination with the behavior model, evaluates the current state of the crusher, and outputs key parameters that are abnormal; a fault diagnosis module that collects fault signals of abnormal key parameters from the data acquisition layer, denoises the fault signals using wavelet thresholding, identifies the fault signals using a random forest algorithm, and outputs the fault type; and a predictive maintenance module that correlates the fault type output by the fault diagnosis module with multiple possible causes of the fault, finds the root cause from the multiple possible causes, and provides a maintenance solution based on device operation specifications and maintenance knowledge.

[0009] Further, the fault diagnosis module further comprises: frequency domain feature extraction of the vibration signal of the crusher, and deep convolutional neural network analysis of the vibration spectrum features to identify early signs of failure of the crusher.

[0010] Further, the predictive maintenance module further comprises: prediction of the remaining life of the relevant key components based on the survival analysis theory according to the fault type output by the fault diagnosis module.

[0011] Furthermore, key parameters include the ratio of the crusher's main shaft speed, discharge port size, and feeding rate. Correspondingly, the execution layer includes: a frequency converter, connected to the crusher's drive motor, used to adjust the main shaft speed; a drive motor, connected to the crusher's vibrating feeder, used to adjust the feeding rate ratio; and a lower-level machine, connected to the crusher's hydraulic adjustment mechanism, used to adjust the discharge port size.

[0012] Furthermore, it includes a performance optimization layer, which uses a near-end policy optimization algorithm to optimize and control key parameters. Here, the policy network receives state data collected by the data acquisition layer and the set theoretical key parameters of the crusher, and outputs the corresponding adjustment amounts of the key parameters; the value network is responsible for evaluating the current state of the crusher and guides the learning direction of the policy network through a multi-objective reward function that includes the crusher's energy consumption, throughput, and granularity indicators; and a pruning mechanism is used to limit the range of change between the old and new policies.

[0013] Furthermore, it also includes a user interaction layer, which displays key parameters of the crusher in real time during operation, as well as the results of fault diagnosis and corresponding maintenance solutions.

[0014] Compared with the prior art, the present invention can achieve the following beneficial effects: The intelligent operation and maintenance system for crushers described in this invention constructs a high-fidelity digital twin model of the crusher and integrates industrial Internet of Things, big data analysis, machine learning, and automatic control technologies based on the data acquisition layer. This enables real-time status monitoring, fault early warning, performance optimization, and predictive maintenance of the crusher, effectively solving technical problems such as delayed response, reliance on human experience, and inability to predict hidden faults in traditional crusher operation and maintenance. Attached Figure Description

[0015] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 A schematic diagram of the intelligent operation and maintenance system for the crusher described in the embodiments of the present invention; Figure 2 A schematic diagram of the eccentric roller crusher described in the embodiment of the present invention; Figure 2 (a) A cross-sectional view of the eccentric roller crusher described in the embodiment of the present invention; Figure 2 (b) A top view of the eccentric roller crusher described in the embodiment of the present invention; Figure 3This is a schematic diagram showing the relationship between the components and key parameters in the execution layer of the present invention, as described in an embodiment of the invention.

[0016] Explanation of reference numerals in the attached figures: 1. Hydraulic adjustment mechanism; 2. Adjustment block; 3. Upper fixed jaw plate; 4. Lower fixed jaw body; 5. Rolling bearing; 6. Main shaft; 7. Material receiving box; 8. Vibrating feeder; 9. Vibrating motor; 10. Crushing roller body; 11. Bearing seat; 12. Lower fixed jaw body support device. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0019] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, 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, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., 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," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0020] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] likeFigure 1 As shown in the embodiment of the present invention, the intelligent operation and maintenance system for a crusher includes a data acquisition layer, a data modeling layer, a fault analysis layer, and an execution layer. The data acquisition layer collects key parameters of the crusher in real time during operation; the data modeling layer establishes a behavioral model and a rule model for the crusher based on these key parameters; the fault analysis layer performs state monitoring and coupled analysis based on the behavioral model established by the data modeling layer, and combines this with the rule model to diagnose faults in the crusher and predict maintenance plans based on the faults. The execution layer adjusts the crusher according to the maintenance plan obtained from the fault analysis layer.

[0023] In this embodiment of the invention, intelligent operation and maintenance are implemented for the PGP2530 eccentric roller crusher of Northern Heavy Industries Group Co., Ltd. The structure of the eccentric roller crusher is as follows: Figure 2 As shown, Figure 2 (a) shows a cross-sectional view of an eccentric roller crusher. Figure 2 (b) Top views of the eccentric roll crusher are shown. The eccentric roll crusher includes a hydraulic adjustment mechanism 1, an adjustment block 2, an upper fixed jaw plate 3, a lower fixed jaw body 4, a rolling bearing 5, a main shaft 6, a receiving box 7, a vibrating feeder 8, a vibrating motor 9, a crushing roll body 10, a bearing seat 11, and a lower fixed jaw body support device 12. The hydraulic adjustment mechanism 1 is connected to the lower fixed jaw body support device 12 and is used to adjust the discharge port size. The adjustment block 2, upper fixed jaw plate 3, lower fixed jaw body 4, rolling bearing 5, main shaft 6, receiving box 7, vibrating feeder 8, crushing roll body 10, bearing seat 11, and lower fixed jaw body support device 12 together constitute the main structure of the crusher. The upper fixed jaw plate 3 and the lower fixed jaw body 4 form the crushing chamber of the eccentric roll crusher, and the lower fixed jaw body 4 is positioned and adjusted by the adjustment block 2 and the lower fixed jaw body support device 12. The crushing roller 10 is mounted on the main shaft 6, which is in turn mounted in the bearing housing 11 via rolling bearings 5, forming a rotatable eccentric crushing mechanism. The bearing housing 11 supports the rotation of the main shaft 6 and the crushing roller 10. Vibration motors 9 are located on both sides of the vibrating feeder 8, providing vibration to the feeder 8, which is positioned above the crusher body. Figure 2 (a) shows the data acquisition layer, which is used to feed materials into the crushing chamber for crushing. The crushed material is discharged through the receiving box 7. In some embodiments, the data acquisition layer includes a triaxial accelerometer, a temperature sensor, an industrial camera, an ultrasonic thickness gauge, and a current sensor. The triaxial accelerometer is used to monitor the vibration of the crusher during operation, the temperature sensor is used to monitor the operating temperature of the crusher in real time, the industrial camera is used to periodically capture and monitor the discharge of the crusher, the ultrasonic thickness gauge is used to periodically measure the thickness of the lower jaw plate in the crusher, and the current sensor is used to monitor the operating current of the crusher's motor in real time.

[0024] In this embodiment of the invention, a triaxial accelerometer is installed on the bearing housing 11 to collect vibration acceleration signals of the crusher in real time during operation, reflecting the impact of changes in spindle speed, the force state of materials in the crushing chamber, and the feed rate on the crushing load. Changes in spindle speed, discharge port size, and feed rate will all cause corresponding changes in the amplitude and spectral characteristics of the vibration signal. Temperature sensors are installed in key locations, such as the motor housing, bearing housing 11, and crushing chamber, to monitor the operating temperature of the equipment. The operating temperature is related to the spindle speed and crushing load level, and can reflect the changes in thermal load caused by changes in discharge port size and feed rate. An industrial camera is installed below the discharge port, equipped with an illumination system and a filter, to collect images of the discharged ore. By analyzing the collected images, the discharge particle size distribution parameters can be obtained. The discharge particle size distribution is related to the discharge port size and spindle speed. An ultrasonic thickness gauge is used to periodically measure the thickness of the middle and lower jaw plates, and automatically evaluates the wear state of the liner plates by combining digital image correlation technology. The wear state is related to long-term operating conditions and the discharge port adjustment frequency. A Hall effect current sensor is installed at the output of the frequency converter to monitor the motor operating current in real time. It reads the voltage signal via RS-485 communication to calculate the motor power, which is directly related to the feeding rate and crushing resistance. In this embodiment, the data acquisition layer also includes a piezoresistive pressure sensor, installed between the adjusting block 2 and the lower jaw support device 12, and at the bottom of the receiving box 7. This sensor detects the crushing force and material weight; the crushing force is used to characterize the adjustment status of the discharge port size. By comprehensively analyzing the data from these multi-source sensors, a correspondence between the crusher's main shaft speed, discharge port size, feeding rate, and operating parameters can be established, providing a basis for adjusting and optimizing the crusher's operating parameters.

[0025] In some embodiments, in the data modeling layer: a geometric model of the crusher is constructed using 3D modeling software; the physical field distribution of the crusher is analyzed based on the geometric model to simulate the physical model of the crusher; based on the physical model, the operating rules of the crusher are learned through an LSTM neural network to construct a behavioral model of the crusher; based on the behavioral model, the equipment operation specifications and maintenance knowledge of the crusher are encoded to obtain a rule model.

[0026] In this embodiment of the invention, for the geometric model: based on CAD design drawings, a precise geometric model of the eccentric roller crusher is constructed using 3D modeling software such as SolidWorks, including all major components and assembly relationships. A refined geometric model is constructed for high-wear components to accurately reflect the tooth structure in the crusher. In constructing the behavioral model, time-series data of key data such as vibration, temperature, and current during crusher operation are used as input, and the crusher's state score is used as output. An LSTM neural network is used to learn the crusher's operating rules. In constructing the rule model, this embodiment of the invention integrates and encodes the crusher's equipment operation specifications, maintenance knowledge, and empirical rules from historical maintenance records to obtain a rule model. The geometric model, physical model, behavioral model, and rule model constitute a digital twin of the crusher.

[0027] In some embodiments, the physical model includes a force field model, a temperature field model, and a wear prediction model, wherein: a force field model is established by performing finite element analysis on the geometric model; a temperature field model is established to predict the temperature rise of key components of the crusher by simulating different ore hardness and material properties; and a wear prediction model is obtained to correct the liner thickness based on Arcard wear theory and combined with periodically measured liner thickness of the crusher.

[0028] In this embodiment of the invention, a temperature field model is established by simulating different ore hardness and material properties, while taking into account the influence of ambient temperature and heat dissipation conditions. The temperature field model is then used to predict the temperature rise of key components such as bearings and motors in the eccentric roller crusher.

[0029] In some embodiments, the fault analysis layer includes a condition monitoring module, a fault diagnosis module, and a predictive maintenance module. The condition monitoring module receives data from the data acquisition layer, performs statistical analysis on key parameters using a behavioral model, assesses the current state of the crusher, and outputs the key parameters indicating anomalies. The fault diagnosis module collects fault signals of the key parameters indicating anomalies from the data acquisition layer, performs wavelet threshold denoising on the fault signals, uses a random forest algorithm to identify faults in the denoised fault signals, and outputs the fault type. In some embodiments, the fault diagnosis module further includes: extracting frequency domain features from the crusher's vibration signals, using a deep convolutional neural network to analyze the vibration spectrum features, and identifying early fault symptoms of the crusher. The predictive maintenance module calls a rule model to associate the fault type output by the fault diagnosis module with multiple possible causes of the fault, finds the root cause from the multiple possible causes, and then provides a maintenance plan based on equipment operation specifications and maintenance knowledge. In some embodiments, the predictive maintenance module further includes: predicting the remaining lifespan of relevant key components based on survival analysis theory according to the fault type output by the fault diagnosis module.

[0030] In this embodiment of the invention, the process by which the status monitoring module assesses the current state of the crusher and outputs key parameters indicating anomalies specifically includes: performing statistical analysis on the key parameters, calculating the mean, peak value, kurtosis, and other indicators of various key parameters, and assessing the current state of the crusher; if a key parameter exceeds a preset normal range, the key parameter exceeding the normal range is recorded as an abnormal parameter, and the event that caused the abnormal parameter is recorded. The fault diagnosis module collects fault signals corresponding to the abnormal parameters from the data acquisition layer, performs wavelet threshold denoising on the fault signals, uses a random forest algorithm to identify faults in the denoised fault signals, and outputs the fault type; simultaneously, it performs wavelet threshold denoising and time-domain and frequency-domain feature extraction on the vibration signal, uses a deep convolutional neural network to analyze the vibration spectrum features, and identifies early fault signs. For example, when the status monitoring module detects that the peak value of the current value in a key parameter exceeds the preset normal range at a certain moment, it marks the current value as an abnormal parameter and records the event that occurred when the crusher's current value exceeded the preset normal range, recording this event as an abnormal event. The fault diagnosis module collects the fault current signal corresponding to the current value from the data acquisition layer, performs wavelet threshold denoising on the fault current signal, uses the random forest algorithm to identify the fault after denoising the fault current signal, and outputs the specific fault type that caused the abnormal current value. At the same time, it performs wavelet threshold denoising and time-domain and frequency-domain feature extraction on the vibration signal, and uses a deep convolutional neural network to analyze the vibration spectrum characteristics to identify early fault signs. The predictive maintenance module calls the rule model to associate the fault type output by the fault diagnosis module with multiple possible causes of the fault, finds the root cause from multiple possible causes, and then predicts the remaining life of relevant key components based on survival analysis theory. To balance production and maintenance needs, it integrates the root cause, remaining life, crusher operating time, and operating condition severity, calls the rule model to generate a maintenance plan, and generates a maintenance work order, including a list of parts to be replaced, tool preparation, and operation procedure instructions.

[0031] In this embodiment of the invention, the YOLOv5s model is also used to identify the discharge images captured by the industrial camera, output the material particle size distribution curve, and then input the material particle size distribution curve into a deep convolutional neural network to identify the corresponding fault.

[0032] In some embodiments, key parameters include the ratio of spindle speed, discharge port size, and feeding rate. Correspondingly, the execution layer includes a frequency converter, a drive motor, and a lower-level machine. In this embodiment, a lower-level machine PLC1200 is specifically used. The frequency converter is connected to the crusher's drive motor to adjust the spindle speed; the drive motor is connected to the crusher's vibrating feeder to adjust the feeding rate ratio; and the lower-level machine is connected to the crusher's hydraulic adjustment mechanism to adjust the discharge port size. The relationship between each component in the execution layer and the key parameters is as follows: Figure 3 As shown.

[0033] In some embodiments, the performance optimization layer employs a reinforcement learning algorithm for optimizing key parameters, where the reward function considers the balance between the crusher's energy consumption, throughput, and granularity indicators. Specifically, the Proximal Policy Optimization (PPO) algorithm is used. This is a DRL algorithm based on an Actor-Critic (policy-value) architecture, highly suitable for industrial control scenarios with high stability and reliability requirements due to its stable training, simple implementation, and ability to effectively handle continuous action spaces. In the PPO algorithm: the policy network receives state data collected by the data acquisition layer, along with the set theoretical key parameters of the crusher, and outputs the corresponding adjustment amounts for the key parameters; the value network evaluates the current state of the crusher and guides the learning direction of the policy network through a multi-objective reward function that includes the crusher's energy consumption, throughput, and granularity indicators; a pruning mechanism is used to limit the range of change between the old and new policies, preventing excessively large policy update steps that could lead to collapse, thus ensuring smooth convergence of the learning process.

[0034] In this embodiment of the invention, the strategy network receives data from the data acquisition layer, including motor current, crushing chamber pressure, bearing vibration, key component temperature, and real-time discharge particle size, as well as the set theoretical rotational speed, theoretical discharge port size, and theoretical feed rate. It then outputs corresponding adjustments to the key parameters, including adjustments to the spindle speed, discharge port size, and feed rate. The value network is responsible for evaluating the current state of the crusher and guides the learning direction of the strategy network through the following multi-objective reward function: ; Where Rt represents multi-objective reward, RQ represents processing capacity reward, RE represents energy consumption reward, Rd80 represents granularity compliance reward, PI represents overload penalty, and PV represents vibration penalty. ω represents the operational stability penalty, and ω represents the weight coefficient of each item.

[0035] The performance optimization module implements intelligent control through the following steps: First, the sensor network collects data in real time. The data is then transmitted to the performance optimization module. The fixed PPO network receives the data. Output the optimal parameter adjustment amount The digital twin model then performs safety verification on the parameters recommended by the PPO to ensure that it will not lead to overload or excessive vibration. After successful verification, control commands are sent to the actuators via the PLC and frequency converter to adjust the discharge port size, spindle speed, and feeding rate in real time. Actual operating results and reward signal feedback are used to dynamically calibrate the parameters of the digital twin model and control algorithm.

[0036] In other embodiments, the performance optimization layer can also obtain the optimal combination of key parameters using a genetic algorithm. Specifically, an objective function is constructed based on the crusher's performance indicators; based on the objective function, a genetic algorithm is used to search for the optimal combination of key parameters; and the crusher is adjusted according to the optimal combination of key parameters.

[0037] In this embodiment of the invention, the objective function includes minimizing the energy consumption of the crusher, maximizing the throughput, and maximizing the particle size qualification rate, as detailed below: ; Where F represents the objective function, Q represents the crusher's processing capacity, L represents the crusher's service life, E represents the crusher's specific energy consumption, and d 80 This represents the particle size corresponding to a sample when the cumulative particle size distribution percentage reaches 80%. Based on the objective function, a genetic algorithm is used to search for the optimal combination of key parameters. The optimal combination of key parameters is applied to the geometric model, and a simulation experiment is conducted on the geometric model to ensure its safety. Then, the crusher is adjusted according to the optimal combination of key parameters.

[0038] In some embodiments, the intelligent operation and maintenance system for the crusher also includes a user interaction layer, which is used to display key parameters of the crusher in real time during operation, and to display the results of fault diagnosis and the corresponding maintenance plan.

[0039] In this embodiment of the invention, the user interaction layer macroscopically displays the overall operating status and key performance indicators of the crusher from a global monitoring view. Specifically, the crusher's geometric model can be rotated, scaled, and sectioned 720 degrees for viewing. Sensor measuring points are displayed as dynamic icons on the geometric model, with color changes representing real-time status and detailed data curves available for viewing. Real-time vibration, temperature, and pressure data streams of key components are displayed in the form of dashboards, line graphs, etc. The health status of each component is intuitively displayed through health rating bars or radar charts. When a fault is predicted, the warning component is marked on the geometric model by highlighting or flashing, and a warning information window pops up, detailing the fault type, possible causes, and suggested measures. The fault analysis view provides in-depth data analysis and fault tracing tools. Any time point can be selected on the timeline to replay the equipment's operating status and data at that time, helping to analyze the fault evolution process. Professional analysis charts such as vibration spectrum diagrams, time-domain waveform diagrams, and temperature trend diagrams are provided, and multi-chart linkage analysis is supported. Equipment components, fault modes, causes, and maintenance plans are visualized and associated in the form of a knowledge graph.

[0040] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0041] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An intelligent operation and maintenance system of a crusher, characterized in that, The application relates to a crusher fault diagnosis and predictive maintenance system. The system includes: a data acquisition layer that collects key parameters in real time when the crusher is working; a data modeling layer that establishes a behavior model and a rule model of the crusher according to the key parameters; a fault analysis layer that performs state monitoring and coupling analysis according to the behavior model established by the data modeling layer, performs fault diagnosis on the crusher in combination with the rule model, and predicts a maintenance scheme for the fault; 2.The intelligent operation and maintenance system of the crusher of claim 1, characterized in that, and an execution layer that adjusts the crusher according to the maintenance scheme obtained by the fault analysis layer. The data acquisition layer includes: a three-axis accelerometer for monitoring the vibration of the crusher when it is working; a temperature sensor for monitoring the working temperature of the crusher in real time; an industrial camera for periodically shooting and monitoring the discharge of the crusher; an ultrasonic thickness gauge for periodically measuring the thickness of the lower fixed jaw plate of the crusher; 3.The intelligent operation and maintenance system of the crusher of claim 1, characterized in that, and a current sensor for monitoring the motor working current of the crusher in real time. In the data modeling layer: a three-dimensional modeling software is used to construct a geometric model of the crusher; a physical model of the crusher is simulated by analyzing the physical field distribution of the crusher based on the geometric model; the operation law of the crusher is learned through an LSTM neural network based on the physical model, and a behavior model of the crusher is constructed; 4. The intelligent operation and maintenance system of the crusher according to claim 3, characterized in that, and a rule model is obtained by encoding the equipment operation specifications and maintenance knowledge of the crusher based on the behavior model. The physical model includes a force field model, a temperature field model and a wear prediction model, wherein: the force field model is established by performing finite element analysis on the geometric model; the temperature field model is established by simulating different ore hardness and material properties to predict the temperature rise of the key components of the crusher; 5. The intelligent operation and maintenance system of the crusher according to claim 4, characterized in that, and the wear prediction model is obtained by combining the Archard wear theory and the periodically measured liner thickness to correct the liner thickness. The fault analysis layer includes: a state monitoring module that receives data from the data acquisition layer, statistically analyzes the key parameters in combination with the behavior model, evaluates the current state of the crusher, and outputs the key parameters that are abnormal; a fault diagnosis module that collects fault signals of the key parameters that are abnormal from the data acquisition layer, performs wavelet threshold denoising on the fault signals, uses a random forest algorithm to identify the fault signals after denoising, and outputs the fault type; 6.The intelligent operation and maintenance system of the crusher of claim 5, characterized in that, and a predictive maintenance module that correlates the fault type output by the fault diagnosis module with multiple possible causes of the fault, finds the root cause from the multiple possible causes, and provides a maintenance scheme according to the equipment operation specifications and maintenance knowledge. The fault diagnosis module further includes:

7. The intelligent operation and maintenance system of the crusher according to claim 5, characterized in that, characteristic extraction in the frequency domain of the vibration signal of the crusher, and analysis of the vibration frequency spectrum characteristics by using a deep convolutional neural network to identify early fault signs of the crusher. The predictive maintenance module further includes: 8.The intelligent operation and maintenance system of the crusher of claim 1, characterized in that, prediction of the remaining life of the related key components based on the survival analysis theory according to the fault type output by the fault diagnosis module. The key parameters include the main shaft speed, the discharge port size and the feeding rate ratio of the crusher, and the corresponding execution layer includes: a frequency converter connected with the driving motor of the crusher for adjusting the main shaft speed; a driving motor connected with the vibrating feeder of the crusher for adjusting the feeding rate ratio; and a lower computer connected with the hydraulic adjustment mechanism of the crusher for adjusting the discharge port size. 9.The intelligent operation and maintenance system of the crusher of claim 1, characterized in that, A performance optimization layer is also included, which uses a proximal policy optimization algorithm to optimize the control of key parameters, wherein: The policy network receives the state data collected by the data collection layer and the theoretical key parameters of the crusher set, and outputs the corresponding adjustment amount of the key parameters; The value network is responsible for evaluating the current state of the crusher and guiding the learning direction of the policy network through a multi-objective reward function including the energy consumption, processing capacity and granularity indicators of the crusher; A clipping mechanism is used to limit the change amplitude between the new and old policies. 10.The intelligent operation and maintenance system of the crusher of claim 1, characterized in that, Also includes: A user interaction layer is used to display the key parameters of the crusher in real time when it is working, and to display the results of fault diagnosis and the corresponding maintenance scheme.

Images