CN115081324A - Performance prediction optimization and fault diagnosis system and method for underground engineering equipment - Google Patents

Abstract

The system comprises a physical object, a data acquisition and analysis module, a data storage module and a digital twin module, wherein the physical object, the data acquisition and analysis module, the data storage module and the digital twin module are connected in series to form a closed loop, and real-time operation data acquisition, health state analysis and fault early warning and diagnosis are carried out on the underground engineering equipment. The invention has the advantages of being capable of being used for the construction site of underground engineering equipment, providing real-time key performance prediction and optimization for the underground engineering equipment, monitoring the operation state and carrying out fault diagnosis, improving the state and performance of the underground engineering equipment, adjusting and optimizing the construction scheme in time, improving the construction efficiency and the service life of equipment, pre-tightening potential faults and reducing the construction risk.

Description

Performance prediction optimization and fault diagnosis system and method for underground engineering equipment

Technical Field

The invention relates to the technical field of underground engineering equipment, in particular to a system and a method for performance prediction optimization and fault diagnosis of underground engineering equipment.

Background

With the increasing investment of large-scale infrastructure, tunnel and underground engineering construction has come to an unprecedented rapid development, and the demand for underground engineering equipment is continuously increased. Compared with the traditional mode which mainly adopts manpower, the underground engineering equipment represented by shield method equipment (a shield machine), drilling and blasting method equipment (a rock drilling trolley), coal mine digging and anchoring equipment and the like has the remarkable advantages of high construction efficiency, higher construction quality and safety and the like. Meanwhile, the development trend of less humanization and no humanization in implementation brought by the rising of the labor cost also puts higher requirements on the intelligent level of the equipment.

The geological environment for underground engineering equipment construction is extremely severe, the equipment often faces complex geological risks such as collapse and rock burst, the equipment structure is huge, the maintenance is difficult, and once a fault occurs, normal construction is affected, and great loss is often caused. Therefore, key performance prediction and fault diagnosis are carried out on underground engineering equipment in the underground construction process, equipment construction parameters are optimized in time on the basis, potential faults are subjected to early warning treatment, major construction quality accidents can be effectively avoided, and the method has important significance for safe and efficient operation of the equipment and the like.

The prediction technology and the diagnosis technology of underground engineering equipment in the prior art are based on a traditional mode, how to ensure the real-time performance and the accuracy of calculation is not described, and a result feedback closed loop is not formed. The technical concept of digital twinning is not introduced in the prior art, the characteristics of difficult data collection/transmission, few digital samples and the like of underground engineering equipment are difficult to adapt, and the engineering application value is not high.

With the continuous expansion of the digitalization and intelligent transformation and upgrading of the equipment industry, the innovative concepts such as the digital twin technology are gradually applied to various industries, and the technical schemes of virtual-real synchronization, virtual-real interactive driving and the like formed by physical equipment operation and virtual equipment are widely mentioned. However, the application of the digital twin technology in various industries is basically in an exploration stage, the application is far from mature, and the digital twin technology mainly faces several problems: the prior art fails to solve the contradiction problem of the digital twin application on the model precision and the calculation efficiency, and real-time calculation and real-time feedback cannot be carried out, so that the digital twin application effect is greatly reduced; network delay is often faced in the development of a digital twin model in practical application, and the on-line cloud platform or remote deployment is not reliable, so that the landing application of the digital twin in underground engineering equipment is greatly limited.

In view of the foregoing, there is a need for a system and method for performance prediction optimization and fault diagnosis of underground engineering equipment to solve the problems in the prior art.

Disclosure of Invention

The invention aims to provide a system and a method for predicting and optimizing the performance of underground engineering equipment and diagnosing faults, and the specific technical scheme is as follows:

a performance prediction optimization and fault diagnosis system of underground engineering equipment comprises a physical object, a data acquisition and analysis module, a data storage module and a digital twin module;

the data acquisition and analysis module is used for acquiring original operation data of the physical object in real time and processing the original operation data, and the data acquisition and analysis module is installed at the edge end of the physical object;

the data storage module is used for storing original operation data and processed operation data;

the digital twin module is used for calculating the data in the data storage module, establishing a digital twin model of the physical object, and correcting the digital twin model according to the calculation result precision and the calculation speed;

the data acquisition and analysis module deploys the modified digital twin model, and inputs real-time operation data into the digital twin model to perform performance prediction optimization and fault diagnosis on the physical object.

Specifically, the data acquisition and analysis module comprises a sensor unit, a data acquisition and storage unit and a data processing and analysis unit;

wherein the sensor unit is used for acquiring raw operating data of a physical object;

the data acquisition and storage unit is used for receiving the operation data and transmitting the operation data to the data processing and analysis unit and the data storage module;

the data processing and analyzing unit cleans and classifies the original operation data and transmits the processed operation data to the data storage module.

Specifically, the sensor unit includes a sensor for acquiring operation data of the physical object, and the sensor is disposed on the physical object.

Specifically, the data storage module comprises an online data platform and an offline local database, the online data platform transmits data through a network, and the offline database copies data through a mobile storage medium and transmits the data.

Specifically, the digital twinning module comprises a data interface unit, a data processing/storing unit, a digital twinning analyzing unit, a result visualization unit and a model modification unit;

the data interface unit is used for receiving original operation data and processed operation data of the data storage module;

the data processing/storing unit carries out secondary processing on the data, eliminates abnormal and unnecessary data and stores the data after secondary processing;

the digital twin analysis unit establishes a performance prediction optimization model and a fault diagnosis model, and inputs the secondarily processed data into the performance prediction optimization model and the fault diagnosis model to obtain a calculation result;

the result visualization unit is used for receiving the calculation result of the digital twin analysis unit and performing visualization processing on the calculation result to form a digital twin model;

the model correction unit corrects or does not correct the digital twin model based on the accuracy of the calculation result and the calculation speed.

Specifically, the calculation result comprises a prediction result and a diagnosis result;

the prediction result is obtained by inputting data subjected to secondary processing into a performance prediction optimization model, and the prediction result comprises optimization parameters;

the diagnosis result is data input fault diagnosis module for secondary processing, and the diagnosis result comprises health state and fault early warning information of the physical object.

The method is characterized in that the physical object is underground engineering equipment and comprises a mechanical unit, a hydraulic unit, an electrical control unit and a display unit;

the electric control unit receives a control command and operates the mechanical unit and the hydraulic unit according to the control command;

the display unit presents the calculation result in the form of data, a graph or a curve.

The invention also provides a performance prediction optimization and fault diagnosis method of underground engineering equipment, which is applied to the performance prediction optimization and fault diagnosis system and comprises the following specific steps:

step S1: acquiring operation data of a physical object, specifically, acquiring original operation data of a physical entity in real time by a data acquisition and analysis module, preprocessing the original operation data, and transmitting the original operation data and the preprocessed operation data to a data storage module;

step S2: establishing a performance prediction optimization model and a fault diagnosis model, specifically, establishing a reduced performance prediction optimization model and a reduced fault diagnosis model, wherein the performance prediction optimization model and the reduced fault diagnosis model are both simulation models and are stored in a digital twin module;

step S3: establishing a digital twin model of the underground engineering equipment, specifically, inputting data in the storage module in the step S1 into the performance prediction optimization model and the fault diagnosis model in the step S2 for calculation to obtain a calculation result, wherein the calculation result comprises optimization parameters, health states and fault early warning information of the underground engineering equipment, and feeds back model calculation time;

step S4: checking whether the digital twin model is qualified, specifically, calculating the precision of the calculation result in the step S3, if the precision exceeds a preset precision threshold and the model calculation time in the step S3 is less than the underground engineering equipment fault early warning handling time, the digital twin model meets the engineering requirement, otherwise, returning to the step S2 to correct the digital twin model;

step S5: deploying a digital twin model, specifically deploying the digital twin model meeting the engineering requirements in step S4 into a data acquisition and analysis module at an edge end, performing real-time simulation calculation according to real-time acquired operating data, performing performance prediction optimization and fault diagnosis on a physical entity, and acquiring optimization parameters, health status and fault early warning information of the physical entity;

step S6: and (4) feeding back the optimization parameters and the fault early warning information, specifically, displaying the optimization parameters, the health state and the fault early warning information in the step S5 in an engineering field, sending a control instruction to the underground engineering equipment physical entity by the data acquisition and analysis module based on the optimization parameters and the fault early warning information, and receiving the control instruction by the underground engineering equipment physical entity to optimize the construction parameters.

Specifically, in step S2, the performance prediction optimization model and the fault diagnosis model are established in the following specific manner:

establishing a simulation model based on the attribute of the physical entity, and applying data fitting, deep learning or proxy model to the simulation model to perform order reduction processing to obtain a performance prediction optimization model and a fault diagnosis model;

or, establishing a performance prediction optimization model and a fault diagnosis model by applying a theoretical/empirical formula;

or, by collecting historical data samples of underground engineering equipment, performing correlation analysis on the historical data samples by applying a machine learning algorithm to establish a performance prediction optimization model and a fault diagnosis model.

The technical scheme of the invention has the following beneficial effects:

the performance prediction optimization and fault diagnosis system and method can be used for the engineering site of underground engineering equipment, provide key performance prediction and optimization for the underground engineering equipment, monitor the operation state and carry out fault diagnosis, improve the state and performance of the underground engineering equipment, adjust and optimize the construction scheme in time, improve the construction efficiency and prolong the service life of equipment, pre-tighten potential faults and reduce the construction risk.

In addition to the objects, features and advantages described above, other objects, features and advantages of the present invention are also provided. The present invention will be described in further detail below with reference to the drawings.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:

FIG. 1 is a system block diagram of a performance prediction optimization and fault diagnosis system;

FIG. 2 is a flow chart diagram of a performance prediction optimization and fault diagnosis method;

FIG. 3 is a flow chart of a method for performance prediction optimization and fault diagnosis of a shield tunneling machine;

FIG. 4 is a flow chart of a method of performance prediction optimization and fault diagnosis of the drill jumbo;

FIG. 5 is a flow chart of a performance prediction optimization and fault diagnosis method for the all-in-one machine.

Detailed Description

Embodiments of the invention will be described in detail below with reference to the drawings, but the invention can be implemented in many different ways, which are defined and covered by the claims.

Example 1:

referring to fig. 1, the embodiment discloses a performance prediction optimization and fault diagnosis system for underground engineering equipment, which includes a physical object, a data acquisition and analysis module, a data storage module and a digital twin module. The physical object is underground engineering equipment, the data acquisition and analysis module can acquire original operation data of the underground engineering equipment in real time and analyze and process the operation data, the data storage module can store the original operation data and the operation data processed by the data acquisition and analysis module, the digital twin module can calculate the data in the data storage module and establish a digital twin model of the physical object, and the digital twin model is corrected according to the calculation result precision and the calculation speed.

Specifically, the preferred data acquisition and analysis module of this embodiment includes a sensor unit, a data acquisition and storage unit, and a data processing and analysis unit; the sensor unit is connected with the physical object, and the data acquisition and storage unit and the data processing and analysis unit are connected with the data storage module;

specifically, the sensor unit is used for collecting raw operation data of a physical object, and the preferred sensor unit of this embodiment includes a vibration sensor, a pressure sensor and a current sensor, and in addition, other sensors may be provided based on the attribute of the physical entity or performance prediction requirements.

Specifically, the data acquisition and storage unit transmits original operation data to the data processing and analysis unit and the data storage module;

specifically, the data processing and analyzing unit processes the operating data and transmits the processed operating data to the data storage module, and the data processing and analyzing unit mainly cleans and classifies part of data which has high sampling frequency and is difficult to meet real-time transmission so as to reduce the data transmission amount and improve the operating efficiency of the system.

Specifically, the preferred data storage module of this embodiment includes an online data platform and an offline local database, where the online data platform transmits data through a network, and the offline database copies and transmits data through a mobile storage medium, where the preferred network transmission in this embodiment is 5G signal transmission, and the mobile storage medium may be one or more of a usb disk, a removable hard disk, a floppy disk, an optical disk, and a memory card.

Specifically, the digital twin module preferred in this embodiment includes a data interface unit, a data processing/storing unit, a digital twin analyzing unit, a result visualizing unit, and a model modifying unit;

the data interface unit is used for receiving the original operation data and the processed operation data of the data storage module;

the data processing/storing unit is used for carrying out secondary processing on the original operation data and the processed operation data, eliminating abnormal and unnecessary data and storing the data;

the digital twin analysis unit establishes a performance prediction optimization model and a fault diagnosis model, and inputs the secondarily processed data into the performance prediction optimization model and the fault diagnosis model to obtain a calculation result;

the result visualization unit is used for receiving the calculation result of the digital twin analysis unit and performing visualization processing on the calculation result to form a digital twin model;

the model correction unit corrects or does not correct the digital twin model based on the accuracy of the calculation result and the calculation speed.

Further, the calculated results include predicted results and diagnostic results;

the prediction result is obtained by inputting data subjected to secondary processing into a performance prediction optimization model, and the prediction result comprises optimization parameters;

the diagnosis result is data input fault diagnosis module for secondary processing, and the diagnosis result comprises health state and fault early warning information of the physical object.

Specifically, the physical object is underground engineering equipment which can be a shield machine, a rock drilling jumbo or an all-in-one machine for digging and anchoring, and the underground engineering equipment comprises a mechanical unit, a hydraulic unit, an electrical control unit and a display unit;

the electrical control unit receives a control instruction and operates the mechanical unit and the hydraulic unit according to the control instruction, so that the key performance optimization of underground engineering equipment is realized;

the display unit preferably comprises a display, the display presents the analysis result in the form of data, a chart or a curve, workers on a construction site can visually see the real-time condition of underground engineering equipment, the construction scheme can be adjusted and optimized in time, the construction efficiency is improved, the service life of the equipment is prolonged, the workers can also directly see equipment faults from the display, and the potential fault elimination of the underground engineering equipment on the construction site and the construction risk reduction are facilitated.

The performance prediction optimization and fault diagnosis system for underground engineering equipment disclosed by the embodiment drives the digital twin model to carry out checking calculation based on-line or off-line data, corrects the model, carries out performance prediction optimization and fault diagnosis in real time after the data processing and analyzing module on the construction site checks the digital twin model meeting the requirements, feeds back or controls a physical object, and the digital twin module, the data storage module, the data acquisition and analysis module and the physical object are connected in series to form a closed loop.

In addition, the embodiment also discloses a performance prediction optimization and fault diagnosis method based on the performance prediction optimization and fault diagnosis system, as shown in fig. 2, the specific steps are as follows:

step S1: acquiring operation data of a physical object, specifically, acquiring original operation data of a physical entity in real time by a data acquisition and analysis module, preprocessing the original operation data, and transmitting the original operation data and the preprocessed operation data to a data storage module;

step S2: establishing a performance prediction optimization model and a fault diagnosis model, specifically, establishing a performance prediction optimization model and a fault diagnosis model after reduction based on the attribute of a physical object, wherein the performance prediction optimization model and the fault diagnosis model are both simulation models and are stored in a digital twin module;

it should be noted that, the specific way of establishing the performance prediction optimization model and the fault diagnosis model is as follows:

establishing a simulation model based on the attribute of the physical entity, and applying data fitting, deep learning or proxy model to the simulation model to perform order reduction processing to obtain a performance prediction optimization model and a fault diagnosis model;

or, establishing a performance prediction optimization model and a fault diagnosis model by applying a theoretical/empirical formula;

or, by collecting historical data samples of underground engineering equipment, performing correlation analysis on the historical data samples by applying a machine learning algorithm to establish a performance prediction optimization model and a fault diagnosis model.

Step S3: establishing a digital twin model of the underground engineering equipment, specifically, inputting data in the storage module in the step S1 into the performance prediction optimization model and the fault diagnosis model in the step S2 for calculation to obtain a calculation result, wherein the calculation result comprises a key performance prediction result, an optimization parameter, a health state and fault early warning information of the underground engineering equipment, and feeds back model calculation time;

step S4: checking whether the digital twin model is qualified, specifically, calculating the precision of the calculation result in the step S3, if the precision exceeds a preset precision threshold and the model calculation time in the step S3 is less than the underground engineering equipment fault early warning handling time, enabling the digital twin model to meet engineering requirements, and otherwise, returning to the step S2 to modify the digital twin model by using a model modification unit;

step S5: deploying a digital twin model, specifically deploying the digital twin model meeting the engineering requirements in step S4 into an edge-end data processing and analyzing unit, performing real-time simulation calculation according to real-time acquired operating data, performing performance prediction optimization and fault diagnosis on a physical entity, and acquiring optimization parameters and fault early warning information;

step S6: and feeding back the optimized parameters and the fault early warning information, specifically, displaying the optimized parameters and the fault early warning information through a display unit, sending a control instruction to underground engineering equipment by a data processing and analyzing unit based on the optimized parameters and the fault early warning information, and receiving the control instruction by the underground engineering equipment to optimize the construction parameters.

The performance prediction optimization and fault diagnosis method for underground engineering equipment disclosed by the embodiment is combined with a conventional modeling method, a digital twin model is constructed by utilizing technologies such as model order reduction and the like, the model calculation efficiency is improved, and then the model is checked and corrected by utilizing a digital twin system, so that the accuracy and the real-time performance of the model are ensured to meet the engineering application requirements.

Example 2:

as shown in fig. 3, the present embodiment discloses a digital twin performance prediction optimization and fault diagnosis based method for a shield machine, which is based on the performance prediction optimization and fault diagnosis system and method described in embodiment 1, and the specific steps are as follows:

(1) the method comprises the following steps of collecting operation data of the shield machine, constructing a shield machine performance prediction optimization model and a shield machine fault diagnosis model, and specifically comprising the following steps:

(1.1) acquiring original operation data of the shield tunneling machine, wherein the original operation data of the shield tunneling machine comprises thrust, torque, oil cylinder pressure, speed, cutter head rotating speed, main bearing/speed reducer acceleration, shield body stress, soil pressure, driving motor current, oil temperature, muck improvement parameters and spiral conveying motor torque, cleaning and classifying the original operation data, and storing the original operation data and the processed operation data in a data storage unit;

(1.2) constructing a shield machine performance prediction optimization model, wherein the preferable shield machine performance prediction optimization model of the embodiment comprises a cutter head and shield body structure stress deformation prediction model, a main drive bearing service life prediction model and a geological type prediction and muck improvement decision optimization model, and the specific construction mode is as follows:

1) establishing a detailed three-dimensional structure simulation model of the cutter head and the shield body by using finite element software, performing order reduction treatment on the simulation model by adopting an agent model technology, and establishing an order reduction model for rapidly calculating the stress and deformation of the cutter head and the shield body according to the load working conditions such as shield thrust, torque, soil pressure and the like, namely a stress deformation prediction model of the cutter head and the shield body;

2) establishing a main drive bearing service life prediction model by adopting a theoretical/empirical formula, for example, establishing a fatigue stress system correction model based on the ISO281 standard aiming at a main bearing of a large shield machine;

3) and collecting historical data samples of the geological type of the construction working face in front of the shield tunneling machine and the muck improvement parameter setting, and establishing a shield tunneling machine geological type prediction and muck improvement decision optimization model by using a machine learning algorithm.

(1.3) constructing a shield machine fault diagnosis model, wherein the preferred shield machine fault diagnosis model in the embodiment is a shield machine rotating mechanical part monitoring and fault diagnosis model, and the specific construction mode is as follows:

1) collecting historical data samples related to the running states and faults of key systems or components such as a main driving bearing of the shield machine, a spiral shaft of a spiral conveyor and the like, and establishing a shield machine rotating mechanical component monitoring and fault diagnosis model by utilizing a machine learning algorithm;

2) a theoretical method/empirical formula is adopted to establish a shield machine rotating mechanical part monitoring and fault diagnosis model, and the empirical formula adopted by the embodiment is used for evaluating the main drive bearing fault characteristic frequency and the fault characteristic amplitude based on vibration signals.

(2) Inputting parameters such as thrust, torque, soil pressure and the like into a cutter head and shield body structure stress deformation prediction model, calculating output stress and deformation displacement results, inputting parameters such as thrust, torque, cutter head rotating speed and the like into a main drive bearing service life prediction model, calculating the residual service life of an output main bearing, inputting parameters such as thrust, torque, oil cylinder pressure, drive motor current, muck improvement and the like into a geological type prediction and muck improvement decision optimization model, and calculating a geological type and muck improvement decision optimization scheme in front of a tunnel face of a shield machine; inputting parameters such as the processed vibration acceleration, the current of a driving motor, the oil temperature and the like into a shield machine rotating mechanical part monitoring and fault diagnosis model, quickly calculating and outputting health state scores (including health levels such as health, sub-health and fault) and fault early warning information (fault type, position and fault development degree) of a main bearing and a speed reducer, inputting data such as the processed torque and the vibration acceleration of a spiral conveying motor into a spiral conveyer spiral shaft monitoring and fault diagnosis model, and quickly calculating and outputting the health state scores (including health levels such as health, sub-health and fault) and the fault early warning information (fault type, position and fault development degree) of the spiral conveyer spiral shaft.

Note that, in the calculation process in (2), the feedback model calculation time is synchronized.

(3) And (3) evaluating and analyzing the precision and the calculation time of the calculation result in the step (2) to see whether the deployment requirement of the twin model engineering field is met. The deployment requirement of the engineering field mainly comprises two aspects:

1) whether the precision or accuracy of the twin model calculation result meets the engineering requirements or not: stress data acquired in the operation process are compared with stress results calculated by a cutter head and shield structure stress deformation prediction model, and the accuracy rate reaches over 90 percent; comparing the actual geological type obtained on site in the shield running process with the calculation result of the geological type prediction model, wherein the accuracy rate of the geological type prediction mileage reaches 90%; comparing the muck improvement parameters with the calculation results of the muck improvement decision optimization model in the shield running process, wherein the preparation rate of the foam stock solution flow, the mixed solution flow and the air flow parameters is more than 90%; the actual health state and fault information (fault time and type) of the main bearing, the speed reducer and the screw shaft of the screw conveyer are compared with the calculation result of the monitoring and fault diagnosis model in the shield running process, and the coincidence rate is more than 90%;

2) the calculation time of the twin model under the condition of equivalent calculation force with equipment on the engineering site is shorter than the actual physical site process data acquisition time or the time identified in advance relative to the fault occurrence node is longer than the time required by site shield machine fault early warning treatment. For example, the acquisition time of thrust and torque data in the shield tunneling process is 1s, and the single calculation time of a reduced model of stress and deformation of a cutter head and a shield body structure is less than or equal to 1 s; the on-site fault early warning handling time of the screw shaft of the screw conveyor is 10 minutes, and the monitoring and fault diagnosis model of the screw shaft of the screw conveyor calculates by using on-site data and needs to identify the fault type 10 minutes ahead of the fault occurrence time point.

(4) And (3) when the calculation result progress and the calculation time in the step (2) meet the requirements in the step (3), entering the next step, and if the calculation result progress and the calculation time do not meet the requirements in the step (3), returning to the step (1) to correct the performance prediction optimization model, the monitoring and fault diagnosis model (including optimizing model parameters, modeling methods, reducing the dimensionality/order of the model and the like) until the requirements of the step (3) on the model calculation precision and the time are met.

(5) And deploying the shield tunneling machine digital twin model to a data processing and analyzing unit of a server or an industrial personal computer on the shield tunneling machine engineering site.

(6) The method comprises the steps of driving a shield machine digital twin model deployed on an engineering site to perform real-time simulation calculation by utilizing shield machine operation data acquired in real time, predicting the key performance of the shield machine in real time, feeding back optimization parameters or evaluating and monitoring the health state of a shield machine key system/component in real time, and feeding back fault early warning information.

(7) Feeding back the real-time calculation result to a shield machine display system, displaying in the form of data, a chart, a text description, a curve and the like, and performing corresponding operation or treatment by site constructors according to the real-time result of the twin model; or sending a control instruction to a shield machine control system by a server or an industrial personal computer on the engineering site according to a real-time calculation result, optimizing the construction parameters of the shield machine, and realizing the autonomous decision optimization of the shield machine construction based on the twin model.

Example 3:

as shown in fig. 4, the present embodiment discloses a digital twin performance prediction optimization and fault diagnosis based method for a rock drilling jumbo based on the performance prediction optimization and fault diagnosis system and method described in embodiment 1, which includes the following specific steps:

(1) the method comprises the following steps of collecting operation data of the drill jumbo, and constructing a drill jumbo performance prediction optimization model and a drill jumbo fault diagnosis model, wherein the method specifically comprises the following steps:

(1.1) acquiring data such as arm support posture, oil cylinder pressure, speed, flow, vibration, stress, motor current, oil temperature, drilling parameters of a rock drill and the like in the operation process by using a sensor unit arranged on the rock drill trolley, and cleaning, classifying and the like the operation data of the rock drill trolley;

(1.2) constructing a drilling jumbo performance prediction optimization model, wherein the drilling jumbo performance prediction optimization model preferred in the embodiment is an arm support stress and tail end deformation prediction model, an arm support tail end joint freedom degree prediction model, a drilling jumbo construction geological type prediction model, a drilling speed prediction model and a drilling tool loss optimization model, and the concrete construction mode is as follows:

1) establishing a detailed three-dimensional structure simulation model of the arm support and a three-dimensional dynamics and one-dimensional hydraulic system combined simulation model by using finite element software, performing order reduction processing on the simulation model by adopting an agent model or a machine learning technology and the like, and establishing a prediction model for rapidly calculating the stress of the arm support and the deformation of the tail end under load working conditions such as the attitude of the arm support, the rock drilling reaction force and the like;

2) establishing an arm support kinematics simulation model, namely an arm support tail end joint freedom degree prediction model, by adopting a theoretical formula, facilitating displacement compensation of tail end deformation calculated by an arm support reduced model, and predicting each joint freedom degree parameter under an arm support tail end target coordinate;

3) historical data samples of the geological type of the construction working face in front of the drill jumbo, the drilling speed and the drilling tool loss are collected, and a drill jumbo construction geological type prediction model, a drilling speed prediction model and a drilling tool loss optimization model are established by utilizing a machine learning algorithm.

(1.3) constructing a fault diagnosis model of the drill jumbo, wherein the fault diagnosis model of the drill jumbo preferred in the embodiment is a rock drill and engine monitoring and fault diagnosis model and a rock drill jumbo rotating mechanical part monitoring and fault diagnosis model, and the specific construction mode is as follows:

1) collecting historical data samples related to the running states and faults of key systems or components such as a rock drill of a rock drilling jumbo, an engine and the like, and establishing a monitoring and fault diagnosis model of the rock drill and the engine by utilizing a machine learning algorithm;

2) a theoretical method/empirical formula is adopted to establish a monitoring and fault diagnosis model of the rotary mechanical part of the drilling jumbo, and the optimal empirical formula of the embodiment is the evaluation of the fault characteristic frequency and the fault characteristic amplitude of the gear of the speed reducer based on vibration signals.

(2) Inputting the processed data into a drill jumbo performance prediction optimization model, inputting data such as boom posture, cylinder pressure, flow and the like into a boom stress and tail end deformation prediction model, quickly calculating output stress and a tail end deformation result, inputting a blasthole coordinate parameter and a boom tail end deformation result into a boom tail end joint degree of freedom prediction model, quickly calculating parameters of each joint degree of freedom of a boom, inputting drilling parameters of a rock drill into a geological type prediction model, a drilling speed prediction model and a drilling tool loss optimization model, and quickly calculating the front geological type of a rock drill jumbo working face, the optimal drilling speed and the drilling tool loss; and inputting the processed vibration acceleration, motor current, oil temperature, drilling parameters of the rock drill and the like into a monitoring and fault diagnosis model of the rotary mechanical part of the rock drill, and quickly calculating and outputting health state scores (including health levels such as health, sub-health and fault) and fault early warning information (fault type, position and fault development degree) of key parts of the rock drill.

Note that, in the calculation process of (2), the feedback model is synchronized in calculation time.

(3) And (3) evaluating and analyzing the precision and the calculation time of the calculation result obtained in the step (2) to see whether the deployment requirement of the twin model engineering field is met. The deployment requirement of the engineering field mainly comprises two aspects:

1) whether the precision or accuracy of the twin model calculation result meets the engineering requirements or not: stress data acquired in the operation process are compared with stress results calculated by a cantilever crane stress and tail end deformation prediction model, and the accuracy rate reaches over 90 percent; when rock drilling is carried out according to a shot hole diagram, the actual positioning coordinate of the tail end of the arm support is compared with the target shot hole coordinate, and the distance error is not more than 5 cm; comparing an actual geological type obtained on site in the construction process of the drilling jumbo with a calculation result of a geological type prediction model, wherein the accuracy rate of the geological type prediction mileage reaches over 90%; actual parameters such as drilling speed, drilling tool loss and the like in the construction process are compared with the calculation parameters of the drilling speed prediction model and the drilling tool loss optimization model, and the accuracy rate is over 90 percent; the actual health states and fault information (fault time and type) of the rock drill and the speed reducer in the construction process of the rock drilling jumbo are compared with the calculation results of the monitoring and fault diagnosis model, and the coincidence rate is more than 90%;

2) the calculation time of the twin model under the equivalent calculation force condition of equipment on the engineering site is less than the actual physical site process data acquisition time or the time identified in advance relative to the fault occurrence node is more than the time required by the site rock drilling jumbo fault early warning treatment. For example, the boom attitude data acquisition time in the construction process of the drill jumbo is 1s, and the single calculation time of the boom stress and tail end deformation prediction model is less than or equal to 1 s; and the handling time of the rock drill field fault is 5 minutes, and the rock drill fault diagnosis model calculates by using field data and needs to identify the fault type 5 minutes ahead of the fault occurrence time point.

(4) And (3) when the calculation result progress and the calculation time in the step (2) meet the requirements of the step (3), entering the next step, and if the calculation result progress and the calculation time do not meet the requirements, returning to the step (1) to correct the performance prediction optimization model, the monitoring and fault diagnosis model (including optimizing model parameters, modeling methods, improving theoretical calculation formulas, reducing dimensionality/order of the models and the like) until the requirements of the step (3) on the model calculation accuracy and time are met.

(5) And deploying the twin model to a data processing and analyzing unit of a server or an industrial personal computer on the engineering site of the drilling jumbo.

(6) The method comprises the steps of driving a twin model deployed on an engineering site to perform real-time simulation calculation by utilizing the real-time collected rock drilling jumbo operation data, predicting the key performance of the rock drilling jumbo in real time, feeding back optimization parameters or evaluating and monitoring the health state of key systems/components of the rock drilling jumbo in real time, and feeding back fault early warning information.

(7) Feeding back the real-time calculation result to a drilling trolley display system, displaying in the form of data, a chart, a text description, a curve and the like, and performing corresponding operation or treatment by site constructors according to the real-time result of the twin model; or sending a control instruction to the drill jumbo control system by a server or an industrial personal computer on the engineering site according to the real-time calculation result, optimizing the construction parameters of the drill jumbo, and realizing the autonomous decision optimization of the construction of the drill jumbo based on the twin model.

Example 4:

as shown in fig. 5, the present embodiment discloses a digital twin performance prediction optimization and fault diagnosis based method for a mining and anchoring all-in-one machine based on the performance prediction optimization and fault diagnosis system and method described in embodiment 1, which includes the following specific steps:

(1) the method comprises the following steps of collecting operation data of the tunneling and anchoring all-in-one machine, constructing a performance prediction optimization model of the tunneling and anchoring all-in-one machine and a fault diagnosis model of the tunneling and anchoring all-in-one machine, and specifically comprising the following steps:

(1.1) data such as cutting height, angle, depth, cutting drum rotating speed, oil cylinder pressure, speed, flow, vibration, stress, motor current, torque, oil temperature and the like in the operation process are collected by utilizing a plurality of types of sensors arranged on the tunneling and anchoring all-in-one machine, and the operation data of the tunneling and anchoring all-in-one machine is cleaned, classified and the like.

(1.2) constructing a performance prediction optimization model of the tunneling and anchoring all-in-one machine, wherein the performance prediction optimization model of the tunneling and anchoring all-in-one machine is a prediction and optimization model of the type of coal rock and the tooth breakage rate of construction of the tunneling and anchoring all-in-one machine:

1) establishing a detailed three-dimensional structure simulation model of the cutting large arm of the tunneling and anchoring all-in-one machine by using finite element software, performing order reduction treatment on the simulation model by adopting a proxy model technology and the like, and establishing an order reduction model for rapidly calculating the stress and deformation of the cutting large arm under the working condition of coal excavation stress of a cutting drum;

2) establishing a load calculation model of the cutting drum aiming at different coal rock drivages by adopting a cutting drum load theory/empirical formula;

3) historical data samples related to construction coal rock types and cutting tooth breakage rate performance in front of the coal digging all-in-one machine are collected, and a machine learning algorithm is utilized to establish a prediction and optimization model of the construction coal rock types and the cutting tooth breakage rates of the coal digging and anchoring all-in-one machine.

(1.3) constructing a monitoring and fault diagnosis model of key components of the tunneling and anchoring all-in-one machine, wherein the monitoring and fault diagnosis model of key components of the tunneling and anchoring all-in-one machine is preferably a monitoring and fault diagnosis model of rotating mechanical components of the tunneling and anchoring all-in-one machine:

1) collecting running state and fault related historical data samples of a scraper chain of the tunneling and anchoring all-in-one machine, and establishing a monitoring and fault diagnosis model of a rotating mechanical part of the tunneling and anchoring all-in-one machine by utilizing a machine learning algorithm;

2) a theoretical method/empirical formula is adopted to establish a monitoring and fault diagnosis model of the rotating mechanical part of the digging and anchoring integrated machine, such as empirical formulas for evaluating the fault characteristic frequency and the fault characteristic amplitude of a speed reducer gear based on a vibration signal.

(2) Inputting the processed data into a construction coal rock type and tooth breakage rate prediction and optimization model of the tunneling and anchoring all-in-one machine, wherein parameters such as cutting oil cylinder pressure, cutting motor torque, cutting depth and the like are input into the construction coal rock type prediction model, the coal rock type in front of the tunneling and anchoring all-in-one machine is rapidly calculated, the cutting height, the cutting depth, the cutting drum rotating speed and the coal rock type are input into a load calculation model, the cutting drum stress working condition is rapidly calculated, parameters such as cutting drum stress working condition, cutting height, cutting angle and the like are input into a cutting large arm stress and deformation reduction model, output stress and deformation results are rapidly calculated, parameters such as the coal rock type, the cutting height, the cutting depth, the cutting rotating speed, the cutting motor torque and the like are input into the tooth breakage rate prediction and optimization model, and tooth breakage rate and optimal construction parameters are rapidly calculated; and inputting the processed parameters such as vibration acceleration, motor current, oil temperature and the like into a monitoring and fault diagnosis model of the rotating mechanical part of the tunneling and anchoring integrated machine, and quickly calculating and outputting health state scores (including health levels such as health, sub-health and fault) and fault early warning information (fault type, position and fault development degree) of key parts of the tunneling and anchoring integrated machine. The above simultaneous feedback model calculates time.

(3) And (3) evaluating and analyzing the precision and the calculation time of the calculation result obtained in the step (2) to see whether the deployment requirement of the twin model engineering field is met. The deployment requirement of the engineering field mainly comprises two aspects:

1) whether the precision or accuracy of the twin model calculation result meets the engineering requirements or not: calculating cutting by utilizing an actual coal rock type in front of the construction of the tunneling and anchoring all-in-one machine and a construction coal rock type prediction model for comparison, wherein the accuracy rate of predicted mileage reaches more than 90%; the stress data collected in the application process is compared with the stress result calculated by the cutting big arm stress and deformation reduced model, and the accuracy rate reaches more than 90%; the actual tooth breaking rate and tooth breaking rate prediction of the cutting drum in the construction process are compared with the tooth breaking rate calculated by the optimization model, and the accuracy rate reaches more than 90%; the actual health states and fault information (fault time and type) of a scraper chain and a cutting speed reducer in the construction process are compared with the calculation results of a monitoring and fault diagnosis model of the rotating mechanical part of the tunneling and anchoring all-in-one machine, and the coincidence rate is more than 90%;

2) the calculation time of the twin model under the condition of equivalent calculation force with equipment on the engineering site is shorter than the actual physical site process data acquisition time or the time identified in advance relative to the fault occurrence node is longer than the time required by the fault early warning treatment of the site tunneling and anchoring all-in-one machine. For example, in the construction process of the tunneling and anchoring all-in-one machine, the data acquisition time of cutting height, angle, depth and the like is 1s, and the single calculation time of the cutting large arm stress and deformation reduced model is less than or equal to 1 s; and if the field fault handling time of the cutting speed reducer of the tunneling and anchoring integrated machine is 5 minutes, the monitoring and fault diagnosis model of the rotating mechanical part of the tunneling and anchoring integrated machine calculates by using field data and needs to identify the fault type 5 minutes ahead of the fault occurrence time point.

(4) And (3) when the calculation result progress and the calculation time in the step (2) meet the requirements in the step (3), entering the next step, and if the calculation result progress and the calculation time do not meet the requirements in the step (3), returning to the step (1) to correct the performance prediction optimization model and the fault diagnosis model (including optimizing model parameters, modeling methods, improving theoretical calculation formulas, reducing dimensionality/order of the models and the like) until the requirements on the model calculation accuracy and time in the step (3) are met.

(5) And deploying the twin model to a data processing and analyzing unit of a server or an industrial personal computer on the engineering site of the tunneling and anchoring all-in-one machine.

(6) And driving a twin-organism model deployed on a project site to perform real-time simulation calculation by utilizing the real-time collected operation data of the tunneling and anchoring all-in-one machine, predicting the key performance of the tunneling and anchoring all-in-one machine in real time, feeding back optimized parameters or evaluating and monitoring the health state of a key system/component of the tunneling and anchoring all-in-one machine in real time, and feeding back fault early warning information.

(7) Feeding back the real-time calculation result to a display system of the tunneling and anchoring all-in-one machine, displaying in the forms of data, graphs, written descriptions, curves and the like, and performing corresponding operation or treatment by site constructors according to the real-time result of the twin model; or sending a control instruction to a control system of the tunneling and anchoring all-in-one machine by a server or an industrial personal computer on the engineering site according to a real-time calculation result, optimizing construction parameters of the tunneling and anchoring all-in-one machine, and realizing autonomous decision optimization of the tunneling and anchoring all-in-one machine construction based on the twin model.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (9)

1. A performance prediction optimization and fault diagnosis system of underground engineering equipment is characterized by comprising a physical object, a data acquisition and analysis module, a data storage module and a digital twin module;

the data acquisition and analysis module is installed at the edge end of the physical object and is used for acquiring original operation data of the physical object in real time and processing the original operation data;

the data storage module is used for storing original operation data and processed operation data;

the digital twin module is used for calculating the data in the data storage module, establishing a digital twin model of the physical object and correcting the digital twin model according to the calculation result precision and the calculation speed;

the data acquisition and analysis module deploys the modified digital twin model, and inputs real-time operation data into the digital twin model to perform performance prediction optimization and fault diagnosis on the physical object.

2. The system of claim 1, wherein the data collection and analysis module comprises a sensor unit, a data collection and storage unit, and a data processing and analysis unit;

wherein the sensor unit is used for acquiring raw operating data of a physical object;

the data acquisition and storage unit is used for receiving the operation data and transmitting the operation data to the data processing and analysis unit and the data storage module;

the data processing and analyzing unit cleans and classifies the original operation data and transmits the processed operation data to the data storage module.

3. The system of claim 2, wherein the sensor unit comprises a sensor for collecting operational data of a physical object, the sensor being disposed on the physical object.

4. The system of claim 1, wherein the data storage module comprises an online data platform and an offline local database, the online data platform transmits data through a network, and the offline database performs data copy transmission through a mobile storage medium.

5. The system of claim 1, wherein the digital twin module comprises a data interface unit, a data processing/storage unit, a digital twin analysis unit, a result visualization unit, and a model modification unit;

the data interface unit is used for receiving original operation data and processed operation data of the data storage module;

the data processing/storing unit carries out secondary processing on the data, eliminates abnormal and unnecessary data and stores the data after the secondary processing;

the digital twin analysis unit establishes a performance prediction optimization model and a fault diagnosis model, and inputs the secondarily processed data into the performance prediction optimization model and the fault diagnosis model to obtain a calculation result;

the result visualization unit is used for receiving the calculation result of the digital twin analysis unit and performing visualization processing on the calculation result to form a digital twin model;

the model correction unit corrects or does not correct the digital twin model based on the accuracy of the calculation result and the calculation speed.

6. The performance prediction optimization and fault diagnosis system of claim 5, wherein the calculated results comprise predicted results and diagnosed results;

the prediction result is obtained by inputting data subjected to secondary processing into a performance prediction optimization model, and the prediction result comprises optimization parameters;

the diagnosis result is data input fault diagnosis module for secondary processing, and the diagnosis result comprises health state and fault early warning information of the physical object.

7. The system of any one of claims 5 or 6, wherein the physical object is underground construction equipment comprising a mechanical unit, a hydraulic unit, an electrical control unit and a display unit;

the electric control unit receives a control command and operates the mechanical unit and the hydraulic unit according to the control command;

the display unit presents the calculation result in the form of data, a graph or a curve.

8. A method for performance prediction optimization and fault diagnosis of underground engineering equipment, which is characterized by applying the system for performance prediction optimization and fault diagnosis according to any one of claims 1 to 7, and comprises the following specific steps:

step S1: acquiring operation data of a physical object, specifically, acquiring original operation data of a physical entity in real time by a data acquisition and analysis module, preprocessing the original operation data, and transmitting the original operation data and the preprocessed operation data to a data storage module;

step S2: establishing a performance prediction optimization model and a fault diagnosis model, specifically, establishing a reduced performance prediction optimization model and a reduced fault diagnosis model, wherein the performance prediction optimization model and the reduced fault diagnosis model are both simulation models and are stored in a digital twin module;

step S3: establishing a digital twin model of the underground engineering equipment, specifically, inputting data in the storage module in the step S1 into the performance prediction optimization model and the fault diagnosis model in the step S2 for calculation to obtain a calculation result, wherein the calculation result comprises optimization parameters, health states and fault early warning information of the underground engineering equipment, and feeds back model calculation time;

step S4: checking whether the digital twin model is qualified, specifically, calculating the precision of the calculation result in the step S3, if the precision exceeds a preset precision threshold and the model calculation time in the step S3 is less than the underground engineering equipment fault early warning handling time, the digital twin model meets the engineering requirement, otherwise, returning to the step S2 to correct the digital twin model;

step S5: deploying a digital twin model, specifically deploying the digital twin model meeting the engineering requirements in step S4 into a data acquisition and analysis module at an edge end, performing real-time simulation calculation according to real-time acquired operating data, performing performance prediction optimization and fault diagnosis on a physical entity, and acquiring optimization parameters, health status and fault early warning information of the physical entity;

step S6: and (4) feeding back the optimization parameters and the fault early warning information, specifically, displaying the optimization parameters, the health state and the fault early warning information in the step S5 in an engineering field, sending a control instruction to the underground engineering equipment physical entity by the data acquisition and analysis module based on the optimization parameters and the fault early warning information, and receiving the control instruction by the underground engineering equipment physical entity to optimize the construction parameters.

9. The method for performance prediction optimization and fault diagnosis according to claim 8, wherein in step S2, the performance prediction optimization model and the fault diagnosis model are established in the following specific manner:

establishing a simulation model based on the attributes of the physical entity, and applying data fitting, deep learning or proxy model to the simulation model to perform order reduction processing to obtain a performance prediction optimization model and a fault diagnosis model;

or, establishing a performance prediction optimization model and a fault diagnosis model by applying a theoretical/empirical formula;

or, by collecting historical data samples of underground engineering equipment, performing correlation analysis on the historical data samples by applying a machine learning algorithm to establish a performance prediction optimization model and a fault diagnosis model.

Images