Method and system for detecting running state of digging and anchoring all-in-one machine based on multi-sensor fusion

Abstract

The invention discloses a multi-sensor fusion-based driving and anchoring all-in-one machine operation state detection method and system, and relates to the technical field of mining equipment, and the method comprises the steps: monitoring a driving and anchoring all-in-one machine in real time through a multi-sensor terminal, and obtaining hydraulic, mechanical and electrical monitoring sequences with operation node identifiers; constructing a credible state grid space according to the operation control decision; guiding anomaly detection according to the node identifier to obtain a ternary anomaly detection result; after historical fault deduction is combined, a bolter miner coupling fault prediction model is introduced for prediction, and a detection map is obtained. The technical problems that in the operation process of an existing digging and anchoring all-in-one machine, multi-sensing data are isolated, the anomaly detection precision is low, and multi-dimensional fault trend deduction cannot be achieved are solved, multi-source data fusion analysis of the digging and anchoring all-in-one machine is achieved, the anomaly detection accuracy of the digging and anchoring all-in-one machine is improved, and the fault probability of the digging and anchoring all-in-one machine is lowered. And the technical effect of accurate fault prediction of the digging and anchoring all-in-one machine is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mining equipment, in particular to a method and system for detecting the running state of a combined excavating and anchoring machine based on multi-sensor fusion. BACKGROUND

[0002] In mining operations, the stable operation of a combined excavating and anchoring machine is crucial to mining efficiency and job safety, and accurate detection of its running state is key to ensuring production. In the prior art, detection of the running state of a combined excavating and anchoring machine mostly uses a single or a small number of sensors to collect data, and the device state is determined through simple analysis. These methods can play a certain role in scenarios where the device working condition is stable, but as the complexity of mining operations increases, the traditional detection technology has limitations in application. Due to the complex mining environment, the combined excavating and anchoring machine has strong coupling among systems such as hydraulic, mechanical, and electrical systems, and traditional methods cannot fully capture multi-dimensional state information of the device, resulting in one-sided detection data and insufficient accuracy, which makes it difficult to meet the demand for accurate assessment and fault warning of the running state of the combined excavating and anchoring machine. SUMMARY

[0003] The present application provides a method and system for detecting the running state of a combined excavating and anchoring machine based on multi-sensor fusion, to solve the technical problems of isolated multi-sensor data, low abnormality detection accuracy, and inability to realize multi-dimensional fault trend deduction during the operation of the combined excavating and anchoring machine.

[0004] In a first aspect, the present application provides a method for detecting the running state of a combined excavating and anchoring machine based on multi-sensor fusion, which comprises: monitoring the combined excavating and anchoring machine in real time through a multi-sensor terminal to obtain a combined excavating and anchoring machine monitoring sequence with a job node identifier, the combined excavating and anchoring machine monitoring sequence including a hydraulic monitoring sequence, a mechanical monitoring sequence, and an electrical monitoring sequence; fitting and analyzing the trusted state features of the combined excavating and anchoring machine according to the combined excavating and anchoring job control decision to construct a trusted state grid space; guiding the trusted state grid space to perform abnormality detection on the combined excavating and anchoring machine monitoring sequence according to the job node identifier to obtain a three-element abnormality detection result; performing multi-dimensional fault trend deduction on the three-element abnormality detection result according to a historical fault event set of the combined excavating and anchoring machine to obtain a first fault deduction result, a second fault deduction result, and a third fault deduction result; introducing a combined excavating and anchoring machine coupling fault prediction model to perform coupling fault prediction on the first fault deduction result, the second fault deduction result, and the third fault deduction result to obtain a combined excavating and anchoring machine detection graph.

[0005] In a possible implementation, a three-dimensional model is established according to the integrated excavating and anchoring machine, to obtain an excavating and anchoring machine model; the excavating and anchoring machine model is simulated and controlled multiple times according to the excavating and anchoring operation control decision, to obtain multiple fitted state data sets; hydraulic feature credible analysis is performed according to the multiple fitted state data sets, to construct a hydraulic credible grid space; mechanical feature credible analysis is performed according to the multiple fitted state data sets, to construct a mechanical credible grid space; electrical feature credible analysis is performed according to the multiple fitted state data sets, to construct an electrical credible grid space; and the hydraulic credible grid space, the mechanical credible grid space, and the electrical credible grid space are used to generate the credible state grid space.

[0006] In a possible implementation, hydraulic feature extraction is performed according to the multiple fitted state data sets, to obtain multiple hydraulic response fitted sets; time sequence synchronization correlation is performed according to the multiple hydraulic response fitted sets, to obtain each time sequence hydraulic response set; multi-dimensional centralized trend analysis is performed according to the each time sequence hydraulic response set, to construct each time sequence hydraulic feature credible sequence; and grid processing is performed according to the each time sequence hydraulic feature credible sequence, to generate the hydraulic credible grid space.

[0007] In a possible implementation, node mapping is performed on the credible state grid space according to the operation node identifier, to obtain a matching hydraulic grid space, a matching mechanical grid space, and a matching electrical grid space; hydraulic abnormality detection is performed on the hydraulic monitoring sequence according to the matching hydraulic grid space, to obtain a hydraulic abnormality detection result; mechanical abnormality detection is performed on the mechanical monitoring sequence according to the matching mechanical grid space, to obtain a mechanical abnormality detection result; electrical abnormality detection is performed on the electrical monitoring sequence according to the matching electrical grid space, to obtain an electrical abnormality detection result; and the three-abnormality detection result is generated by combining the hydraulic abnormality detection result and the mechanical abnormality detection result.

[0008] In a possible implementation, the historical fault event set is classified, to obtain a hydraulic fault event set, a mechanical fault event set, and an electrical fault event set; hydraulic fault trend deduction is performed on the hydraulic abnormality detection result according to the hydraulic fault event set based on a fault trigger sensitive threshold, to generate the first fault deduction result; mechanical fault trend deduction is performed on the mechanical abnormality detection result according to the mechanical fault event set based on the fault trigger sensitive threshold, to generate the second fault deduction result; and electrical fault trend deduction is performed on the electrical abnormality detection result according to the electrical fault event set based on the fault trigger sensitive threshold, to generate the third fault deduction result.

[0009] In a possible implementation, according to the set of hydraulic fault events, fault type identification is performed to obtain each fault top-level factor; according to the set of hydraulic fault events, direct factor tracing is performed on the each fault top-level factor to obtain each fault direct factor; according to the set of hydraulic fault events, bottom factor tracing is performed on the each fault direct factor to obtain each fault bottom factor; Boolean logic association is performed according to the each fault top-level factor, the each fault direct factor and the each fault bottom factor to obtain a hydraulic fault deduction channel; the hydraulic anomaly detection result is input into the hydraulic fault deduction channel to obtain a plurality of hydraulic fault deduction paths; and according to the fault trigger sensitive threshold, trigger sensitive optimization is performed on the plurality of hydraulic fault deduction paths to obtain the first fault deduction result.

[0010] In a possible implementation, according to the plurality of hydraulic fault deduction paths, trigger factor identification is performed to obtain a plurality of hydraulic fault factors; according to the hydraulic anomaly detection result, mapping comparison is respectively performed on the plurality of hydraulic fault factors to obtain a plurality of fault trigger sensitive coefficients; and based on the plurality of fault trigger sensitive coefficients, the plurality of hydraulic fault deduction paths are optimized and screened according to the fault trigger sensitive threshold to generate the first fault deduction result.

[0011] In a possible implementation, according to the plurality of sensing terminals, a tunnel anchor machine monitoring data set is obtained; the tunnel anchor machine monitoring data set is cleaned and classified to generate the tunnel anchor machine monitoring sequence.

[0012] In a possible implementation, according to the tunnel anchor machine detection graph, a tunnel anchor machine early warning signal is generated.

[0013] In a second aspect of the present application, a multi-sensor fusion-based running state detection system for a combined excavator-anchor machine is provided, which comprises: an excavator-anchor machine monitoring sequence acquisition module, configured to monitor the combined excavator-anchor machine in real time through a multi-sensor terminal, and acquire an excavator-anchor machine monitoring sequence with a work node identifier, wherein the excavator-anchor machine monitoring sequence comprises a hydraulic monitoring sequence, a mechanical monitoring sequence, and an electrical monitoring sequence; a trusted state grid space construction module, configured to fit and analyze the trusted state characteristics of the combined excavator-anchor machine according to an excavator work control decision, and construct a trusted state grid space; a ternary anomaly detection result acquisition module, configured to guide the trusted state grid space to perform anomaly detection on the excavator-anchor machine monitoring sequence according to the work node identifier, and acquire a ternary anomaly detection result; a fault deduction result acquisition module, configured to perform multi-dimensional fault trend deduction on the ternary anomaly detection result according to a historical fault event set of the combined excavator-anchor machine, and acquire a first fault deduction result, a second fault deduction result, and a third fault deduction result; and an excavator-anchor machine detection atlas acquisition module, configured to introduce an excavator-anchor machine coupling fault prediction model to perform coupling fault prediction on the first fault deduction result, the second fault deduction result, and the third fault deduction result, and acquire an excavator-anchor machine detection atlas.

[0014] The one or more technical solutions provided in the present application have at least the following technical effects or advantages: The present application monitors the combined excavator-anchor machine in real time through a multi-sensor terminal, acquires hydraulic, mechanical, and electrical monitoring sequences with a work node identifier, constructs a trusted state grid space according to an excavator work control decision, performs anomaly detection on the monitoring sequences under the guidance of the work node identifier to obtain a ternary anomaly detection result, performs multi-dimensional fault trend deduction in combination with a historical fault event set, introduces a coupling fault prediction model for prediction, and acquires an excavator-anchor machine detection atlas, thereby realizing accurate detection of the running state of the combined excavator-anchor machine, making the equipment running state detection result in a mining scene more comprehensive and reliable, achieving multi-source data fusion analysis of the combined excavator-anchor machine, improving the accuracy of excavator-anchor machine anomaly detection, and realizing accurate fault prediction of the combined excavator-anchor machine. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0016] Figure 1 FIG. 1 is a flowchart of a multi-sensor fusion-based running state detection method for a combined excavator-anchor machine provided by an embodiment of the present application.

[0017] Figure 2FIG. 1 is a structural schematic diagram of a running state detection system of a tunneling-anchor integrated machine based on multi-sensor fusion provided by an embodiment of the present application.

[0018] Legend: tunnel-anchor machine monitoring sequence acquisition module 1, trusted state grid space construction module 2, ternary anomaly detection result acquisition module 3, fault deduction result acquisition module 4, and tunnel-anchor machine detection atlas acquisition module 5. DETAILED DESCRIPTION

[0019] The present application provides a running state detection method and system of a tunnel-anchor integrated machine based on multi-sensor fusion, which is used to solve the technical problems of isolated multi-sensor data, low anomaly detection accuracy, and inability to realize multi-dimensional fault trend deduction during the running of the existing tunnel-anchor integrated machine.

[0020] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0021] It should be noted that the terms "first", "second", and the like in the specification and the above drawings of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server including a series of steps or units does not have to be limited to only those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices.

[0022] Embodiment one, as shown in FIG. 1, a running state detection method of a tunnel-anchor integrated machine based on multi-sensor fusion, wherein the method comprises: Figure 1 Step A100: Real-time monitoring of the tunnel-anchor integrated machine by a multi-sensor terminal to obtain a tunnel-anchor machine monitoring sequence with a work node identifier, the tunnel-anchor machine monitoring sequence including a hydraulic monitoring sequence, a mechanical monitoring sequence, and an electrical monitoring sequence.

[0023] In the embodiments of the present application, the tunnel-anchor integrated machine is a key device for tunneling and anchoring operations in a mining scene, and its running state involves multiple systems such as hydraulic, mechanical, and electrical systems. It is necessary to detect the running state through multi-sensor fusion and other technologies to ensure stable operation and safe operation in mining operations. ​

[0024] Specifically, in the mining scene, in order to comprehensively capture the running state of the integrated machine, multiple sensing terminals need to be deployed at key positions of the device. The multiple sensing terminals can collect real-time data of parameters such as pressure and flow of the hydraulic system, vibration and displacement of the mechanical structure, and current and voltage of the electrical system, and the specific steps are described in detail in A110-A120.

[0025] During the collection process, each sensing terminal synchronously records the work node identifier corresponding to the monitored position, which is associated with the specific work stage or position of the integrated machine, ensuring that each set of monitoring data can accurately correspond to the actual running scene of the device.

[0026] Subsequently, the original monitoring data set obtained by the multiple sensing terminals is processed, invalid or interference data is removed through cleaning, and then classified and arranged according to the three system categories of hydraulic, mechanical and electrical, and finally the hydraulic monitoring sequence, mechanical monitoring sequence and electrical monitoring sequence with work node identifier are formed.

[0027] Through the deployment of multiple sensing terminals, real-time data collection, association of work node identifier and classification processing, a comprehensive and scene-specific integrated machine monitoring sequence is obtained, providing a reliable data basis for subsequent state detection.

[0028] Step A200: fitting analysis of the trusted state characteristics of the integrated machine according to the integrated work control decision, and constructing a trusted state grid space.

[0029] Optionally, the integrated machine is three-dimensionally modeled to obtain an integrated machine model, multiple fitting state data sets are obtained by repeatedly simulating and controlling the model according to the integrated work control decision, and then the hydraulic, mechanical and electrical characteristics are analyzed respectively to construct respective trusted state grid spaces, and finally the trusted state grid space is integrated and generated, and the specific steps are described in detail in A210-A260.

[0030] Step A300: abnormality detection of the integrated machine monitoring sequence according to the work node identifier guided by the trusted state grid space, and obtaining a three-element abnormality detection result.

[0031] In an embodiment of the present application, the trusted state grid space is node-mapped according to the work node identifier to obtain a matched three-element grid space, the hydraulic, mechanical and electrical monitoring sequences are respectively subjected to abnormality detection to obtain corresponding results, and the three-element abnormality detection result is generated by combining, and the specific steps are described in detail in A310-A340.

[0032] Step A400: multi-dimensional fault trend inference of the three-element abnormality detection result according to the historical fault event set of the integrated machine, and obtaining a first fault inference result, a second fault inference result and a third fault inference result.

[0033] Specifically, the historical fault event set is classified into three fault event sets. Based on the fault trigger sensitivity threshold, trend deduction is performed on the hydraulic, mechanical, and electrical anomaly detection results to generate the first, second, and third fault deduction results. The specific steps are described in detail in A410-A440.

[0034] Step A500: introducing a coupled fault prediction model for an anchor miner to perform coupled fault prediction on the first fault deduction result, the second fault deduction result, and the third fault deduction result to obtain a detection map of the anchor miner.

[0035] Specifically, we first constructed a coupled fault prediction model for anchor miners by defining its input and output elements. The input consisted of the results of the first, second, and third fault deductions. Key features were extracted from these models, including the fault type of each system (e.g., hydraulic cylinder leakage, mechanical gear wear), the correlation strength of the fault path, and the temporal relationship of the fault occurrence. These features together formed the model's input vector. The output consisted of the possible coupled fault types and their corresponding likelihood of occurrence, providing core information for the subsequent generation of anchor miner detection maps.

[0036] The coupled fault prediction model for anchor miners utilizes a machine learning architecture, such as a graph neural network, that can capture the interdependencies between multiple systems. Its structure comprises an input layer, a graph convolutional layer, a fully connected layer, and an output layer. These layers receive extracted feature vectors, learn correlation weights between system faults, strengthen the influence of related features, integrate the processed features, and output a prediction result for coupled faults. Furthermore, an attention mechanism is introduced to enable the model to assign higher weight to highly correlated fault features, such as the strong correlation between hydraulic and mechanical power transmission, improving prediction accuracy.

[0037] When training the coupled fault prediction model for bolter miners, historical coupled fault records are used as the foundational data. Data preprocessing is performed to convert fault types into a computable encoding format and standardize features such as correlation strength to ensure the data format meets model input requirements. The training process is carried out in stages: Initial training is performed using a portion of the data. Network parameters are adjusted using a loss function to enable the model to initially learn fault correlation patterns. The model is then validated using another portion of the data, optimizing the network structure, such as adjusting the number of neurons in the graph convolution layer to avoid overfitting. Finally, the model is tested using the remaining data to ensure stable prediction capabilities for unseen coupled faults.

[0038] When the coupling fault prediction model of the tunneling and anchoring machine is applied, the feature vectors of the first, second and third fault deduction results are input into the trained model. The model calculates the correlation weight between each system fault through a graph convolution layer, such as the correlation weight between the hydraulic fault and the mechanical fault due to power transmission. After processing by a fully connected layer, the prediction results of each type of coupling fault are output, including the fault type, the involved system and the occurrence probability. These results are integrated into a tunneling and anchoring machine detection map, which clearly presents the correlation path and risk degree of each coupling fault.

[0039] By constructing and training the coupling fault prediction model of the tunneling and anchoring machine, the fault deduction results of the three systems are analyzed and coupled, and finally the tunneling and anchoring machine detection map is obtained, which can comprehensively reflect the fault linkage relationship between multiple systems, and provide accurate and systematic decision basis for the coupling fault early warning and targeted maintenance of the tunneling and anchoring integrated machine.

[0040] Further, step A200 in the method provided by the embodiment of the application comprises: A210: three-dimensional modeling is performed according to the tunneling and anchoring integrated machine to obtain a tunneling and anchoring machine model.

[0041] A220: multiple simulation controls are performed on the tunneling and anchoring machine model according to the tunneling operation control decision to obtain multiple fitted state data sets.

[0042] A230: hydraulic feature credible analysis is performed according to the multiple fitted state data sets to construct a hydraulic credible grid space.

[0043] A240: mechanical feature credible analysis is performed according to the multiple fitted state data sets to construct a mechanical credible grid space.

[0044] A250: electrical feature credible analysis is performed according to the multiple fitted state data sets to construct an electrical credible grid space.

[0045] A260: the credible state grid space is generated according to the hydraulic credible grid space, the mechanical credible grid space and the electrical credible grid space.

[0046] Specifically, to construct the credible state grid space, first, three-dimensional modeling is performed on the tunneling and anchoring integrated machine. The overall structure, size and relative position of the key components of the tunneling and anchoring integrated machine are scanned by a laser scanning device to obtain point cloud data with an accuracy of 0.1 mm. Then, a three-dimensional modeling software is used to process the data to reconstruct a three-dimensional model containing the tunneling arm, the anchor drilling mechanism, the hydraulic pipeline, the motor and other key components. The geometric parameter error between the model and the actual object is controlled within ±1 mm, ensuring that the model can accurately reflect the physical structure and motion characteristics of the equipment.

[0047] Based on the constructed three-dimensional model of the roof bolter, combined with the control decision of the roof bolting operation, multiple simulation controls are carried out. The control decision of the roof bolting operation covers parameter settings under different operation scenarios, for example, when the coal seam hardness is f = 3-5, the tunneling speed is set to 0.8 m / min, the oil cylinder thrust is set to 20 kN during anchoring operation, and the working pressure of the hydraulic system is set to 20 MPa, etc. The specific decision is determined by a person skilled in the art according to the actual operation scenario. For these decisions, different working conditions are simulated in the simulation system, and more than 200 simulation runs are carried out, each simulation lasting 1 hour. The hydraulic system flow, mechanical component vibration frequency, motor operating current and other data at each time are recorded, and finally multiple fitted state data sets are formed, each data set containing more than 10,000 time series data, covering the key operating characteristics of the hydraulic, mechanical and electrical systems.

[0048] Among them, the construction of the above simulation system takes the three-dimensional model of the roof bolter as the core basis, integrates the parameter settings of different working conditions in the roof bolting operation control decision, such as the tunneling speed under different coal seam hardness, the oil cylinder thrust during anchoring operation, and the working pressure of the hydraulic system, etc. Through programming, the driving control of the three-dimensional model is realized, so that it can simulate the running process of the equipment under various operation scenarios; at the same time, the system has a built-in data acquisition module, which can record the key parameters such as hydraulic system flow, mechanical component vibration frequency, motor operating current in real time during the simulation process, support the simulation operation of different working conditions, and generate fitted state data sets containing a large amount of time series data, covering the key operating characteristics of each system of the equipment.

[0049] Next, hydraulic features are extracted from multiple fitted state data sets to obtain multiple hydraulic response fitting sets, and each time series hydraulic response set is obtained through time series synchronization correlation. Then, through multi-dimensional centralized trend analysis, each time series hydraulic feature credible sequence is constructed, and finally the hydraulic credible grid space is generated through gridding processing. The specific steps are described in detail in A231-A234.

[0050] Then, mechanical features are extracted from multiple fitted state data sets to obtain multiple mechanical response fitting sets. The fitted state data set contains key operating parameters of the mechanical system, from which feature parameters such as tunneling arm vibration acceleration, anchor drilling mechanism displacement, and transmission component speed are selected, and redundant information is removed through feature extraction algorithm. For example, the mechanical response fitting set during tunneling operation contains time series data of vibration acceleration 0.5-3g and speed 100-300r / min, and the time series data of displacement 50-200mm during anchoring operation. Each fitting set accurately corresponds to the mechanical system response characteristics under a specific working condition.

[0051] The multiple mechanical response fitting sets are time-synchronously associated, taking the operation action triggering time of the anchor integrated machine as the time reference, the time stamp alignment algorithm is used to unify the time sequence data of each fitting set to the same time axis, and the sampling interval is set to 50 ms. After synchronization, the mechanical parameters at each time node can correspond to each other, forming each time sequence mechanical response set, for example, the vibration acceleration, displacement and speed data at the 20th second are included at the same time, realizing the time sequence consistency of mechanical data under different working conditions.

[0052] Based on each time sequence mechanical response set, multi-dimensional centralized trend analysis is performed, and each time sequence mechanical feature reliable sequence is constructed. For the multi-dimensional mechanical parameters of each time sequence node, the mean and standard deviation of the parameters under multiple fitting states are calculated, and the normal fluctuation range is determined. For example, the vibration acceleration mean of a certain time sequence node is 1.5 g, and the standard deviation is 0.3 g, the reliable interval is set to 0.6-2.4 g, the abnormal values exceeding the range are removed, the parameters meeting the centralized trend are retained, the mechanical feature reliable sequence of the time sequence is formed, and it is ensured that the sequence can represent the normal operation state of the mechanical system.

[0053] According to each time sequence mechanical feature reliable sequence, grid processing is performed, taking the time axis as the X axis, and taking each 1 minute as a grid unit, taking the reliable interval of the mechanical parameter as the Y axis, for example, the vibration acceleration is set as an interval of 0.5 g, the reliable sequence data of each time sequence node is mapped into the grid unit, the frequency and probability of the parameters appearing in the range are recorded, and a mechanical reliable grid space is generated, which directly presents the normal operation state distribution of the mechanical system in different time and different parameter ranges.

[0054] Then, the electrical features are extracted from the multiple fitting state data sets to obtain multiple electrical response fitting sets. The fitting state data set contains key parameters of the electrical system, and the feature extraction algorithm is used to process the feature parameters such as driving motor current, control loop voltage and power factor to form electrical response fitting sets corresponding to different working conditions. For example, the electrical response fitting set under heavy load operation contains time sequence data of current 80-150 A and voltage 380 V±10%, and the time sequence data of current 60-100 A under light load operation, which accurately reflects the response characteristics of the electrical system under different working conditions.

[0055] The multiple electrical response fitting sets are time-synchronously associated, taking the operation action triggering time as the time reference, the time stamp alignment algorithm is used to unify the time axis, the sampling interval is set to 50 ms, the time sequence data of each fitting set is corresponded at the same time node, and each time sequence electrical response set is formed. For example, at the 30th second, the current, voltage and power factor data at the same time are included in the set, realizing the consistency of electrical data in time sequence.

[0056] Based on the set of each time sequence electrical response, multi-dimensional centralized trend analysis is performed to construct a reliable sequence of each time sequence electrical characteristic. For the multi-dimensional electrical parameters of each time sequence node, the mean and standard deviation of the parameters in multiple fitting states are calculated to determine the normal fluctuation range. For example, the mean of the current of a time sequence node is 110 A, the standard deviation is 10 A, and the reliable interval is set to 80-140 A. After removing abnormal values, the parameters conforming to the centralized trend are retained to form the reliable sequence of the electrical characteristics of the time sequence, ensuring that the sequence can represent the normal operation state of the electrical system.

[0057] According to the reliable sequence of each time sequence electrical characteristic, grid processing is performed. The time axis is used as the X axis, and each 1 minute is used as a grid unit. The reliable interval of the electrical parameter is used as the Y axis, such as 20 A for the current. The reliable sequence data of each time sequence node is mapped into the grid unit, the occurrence frequency and probability of the parameter in the range are recorded, and the electrical reliable grid space is generated to intuitively present the normal operation state distribution of the electrical system in different time and different parameter ranges.

[0058] Finally, the hydraulic reliable grid space, the mechanical reliable grid space, and the electrical reliable grid space are aligned in the time dimension to establish the association mapping of the parameters of each system at the same time node, for example, the corresponding relationship between the hydraulic pressure value, the mechanical vibration acceleration, and the electrical current value at a certain time node. Finally, the reliable state grid space covering the hydraulic system, the mechanical system, and the electrical system is integrated to realize the reliable representation of the overall operation state of the tunnel-anchor integrated machine.

[0059] By performing reliable analysis on the mechanical and electrical characteristics to construct the reliable grid space of each system, and integrating with the hydraulic reliable grid space, the reliable state grid space reflecting the normal operation state of each system of the tunnel-anchor integrated machine is formed, providing a unified and reliable reference benchmark for subsequent abnormal detection based on the space.

[0060] Further, step A230 in the method provided by the embodiment of the application includes: A231: hydraulic feature extraction is performed according to the multiple fitting state data sets to obtain multiple hydraulic response fitting sets.

[0061] A232: time sequence synchronization association is performed according to the multiple hydraulic response fitting sets to obtain each time sequence hydraulic response set.

[0062] A233: multi-dimensional centralized trend analysis is performed according to the each time sequence hydraulic response set to construct a reliable sequence of each time sequence hydraulic characteristic.

[0063] A234: grid processing is performed according to the each time sequence hydraulic characteristic reliable sequence to generate the hydraulic reliable grid space.

[0064] In the embodiments of the present application, the hydraulic response fitting set is obtained by extracting hydraulic features from a plurality of fitting state data sets. The fitting state data set contains the response data of key parameters such as pressure and flow rate of the hydraulic system, and can reflect the response characteristics of the hydraulic system under the corresponding working condition.

[0065] Optionally, when constructing the hydraulic credible grid space, first, hydraulic features are extracted from a plurality of fitting state data sets. The fitting state data set contains various operating parameters of the hydraulic system, from which key hydraulic feature parameters such as pressure, flow rate, and cylinder extension amount are screened out, and redundant information is removed by principal component analysis feature extraction algorithm to obtain a plurality of hydraulic response fitting sets. Each fitting set corresponds to a typical working condition, for example, the hydraulic response fitting set during tunneling contains time series data of pressure 18-22 MPa and flow rate 80-100 L / min, and the hydraulic response fitting set during anchoring contains time series data of pressure 20-24 MPa and flow rate 60-80 L / min, which ensures that each fitting set accurately reflects the response characteristics of the hydraulic system under the corresponding working condition.

[0066] Next, the plurality of hydraulic response fitting sets are time-synchronously associated. Since the sampling start time and interval of different fitting sets may differ, the time of triggering the operation action of the tunnel-anchor integrated machine is taken as the time reference, and the time series data of each fitting set is unified to the same time axis by timestamp alignment algorithm, and the sampling interval is unified to 50 ms. After synchronization, the hydraulic parameters at each time node can correspond to each other, forming each time series hydraulic response set, for example, at the 10th second, the set contains the pressure value, flow rate value and cylinder extension amount at the moment, realizing the time series consistency of hydraulic data under different working conditions.

[0067] Then, based on each time series hydraulic response set, multi-dimensional centralized trend analysis is performed to construct each time series hydraulic feature credible sequence. For the multi-dimensional hydraulic parameters at each time node, i.e., pressure, flow rate, and cylinder extension amount, the mean, median and standard deviation under a plurality of fitting states are calculated to determine the normal fluctuation range of the parameters. For example, the mean pressure of a certain time node is 20 MPa, and the standard deviation is 1 MPa, then 17-23 MPa is taken as the credible interval of the pressure of the node, and the abnormal values exceeding this range are removed, and the parameters conforming to the centralized trend are retained to form the hydraulic feature credible sequence of the time series, ensuring that the data in the sequence can represent the normal operating state of the hydraulic system.

[0068] Finally, the credible sequence of each time sequence hydraulic characteristic is grid processed. The time axis is taken as the X axis, the credible interval of the hydraulic parameter is taken as the Y axis, the credible sequence data of each time sequence node is mapped into a grid unit, each grid corresponds to a specific time interval and parameter range, for example, 1 minute is a grid unit on the X axis, and 2 MPa is an interval on the Y axis. The frequency and probability of the parameter in the range are recorded in the grid to form a hydraulic credible grid space covering the whole operation cycle, and the normal operation state distribution of the hydraulic system in different time and different parameter ranges is directly presented.

[0069] Through the above steps, from hydraulic feature extraction to grid processing, the grid space accurately reflecting the credible state of the hydraulic system is constructed, and a reliable reference benchmark is provided for subsequent hydraulic abnormality detection based on the space.

[0070] Further, the step A300 in the method provided by the embodiment of the application comprises: A310: node mapping is performed on the credible state grid space according to the operation node identifier to obtain a matching hydraulic grid space, a matching mechanical grid space and a matching electrical grid space.

[0071] A320: abnormality detection is performed on the hydraulic monitoring sequence according to the matching hydraulic grid space to obtain a hydraulic abnormality detection result.

[0072] A330: abnormality detection is performed on the mechanical monitoring sequence according to the matching mechanical grid space to obtain a mechanical abnormality detection result.

[0073] A340: abnormality detection is performed on the electrical monitoring sequence according to the matching electrical grid space to obtain an electrical abnormality detection result, and the three-element abnormality detection result is generated in combination with the hydraulic abnormality detection result and the mechanical abnormality detection result.

[0074] Specifically, when the credible state grid space is node mapped according to the operation node identifier, first, the operation node in which the tunnel-anchor integrated machine is currently located is determined, such as a tunneling node, an anchoring node or a transfer node, and each node corresponds to a specific device operating parameter range, as shown in Table 1.

[0075] Table 1: device operating parameter range table corresponding to each operation node of the tunnel-anchor integrated machine Based on these node characteristics, the subspaces matched with the current node parameter range are filtered from the credible state grid space, that is, the matching hydraulic grid space containing the credible interval of the hydraulic parameter under the node, the matching mechanical grid space containing the credible interval of the mechanical parameter under the node, and the matching electrical grid space containing the credible interval of the electrical parameter under the node, so as to realize accurate correspondence between the grid space and the operation node.

[0076] Next, according to the matching hydraulic grid space, the hydraulic monitoring sequence is detected for abnormality, and the real-time collected hydraulic monitoring data (such as pressure, flow, oil cylinder extension speed, etc.) is compared with the trusted parameter range in the matching hydraulic grid space one by one. For example, the pressure trusted interval of the tunneling node in the matching hydraulic grid space is 18-22 MPa, if the real-time monitoring pressure is 24.5 MPa at a certain time, and exceeds the range for 3 consecutive sampling periods (each period is 50 ms), it is determined that the hydraulic pressure is abnormal, and is recorded as pressure overrun in the hydraulic abnormality detection result.

[0077] According to the matching mechanical grid space, the mechanical monitoring sequence is detected for abnormality, and the real-time collected mechanical parameters such as tunneling arm vibration acceleration, anchor drilling mechanism speed, transmission gear temperature, etc. are compared with the trusted threshold in the matching mechanical grid space. Assuming that the vibration acceleration trusted interval of the anchoring node in the matching mechanical grid space is 0.5-1.5g, if the real-time monitoring vibration acceleration reaches 2.3g and lasts for 2 sampling periods, it is determined that the mechanical abnormality is mechanical abnormality, and the anchor mechanism vibration abnormality is marked in the mechanical abnormality detection result.

[0078] According to the matching electrical grid space, the electrical monitoring sequence is detected for abnormality, and the real-time collected electrical parameters (such as motor working current, control loop voltage, power factor, etc.) are compared with the standard range in the matching electrical grid space. For example, the motor current trusted interval of the transfer node in the matching electrical grid space is 80-110A, if the current suddenly increases to 140A and does not recover within 1 sampling period, it is determined that the electrical abnormality is electrical abnormality, and the transfer motor overcurrent is recorded in the electrical abnormality detection result. Then, the hydraulic, mechanical and electrical three types of abnormality detection results are integrated, the abnormal types, occurrence time and parameter deviation value of each system are determined, and the ternary abnormality detection result containing three types of abnormality information is generated.

[0079] Through the grid space matching guided by the working node identification and the sub-system abnormality detection, the precise identification and integration of the hydraulic, mechanical and electrical system abnormal states of the tunnel-anchor integrated machine are realized, which provides accurate basic data for the subsequent fault trend deduction based on the ternary abnormality detection result.

[0080] Further, the step A400 in the method provided by the embodiment of the application comprises: A410: classifying the set of historical fault events to obtain a hydraulic fault event set, a mechanical fault event set and an electrical fault event set.

[0081] A420: based on a fault trigger sensitive threshold, performing hydraulic fault trend deduction on the hydraulic abnormality detection result according to the hydraulic fault event set to generate the first fault deduction result.

[0082] A430: based on the fault trigger sensitive threshold, performing mechanical fault trend inference on the mechanical anomaly detection result according to the set of mechanical fault events, to generate the second fault inference result.

[0083] A440: based on the fault trigger sensitive threshold, performing electrical fault trend inference on the electrical anomaly detection result according to the set of electrical fault events, to generate the third fault inference result.

[0084] In the embodiments of the present application, the fault trigger sensitive threshold is a critical value for measuring the possibility of fault occurrence. In the fault trend inference of the hydraulic, mechanical and electrical systems, the fault inference path that is more likely to occur is selected by comparing with the fault trigger sensitive coefficient, so as to optimize the fault inference result.

[0085] Specifically, before classifying the historical fault event set of the combined tunneling and anchoring machine, a sufficient number of historical fault records are first collected, which cover various faults occurring since the device is put into use, including fault occurrence time, fault phenomenon, involved components, maintenance scheme and fault cause, etc. For example, collect fault events in the past 5 years to form a corresponding historical fault event set, and each record describes specific content such as sudden pressure drop of the hydraulic system during tunneling operation, which is caused by aging and damage of the oil cylinder sealing element, anchoring operation, anchor drill mechanism jamming, and the reason is excessive wear of the transmission gear, etc.

[0086] During the classification process, clear standards are formulated according to the system type to which the fault belongs: hydraulic fault events refer to faults related to the hydraulic system, such as oil cylinder leakage, insufficient hydraulic pump pressure, hydraulic pipeline blockage, etc.; mechanical fault events refer to faults of mechanical structures and transmission components, such as gear wear, bearing damage, tunneling arm deformation, etc.; electrical fault events involve electrical control systems and motors, such as motor overload, line short circuit, sensor failure, etc. All historical fault records are identified one by one according to the standards, for example, hydraulic valve jamming leading to slow action is classified into the hydraulic fault event set, transmission belt roller bearing burnout is classified into the mechanical fault event set, and control loop voltage instability causing shutdown is classified into the electrical fault event set.

[0087] After the classification is completed, each set needs to be verified by cross-checking the component attribution and system association in the fault maintenance record to ensure the accuracy of the classification. Finally, the hydraulic fault event set, the mechanical fault event set and the electrical fault event set are obtained, and the fault events in each set have clear system properties and characteristic descriptions.

[0088] Next, based on the fault trigger sensitive threshold, the hydraulic fault event set is subjected to fault top factor identification, direct and bottom factor tracing, and Boolean logic association to obtain a hydraulic fault deduction channel. The first fault deduction result is generated after optimization of the multiple paths obtained by inputting the hydraulic anomaly detection result, and the specific steps are described in detail in A421-A426.

[0089] Then, based on the mechanical fault event set, the mechanical fault trend is deduced. First, the type of the fault event in the set is identified. For example, the mechanical fault event set is analyzed by statistical analysis to extract four main fault types: gear wear, bearing damage, excavation arm deformation, and anchor drill mechanism jam, as the top layer factors of mechanical failure. Each top layer factor corresponds to dozens of specific event records.

[0090] Direct factor tracing is performed on each mechanical fault top layer factor. For example, for the top layer factor of gear wear, it is analyzed from the corresponding records that the three direct factors of insufficient lubrication, excessive load, and insufficient gear material strength are obtained; for bearing damage, the direct factors of installation deviation and excessive axial force are traced, and each top layer factor corresponds to several direct factors on average.

[0091] Further bottom factor tracing is performed on each direct factor. For example, the bottom factors of insufficient lubrication include insufficient lubricating oil filling amount, lubricating oil pipeline blockage, and long lubrication period; the bottom factors of excessive load include sudden increase in coal seam hardness and unreasonable operation parameter setting, and each direct factor corresponds to several bottom factors.

[0092] Boolean logic association is performed according to the mechanical fault top layer factor, direct factor and bottom factor to construct the mechanical fault deduction channel. For example, the trigger logic of the top layer factor of gear wear is: when the direct factor of insufficient lubrication and (the bottom factor of insufficient lubricating oil filling amount or the bottom factor of lubricating oil pipeline blockage) or the direct factor of excessive load and the bottom factor of sudden increase in coal seam hardness occur, a complete causal chain channel is formed.

[0093] The mechanical anomaly detection result is input into the mechanical fault deduction channel, and the abnormal features are matched with the channel factors to obtain multiple mechanical fault deduction paths such as coal seam hardness sudden increase → excessive load → gear wear, lubricating oil pipeline blockage → insufficient lubrication → gear wear, etc.

[0094] Trigger factor identification is performed on the multiple mechanical fault deduction paths to obtain trigger factors such as sudden increase in coal seam hardness and lubricating oil pipeline blockage. According to the mapping and comparison between the mechanical anomaly detection result and each trigger factor, the fault trigger sensitive coefficient is calculated: if the current gear vibration exceeds the standard degree and the similarity of the historical coal seam hardness sudden increase caused by wear case is 0.7 and the similarity of the lubricating oil pipeline blockage case is 0.6, the corresponding coefficient is obtained. Set the fault trigger sensitive threshold to 0.5, and select the paths with a coefficient ≥ 0.5 to generate the second fault deduction result.

[0095] After that, the electrical fault trend deduction process is the same as the mechanical. From the electrical fault event set, identify electrical fault top-level factors such as motor overload, line short circuit, sensor failure, etc. For motor overload, trace back the direct factors of excessive load and heat dissipation system failure, and then trace back the bottom-level factors of excessive load corresponding to sudden increase of work resistance, electrical parameter setting error, heat dissipation system failure corresponding to heat dissipation fan damage and heat dissipation hole blockage.

[0096] Build an electrical fault deduction channel through Boolean logic association, for example, motor overload (top-level factor) is triggered when the load is too large (direct factor) and the work resistance suddenly increases (bottom-level factor) or the electrical parameter setting is wrong (bottom-level factor). Input the electrical abnormality detection result into the channel to obtain multiple electrical fault deduction paths such as sudden increase of work resistance → excessive load → motor overload, heat dissipation fan damage → heat dissipation system failure → motor overload, etc.

[0097] Identify the triggering factors of each path, such as sudden increase of work resistance and heat dissipation fan damage, and map and compare them with the electrical abnormality detection result to calculate the sensitivity coefficient: for example, the similarity of the current exceeding the standard and the case of sudden increase of work resistance is 0.65, and the similarity of the current exceeding the standard and the case of heat dissipation fan damage is 0.55. Based on the threshold value 0.5 screening, the paths that meet the conditions are retained to generate the third fault deduction result.

[0098] Through the same hierarchical factor tracing, logical association and path screening method as the hydraulic fault trend deduction, the fault trend deduction of the mechanical and electrical systems is completed respectively, and the second and third fault deduction results generated accurately retain the high-possibility fault paths highly related to the abnormality detection result, providing systematic and reliable basic data for subsequent coupled fault prediction.

[0099] Further, step A420 in the method provided by the embodiment of the application comprises: A421: performing fault type identification according to the hydraulic fault event set to obtain each fault top-level factor.

[0100] A422: performing direct factor tracing on the each fault top-level factor according to the hydraulic fault event set to obtain each fault direct factor.

[0101] A423: performing bottom-level factor tracing on the each fault direct factor according to the hydraulic fault event set to obtain each fault bottom-level factor.

[0102] A424: performing Boolean logic association according to the each fault top-level factor, the each fault direct factor and the each fault bottom-level factor to obtain a hydraulic fault deduction channel.

[0103] A425: Input the hydraulic anomaly detection result into the hydraulic fault deduction channel to obtain multiple hydraulic fault deduction paths.

[0104] A426: Perform trigger-sensitive optimization on the multiple hydraulic fault deduction paths according to the fault trigger-sensitive threshold to obtain the first fault deduction result.

[0105] Specifically, when identifying fault types based on a collection of hydraulic fault events, the fault events within the collection are first summarized and representative fault categories are extracted as top-level factors. Taking the hydraulic fault event collection as an example, by statistically analyzing the frequency and common characteristics of each fault phenomenon, five major fault types were identified: cylinder leakage, insufficient hydraulic pump pressure, stuck hydraulic valve, pipe blockage, and excessive oil temperature. These were then determined as top-level factors for each fault. Each top-level factor corresponds to dozens of specific fault event records.

[0106] Next, when tracing the direct factors of each top-level fault factor, the direct causes of the top-level fault were extracted from the hydraulic fault event set. For example, for the top-level factor of cylinder leakage, analysis of dozens of corresponding fault records revealed three direct causes of leakage: seal aging and damage, cylinder barrel scratches, and seal misalignment during installation. These factors served as direct failure factors for this top-level factor. For the top-level factor of insufficient hydraulic pump pressure, direct factors such as wear of pump components, air leakage in the oil suction line, and inconsistent oil viscosity were traced. On average, each top-level factor corresponded to several direct factors.

[0107] Next, we traced the underlying factors of each direct failure factor to further uncover the root causes. For example, analysis of the direct factor of seal aging and damage revealed three underlying factors: insufficient heat resistance of the seal material, operating temperature continuously exceeding 35°C, and excessive impurities in the oil. For the direct factor of pump component wear, we traced underlying factors such as insufficient component material hardness, poor lubrication, and prolonged overload operation. Each direct factor typically corresponds to several underlying factors.

[0108] Next, when performing Boolean logic associations based on each fault's top-level factor, direct factor, and bottom-level factor, the causal relationship between the three must be clarified. For example, the logic for a top-level factor cylinder leak is triggered by aging and damage of the direct factor seal (the bottom-level factor seal material lacks heat resistance, the bottom-level factor operating environment temperature continuously exceeds 35°C, or the bottom-level factor oil contains excessive impurities), or when the direct factor cylinder barrel is scratched and foreign matter enters during bottom-level assembly. This logical association constructs a complete hydraulic fault deduction channel, each containing a causal chain from the bottom-level factor to the top-level factor.

[0109] The hydraulic abnormality detection result is then input into the hydraulic fault deduction channel, and the abnormal features are matched with the factors in the channel. For example, if the hydraulic abnormality detection result is that the pressure in the oil cylinder area suddenly drops and the oil temperature rises to 40°C during the tunneling stage, the abnormality will trigger the logic chain related to the aging and damage of the sealing element in the oil cylinder leakage deduction channel, generating multiple hydraulic fault deduction paths such as sealing element material temperature resistance deficiency→sealing element aging and damage→oil cylinder leakage, working environment temperature too high→sealing element aging and damage→oil cylinder leakage, and each path corresponds to a possible fault development process.

[0110] Finally, the hydraulic fault factors are obtained by identifying the trigger factors from the multiple hydraulic fault deduction paths, the fault trigger sensitive coefficients are compared with the hydraulic abnormality detection result, and the paths are screened based on the coefficients and the fault trigger sensitive threshold to generate the first fault deduction result. The specific steps are described in detail in A426-1-A426-3.

[0111] Through factor hierarchical tracing and logical association of the hydraulic fault event set, the hydraulic fault deduction channel is constructed, and multiple fault deduction paths are generated in combination with the hydraulic abnormality detection result, realizing systematic deduction from abnormal phenomena to fault root causes, and laying a foundation for subsequent path optimization based on the fault trigger sensitive threshold.

[0112] Further, the method provided in the embodiments of the present application includes the following steps A426. A426-1: According to the multiple hydraulic fault deduction paths, trigger factor identification is performed to obtain multiple hydraulic fault factors.

[0113] A426-2: According to the hydraulic abnormality detection result, the multiple hydraulic fault factors are respectively mapped and compared to obtain multiple fault trigger sensitive coefficients.

[0114] A426-3: Based on the multiple fault trigger sensitive coefficients, the multiple hydraulic fault deduction paths are screened according to the fault trigger sensitive threshold to generate the first fault deduction result.

[0115] In one embodiment, first, when trigger factor identification is performed according to the multiple hydraulic fault deduction paths, the key factors directly leading to fault occurrence are extracted from each path as hydraulic fault factors. For example, if there are three hydraulic fault deduction paths: sealing element material temperature resistance deficiency→sealing element aging and damage→oil cylinder leakage, working environment temperature too high→sealing element aging and damage→oil cylinder leakage, and oil impurity content exceeding standard→sealing element wear→oil cylinder leakage, three hydraulic fault factors of sealing element material temperature resistance deficiency, working environment temperature too high, and oil impurity content exceeding standard are identified, and each factor corresponds to the initial trigger condition in the path that causes the fault.

[0116] Next, a similarity comparison was performed on multiple hydraulic failure factors based on the hydraulic anomaly detection results. The specific parameters of the anomaly detection results were first identified, and then their similarity with each factor was measured through quantitative calculations. For example, the hydraulic anomaly detection results showed a sustained oil temperature of 38°C (exceeding the normal range of 35°C) in the cylinder area during excavation (exceeding the normal range of 35°C), seal wear of 0.6 (normal threshold ≤ 0.5), and oil impurity content of 0.1 mg / L (normal threshold ≤ 0.2 mg / L). The similarity calculation used normalization: For the excessive working environment temperature factor, the historical data associated with oil temperature exceeding the standard was 3-8°C. The current data exceeded the standard by 3°C, resulting in a similarity of 3 / 8 = 0.375. For the insufficient seal material temperature resistance factor, historically, this factor was often associated with seal wear exceeding the threshold by 0.1-0.3. The current data exceeded 0.1, resulting in a similarity of 0.1 / 0.3 = 0.333. For the excessive oil impurity factor, the current data showed no impurities exceeding the standard, resulting in a similarity of 0. Finally, three fault trigger sensitivity coefficients are obtained: 0.375, 0.333, and 0.

[0117] Finally, based on multiple fault trigger sensitivity coefficients, a reasonable threshold (such as 0.3) is set when optimizing and screening paths according to the fault trigger sensitivity threshold. By comparing each coefficient with the threshold, paths corresponding to excessively high operating temperature (0.375 ≥ 0.3) and insufficient seal material temperature resistance (0.333 ≥ 0.3) are retained, while paths corresponding to excessive oil impurities (0 < 0.3) are eliminated. This ultimately generates a first fault deduction result containing two highly likely paths.

[0118] Through trigger factor identification, quantitative similarity comparison and threshold screening, the fault deduction path that is highly correlated with the hydraulic anomaly detection results is accurately retained, effectively improving the pertinence and reliability of the first fault deduction result.

[0119] Furthermore, step A100 in the method provided in the embodiment of the present application includes: A110: Obtain a bolter mining machine monitoring data set according to the multi-sensor terminal.

[0120] A120: Clean and classify the bolter miner monitoring data set to generate the bolter miner monitoring sequence.

[0121] Optionally, in mining operations, real-time monitoring of the anchor miner requires data collection using a multi-sensor terminal. This terminal typically includes pressure sensors deployed in the miner's hydraulic lines, vibration sensors in the mechanical joints, and current sensors in the electrical control cabinet. These sensors collect data every 100 milliseconds at a preset frequency, forming a monitoring dataset for the miner. This dataset contains a variety of raw information, including hydraulic system pressure, mechanical structure vibration acceleration, and electrical system current.

[0122] After obtaining the monitoring data set of the tunneling and anchoring machine, cleaning and classification processing is required. In the cleaning process, abnormal values caused by transient sensor failure, such as invalid data with hydraulic pressure suddenly jumping to 50 MPa, are removed, and a small amount of missing values caused by signal interruption are filled by using the average of the adjacent 5 data to ensure the integrity and effectiveness of the data. In the classification stage, the cleaned data is divided into three categories according to the corresponding system attributes: pressure and flow data related to hydraulic pumps and hydraulic cylinders are classified into hydraulic monitoring sequences; vibration and displacement data related to tunneling arms and anchor drilling mechanisms are classified into mechanical monitoring sequences; current and voltage data related to drive motors and control systems are classified into electrical monitoring sequences.

[0123] Through the above steps, the original data is collected from the multi-sensor terminal, the invalid information is removed through cleaning, and the data is classified to determine the ownership, and finally the hydraulic, mechanical and electrical monitoring sequences with clear structure are generated, which provides accurate and classified basic data for subsequent anomaly detection based on the credible state grid space, and ensures the smooth development of the entire detection process.

[0124] Further, the step A500 in the method provided by the embodiment of the application comprises: A510: generating a tunneling and anchoring machine early warning signal according to the tunneling and anchoring machine detection graph.

[0125] In one embodiment, when generating the early warning signal according to the tunneling and anchoring machine detection graph, the core information in the graph needs to be analyzed first. The detection graph contains various coupling fault types, corresponding occurrence probabilities and associated system information, such as the coupling fault of hydraulic leakage-mechanical load imbalance-electrical overload and its occurrence probability, and the specific component correlation involved, such as hydraulic pipeline, tunneling arm and drive motor.

[0126] Then, multi-level early warning thresholds are set according to the severity and occurrence probability of the fault. Generally, three levels are divided according to the probability: the probability higher than 70% is the first level of early warning, which needs to be handled urgently, the probability of 50%-70% is the second level of early warning, which needs to be paid attention to, the probability of 30%-50% is the third level of early warning, which needs to be checked regularly, and the probability lower than 30% does not trigger the early warning. At the same time, combined with the influence degree of the fault on the equipment operation, the threshold is appropriately reduced for the faults that may cause shutdown or safety accidents (such as motor overload causing line short circuit), to ensure early warning.

[0127] The coupling fault probability in the tunneling and anchoring machine detection graph is compared with the early warning threshold to match the corresponding early warning level. If the coupling fault probability is 75% and involves the key power system, the first level of early warning is triggered; if the mechanical structure correlation fault probability is 60%, the second level of early warning is triggered. For the faults triggering the early warning, the core features, including the involved system, key components and possible development trend, are extracted as the core content of the early warning signal.

[0128] The generated early warning signal needs to be delivered through multiple channels, including sound and light alarm on the device operation panel, red sound and light for first-level early warning, yellow sound and light for second-level early warning, blue prompt for third-level early warning, real-time pop-up window of the central control system, mobile terminal push of the management personnel, etc. The signal clearly marks the early warning level, fault type, associated components and preliminary processing suggestions, for example, the first-level early warning signal displays: emergency: hydraulic leakage may cause motor overload, suggest checking the oil cylinder seal and drive motor immediately.

[0129] After the early warning signal is generated, the running log of the associated device is synchronized, and the early warning occurrence time, fault information and processing status are automatically recorded to provide a basis for subsequent tracing and analysis. At the same time, the system automatically adjusts the device running parameters according to the early warning level, such as appropriately reducing the tunneling speed during second-level early warning to reduce the risk of fault deterioration.

[0130] By analyzing the tunneling-anchor integrated machine detection graph, setting threshold comparison and generating multi-level and multi-channel early warning signals, timely identification and early warning of potential coupling faults of the tunneling-anchor integrated machine are realized, which saves processing time for the workers and effectively reduces the downtime loss and safety risk caused by fault expansion.

[0131] In summary, the method for detecting the running state of the tunneling-anchor integrated machine based on multi-sensor fusion provided by the embodiments of the present application has the following technical effects: The present application monitors the tunneling-anchor integrated machine in real time through a multi-sensor terminal, obtains a hydraulic, mechanical and electrical monitoring sequence with a work node identifier, performs abnormality detection under the guidance of the work node identifier to obtain a ternary abnormality detection result, performs multi-dimensional fault trend deduction in combination with a historical fault event set, introduces a coupling fault prediction model to obtain a tunneling-anchor machine detection graph, thereby accurately detecting the running state of the tunneling-anchor integrated machine, making the detection result of the running state of the device in the mining environment more comprehensive, accurate and reliable, achieving multi-source data fusion analysis of the tunneling-anchor integrated machine, improving the abnormality detection accuracy of the tunneling-anchor integrated machine, and realizing accurate fault prediction of the tunneling-anchor integrated machine.

[0132] Embodiment Two, as Figure 2 shown, based on the same inventive concept as the aforementioned Embodiment One, the embodiments of the present application provide a system for detecting the running state of a tunneling-anchor integrated machine based on multi-sensor fusion, which comprises: A tunneling-anchor machine monitoring sequence acquisition module 1 is configured to monitor the tunneling-anchor integrated machine in real time through a multi-sensor terminal, and obtain a tunneling-anchor machine monitoring sequence with a work node identifier, wherein the tunneling-anchor machine monitoring sequence comprises a hydraulic monitoring sequence, a mechanical monitoring sequence and an electrical monitoring sequence.

[0133] A trusted state grid space construction module 2 is configured to perform the following steps:

[0134] A ternary anomaly detection result acquisition module 3 is configured to perform the following steps:

[0135] A fault deduction result acquisition module 4 is configured to perform the following steps:

[0136] A drill-anchor machine detection atlas acquisition module 5 is configured to perform the following steps:

[0137] Further, the trusted state grid space construction module 2 is configured to perform the following steps: According to the three-dimensional modeling of the drill-anchor machine, a drill-anchor machine model is obtained; according to the drill-anchor operation control decision, the drill-anchor machine model is simulated and controlled multiple times to obtain multiple fitting state data sets; according to the multiple fitting state data sets, hydraulic feature trusted analysis is performed to construct a hydraulic trusted grid space; according to the multiple fitting state data sets, mechanical feature trusted analysis is performed to construct a mechanical trusted grid space; according to the multiple fitting state data sets, electrical feature trusted analysis is performed to construct an electrical trusted grid space; and according to the hydraulic trusted grid space, the mechanical trusted grid space, and the electrical trusted grid space, the trusted state grid space is generated.

[0138] Further, the trusted state grid space construction module 2 is configured to perform the following steps: According to the multiple fitting state data sets, hydraulic feature extraction is performed to obtain multiple hydraulic response fitting sets; according to the multiple hydraulic response fitting sets, time sequence synchronization correlation is performed to obtain each time sequence hydraulic response set; according to the each time sequence hydraulic response set, multi-dimensional centralized trend analysis is performed to construct each time sequence hydraulic feature trusted sequence; and according to the each time sequence hydraulic feature trusted sequence, grid processing is performed to generate the hydraulic trusted grid space.

[0139] Further, the ternary anomaly detection result acquisition module 3 is configured to perform the following steps: According to the job node identification, the credible state grid space is node-mapped to obtain a matching hydraulic grid space, a matching mechanical grid space and a matching electrical grid space; according to the matching hydraulic grid space, the hydraulic monitoring sequence is subjected to abnormality detection to obtain a hydraulic abnormality detection result; according to the matching mechanical grid space, the mechanical monitoring sequence is subjected to abnormality detection to obtain a mechanical abnormality detection result; according to the matching electrical grid space, the electrical monitoring sequence is subjected to abnormality detection to obtain an electrical abnormality detection result; the three abnormality detection results are combined with the hydraulic abnormality detection result and the mechanical abnormality detection result to generate the ternary abnormality detection result.

[0140] Further, the fault deduction result acquisition module 4 is configured to perform the following steps: The historical fault event set is classified to obtain a hydraulic fault event set, a mechanical fault event set and an electrical fault event set; based on a fault trigger sensitive threshold, the hydraulic abnormality detection result is subjected to hydraulic fault trend deduction according to the hydraulic fault event set to generate the first fault deduction result; based on the fault trigger sensitive threshold, the mechanical abnormality detection result is subjected to mechanical fault trend deduction according to the mechanical fault event set to generate the second fault deduction result; based on the fault trigger sensitive threshold, the electrical abnormality detection result is subjected to electrical fault trend deduction according to the electrical fault event set to generate the third fault deduction result.

[0141] Further, the fault deduction result acquisition module 4 is configured to perform the following steps: According to the hydraulic fault event set, fault type identification is performed to obtain each fault top-level factor; according to the hydraulic fault event set, direct factor tracing is performed on the each fault top-level factor to obtain each fault direct factor; according to the hydraulic fault event set, bottom factor tracing is performed on the each fault direct factor to obtain each fault bottom factor; according to the each fault top-level factor, the each fault direct factor and the each fault bottom factor, Boolean logic association is performed to obtain a hydraulic fault deduction channel; the hydraulic abnormality detection result is input into the hydraulic fault deduction channel to obtain a plurality of hydraulic fault deduction paths; according to the fault trigger sensitive threshold, trigger sensitive optimization is performed on the plurality of hydraulic fault deduction paths to obtain the first fault deduction result.

[0142] Further, the fault deduction result acquisition module 4 is configured to perform the following steps: According to the plurality of hydraulic fault deduction paths, trigger factor identification is performed to obtain a plurality of hydraulic fault factors; according to the hydraulic anomaly detection result, the plurality of hydraulic fault factors are respectively mapped and compared to obtain a plurality of fault trigger sensitive coefficients; based on the plurality of fault trigger sensitive coefficients, the plurality of hydraulic fault deduction paths are screened according to the fault trigger sensitive threshold optimization to generate the first fault deduction result.

[0143] Further, the excavator-anchor machine monitoring sequence acquisition module 1 is configured to perform the following steps: According to the plurality of sensing terminals, an excavator-anchor machine monitoring data set is obtained; the excavator-anchor machine monitoring data set is cleaned and classified to generate the excavator-anchor machine monitoring sequence.

[0144] Further, the excavator-anchor machine detection atlas acquisition module 5 is configured to perform the following steps: According to the excavator-anchor machine detection atlas, an excavator-anchor machine early warning signal is generated.

[0145] The excavator-anchor integrated machine operation state detection system based on multi-sensing fusion provided in the embodiments of the present application can perform the excavator-anchor integrated machine operation state detection method based on multi-sensing fusion provided in any embodiment of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0146] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module included is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy mutual differentiation, and do not limit the protection scope of the present application.

[0147] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, 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 the present application shall be included in the protection scope of the present application. In some cases, the actions or steps described in the present application can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A method for detecting the running state of a combined excavator-anchor machine based on multi-sensor fusion, characterized in that, The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises:

2. The multi-sensor fusion-based running state detection method for a combined anchor-excavation machine according to claim 1, characterized in that, The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises:

3. The method according to claim 2, wherein the method comprises the steps of: determining the running state of the integrated anchor and tunneling machine based on the fusion of the multi-sensor data. The method comprises: The method comprises: The method comprises: The method comprises: The method comprises:

4. The multi-sensor fusion-based running state detection method for a combined anchor-excavation machine according to claim 1, characterized in that, The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: The method comprises: According to the matching electrical grid space, the electrical monitoring sequence is subjected to anomaly detection to obtain an electrical anomaly detection result, and the hydraulic anomaly detection result and the mechanical anomaly detection result are combined to generate the ternary anomaly detection result.

5. The multi-sensor fusion-based running state detection method for a combined anchor-excavation machine according to claim 1, characterized in that, According to the historical fault event set of the combined tunneling-anchor machine, the ternary anomaly detection result is subjected to multi-dimensional fault trend deduction, including: The historical fault event set is classified to obtain a hydraulic fault event set, a mechanical fault event set, and an electrical fault event set; Based on a fault trigger sensitive threshold, the hydraulic anomaly detection result is subjected to hydraulic fault trend deduction according to the hydraulic fault event set to generate the first fault deduction result; Based on the fault trigger sensitive threshold, the mechanical anomaly detection result is subjected to mechanical fault trend deduction according to the mechanical fault event set to generate the second fault deduction result; Based on the fault trigger sensitive threshold, the electrical anomaly detection result is subjected to electrical fault trend deduction according to the electrical fault event set to generate the third fault deduction result.

6. The multi-sensor fusion-based running state detection method for a combined anchor-excavation machine according to claim 5, characterized in that, Based on a fault trigger sensitive threshold, the hydraulic anomaly detection result is subjected to hydraulic fault trend deduction according to the hydraulic fault event set to generate the first fault deduction result, including: According to the hydraulic fault event set, a fault type is identified to obtain each fault top-level factor; According to the hydraulic fault event set, each fault top-level factor is subjected to direct factor tracing to obtain each fault direct factor; According to the hydraulic fault event set, each fault direct factor is subjected to bottom-level factor tracing to obtain each fault bottom-level factor; According to the Boolean logic association of the each fault top-level factor, the each fault direct factor, and the each fault bottom-level factor, a hydraulic fault deduction channel is obtained; The hydraulic anomaly detection result is input into the hydraulic fault deduction channel to obtain a plurality of hydraulic fault deduction paths; According to the fault trigger sensitive threshold, the plurality of hydraulic fault deduction paths are subjected to trigger sensitive optimization to obtain the first fault deduction result.

7. The multi-sensor fusion-based running state detection method for a combined anchor-excavation machine according to claim 6, characterized in that, According to the fault trigger sensitive threshold, the plurality of hydraulic fault deduction paths are subjected to trigger sensitive optimization to obtain the first fault deduction result, including: According to the plurality of hydraulic fault deduction paths, a trigger factor is identified to obtain a plurality of hydraulic fault factors; According to the hydraulic anomaly detection result, the plurality of hydraulic fault factors are respectively subjected to mapping comparison to obtain a plurality of fault trigger sensitive coefficients; Based on the plurality of fault trigger sensitive coefficients, the plurality of hydraulic fault deduction paths are optimized and screened according to the fault trigger sensitive threshold to generate the first fault deduction result. 8.The method of claim 1, wherein, A combined tunneling-anchor machine is subjected to real-time monitoring by a multi-sensing terminal, including: According to the multi-sensing terminal, a tunnel-anchor machine monitoring data set is obtained; The tunnel-anchor machine monitoring data set is subjected to cleaning and classification to generate the tunnel-anchor machine monitoring sequence. 9.The multi-sensor fusion based running state detection method for the combined anchor-excavation machine according to claim 1, wherein, According to the tunnel-anchor machine detection map, a tunnel-anchor machine early warning signal is generated.

10. A running state detection system for a combined excavator-anchor machine based on multi-sensor fusion, characterized in that, A system for implementing the combined tunneling-anchor machine operation state detection method based on multi-sensing fusion according to any one of claims 1-9, the system comprising: The anchor machine monitoring sequence acquisition module (1) is used for real-time monitoring of the anchor machine by a multi-sensing terminal, and obtaining an anchor machine monitoring sequence with a work node identifier, wherein the anchor machine monitoring sequence comprises a hydraulic monitoring sequence, a mechanical monitoring sequence and an electrical monitoring sequence; The trusted state grid space construction module (2) is used for fitting and analyzing the trusted state characteristics of the anchor machine according to an anchor work control decision, and constructing a trusted state grid space; The ternary anomaly detection result acquisition module (3) is used for guiding the trusted state grid space to perform anomaly detection on the anchor machine monitoring sequence according to the work node identifier, and obtaining a ternary anomaly detection result; The fault deduction result acquisition module (4) is used for performing multi-dimensional fault trend deduction on the ternary anomaly detection result according to a historical fault event set of the anchor machine, and obtaining a first fault deduction result, a second fault deduction result and a third fault deduction result; The anchor machine detection atlas acquisition module (5) is used for introducing an anchor machine coupling fault prediction model to perform coupling fault prediction on the first fault deduction result, the second fault deduction result and the third fault deduction result, and obtaining an anchor machine detection atlas.

Images