Intelligent fault diagnosis method and system for main driving device of shield tunneling machine

Abstract

The invention discloses an intelligent fault diagnosis method and system for a main driving device of a shield tunneling machine. The method comprises the following steps: acquiring multi-source monitoring information of the main driving device of the shield tunneling machine; feature indexes are extracted based on the multi-source monitoring information, and corresponding multi-source features are obtained through normalization processing; performing feature fusion based on the multi-source features to obtain fused features; inputting the fusion features into a trained deep belief network to obtain a fault category; according to the method, adaptive fusion of multi-source monitoring data can be realized, fault category judgment is realized in combination with the deep belief network, the accuracy of fault recognition is improved, more sufficient data can be mastered in the early stage, and early warning is further realized.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and more specifically to an intelligent fault diagnosis method and system for the main drive device of a tunnel boring machine. Background Technology

[0002] In the field of condition monitoring and fault diagnosis of tunnel boring machine (TBM) main drive systems, existing technologies mainly face the dilemma of limited monitoring methods and insufficient diagnostic reliability. Traditional diagnostic schemes often rely solely on the analysis of single-type signals such as vibration, current, or hydraulic pressure. Due to the complex and harsh operating conditions of TBMs, single signals are easily interfered with and cannot comprehensively reflect the system's health status, easily leading to false alarms or missed alarms. Although some technologies attempt to collect multi-source signals, most remain at the level of simple data superposition or threshold comparison, lacking the ability to effectively unify and deeply integrate heterogeneous signals such as hydraulic, electrical, and vibration signals, and failing to delve into the complex mapping relationship between faults and multi-source information. Furthermore, existing diagnostic methods based on traditional machine learning models have limited ability to represent complex nonlinear faults, and their performance is highly dependent on complete historical fault data. In practical applications, it is difficult to obtain sufficient samples, resulting in insufficient model generalization ability and reasoning ability for unknown faults, thus restricting the accuracy of diagnosis and early warning effectiveness.

[0003] Therefore, how to improve the accuracy of fault detection and achieve early warning is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method and system for intelligent diagnosis of faults in the main drive device of a tunnel boring machine that overcomes or at least partially solves the above problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A method for intelligent fault diagnosis of the main drive unit of a tunnel boring machine includes the following steps: Acquire multi-source monitoring information of the main drive unit of the tunnel boring machine; Based on the multi-source monitoring information, feature indicators are extracted, and corresponding multi-source features are obtained through normalization processing. Based on the multi-source features, feature fusion is performed using adaptive weights to obtain fused features; The fused features are input into a trained deep belief network to obtain diagnostic results.

[0007] Preferably, the steps further include: performing a secondary assessment of the confidence level of the diagnostic results through fuzzy Bayesian inference.

[0008] Preferably, the steps of the fuzzy Bayes inference include: The fuzzy membership degree is calculated based on the preset membership function and the fault probability distribution in the diagnostic results; A Bayesian confidence matrix is ​​constructed based on the prior probability of occurrence of various faults, and the corrected confidence value is obtained by jointly calculating the fuzzy membership degree and the prior probability.

[0009] Preferably, a multi-source feature fusion model is established to perform the feature fusion, and the multi-source feature fusion model is as follows:

[0010] in, , and For different adaptive weight coefficients, Characteristics of hydraulic signals, Characteristics of current signals, These are characteristics of vibration signals.

[0011] Preferably, the extracted feature indicators include:

[0012] Where X is the characteristic index. This represents the mean. It is the standard deviation; The root mean square value represents the overall energy level of the signal. It is kurtosis, used to indicate the steepness of the distribution shape of a signal waveform; It represents the wavelet packet energy, reflecting the energy distribution of the signal in a specific frequency band; Peak frequency is used to represent the frequency component with the highest energy in the spectrum.

[0013] A fault intelligent diagnosis system for the main drive unit of a tunnel boring machine includes: The data acquisition module is used to acquire multi-source monitoring information of the main drive unit of the tunnel boring machine; The feature extraction module is used to extract feature indicators based on the multi-source monitoring information and obtain the corresponding multi-source features through normalization processing; The feature fusion module performs feature fusion based on the multi-source features to obtain fused features; it adopts an improved weighted principal component analysis and multi-domain feature fusion model to uniformly map the feature vectors from hydraulic, electrical and vibration signals to a low-dimensional feature space. The fault identification module is used to determine the fault category based on the fused features using a trained deep belief network.

[0014] Preferably, the data acquisition module includes a pressure sensor, a flow sensor, a vibration acceleration sensor, and a main motor current sensor.

[0015] Preferably, it also includes a fuzzy inference module; the fuzzy inference module is used to perform a secondary evaluation of the confidence of the diagnostic results through fuzzy Bayesian inference.

[0016] Preferably, the fuzzy inference module includes: The inference configuration submodule is used to set the fuzzy level and configure the corresponding membership function for each fuzzy level; it is also used to obtain the prior probability of various faults and construct the Bayesian confidence matrix. The confidence correction submodule is used to receive the diagnostic results and confirm the fuzzy membership degree according to the membership function; and to calculate the corrected confidence value according to the fuzzy membership degree and the Bayesian confidence matrix.

[0017] Preferably, it also includes an early warning module, which is used to determine the confidence level of the fault category and issue an early warning when the confidence level exceeds a preset threshold.

[0018] As can be seen from the above technical solution, compared with the prior art, the present invention discloses an intelligent diagnostic system and method for the main drive device of a tunnel boring machine. By adopting an improved weighted principal component analysis (WPCA) and a multi-domain feature fusion model, heterogeneous feature vectors from hydraulic, electrical and vibration sources are uniformly mapped to a low-dimensional space, thereby achieving effective fusion of multi-source information at the feature layer and overcoming the shortcomings of traditional methods, such as shallow information fusion level and difficulty in collaborative analysis. At the decision layer, deep belief networks (DBN) and fuzzy Bayes inference are combined to perform multi-classification decision-making, which enhances the representation and reasoning ability of complex nonlinear fault modes, reduces the dependence on complete historical fault data, and improves the generalization and accuracy of the diagnostic model. Ultimately, it achieves accurate and early warning of typical faults such as hydraulic system leakage, motor overload, and bearing wear. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of an intelligent fault diagnosis method for the main drive system of a tunnel boring machine provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of a fault intelligent diagnosis system for the main drive device of a tunnel boring machine provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1 like Figure 1 This invention discloses an intelligent fault diagnosis method for the main drive device of a tunnel boring machine, comprising the following steps: S1: Acquire multi-source monitoring information of the tunnel boring machine's main drive unit; S2: Extract feature indicators based on multi-source monitoring information and obtain the corresponding multi-source features through normalization processing; S3: Feature fusion based on multi-source features to obtain fused features; S4: Input the fused features into the trained deep belief network to obtain the fault category.

[0023] In one implementation, the multi-source monitoring information acquired in S1 includes data such as pressure, current, vibration, and rotational speed; the collected data needs to be preprocessed before extracting feature indicators.

[0024] Specifically, this includes data synchronization, noise filtering, normalization, and temporal resampling. Furthermore, at the edge, simple feature extraction can be performed, such as mean, variance, and envelope spectral energy, to reduce the amount of data transmitted to the host computer.

[0025] In one implementation, the fault types include typical faults such as hydraulic system leakage, motor overload, bearing wear, main reducer jamming, and sensor drift.

[0026] To further implement the above technical solution, in S4, the deep belief network is also used to generate the confidence level of the fault category, and to issue an early warning when the confidence level exceeds a preset threshold.

[0027] To further implement the above technical solution, a multi-source feature fusion model is established in S3 for feature fusion. The multi-source feature fusion model is as follows:

[0028] in, , and For different adaptive weight coefficients, Characteristics of hydraulic signals, Characteristics of current signals, These are characteristics of vibration signals.

[0029] Furthermore, the characteristic indicators include:

[0030] Wherein, X is a characteristic index, which in this embodiment includes the hydraulic signal characteristics, current signal characteristics and vibration signal characteristics mentioned above; This represents the mean. It is the standard deviation; The root mean square value represents the overall energy level of the signal. It is kurtosis, used to indicate the steepness of the distribution shape of a signal waveform; It represents the wavelet packet energy, reflecting the energy distribution of the signal in a specific frequency band; Peak frequency is used to represent the frequency component with the highest energy in the spectrum.

[0031] In this embodiment, the adaptive weight coefficients of various signal features are obtained through principal component analysis. During the acquisition process, principal component analysis is used to quantitatively evaluate the data integrity, fluctuation stability, and correlation with typical fault modes. Features with higher signal quality and greater sensitivity to fault changes receive higher weights in the principal component analysis process, while features with higher signal noise or weaker fault indication capabilities have their weights reduced accordingly. This allows the weights to be dynamically adjusted according to changes in operating conditions and data characteristics.

[0032] To further implement the above technical solution, the steps also include: S5: Performing a secondary evaluation of the confidence level of the diagnostic results using fuzzy Bayesian inference. The steps of fuzzy Bayesian inference include: S51: Calculate the fuzzy membership degree based on the preset membership function and the fault probability distribution in the diagnostic results.

[0033] S52: Construct a Bayesian confidence matrix based on the prior probability of occurrence of various faults, and obtain the corrected confidence value through the joint calculation of the fuzzy membership degree and the prior probability.

[0034] Specifically, firstly, the fault probability values ​​output by the deep belief network are converted into fuzzy membership degrees, and triangular or trapezoidal membership functions are used to describe the three fuzzy levels of "low confidence," "medium confidence," and "high confidence." Then, a Bayesian confidence matrix is ​​constructed based on the prior probability of each type of fault. The corrected confidence value is obtained through the joint calculation of the fuzzy membership degrees and prior probabilities. This corrected confidence value is given by the following formula:

[0035] in, A fault output by a deep belief network Fuzzy membership degree, This corresponds to the prior probability of the fault. Finally, the corrected confidence level is compared with a set threshold, such as... If a clear warning is issued, the status will be considered suspicious and monitoring will continue. This fuzzy Bayesian secondary evaluation effectively corrects identification biases caused by imbalanced samples or weak fault features, improving the reliability of fault diagnosis results.

[0036] Example 2 like Figure 2 Based on the same inventive concept, this invention discloses an intelligent fault diagnosis system for the main drive unit of a tunnel boring machine, comprising: The data acquisition module is used to acquire multi-source monitoring information of the main drive unit of the tunnel boring machine; The feature extraction module is used to extract feature indicators based on the multi-source monitoring information and obtain the corresponding multi-source features through normalization processing; The feature fusion module performs feature fusion based on the multi-source features to obtain fused features; it adopts an improved weighted principal component analysis and multi-domain feature fusion model to uniformly map the feature vectors from hydraulic, electrical and vibration signals to a low-dimensional feature space. The fault identification module is used to determine the fault category based on the fused features using a trained deep belief network.

[0037] To further implement the above technical solution, an edge preprocessing module is also included. The edge preprocessing module completes data synchronization, noise filtering, normalization and temporal resampling through an embedded acquisition unit, and performs simple feature extraction, such as mean, variance, envelope spectrum energy, etc., to reduce the amount of data transmission to the host computer.

[0038] To further implement the above technical solution, the data acquisition module includes a pressure sensor, a flow sensor, a temperature sensor, a vibration acceleration sensor, and a main motor current and speed detection unit. To further implement the above technical solution, a fuzzy inference module is also included; the fuzzy inference module is used to perform a secondary evaluation of the confidence of the diagnostic results through fuzzy Bayesian inference.

[0039] To further implement the above technical solution, an early warning module is also included. The early warning module is used to determine the confidence level of the fault category and to issue an early warning when the confidence level exceeds a preset threshold.

[0040] Furthermore, it also includes a human-computer interaction and visualization module, which, based on the DeskSim or LabVIEW interface, displays the trends, diagnostic results, and confidence indices of each monitoring signal in real time; and can upload the confidence judgment results to the shield tunnel remote monitoring center via TCP / IP protocol to achieve remote early warning and expert review.

[0041] This invention utilizes a dual deep fusion mechanism of feature layer and decision layer to perform collaborative analysis and intelligent reasoning of traditionally independent hydraulic, electrical and vibration signals. This fundamentally overcomes the drawbacks of poor reliability and susceptibility to interference in diagnosis from a single signal source, and significantly improves the diagnostic accuracy for complex faults such as hydraulic system leakage, motor overload, and bearing wear.

[0042] Based on the intelligent model of improved weighted principal component analysis (WPCA) and deep belief network (DBN), it is possible to deeply mine the subtle features and early signs of faults from multi-source signals, realize early warning of potential faults, and provide valuable time for preventive maintenance.

[0043] The decision-level fusion method combining deep belief networks and fuzzy Bayesian inference enhances the ability to represent and reason about complex nonlinear fault modes, reduces the dependence on a large amount of complete historical fault data, and enables the system to maintain stable and reliable diagnostic performance even when the data is incomplete or when facing new operating conditions.

[0044] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0045] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for intelligent fault diagnosis of the main drive unit of a tunnel boring machine, characterized in that, Includes the following steps: Acquire multi-source monitoring information of the main drive unit of the tunnel boring machine; Based on the multi-source monitoring information, feature indicators are extracted, and corresponding multi-source features are obtained through normalization processing. Based on the multi-source features, feature fusion is performed using adaptive weights to obtain fused features; The fused features are input into a trained deep belief network to obtain diagnostic results.

2. The intelligent fault diagnosis method for the main drive system of a tunnel boring machine according to claim 1, characterized in that, The steps also include: conducting a secondary assessment of the confidence level of the diagnostic results through fuzzy Bayesian inference.

3. The intelligent fault diagnosis method for the main drive system of a tunnel boring machine according to claim 2, characterized in that, The steps of the fuzzy Bayes inference include: The fuzzy membership degree is calculated based on the preset membership function and the fault probability distribution in the diagnostic results; A Bayesian confidence matrix is ​​constructed based on the prior probability of occurrence of various faults, and the corrected confidence value is obtained by jointly calculating the fuzzy membership degree and the prior probability.

4. The intelligent fault diagnosis method for the main drive device of a tunnel boring machine according to claim 1, characterized in that, A multi-source feature fusion model is established to perform the feature fusion. The multi-source feature fusion model is as follows: in, , and For different adaptive weight coefficients, Characteristics of hydraulic signals, Characteristics of current signals, These are characteristics of vibration signals.

5. A method for intelligent fault diagnosis of the main drive device of a tunnel boring machine according to claim 1 or 4, characterized in that, The extracted feature indicators include: Where X is the characteristic index. This represents the mean. It is the standard deviation; The root mean square value represents the overall energy level of the signal. It is kurtosis, used to indicate the steepness of the distribution shape of a signal waveform; It represents the wavelet packet energy, reflecting the energy distribution of the signal in a specific frequency band; Peak frequency is used to represent the frequency component with the highest energy in the spectrum.

6. A fault intelligent diagnosis system for the main drive unit of a tunnel boring machine, characterized in that, include: The data acquisition module is used to acquire multi-source monitoring information of the main drive unit of the tunnel boring machine; The feature extraction module is used to extract feature indicators based on the multi-source monitoring information and obtain the corresponding multi-source features through normalization processing; The feature fusion module performs feature fusion based on the multi-source features to obtain fused features; An improved weighted principal component analysis and multi-domain feature fusion model is adopted to uniformly map feature vectors from hydraulic, electrical and vibration signals to a low-dimensional feature space; The fault identification module is used to determine the fault category based on the fused features using a trained deep belief network.

7. The intelligent fault diagnosis system for the main drive unit of a tunnel boring machine according to claim 6, characterized in that, The data acquisition module includes a pressure sensor, a flow sensor, a vibration acceleration sensor, and a main motor current sensor.

8. The intelligent fault diagnosis system for the main drive unit of a tunnel boring machine according to claim 6, characterized in that, It also includes a fuzzy inference module; the fuzzy inference module is used to perform a secondary evaluation of the confidence of the diagnostic results through fuzzy Bayesian inference.

9. A fault intelligent diagnosis system for the main drive unit of a tunnel boring machine according to claim 6, characterized in that, The fuzzy reasoning module includes: The inference configuration submodule is used to set the fuzzy level and configure the corresponding membership function for each fuzzy level; it is also used to obtain the prior probability of various faults and construct the Bayesian confidence matrix. The confidence correction submodule is used to receive the diagnostic results and confirm the fuzzy membership degree according to the membership function; and to calculate the corrected confidence value according to the fuzzy membership degree and the Bayesian confidence matrix.

10. A fault intelligent diagnosis system for the main drive unit of a tunnel boring machine according to claim 6, characterized in that, It also includes an early warning module, which is used to determine the confidence level of the fault category and issue an early warning when the confidence level exceeds a preset threshold.

Images