Ship system fault diagnosis method and device based on multi-source information fusion

Abstract

The invention discloses a ship system fault diagnosis method and device based on multi-source information fusion, and belongs to the field of prediction and health management, and the method comprises the steps: collecting monitoring data including voltage, current, vibration acceleration, rotating speed, displacement and noise; preprocessing the monitoring data, respectively inputting each type of preprocessed monitoring data into a single parameter fault diagnosis model, and outputting a single fault diagnosis result; analyzing the plurality of single fault diagnosis results according to a multi-source information fusion diagnosis rule to obtain a fusion diagnosis result; and according to a fusion diagnosis result, determining a fault occurrence part and a fault property. According to the invention, multi-source and multi-dimensional parameters are collected to carry out fault diagnosis on a ship key system and equipment, so that the detection accuracy and credibility are improved.

Description

Technical Field

[0001] The present invention belongs to the field of prediction and health management (PHM) technology, and more specifically, relates to a ship system fault diagnosis method and device based on multi-source information fusion. Background Art

[0002] Ships and other equipment are generally complex systems composed of mechanical, electrical, hydraulic and gas systems. They have complex structures and working principles, and usually have complex fault mechanisms. Existing fault knowledge is very limited, and typical fault samples are insufficient. At the same time, the causes and symptoms of ship equipment failures are not necessarily a definite mapping relationship. Due to the complexity of the diagnosis object, the limitations of the test methods, and the inaccuracy of knowledge expression, the mapping relationship between the fault symptoms and the fault causes is uncertain. To achieve fault diagnosis, correct information acquisition is an important part. Due to many factors, the sample data obtained is often incomplete, inaccurate and multi-source heterogeneous. Therefore, the key to equipment fault diagnosis is to use inference diagnosis of multi-source heterogeneous information under uncertain information conditions. In the actual development of algorithms related to fault diagnosis and health management, these data need to be integrated to ensure the full utilization of data assets of complex equipment systems. Making timely and accurate judgments on the faults that occur, identifying the fault locations, finding the fault causes and troubleshooting methods can greatly reduce the blindness of repairs and improve economy and safety.

[0003] Fault detection of ship systems mostly relies on manual experience and single sensor data monitoring, and fusion analysis and diagnosis remain at the same multi-dimensional parameter level, rarely introducing multiple parameters and other equipment parameters for fusion analysis. This makes fault detection still have many problems such as high redundancy, inconsistency and incompleteness of diagnostic parameters, which affects the accuracy and reliability of fault diagnosis. Summary of the invention

[0004] In view of the above defects or improvement needs of the prior art, the present invention proposes a ship system fault diagnosis method and device based on multi-source information fusion, which improves the detection accuracy and reliability by collecting multi-source and multi-dimensional parameters to perform fault diagnosis of key ship systems and equipment.

[0005] To achieve the above object, according to one aspect of the present invention, a ship system fault diagnosis method based on multi-source information fusion is provided, comprising:

[0006] Collect monitoring data including voltage, current, vibration acceleration, speed, displacement and noise;

[0007] Preprocess the monitoring data, input each type of preprocessed monitoring data into a single parameter fault diagnosis model, and output a single fault diagnosis result. The single parameter fault diagnosis model is trained based on a machine learning model using labeled massive ship system monitoring data;

[0008] Analyze multiple single fault diagnosis results using multi-source information fusion diagnosis rules to obtain fusion diagnosis results, wherein the fusion diagnosis rules are constructed based on DS evidence theory;

[0009] Based on the fusion diagnosis results, determine the location and nature of the fault.

[0010] In some optional implementation schemes, the single parameter fault diagnosis model includes: using a long short-term memory network LSTM as the core diagnosis model, and combining deep learning technology to construct a fault diagnosis model framework; extracting features related to fault diagnosis for each type of monitoring parameter, training the deep learning model in combination with historical fault data sets, and constructing the final single parameter fault diagnosis model.

[0011] In some optional implementation schemes, the multiple single fault diagnosis results are analyzed using multi-source information fusion diagnosis rules to obtain a fusion diagnosis result, including:

[0012] According to the DS combination rule, the individual diagnostic results are fused layer by layer, and the fused diagnostic result is obtained.

[0013] According to another aspect of the present invention, a ship system fault diagnosis device based on multi-source information fusion is provided, comprising:

[0014] Data receiving module, used to collect monitoring data including voltage, current, vibration acceleration, speed, displacement and noise;

[0015] A data preprocessing module is used to preprocess the monitoring data;

[0016] The feature extraction and model training module is used to input various types of pre-processed monitoring data into a single parameter fault diagnosis model and output a single fault diagnosis result. The single parameter fault diagnosis model is trained based on a machine learning model using labeled massive ship system monitoring data;

[0017] A fusion diagnosis module is used to analyze multiple single fault diagnosis results using multi-source information fusion diagnosis rules to obtain fusion diagnosis results, wherein the fusion diagnosis rules are constructed based on DS evidence theory;

[0018] The evaluation and judgment module is used to determine the location and nature of the fault based on the fusion diagnosis results.

[0019] In some optional implementation schemes, the single parameter fault diagnosis model includes: using a long short-term memory network LSTM as the core diagnosis model, and combining deep learning technology to construct a fault diagnosis model framework; extracting features related to fault diagnosis for each type of monitoring parameter, training the deep learning model in combination with historical fault data sets, and constructing the final single parameter fault diagnosis model.

[0020] In some optional implementation schemes, the fusion diagnosis module is used to fuse the individual diagnosis results layer by layer in pairs according to the DS combination rule to obtain a fusion diagnosis result.

[0021] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0022] The present invention aims to improve the safety of ship system operation, ensure the safe operation of each component of the system, and significantly improve the accuracy and reliability of fault detection through multi-source information fusion technology. It can achieve early fault warning and allow maintenance measures to be taken in time, thereby reducing the occurrence of failures, extending the service life of equipment, and reducing maintenance costs. In addition, the present invention also aims to improve maintenance efficiency and optimize the maintenance decision-making process. In general, the present invention will promote the transformation of ship maintenance work to intelligent and predictive maintenance, and contribute to the realization of smart ship equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic flow chart of a ship system fault diagnosis method provided by an embodiment of the present invention;

[0024] Figure 2 It is a schematic diagram of a multi-source information fusion fault diagnosis device applicable to a ship system provided by an embodiment of the present invention;

[0025] in,

[0026] DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0028] The ship system fault diagnosis method and device of the present invention are described in detail below. The execution subject of the method is a ship health management system.

[0029] The present invention provides a ship system fault diagnosis method based on multi-source information fusion, which is used to solve the problems of missed diagnosis and misdiagnosis that are easily caused by relying on manual experience judgment and single parameter diagnosis in the existing ship system fault diagnosis method. The use of a fusion diagnosis strategy avoids the high redundancy and inconsistency defects in the use of multi-source heterogeneous monitoring parameters, thereby improving the accuracy and reliability of fault diagnosis.

[0030] like Figure 1 As shown, the present invention provides a ship system multi-source information fusion fault diagnosis method, comprising:

[0031] Collect monitoring data through various types of sensors installed on the system and its equipment, including but not limited to voltage, current, vibration acceleration, speed, displacement, noise and other parameters;

[0032] Clean and filter the collected data to remove noise and outlier interference;

[0033] The monitoring data are respectively input into a fault diagnosis model of a single monitoring parameter, and a single fault diagnosis result is output, wherein the single parameter fault diagnosis model is obtained by training based on a machine learning model using labeled massive monitoring data of ship systems;

[0034] The diagnosis results of multiple deep learning models are analyzed with multi-source information fusion diagnosis rules to obtain fusion diagnosis conclusions and give fusion diagnosis reports, where the fusion diagnosis rules are constructed based on DS evidence theory;

[0035] Based on the results of fusion diagnosis, the location and nature of the fault are determined.

[0036] Furthermore, sensor selection and deployment include:

[0037] (1) Select sensor type: Select the appropriate sensor based on the common faults and key performance indicators of the ship system and its equipment, including but not limited to voltage sensors, current sensors, vibration acceleration sensors, speed sensors, displacement sensors, noise sensors, etc.

[0038] (2) Determination of deployment location: Analyze the principles and operating mechanisms of each system and its equipment, and determine key monitoring locations such as oil pumps, oil pipelines, and control valves.

[0039] (3) Install sensors: Install sensors at the above pre-selected locations to ensure that they can work stably and collect the required monitoring data regularly or continuously.

[0040] Furthermore, data preprocessing includes:

[0041] Configure a certain filtering algorithm to clean the signals collected by each sensor, mainly including pre-processing operations such as denoising, outlier removal, filtering or signal enhancement.

[0042] Furthermore, feature extraction and model training include:

[0043] (1) Select a suitable deep learning model, such as the long short-term memory network (LSTM), and combine it with other deep learning methods such as convolutional neural network (CNN) and deep neural network (DNN).

[0044] (2) Use labeled massive ship system monitoring data to train the selected deep learning model and extract features related to fault diagnosis.

[0045] (3) Construct a single fault diagnosis model and optimize the model parameters.

[0046] Furthermore, the fusion diagnostic analysis includes:

[0047] (1) The output results of the single fault diagnosis model are taken as input, and the fusion rules based on DS evidence theory are used to perform multi-source information fusion diagnosis.

[0048] (2) Post-process the fused results to improve the consistency and credibility of the diagnostic results.

[0049] The specific steps are as follows:

[0050] Input the monitoring data into the fault diagnosis model of a single monitoring parameter and output the single fault diagnosis result m i , i = 1, 2, ..., n, i represents the output diagnosis results of various monitoring parameters under the same ship system;

[0051] According to the DS combination rule, the single diagnosis result m i Perform layer-by-layer pairwise fusion, the rules are as follows:

[0052]

[0053] In the formula, A i and B i is the parameter in each parameter diagnosis result, and finally the fusion diagnosis result m is obtained.

[0054] Furthermore, fault location and nature determination include:

[0055] (1) Based on the results of fusion diagnosis, determine the location and nature of the fault.

[0056] (2) Output a fusion diagnosis report, providing detailed information on fault location and nature determination.

[0057] like Figure 2 As shown, the present invention also provides a ship system multi-source information fusion fault diagnosis device 8, comprising:

[0058] The data receiving module 9 receives monitoring parameters from various types of sensors arranged on the system and its equipment;

[0059] Among them, monitoring data is collected through various types of sensors installed on the system and its equipment, that is, multi-measurement point parameters 1 include but are not limited to voltage 2, current 3, vibration acceleration 4, speed 5, oil 6, displacement, noise 7 and other types of parameters.

[0060] The data preprocessing module 10 cleans and normalizes the collected data to remove noise interference;

[0061] The feature extraction and model training module 11 uses a deep learning model to extract features of each type of monitoring parameters, uses a historical fault data set to train the deep learning model, and optimizes the model parameters;

[0062] The fusion diagnosis module 12 fuses the diagnosis results of each deep learning model based on the constructed fusion rules, and performs post-processing on the fused results to improve the consistency and credibility of the diagnosis results;

[0063] The evaluation and determination module 13 outputs the location and nature of the fault according to the result of the fusion diagnosis.

[0064] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, and two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0065] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A ship system fault diagnosis method based on multi-source information fusion, characterized in that: include: Collect monitoring data including voltage, current, vibration acceleration, speed, displacement and noise; Preprocess the monitoring data, input each type of preprocessed monitoring data into a single parameter fault diagnosis model, and output a single fault diagnosis result. The single parameter fault diagnosis model is trained based on a machine learning model using labeled massive ship system monitoring data; Analyze multiple single fault diagnosis results using multi-source information fusion diagnosis rules to obtain fusion diagnosis results, wherein the fusion diagnosis rules are constructed based on DS evidence theory; Based on the fusion diagnosis results, determine the location and nature of the fault.

2. The method according to claim 1, characterized in that The single parameter fault diagnosis model includes: using a long short-term memory network (LSTM) as a core diagnosis model, and combining deep learning technology to build a fault diagnosis model framework; extracting features related to fault diagnosis for each type of monitoring parameter, training the deep learning model in combination with historical fault data sets, and building a final single parameter fault diagnosis model.

3. The method according to claim 2, characterized in that The multiple single fault diagnosis results are analyzed using multi-source information fusion diagnosis rules to obtain fusion diagnosis results, including: According to the DS combination rule, the individual diagnostic results are fused layer by layer, and the fused diagnostic result is obtained.

4. A ship system fault diagnosis device based on multi-source information fusion, characterized in that: include: Data receiving module, used to collect monitoring data including voltage, current, vibration acceleration, speed, displacement and noise; A data preprocessing module is used to preprocess the monitoring data; The feature extraction and model training module is used to input various types of pre-processed monitoring data into a single parameter fault diagnosis model and output a single fault diagnosis result. The single parameter fault diagnosis model is trained based on a machine learning model using labeled massive ship system monitoring data; A fusion diagnosis module is used to analyze multiple single fault diagnosis results using multi-source information fusion diagnosis rules to obtain fusion diagnosis results, wherein the fusion diagnosis rules are constructed based on DS evidence theory; The evaluation and judgment module is used to determine the location and nature of the fault based on the fusion diagnosis results.

5. The device according to claim 4, characterized in that The single parameter fault diagnosis model includes: using a long short-term memory network (LSTM) as a core diagnosis model, and combining deep learning technology to build a fault diagnosis model framework; extracting features related to fault diagnosis for each type of monitoring parameter, training the deep learning model in combination with historical fault data sets, and building a final single parameter fault diagnosis model.

6. The device according to claim 5, characterized in that The fusion diagnosis module is used to fuse the individual diagnosis results layer by layer in pairs according to the DS combination rule to obtain a fusion diagnosis result.

Images