A bearing load measurement system based on digital twinning and an application method thereof

Abstract

This invention belongs to the field of ship shafting condition monitoring and fault diagnosis technology, specifically relating to a bearing load measurement system based on digital twins and its application method. The invention constructs a measurement system consisting of a data acquisition module, a data processing module, a digital twin analysis module, a prediction module, and a display and interaction module. Through distributed sensor network deployment, intelligent processing of multi-source data, deep analysis of digital twin models, and hybrid algorithm prediction, it achieves holographic perception and dynamic prediction of bearing loads. Simultaneously, it designs a system application method encompassing sensor network deployment, data system configuration, intelligent model training, system verification and delivery, and the construction of an operation and maintenance optimization system, along with a corresponding preventative maintenance mechanism. This invention effectively improves the accuracy and response speed of bearing load monitoring, enabling real-time measurement of bearing loads under shafting operating conditions and prediction of load change trends and fault risks, significantly improving the safety, reliability, and economy of bearing operation.

Description

Technical Field

[0001] This invention belongs to the field of ship shafting condition monitoring and fault diagnosis technology, specifically relating to a bearing load measurement system based on digital twins and its application method. Background Technology

[0002] As the core transmission component of rotating machinery, the load condition of bearings directly determines the safety and reliability of equipment operation. Especially in the propulsion shafting of ships, the load of intermediate bearings is affected by multiple factors such as ship navigation conditions, marine environment, shaft vibration, etc., and the load condition is complex and variable. Accurate and real-time monitoring of bearings is the key to the safe operation of ship power systems.

[0003] Existing bearing load monitoring technologies mostly employ single-point sensor measurement methods, which can only acquire load data at local bearing locations and cannot comprehensively reflect the overall load distribution of the bearing, resulting in insufficient monitoring completeness. Simultaneously, environmental noise such as mechanical vibration and electromagnetic interference during ship navigation can severely interfere with monitoring signals. Existing technologies have limited anti-interference processing methods, leading to low signal-to-noise ratios and poor accuracy. Furthermore, traditional monitoring technologies can only measure bearing loads under static installation conditions, failing to monitor real-time loads during shaft operation. They also lack effective dynamic prediction methods, making it impossible to anticipate load change trends and fault risks, resulting in slow response times and difficulty meeting the real-time monitoring and early warning needs under complex operating conditions. Moreover, existing monitoring devices have poor adaptability, making it difficult to flexibly adjust to different shaft types and operating conditions, further limiting their application scope.

[0004] Digital twin technology provides a brand-new technical approach to equipment condition monitoring. By constructing a precise mapping between physical entities and digital models, real-time simulation and dynamic analysis of equipment operating status can be achieved. However, at present, digital twin technology has not yet formed a mature application solution in the field of bearing load monitoring, especially in the field of ship shaft bearing load monitoring. It lacks a technical architecture that deeply integrates distributed sensor networks, intelligent data processing algorithms and digital twin models, and cannot achieve the integrated requirements of multi-bearing collaborative monitoring, strong anti-interference, real-time response and intelligent prediction.

[0005] Therefore, there is an urgent need to develop a bearing load measurement system and its application method based on digital twins, to break through the limitations of traditional monitoring technology, realize holographic perception, real-time processing, dynamic prediction and intelligent early warning of bearing load under shaft operation, improve the accuracy and efficiency of bearing load monitoring, and ensure the safe and stable operation of equipment. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a bearing load measurement system and its application method based on digital twins, which can effectively improve the accuracy and response speed of bearing load monitoring, and can measure the bearing load under the shaft system operation state in real time and predict the load change trend and failure risk.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: I. A bearing load measurement system based on digital twin This invention provides a bearing load measurement system based on digital twins, comprising a data acquisition module 101, a data processing module 102, a digital twin analysis module 103, a prediction module 104, and a display and interaction module 105 that transmit data sequentially via wired or wireless networks; the data acquisition module 101 is deployed at a preset mechanical node position of the intermediate bearing of a ship to collect multimodal monitoring data of the bearing; the data processing module 102 is used to synchronize, denoise, standardize, and extract features from the multimodal monitoring data; the digital twin analysis module 103 is used to construct a digital twin model of the bearing and output real-time predicted load data; the prediction module (104) is used to predict the bearing load change trend and assess the failure risk; the display and interaction module (105) is used to visualize the multimodal monitoring data and the bearing load change trend.

[0008] Preferably, the data acquisition module 101 includes a shaft speed sensor 106, a full-bridge strain gauge 107, a strain signal wireless transmitter 108, an angle sensor 109, a pressure sensor 110, and a mobile power supply 111. The shaft speed sensor 106 and angle sensor 109 are used to monitor the shaft speed and torsion angle in real time, respectively. The pressure sensor 110 is used to measure the real-time pressure change of the bearing. The mobile power supply 111 is used to power the strain signal wireless transmitter 108. The strain signal wireless transmitter 108 is used to transmit the shaft strain signal measured by the full-bridge strain gauge 107.

[0009] Preferably, the full-bridge strain gauge 107 is arranged in a cross-symmetric layout to form a sensing array and is installed at a preset mechanical node position of the intermediate bearing to be tested; the pressure sensor 110 is installed on the support surface of the intermediate bearing to be tested.

[0010] Preferably, the data processing module 102 includes an industrial control computer processor 302, which is used to sequentially perform synchronous processing, Kalman filtering noise reduction, data standardization and multi-dimensional feature extraction operations on the multimodal monitoring data collected by the data acquisition module 101 to obtain strain feature data.

[0011] Preferably, the digital twin analysis module 103 includes a data model library 302, which has a built-in deep neural network model for modeling and analyzing the strain feature data and constructing a bearing digital twin mapping.

[0012] Preferably, the prediction module 104 uses a four-stage process to process the predicted load data output by the digital twin analysis module 103 to predict the bearing load change trend; the four-stage process is as follows: data input stage 401, predicted load model stage 402, prediction process execution stage 403, and prediction result output stage 404.

[0013] Preferably, the user interface of the display and interaction module 105 includes six functional blocks, namely, real-time bearing load value block 501, strain monitoring data block 502, trend prediction graph block 503, fault early warning block 504, operation panel block 505, and configuration and alarm block 506.

[0014] II. An Application Method of a Bearing Load Measurement System Based on Digital Twin Based on the same inventive concept, the present invention also provides an application method for the bearing load measurement system based on digital twins as described above, comprising the following steps: S1, Sensor Network Deployment: Data acquisition modules are arranged at the preset mechanical node positions of the intermediate bearing and a wireless self-organizing network is built through an industrial-grade wireless transmitter; S2, Data System Configuration: Build a data acquisition platform based on edge computing architecture, configure adaptive sampling strategies, and establish sensor digital twin mapping relationships; S3, Intelligent Model Training: The predictive model is built using an incremental learning framework. In the initial stage, historical databases are loaded for transfer learning. In the online stage, real-time data drives the fine-tuning of model parameters. At the same time, model performance evaluation indicators are set, and an automatic retraining trigger mechanism is established. When data drift is detected to exceed the threshold, the model is automatically optimized. S4, System Verification and Delivery: Execute unit tests, integration tests and scenario tests in sequence, and ensure that the output of each stage meets the standards through quality gating. After completing all tests, perform final debugging and delivery of the system. S5, Operation and Maintenance Optimization System Construction: Establish a preventive maintenance mechanism, perform sensor zero-point calibration monthly, update model feature engineering quarterly, and conduct a full system health diagnosis every six months.

[0015] Preferably, in step S2, the data system configuration includes: S21, Edge Computing Platform Construction: Build a data acquisition platform based on edge computing architecture and push data preprocessing functions to the edge; S22, Adaptive sampling strategy configuration: Configure the sensor's adaptive sampling frequency according to different navigation conditions of the ship. Increase the sampling frequency in high-speed navigation conditions to capture subtle load changes, and reduce the sampling frequency in low-speed navigation conditions to save energy. S23, Establishment of digital twin mapping relationship: Based on the three-dimensional model and mechanical properties of the bearing, establish a one-to-one mapping relationship between the physical location of the sensor measuring point and the node of the digital model, so as to realize the real-time transmission of physical data to the digital model; S24, Field Calibration: All sensors are calibrated in the field using standard loads.

[0016] Preferably, in step S3, the intelligent model training includes: S31, Historical Data Preprocessing: Collect historical load data and fault data of the bearing throughout its entire life cycle, perform data cleaning, deduplication, and completion processing, remove abnormal data, and build a standardized historical database; S32, Transfer Learning Initialization: Based on the incremental learning framework, a prediction model is built. In the initial stage, the historical database is loaded for transfer learning. The existing data features are used to complete the initial initialization of the model parameters, shortening the model training cycle. S33, Incremental Learning Framework Construction: Build an online incremental learning framework, use the real-time collected bearing load data as new training samples, continuously drive the fine-tuning of model parameters, and make the model adapt to the dynamic changes in the bearing operating state. S34, Model Performance Evaluation: Set performance evaluation indicators for model prediction accuracy, response speed, and fault identification rate, and monitor the model's running status in real time; S35, Automatic Retraining Mechanism Settings: Establish a data drift detection and automatic model retraining trigger mechanism. When the drift between real-time data and model training data exceeds a preset threshold, the system automatically starts model retraining to optimize model parameters and ensure prediction accuracy.

[0017] Compared with the prior art, the present invention has the following main advantages: 1. This invention innovatively integrates strain gauge arrays with digital twin technology. By arranging a distributed strain measurement network in the spatial dimension and combining it with dynamic modeling and in-depth analysis of digital twins in the time dimension, it achieves holographic perception of bearing load status, significantly improving the integrity and accuracy of multi-bearing collaborative monitoring. At the same time, it breaks through the limitation of existing technologies that can only measure static loads, and can accurately monitor the real-time load of bearings under shaft system operation. The anti-interference signal processing flow developed for special ship operating conditions integrates advanced technologies such as improved Kalman filtering, effectively overcoming the influence of environmental noise such as ship vibration and electromagnetic interference, greatly improving measurement accuracy. Moreover, the device has a simple structure, is easy to install, and can be processed and adapted for multiple types of shaft segments, possessing good versatility and adaptability.

[0018] 2. By introducing the LSTM-DNN hybrid model and real-time edge computing technology, this invention has successfully achieved a technological leap from static acquisition of bearing load to dynamic prediction. The system can not only reflect the current load status in real time, but also accurately predict the load change trend within a preset time period in the future, shortening the early warning response time to the millisecond level. At the same time, it outputs a fault risk level assessment based on probability statistics, providing a reliable basis for early warning and handling of bearing faults, and significantly improving the safety and reliability of shaft system operation of ships and other equipment. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall architecture of the bearing load measurement system in an embodiment of the present invention; Figure 2 This is a schematic diagram of the arrangement of the data acquisition module on the shaft segment in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the working principle of the digital twin analysis module in this embodiment of the invention. Figure 4 This is a schematic diagram illustrating the working principle of the prediction module in an embodiment of the present invention; Figure 5 This is a schematic diagram of the user interface of the display and interaction module in an embodiment of the present invention.

[0020] In the diagram: 101-Data Acquisition Module, 102-Data Processing Module, 103-Digital Twin Analysis Module, 104-Prediction Module, 105-Display and Interaction Module, 106-Shaft Speed ​​Sensor, 107-Strain Gauge, 108-Strain Signal Wireless Transmitter, 109-Angle Sensor, 110-Pressure Sensor, 111-Mobile Power Supply, 201-Ship Propulsion Shaft Section, 202-Intermediate Bearing to be Measured, 203-Intermediate Bearing Not to be Measured, 204 - Coordinate paper, 205- Universal bracket, 301- Receiver module, 302- Processor, 303- Data model library, 401- Data input stage, 402- Predictive load model stage, 403- Execute prediction process stage, 404- Output prediction results stage, 501- Bearing load value block, 502- Strain monitoring data block, 503- Trend prediction graph block, 504- Fault early warning block, 505- Operation panel block, 506- Configuration and alarm block. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0022] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0023] In this invention, unless otherwise expressly specified and limited, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise expressly and specifically limited.

[0024] Example 1: This example provides a bearing load measurement system based on digital twins, such as... Figures 1-5 As shown, it mainly includes: a data acquisition module 101, a data processing module 102, a digital twin analysis module 103, a prediction module 104, and a display and interaction module 105, which transmit data sequentially via wired or wireless networks; the data acquisition module 101 is deployed at a preset mechanical node position of the ship's intermediate bearing and is used to collect multimodal monitoring data of the bearing; the data processing module 102 is used to synchronize, denoise, standardize, and extract features from the multimodal monitoring data; the digital twin analysis module 103 is used to construct a digital twin model of the bearing and output real-time predicted load data; the prediction module (104) is used to predict the bearing load change trend and assess the failure risk; the display and interaction module (105) is used to visualize the multimodal monitoring data and the bearing load change trend.

[0025] Furthermore, the data acquisition module 101 includes a shaft speed sensor 106, a full-bridge strain gauge 107, a strain signal wireless transmitter 108, an angle sensor 109, a pressure sensor 110, and a mobile power supply 111. The shaft speed sensor 106 and angle sensor 109 are used to monitor the shaft speed and torsion angle in real time, respectively. The pressure sensor 110 is used to measure the real-time pressure change of the bearing. The mobile power supply 111 is used to power the strain signal wireless transmitter 108. The strain signal wireless transmitter 108 is used to transmit the shaft strain signal measured by the full-bridge strain gauge 107.

[0026] Furthermore, the full-bridge strain gauge 107 adopts a cross-symmetrical layout to form a sensing array and is installed at a preset mechanical node position of the intermediate bearing to be tested; the pressure sensor 110 is installed on the support surface of the intermediate bearing to be tested.

[0027] Furthermore, the data processing module 102 includes an industrial control computer processor 302, which is used to sequentially perform synchronous processing, Kalman filtering noise reduction, data standardization and multi-dimensional feature extraction operations on the multimodal monitoring data collected by the data acquisition module 101 to obtain strain feature data.

[0028] Furthermore, the digital twin analysis module 103 includes a data model library 302, which has a built-in deep neural network model for modeling and analyzing the strain feature data and constructing a bearing digital twin mapping.

[0029] Furthermore, the prediction module 104 uses a four-stage process to process the predicted load data output by the digital twin analysis module 103 to predict the bearing load change trend; the four-stage process is as follows: data input stage 401, predicted load model stage 402, prediction process execution stage 403, and prediction result output stage 404.

[0030] Furthermore, the user interface of the display and interaction module 105 includes six functional blocks, namely, real-time bearing load value block 501, strain monitoring data block 502, trend prediction graph block 503, fault early warning block 504, operation panel block 505, and configuration and alarm block 506.

[0031] Example 2: This example provides a bearing load measurement system based on digital twins, such as... Figure 1 As shown, the system includes a data acquisition module 101, a data processing module 102, a digital twin analysis module 103, a prediction module 104, and a display and interaction module 105. The data acquisition module is installed at a critical location in the ship's intermediate bearing and is equipped with a shaft speed sensor 106, a full-bridge strain gauge 107, a strain signal wireless transmitter 108, an angle sensor 109, a pressure sensor 110, and a power supply 111. The shaft speed sensor and angle sensor are used to monitor the shaft's rotational speed and torsional angle in real time. The strain signal wireless transmitter transmits the real-time strain measured by the strain gauge to the receiver, the pressure sensor measures the real-time pressure changes experienced by the bearing, and the power supply provides real-time power to the strain signal wireless transmitter. The strain gauge measures the strain signal of the shaft segment.

[0032] The data acquisition module 101 transmits the acquired multimodal data to the data processing module 102 via a wired or wireless network. The data processing module 102 first performs synchronous processing on data from different sources to ensure temporal consistency among the sensor data. Then, Kalman filtering is used for noise reduction and purification. Next, data standardization is performed to allow data from different sensors to be analyzed on the same scale. Subsequently, the system extracts features from the processed data, including multi-dimensional features such as the maximum value, variance, peak value, and spectral characteristics of the strain signal, to comprehensively reflect the bearing's operating status.

[0033] In this embodiment, the data acquisition module 101 is installed in the ship's propulsion shaft section 201 at the following location: Figure 2 As shown, select the intermediate bearing 202 to be measured and install the strain gauge 107, strain signal wireless transmitter 108 and mobile power supply 111; at the same time, install the coordinate paper 204 next to the intermediate bearing 203 that is not being measured, and then install the angle sensor 109 on the universal bracket 205, and adjust the bracket angle so that the angle sensor 109 is aligned with the coordinate paper 204 at 90°.

[0034] In this embodiment, the feature data processed by the data processing module 102 is transmitted to the digital twin analysis module 103 to obtain the predicted load data, such as... Figure 3 As shown, the strain data acquired by the data acquisition module 101 is first transmitted to the processor 302 of the data processing module 102 via the receiving module 301 of the industrial control computer for Kalman filtering to obtain feature data; then the feature data is sent to the data model library 302 of the digital twin analysis module 103 for deep neural network analysis to obtain predicted load data.

[0035] In this embodiment, the digital twin analysis module 103 transmits the acquired real-time predicted load data to the prediction module 104 to output the bearing load prediction result, such as... Figure 4 As shown, a four-stage processing flow is adopted: the data input stage 401 integrates multi-source historical load data to establish a training set; the load prediction model stage 402 uses time series analysis, convolutional neural network (CNN) and long short-term memory network (LSTM) to construct a hybrid prediction algorithm, and optimizes feature weights through an attention mechanism; the prediction process execution stage 403 integrates the latest monitoring data in real time and dynamically updates the model parameters using sliding window technology; finally, in the output prediction result stage 404, not only can the load change trend curve within the future preset time period be provided, but also the fault risk level assessment based on probability statistics is also included, realizing a closed-loop processing from data collection to intelligent prediction.

[0036] In this embodiment, the bearing load prediction results generated by the prediction module 104 are displayed in a multi-dimensional visualization through the display and interaction module 105, such as... Figure 5 As shown, the user interface of the display and interaction module 105 includes six functional blocks: the real-time bearing load value block 501 uses dynamic curves to compare and display the data of each measuring point; the strain monitoring data block 502 visually presents the stress distribution of key parts through a three-dimensional heat map; the trend prediction graph block 503 overlays and displays the deviation analysis between the predicted value and the actual monitored value; the fault early warning block 504 uses red, yellow, and green indicator lights to realize risk-level alarms; the operation panel block 505 supports multi-dimensional data filtering and prediction model parameter adjustment; and the configuration and alarm block 506 provides a custom early warning rule setting function.

[0037] In this embodiment, the system can adjust the sensor sampling frequency and data processing algorithm according to the actual operating conditions of the ship. For example, for ships operating at high speeds, the sensor sampling frequency can be set higher to capture more subtle vibrations and pressure changes; for different types of intermediate bearings, the system can select the most suitable feature extraction method and machine learning algorithm to improve the accuracy of monitoring and prediction.

[0038] Furthermore, the system of this invention can also be integrated with other monitoring systems on a ship via a wireless network to achieve more comprehensive monitoring of the ship's operational status. For example, the system can share and coordinate data with the ship's navigation system, power system monitoring system, etc., to achieve collaborative monitoring and fault diagnosis of multiple systems.

[0039] Example 3: Based on the same inventive concept, this example also provides an application method for the bearing load measurement system based on digital twins as described above, including the following steps: Step S1, Sensor Network Deployment S11, Measurement point positioning: Based on the mechanical simulation analysis of the ship's propulsion shaft system, the key mechanical nodes of the intermediate bearing are determined as the core measurement points for sensor placement, ensuring that the measurement points cover the sensitive areas of bearing load changes; S12, Sensor Selection and Installation: Select a suitable high-precision sensor for the measurement point conditions. The full-bridge strain gauge 107 adopts a cross-symmetrical layout to form a sensing array to measure the multi-dimensional strain field of the shaft section. The pressure sensor 110 is installed in close contact with the bearing support surface. The shaft speed sensor 106 and the angle sensor 109 are respectively installed on the rotating part of the shaft system and the fixed support part to ensure that the measurement is interference-free. S13, Self-organizing network construction: All sensors build a wireless self-organizing network system through the industrial-grade strain signal wireless transmitter 108, set the network communication protocol, and ensure the stability of sensor data transmission. S14, Synchronization Accuracy Calibration: Time synchronization is performed on all sensors in the self-organizing network through a time synchronization protocol to ensure that the sampling synchronization accuracy reaches ±0.1ms, providing a foundation for multimodal data fusion.

[0040] Step S2, Data System Configuration S21, Edge Computing Platform Construction: Build a data acquisition platform based on edge computing architecture, push data preprocessing functions to the edge, reduce cloud data transmission pressure, and improve data processing response speed; S22, Adaptive sampling strategy configuration: Configure the sensor's adaptive sampling frequency according to different ship operating conditions (high-speed navigation, low-speed navigation, berthing). Increase the sampling frequency in high-speed conditions to capture subtle load changes, and decrease the sampling frequency in low-speed conditions to save energy. S23, Establishment of digital twin mapping relationship: Based on the three-dimensional model and mechanical properties of the bearing, establish a one-to-one mapping relationship between the physical location of the sensor measuring point and the node of the digital model, and realize the real-time transmission of physical data to the digital model; S24, Field Calibration: All sensors are calibrated in the field using standard loads to eliminate systematic errors in sensor installation and measurement, thereby improving measurement accuracy; S25, Dual-mode transmission channel configuration: Configured with 5G / Wi-Fi6 dual-mode redundant wireless transmission channels. When one transmission mode fails, it automatically switches to the other to ensure that the data is transmitted to the data processing module 102 in real time and stably.

[0041] Step S3, Intelligent Model Training S31, Historical Data Preprocessing: Collect historical load data and fault data of the bearing throughout its entire life cycle, perform data cleaning, deduplication, and completion processing, remove abnormal data, and build a standardized historical database; S32, Transfer Learning Initialization: Based on the incremental learning framework, a prediction model is built. In the initial stage, the historical database is loaded for transfer learning. The existing data features are used to complete the initial initialization of the model parameters, shortening the model training cycle. S33, Incremental Learning Framework Construction: Build an online incremental learning framework, use the real-time collected bearing load data as new training samples, continuously drive the fine-tuning of model parameters, and make the model adapt to the dynamic changes in the bearing operating state. S34, Model Performance Evaluation: Set performance evaluation indicators such as model prediction accuracy, response speed, and fault identification rate, and monitor the model's running status in real time; S35, Automatic Retraining Mechanism Settings: Establish a data drift detection and automatic model retraining trigger mechanism. When the drift between real-time data and model training data exceeds a preset threshold, the system automatically starts model retraining to optimize model parameters and ensure prediction accuracy.

[0042] Step S4, System Verification and Delivery S41, Unit Testing: Perform functional tests on each module of the system to verify the operational stability and functional integrity of a single module, and optimize the module based on the problems found in the tests; S42, Integration Test: Integrate and test the modules to test the data transmission efficiency, communication compatibility and collaborative operation capability between the modules, and verify the overall closed-loop operation effect of the system. S43, Scenario Testing: Simulate different operating scenarios of the ship, such as high-speed navigation, severe sea conditions, start-up and shutdown conditions, and conduct on-site system testing to verify the system's measurement accuracy, anti-interference ability and prediction accuracy under complex operating conditions. S44, Quality Gating: Set quality acceptance standards for each level of testing. Only after passing the previous level of testing and meeting the acceptance standards can the next level of testing be carried out, ensuring that the output of each stage meets the standards. S45, System Delivery: After completing all tests, the system undergoes final commissioning, and an intelligent bearing load monitoring system with self-diagnosis and self-optimization capabilities is delivered to the user, along with system operation and maintenance training.

[0043] Step S5, Construction of Operation and Maintenance Optimization System S51, Routine Calibration and Maintenance: Establish a preventive maintenance mechanism, perform zero-point calibration on all sensors monthly, conduct communication testing on the wireless transmission channel quarterly, and replace aging parts in a timely manner to ensure the operating accuracy of hardware equipment; S52, Model Feature Engineering Update: The model's feature engineering is updated quarterly based on the latest bearing operating data, the feature extraction method is optimized, and the model's ability to represent the bearing's operating status is improved. S53, System-wide Health Diagnosis: A comprehensive health diagnosis is performed on the system every six months, including the operational status detection of hardware devices, software systems, and model algorithms, to promptly identify and resolve potential system faults; S54, Model Version Management: Develop a model version management system to record and save the version of the prediction model at different stages, support iterative updates and rollbacks of the model, and track the performance metrics of each version of the model in real time. S55, Operating Condition Adaptation Optimization: Based on the long-term changes in ship operating conditions, the sensor sampling strategy and model algorithm parameters are dynamically adjusted to ensure that the system always adapts to the actual operating conditions and achieves long-term stable operation.

[0044] Furthermore, all parts of this application that are not described in detail are the same as or implemented using existing technology.

[0045] In summary: 1. This invention innovatively integrates strain gauge arrays with digital twin technology. By arranging a distributed strain measurement network in the spatial dimension and combining it with dynamic modeling and in-depth analysis of digital twins in the time dimension, it achieves holographic perception of bearing load status, significantly improving the integrity and accuracy of multi-bearing collaborative monitoring. At the same time, it breaks through the limitation of existing technologies that can only measure static loads, and can accurately monitor the real-time load of bearings under shaft system operation. The anti-interference signal processing flow developed for special ship operating conditions integrates advanced technologies such as improved Kalman filtering, effectively overcoming the influence of environmental noise such as ship vibration and electromagnetic interference, greatly improving measurement accuracy. Moreover, the device has a simple structure, is easy to install, and can be processed and adapted for multiple types of shaft segments, possessing good versatility and adaptability.

[0046] 2. By introducing the LSTM-DNN hybrid model and real-time edge computing technology, this invention has successfully achieved a technological leap from static acquisition of bearing load to dynamic prediction. The system can not only reflect the current load status in real time, but also accurately predict the load change trend within a preset time period in the future, shortening the early warning response time to the millisecond level. At the same time, it outputs a fault risk level assessment based on probability statistics, providing a reliable basis for early warning and handling of bearing faults, and significantly improving the safety and reliability of shaft system operation of ships and other equipment.

[0047] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0048] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0049] Those skilled in the art will readily understand that the above description is merely 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 within the scope of protection of the present invention.

Claims

1. A bearing load measurement system based on digital twin, characterized in that: The system includes a data acquisition module (101), a data processing module (102), a digital twin analysis module (103), a prediction module (104), and a display and interaction module (105) that transmit data sequentially via wired or wireless networks. The data acquisition module (101) is deployed at a preset mechanical node position of the ship's intermediate bearing to collect multimodal monitoring data of the bearing. The data processing module (102) is used to synchronize, denoise, standardize, and extract features from the multimodal monitoring data. The digital twin analysis module (103) is used to construct a digital twin model of the bearing and output real-time predicted load data. The prediction module (104) is used to predict the bearing load change trend and assess the failure risk. The display and interaction module (105) is used to visualize the multimodal monitoring data and the bearing load change trend.

2. The bearing load measurement system based on digital twin according to claim 1, characterized in that, The data acquisition module (101) includes a shaft speed sensor (106), a full-bridge strain gauge (107), a strain signal wireless transmitter (108), an angle sensor (109), a pressure sensor (110), and a mobile power supply (111). The shaft speed sensor (106) and angle sensor (109) are used to monitor the shaft speed and torsion angle in real time, respectively. The pressure sensor (110) is used to measure the real-time pressure change of the bearing. The mobile power supply (111) is used to power the strain signal wireless transmitter (108). The strain signal wireless transmitter (108) is used to transmit the shaft strain signal measured by the full-bridge strain gauge (107).

3. The bearing load measurement system based on digital twin according to claim 2, characterized in that, The full-bridge strain gauge (107) is arranged in a cross-symmetric layout to form a sensing array and is installed at a preset mechanical node position of the intermediate bearing to be tested; the pressure sensor (110) is installed on the support surface of the intermediate bearing to be tested.

4. The bearing load measurement system based on digital twin according to claim 1, characterized in that, The data processing module (102) includes an industrial control computer processor (302), which is used to perform synchronous processing, Kalman filtering noise reduction, data standardization and multi-dimensional feature extraction operations on the multimodal monitoring data collected by the data acquisition module (101) in sequence to obtain strain feature data.

5. A bearing load measurement system based on digital twin according to claim 4, characterized in that, The digital twin analysis module (103) includes a data model library (302), which has a built-in deep neural network model for modeling and analyzing the strain feature data and constructing a bearing digital twin mapping.

6. The bearing load measurement system based on digital twin according to claim 1, characterized in that, The prediction module (104) uses a four-stage process to process the predicted load data output by the digital twin analysis module (103) to predict the bearing load change trend. The four-stage process is as follows: data input stage (401), predicted load model stage (402), prediction process execution stage (403), and prediction result output stage (404).

7. The bearing load measurement system based on digital twin according to claim 1, characterized in that, The user interface of the display and interaction module (105) includes six functional blocks: real-time bearing load value block (501), strain monitoring data block (502), trend prediction graph block (503), fault warning block (504), operation panel block (505), and configuration and alarm block (506).

8. An application method of the bearing load measurement system based on digital twin as described in any one of claims 1 to 7, characterized in that, Includes the following steps: S1, Sensor Network Deployment: Data acquisition modules are arranged at the preset mechanical node positions of the intermediate bearing and a wireless self-organizing network is built through an industrial-grade wireless transmitter; S2, Data System Configuration: Build a data acquisition platform based on edge computing architecture, configure adaptive sampling strategies, and establish sensor digital twin mapping relationships; S3, Intelligent Model Training: The predictive model is built using an incremental learning framework. In the initial stage, historical databases are loaded for transfer learning. In the online stage, real-time data drives the fine-tuning of model parameters. At the same time, model performance evaluation indicators are set, and an automatic retraining trigger mechanism is established. When data drift is detected to exceed the threshold, the model is automatically optimized. S4, System Verification and Delivery: Execute unit tests, integration tests and scenario tests in sequence, and ensure that the output of each stage meets the standards through quality gating. After completing all tests, perform final debugging and delivery of the system. S5, Operation and Maintenance Optimization System Construction: Establish a preventive maintenance mechanism, perform sensor zero-point calibration monthly, update model feature engineering quarterly, and conduct a full system health diagnosis every six months.

9. A bearing load measurement method based on digital twins according to claim 8, characterized in that... In step S2, the data system configuration includes: S21, Edge Computing Platform Construction: Build a data acquisition platform based on edge computing architecture and push data preprocessing functions to the edge; S22, Adaptive sampling strategy configuration: Configure the sensor's adaptive sampling frequency according to different navigation conditions of the ship. Increase the sampling frequency in high-speed navigation conditions to capture subtle load changes, and reduce the sampling frequency in low-speed navigation conditions to save energy. S23, Establishment of digital twin mapping relationship: Based on the three-dimensional model and mechanical properties of the bearing, establish a one-to-one mapping relationship between the physical location of the sensor measuring point and the node of the digital model, so as to realize the real-time transmission of physical data to the digital model; S24, Field Calibration: All sensors are calibrated in the field using standard loads.

10. A bearing load measurement method based on digital twin according to claim 8, characterized in that... In step S3, the training of the intelligent model includes: S31, Historical Data Preprocessing: Collect historical load data and fault data of the bearing throughout its entire life cycle, perform data cleaning, deduplication, and completion processing, remove abnormal data, and build a standardized historical database; S32, Transfer Learning Initialization: Based on the incremental learning framework, a prediction model is built. In the initial stage, the historical database is loaded for transfer learning. The existing data features are used to complete the initial initialization of the model parameters, shortening the model training cycle. S33, Incremental Learning Framework Construction: Build an online incremental learning framework, use the real-time collected bearing load data as new training samples, continuously drive the fine-tuning of model parameters, and make the model adapt to the dynamic changes in the bearing operating state. S34, Model Performance Evaluation: Set performance evaluation indicators for model prediction accuracy, response speed, and fault identification rate, and monitor the model's running status in real time; S35, Automatic Retraining Mechanism Setting: Establish a data drift detection and automatic model retraining trigger mechanism. When the drift between real-time data and model training data exceeds a preset threshold, the system automatically starts model retraining to optimize model parameters and ensure prediction accuracy.

Images