Unmanned ship state monitoring and fault early warning system based on big data analysis

Abstract

The invention relates to the technical field of fault monitoring and early warning, in particular to an unmanned ship state monitoring and fault early warning system based on big data analysis, and the system comprises a parameter collection module which is used for collecting the operation data of an unmanned ship through a sensor to obtain standardized monitoring data; the coupling analysis module is used for analyzing the change trend of the potential coupling relationship of the operation data and judging the change rule of the drift characteristics of the sensor; the state identification module is used for identifying the characteristic mode deviation between the current operation state and the historical operation state of the unmanned ship and identifying the characteristic mode of the operation state change; the model evaluation module is used for quantifying the application degree of the historical fault early warning model in the current operation state of the unmanned ship; the threshold value correction module is used for correcting a fault early warning threshold value of the current historical fault early warning model and outputting a working condition early warning model; and the early warning output module is used for analyzing the standardized monitoring data by adopting a working condition early warning model and generating early warning information of the operation fault of the unmanned ship.

Description

Technical Field

[0001] This invention relates to the field of fault monitoring and early warning technology, specifically to an unmanned vessel status monitoring and fault early warning system based on big data analysis. Background Technology

[0002] Unmanned surface vessels (USVs) typically operate in complex and variable aquatic environments, such as oceans, rivers, and lakes. Their navigation safety is directly related to the successful completion of missions, the integrity of equipment, and the safety of personnel, such as when carrying personnel or performing rescue missions. Through a fault monitoring and early warning system, key components and systems of USVs, such as power systems, navigation systems, communication systems, sensors, energy systems, and control systems, can be continuously monitored. This allows for the real-time and accurate identification and prediction of various potential faults or anomalies during the USV's mission, enabling timely implementation of corresponding control measures to prevent the faults from escalating into serious consequences.

[0003] Existing unmanned vessel status monitoring and fault early warning systems mostly rely on static threshold settings or rule-based judgments based on single sensor data. These systems are unable to effectively cope with the uncertainties brought about by complex sea conditions, multi-source interference, and dynamic changes in operational status. Consequently, the early warning models have insufficient generalization ability and low accuracy. Furthermore, they lack dynamic modeling and identification of sensor drift phenomena caused by changes in the operating environment, making it impossible to achieve accurate perception and timely response to unmanned vessel fault risks. Summary of the Invention

[0004] To address the technical problems of existing unmanned surface vessel (USV) status monitoring and fault early warning systems, such as their inability to cope with changes in uncertainties, insufficient model generalization ability, low early warning accuracy, inability to dynamically identify USV fault risks, and resulting in untimely responses, this invention aims to provide an USV status monitoring and fault early warning system based on big data analysis. The specific technical solution adopted is as follows:

[0005] The parameter acquisition module is used to: collect operational data of the unmanned vessel through sensors and preprocess it to obtain standardized monitoring data;

[0006] The coupling analysis module is used to: analyze the changing trends of potential coupling relationships in operational data based on standardized monitoring data, and determine the changing patterns of sensor drift characteristics;

[0007] The status recognition module is used to: identify the characteristic pattern deviation between the current and historical operating states of the unmanned vessel based on the changing patterns of sensor drift characteristics, and to identify the characteristic patterns of changes in operating state;

[0008] The model evaluation module is used to: acquire historical fault warning models and quantify the applicability of historical fault warning models in the current operating state of unmanned vessels based on the changing patterns of sensor drift characteristics;

[0009] The threshold correction module is used to: comprehensively adjust the fault warning threshold of the current historical fault warning model based on the characteristic patterns and applicability of the changes in operating status, and output the operating condition warning model.

[0010] The early warning output module is used to: analyze standardized monitoring data using an operational condition early warning model to generate early warning information for unmanned vessel operational failures.

[0011] Preferably, the operational data includes speed parameters, wave environment parameters, and load status parameters.

[0012] Preferably, operational data of the unmanned vessel is collected through sensors and preprocessed to obtain standardized monitoring data, including:

[0013] Based on the sensor data collected during the operation of the unmanned vessel, including speed parameters, wave environment parameters, and load status parameters;

[0014] Abnormal data and / or noisy data in the operation data are removed by an abnormal data monitoring method, and the time series sequence of the operation data is unified by a time series interpolation method to obtain standardized monitoring data.

[0015] Preferably, based on standardized monitoring data, the changing trends of potential coupling relationships in operational data are analyzed to determine the changing patterns of sensor drift characteristics, including:

[0016] Based on standardized monitoring data, nonlinear correlation analysis was used to analyze the potential coupling relationship of operational data and extract the changing trend of operational data under different operating conditions of unmanned vessels.

[0017] A nonlinear feature extraction algorithm is used to identify sensor drift characteristics corresponding to changing trends, and a model of the variation of sensor drift characteristics with operating data is established to obtain the variation law of sensor drift characteristics.

[0018] Preferably, the deviation of characteristic patterns between the current and historical operating states of the unmanned vessel is identified based on the changing patterns of sensor drift characteristics, and the characteristic patterns of changes in operating state are identified, including:

[0019] Based on the changing patterns of sensor drift characteristics, deep clustering analysis is used to perform feature pattern clustering analysis on the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel, and the clustering positions of the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel in the feature space are determined in turn.

[0020] Analyze the deviation in cluster positions to identify characteristic patterns of changes in the operational status of unmanned vessels.

[0021] Preferably, the applicability of the historical fault early warning model to the current operating state of the unmanned vessel is quantified based on the changing patterns of sensor drift characteristics, including:

[0022] Based on the changing patterns of sensor drift characteristics, the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel are obtained and input into the historical fault early warning model to obtain the fault risk probability of the unmanned vessel under the current and historical operating states.

[0023] By analyzing the differences between failure risk probabilities using transfer learning methods, the applicability of historical failure early warning models to the current operating state of unmanned vessels can be determined.

[0024] Preferably, the fault warning threshold of the current historical fault warning model is corrected by comprehensively considering the characteristic patterns and applicability of changes in operating status, and the operating condition warning model is output, including:

[0025] The degree of change in the unmanned vessel's operational status is determined based on the characteristic patterns of the changes.

[0026] The applicability of the model is evaluated based on its applicability to the current operating state of the unmanned vessel, according to the historical fault early warning model.

[0027] By establishing a correlation between the degree of change in the unmanned vessel's operating status and the applicability of the model, the fault warning threshold in the historical fault warning model is corrected, and a working condition warning model applicable to the current operating status of the unmanned vessel is output.

[0028] Preferably, a condition-based early warning model is used to analyze standardized monitoring data to generate early warning information for unmanned surface vessel (USV) operational malfunctions, including:

[0029] Based on the analysis of standardized monitoring data using the operating condition early warning model, the probability of failure risk of the unmanned vessel under the current operating condition is obtained;

[0030] By comparing the corrected fault warning threshold and fault risk probability, it is determined whether the unmanned vessel has reached the fault warning condition under the current operating state. If so, a warning message for the unmanned vessel's operating fault is generated.

[0031] To address the aforementioned problems, the present invention also provides an electronic device comprising a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus, and the processor calls logical instructions in the memory to execute the unmanned vessel status monitoring and fault early warning system based on big data analysis described in any of the preceding claims.

[0032] The present invention has the following beneficial effects:

[0033] By analyzing the changing trends of potential coupling relationships among operational data of unmanned surface vessels (USVs) during operation, sensor drift characteristics are dynamically identified, enhancing the ability to identify nonlinear interference in complex operating environments. Next, characteristic patterns of operational state changes are identified to effectively address monitoring deviations caused by sudden changes in operational state. Then, the applicability of historical fault warning models is evaluated based on the changing patterns of sensor drift characteristics, enabling these models to adapt to new operating conditions. The fault warning threshold is adjusted using the applicability evaluation index (degree of applicability) of the operational state change characteristic patterns and historical fault warning models to generate a condition-based warning model, enhancing its generalization ability and response accuracy. Based on the condition-based warning model, fault risk is determined, enabling early identification and proactive warning of USV operational faults, thereby improving the intelligence level and safety assurance capabilities of USV operations. Attached Figure Description

[0034] To more clearly illustrate the technical solutions and advantages 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic diagram of the structure of an unmanned vessel status monitoring and fault early warning system based on big data analysis, provided as an embodiment of the present invention. Detailed Implementation

[0036] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an unmanned vessel status monitoring and fault early warning system based on big data analysis proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0038] The following description, in conjunction with the accompanying drawings, details a specific solution for an unmanned vessel status monitoring and fault early warning system based on big data analysis provided by this invention.

[0039] To better illustrate, unmanned surface vessels (USVs) refer to surface vessels that do not require onboard personnel and complete predetermined tasks through remote control or autonomous navigation systems. They typically integrate advanced sensors, communication equipment, navigation systems, and artificial intelligence algorithms, enabling them to perform various tasks such as transportation, exploration, patrol, environmental monitoring, and search and rescue in complex aquatic environments. They come in a variety of types, including small remotely operated USVs, autonomously navigated USVs, underwater USVs, and surface USVs, and are widely used in marine engineering, fisheries, military, scientific research, and other fields.

[0040] The purpose of proposing status monitoring and fault early warning is to address various problems encountered during actual operation. Specifically, in actual operation, firstly, unmanned vessels often operate in complex environments far from human monitoring, such as the deep sea, in severe weather, or in remote waters. Once a malfunction occurs, timely human intervention is difficult, which may lead to mission interruption, equipment damage, or even safety accidents. Secondly, as the application scope of unmanned vessels expands, the requirements for their autonomy and intelligence are constantly increasing. Status monitoring and fault early warning are the core technical support for realizing autonomous operation and maintenance of unmanned vessels and improving mission continuity. They can ensure the stable and safe operation of unmanned vessels in complex dynamic environments and guarantee the efficient completion of missions. However, existing status monitoring and fault early warning technologies still have several problems. Therefore, this invention proposes an unmanned vessel status monitoring and fault early warning system based on big data analysis to overcome the shortcomings of existing technologies.

[0041] Please see Figure 1 The diagram illustrates a structural schematic of an unmanned vessel status monitoring and fault early warning system based on big data analysis, according to an embodiment of the present invention. The system includes:

[0042] The parameter acquisition module is used to: collect operational data of the unmanned vessel through sensors and preprocess it to obtain standardized monitoring data;

[0043] The coupling analysis module is used to: analyze the changing trends of potential coupling relationships in operational data based on standardized monitoring data, and determine the changing patterns of sensor drift characteristics;

[0044] The status recognition module is used to: identify the characteristic pattern deviation between the current and historical operating states of the unmanned vessel based on the changing patterns of sensor drift characteristics, and to identify the characteristic patterns of changes in operating state;

[0045] The model evaluation module is used to: acquire historical fault warning models and quantify the applicability of historical fault warning models in the current operating state of unmanned vessels based on the changing patterns of sensor drift characteristics;

[0046] The threshold correction module is used to: comprehensively adjust the fault warning threshold of the current historical fault warning model based on the characteristic patterns and applicability of the changes in operating status, and output the operating condition warning model.

[0047] The early warning output module is used to: analyze standardized monitoring data using an operational condition early warning model to generate early warning information for unmanned vessel operational failures.

[0048] It can be explained that during the operation of the unmanned vessel, the parameter acquisition module, coupling analysis module, status recognition module, model evaluation module, threshold correction module, and early warning output module work together to achieve comprehensive monitoring of the unmanned vessel's status and early warning of potential faults, thus ensuring the safe and stable operation of the unmanned vessel.

[0049] Furthermore, the operational data includes speed parameters, wave environment parameters, and load status parameters.

[0050] The operational data is obtained by the parameter acquisition module. The speed parameter refers to the unmanned vessel's speed and corresponding speed change data during actual operation, including the instantaneous speed, average speed, and acceleration of speed change of the unmanned vessel during operation. The wave environment parameter is the real-time sea state information of the unmanned vessel's operating environment, including the wave height, wave period, and wave direction of the sea area around the unmanned vessel. The load status parameter is the real-time weight information of the equipment carried by the unmanned vessel during actual operation.

[0051] Further, step S1: Collect operational data of the unmanned vessel through sensors, and preprocess it to obtain standardized monitoring data, including:

[0052] Step S11: Collect speed parameters, wave environment parameters, and load status parameters of the unmanned vessel during its operation based on sensors.

[0053] As explained, in this embodiment, operational data is collected sequentially using a speed sensor, a wave environment sensor, and a load status sensor. The speed parameter is measured in real-time by a speed sensor mounted on the hull. For example, when the unmanned vessel is cruising, the speed sensor can record the instantaneous speed of the unmanned vessel at fixed time intervals such as per second or per minute. The wave environment parameter is monitored in real-time by a wave environment sensor installed on the exterior of the unmanned vessel. For example, the wave environment sensor monitors the wave height and wave period in the sea area where the unmanned vessel is located, outputting complete wave height data and corresponding wave period data, while also recording the direction of wave propagation. The load status parameter is obtained in real-time by a load status sensor located inside the unmanned vessel, i.e., by measuring the actual weight of the equipment carried by the unmanned vessel in real-time.

[0054] Step S12: Use anomaly data monitoring methods to remove abnormal and / or noisy data from the operational data, and use time-series interpolation methods to unify the time-series sequence of the operational data to obtain standardized monitoring data.

[0055] As an alternative implementation method, the abnormal data monitoring method is a statistical method, such as a method for determining the data anomaly threshold based on the standard deviation principle.

[0056] Specifically, a data sequence is constructed for each parameter based on the real-time collected operational data. The mean and standard deviation of the data sequences corresponding to the speed parameter, wave environment parameter, and load state parameter are calculated. The upper and lower thresholds of each data sequence are determined, where the upper threshold is the sum of the mean and standard deviation, and the lower threshold is the difference between the mean and standard deviation. Each data point in each data sequence is compared with its corresponding upper and lower thresholds. Data points within the upper and lower thresholds are recorded as normal data. Conversely, data points exceeding the upper and lower thresholds are judged as abnormal data and / or noisy data and are removed. That is, the data cleaning work is completed based on the abnormal data monitoring method.

[0057] A time-series interpolation method is used to align and unify the data sequences corresponding to the speed parameters, wave environment parameters, and load state parameters. In this embodiment, linear interpolation is used for time-series calibration. That is, for each data sequence, based on the numerical linear relationship between two adjacent data points, the missing or unsampled data values ​​are determined, thereby obtaining data points with consistent time intervals on a unified time series, so that the data points of the speed parameters, wave environment parameters, and load state parameters are consistent on the time axis. For example, in actual operation, the speed parameter data of the unmanned vessel is collected once per second, the wave environment parameter is collected once every two seconds, and the load state parameter is collected once every three seconds. Linear interpolation is used to interpolate the wave environment parameters and load state parameters, so that the collection frequency of all wave environment parameters and load state parameters is unified to once per second, completing the time-series calibration. Finally, standardized monitoring data is obtained through abnormal data monitoring methods and time-series interpolation methods.

[0058] Further, step S2: Based on standardized monitoring data, analyze the changing trends of potential coupling relationships in the operational data, and determine the changing patterns of sensor drift characteristics, including:

[0059] Step S21: Based on the standardized monitoring data, use nonlinear correlation analysis to analyze the potential coupling relationship of the operational data and extract the changing trend of the operational data under different operating conditions of the unmanned vessel.

[0060] It is explained that the potential coupling relationship refers to the fact that during the operation of unmanned vessels, the operational data do not change independently, but rather there is a certain implicit, nonlinear interdependence and synergistic change relationship. This may not be directly reflected on the data surface, but can be captured by nonlinear correlation analysis methods to capture the nonlinear potential coupling characteristics between speed parameters, wave environment parameters, and load state parameters, and to reflect the nonlinear influence mechanism between speed parameters, wave environment parameters, and load state parameters in the actual operating environment.

[0061] Specifically, based on the nonlinear correlation analysis method, nonlinear transformation processing is performed on each data sequence corresponding to the speed parameter, wave environment parameter, and load state parameter. That is, each data sequence is transformed using a specific nonlinear function such as an exponential function or a logarithmic function. Each data point in each data sequence is used as the input of the nonlinear function, and the nonlinear function is used to transform it into new data, thus completing the nonlinear transformation processing. Then, the nonlinear correlation measure in the nonlinear correlation analysis method is used to quantitatively analyze the coupling relationship between the speed parameter, wave environment parameter, and load state parameter after the nonlinear transformation processing, and the nonlinear correlation analysis results are obtained.

[0062] It can be noted that the methods for measuring nonlinear correlation include, but are not limited to, nonlinear correlation coefficient calculation methods or entropy-based correlation calculation methods. In this embodiment, the nonlinear correlation coefficient calculation method is adopted. By calculating the covariance between the nonlinearly transformed speed parameter and the wave environment parameter, and dividing it by the product of the standard deviations of the speed parameter and the wave environment parameter, the nonlinear correlation coefficient between the speed parameter and the wave environment parameter is obtained. Similarly, the nonlinear correlation coefficients between the speed parameter and the load state parameter, and between the wave environment parameter and the load state parameter are calculated respectively.

[0063] Then, based on the results of nonlinear correlation analysis, the changing trends of speed parameters, wave environment parameters, and load state parameters under different operating conditions are specifically extracted. That is, by analyzing the law of nonlinear correlation coefficient changes with the operating state of the unmanned vessel, the coupling relationship between speed parameters, wave environment parameters, and load state parameters is determined under different operating conditions such as different speed conditions, different wave environment conditions, and different load state conditions of the unmanned vessel. For example, under low wave height operating conditions, the nonlinear correlation coefficient between speed parameters and load state parameters is small, while under high wave height operating conditions, the nonlinear correlation coefficient between speed parameters and load state parameters increases significantly, thereby determining the trend of parameter changes under different operating conditions.

[0064] Step S22: Use a nonlinear feature extraction algorithm to identify the sensor drift features corresponding to the changing trend, establish a regular model of the sensor drift features changing with the operating data, and obtain the changing law of the sensor drift features.

[0065] Optionally, the nonlinear feature extraction algorithm refers to an algorithm that can effectively extract nonlinear features from data. In this embodiment, the kernel principal component analysis algorithm is used. The kernel function maps the original data to a high-dimensional feature space, and principal component analysis is performed in the high-dimensional space to effectively extract the nonlinear features from the data.

[0066] Specifically, standardized monitoring data is input into a kernel principal component analysis algorithm. The kernel function maps the data to a high-dimensional feature space. Principal component analysis is then performed on the data in the high-dimensional feature space to extract principal component vectors representing sensor drift characteristics. Based on these principal component vectors, a feature subspace is constructed to identify and extract sensor drift characteristics corresponding to the changing trends of speed parameters, wave environment parameters, and load state parameters. For example, when the unmanned vessel's speed parameters are in a high speed range such as exceeding 20 knots per hour, and the wave height in the wave environment parameters is in a large wave range such as exceeding 3 meters, the speed parameters, wave environment parameters, and load state parameters exhibit nonlinear changing trends. Kernel principal component analysis (KPI) is employed to extract sensor drift features. This involves mapping speed parameters, wave environment parameters, and load state parameters to a high-dimensional feature space using a specific kernel function, such as a Gaussian kernel, to obtain a new high-dimensional data distribution. Principal component analysis is then performed in this high-dimensional feature space, selecting the direction with the largest explained variance from multiple principal component directions as the sensor drift feature direction. The maximum value of the principal component load corresponding to the sensor drift feature direction is used to represent and describe the sensor drift features under conditions of high-speed operation and large wave height of the unmanned vessel, thus achieving the quantification and description of sensor drift features.

[0067] Next, based on the sensor drift characteristics, a model is established to show the variation of sensor drift characteristics with speed parameters, wave environment parameters, and load state parameters. Specifically, a nonlinear regression analysis method is used to establish the functional relationship between sensor drift characteristics and speed parameters, wave environment parameters, and load state parameters. With speed parameters, wave environment parameters, and load state parameters as independent variables and sensor drift characteristics as dependent variables, the function coefficients are determined by fitting any nonlinear function in the form of an exponential function, polynomial function, or logarithmic function. This ensures that the fitted function can fully reflect the variation of sensor drift characteristics with speed parameters, wave environment parameters, and load state parameters.

[0068] For example, if, as the speed parameter gradually increases, the sensor drift characteristic is found to increase rapidly and non-linearly with the speed parameter, exhibiting an exponential growth trend, a quantitative relationship model between the speed parameter and the sensor drift characteristic can be established using nonlinear regression analysis. Specifically, the speed parameter is used as the independent variable, the sensor drift characteristic as the dependent variable, and an exponential function is selected as the basic function form. The undetermined coefficients in the exponential function model are determined using nonlinear least squares calculation. Specifically, given initial parameter values ​​for the exponential function, the best-fit parameter values ​​for the exponential function model are determined by minimizing the sum of squares of the differences between the speed parameter and the sensor drift characteristic obtained from the model. Finally, a functional expression for the sensor drift characteristic changing with the speed parameter is established, thus yielding a corresponding model of the sensor drift characteristic's variation with speed parameter, wave environment parameter, and load state parameter.

[0069] Further, step S3: Based on the changing patterns of sensor drift characteristics, identify the characteristic pattern deviation between the current and historical operating states of the unmanned vessel, and identify the characteristic patterns of changes in operating state, including:

[0070] Step S31: Based on the changing patterns of sensor drift characteristics, use deep clustering analysis to perform feature pattern clustering analysis on the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel, and determine the clustering positions of the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel in the feature space.

[0071] Specifically, the speed, wave environment, and load status parameters corresponding to the current and historical operating states of the unmanned surface vessel (USV) are input into a model that illustrates the variation of sensor drift characteristics with operational data. This model calculates the sensor drift characteristics corresponding to the current and historical operating states. For example, inputting the speed, wave environment, and load status parameters of the USV operating at high speeds exceeding 20 knots and wave heights exceeding 3 meters into the model yields the sensor drift characteristics under high-speed operating conditions. Similarly, inputting the speed, wave environment, and load status parameters of the USV historically operating at low speeds below 10 knots and wave heights below 1 meter into the model calculates the sensor drift characteristics under historical operating conditions, representing the sensor drift characteristics under low-speed operating conditions.

[0072] Then, deep clustering analysis is used to perform feature pattern clustering analysis on the sensor drift features corresponding to the current and historical operating states of the unmanned vessel. The sensor drift features corresponding to the current and historical operating states are input into the deep clustering analysis algorithm, and feature dimensionality reduction processing is performed on the sensor drift features through multi-layer nonlinear transformation. The dimensionality reduction method includes, but is not limited to, feature encoding by an autoencoder network to obtain a compressed representation that can represent the original data features. That is, the autoencoder network compresses and obtains a low-dimensional feature representation that can express the sensor drift features by performing continuous nonlinear mapping and reconstruction on the sensor drift features. Subsequently, the compressed data is classified using clustering analysis. That is, specific clustering algorithms such as distance-based clustering are used on the low-dimensional feature representation. The Euclidean distance between each data point and all cluster centers is calculated for the low-dimensional feature representation. The cluster center with the smallest Euclidean distance is selected as the classification result of the data point, and the cluster position of the current and historical operating states of the unmanned vessel is determined.

[0073] Step S32: Analyze the deviation of cluster positions and identify the characteristic patterns of changes in the unmanned vessel's operating status.

[0074] Specifically, firstly, a feature space is constructed. The sensor drift features output by the regularity model are dimensionality-reduced through an autoencoder network to obtain a low-dimensional feature representation. In the feature space, each dimension represents a combination of the features of the speed parameter, wave environment parameter, and load state parameter. The clustering position is determined by the aforementioned step S31, which is represented by the coordinate position of the sensor drift features in the dimensionality-reduced low-dimensional feature space. That is, the position coordinates of the sensor drift features in the low-dimensional feature space under the current operating state and the historical operating state are obtained through cluster analysis to express the operating state feature pattern of the unmanned vessel.

[0075] For example, when the speed parameters of the unmanned surface vessel (USV) under its current operating conditions are 20 to 25 knots per hour and the wave height is 3 to 5 meters, the cluster positions of the sensor drift features obtained through deep cluster analysis are located in a set of coordinate positions in the feature space representing high-speed and high-wave environments. When the speed parameters under historical operating conditions are 10 to 12 knots per hour and the wave height is 0.5 to 1 meter, the cluster positions of the obtained sensor drift features are located in another set of coordinate positions representing low-speed and low-wave environments. The cluster positions corresponding to the current and historical operating conditions are recorded, and the square of the difference between the coordinates of the two cluster positions is calculated. The square root of the sum of the squares of all coordinate differences is then taken to obtain the position deviation distance in the feature space. This deviation between the two cluster positions is then determined to reflect the feature pattern deviation between the current and historical operating conditions of the USV, and the deviation indicates the degree of change in the USV's operating conditions.

[0076] Then, the characteristic patterns of changes in the unmanned vessel's operating state are identified based on the deviation. For example, the deviation distance between the cluster positions of sensor drift features in the current high-speed, high-wave operating state and the cluster positions of sensor drift features in the historical low-speed, low-wave operating state is large, indicating a difference between the two operating states. Thus, the characteristic patterns of changes in the unmanned vessel's operating state are identified and defined; among them, the characteristic pattern is the change in the unmanned vessel's operating state from a historical low-load, low-wave environment, low-speed mode to a high-load, high-wave environment, high-speed mode.

[0077] Further, step S4: Quantifying the applicability of the historical fault early warning model to the current operating state of the unmanned vessel based on the changing patterns of sensor drift characteristics, including:

[0078] Step S41: Based on the changing patterns of sensor drift characteristics, obtain the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel, and input them into the historical fault early warning model to obtain the fault risk probability of the unmanned vessel under the current and historical operating states.

[0079] It is explained that the historical fault early warning model is a fault early warning classification model trained based on the operational data obtained from the historical monitoring of the unmanned vessel and the corresponding fault markers. It can output the probability of fault risk in the historical operating state of the unmanned vessel according to the input sensor drift characteristics.

[0080] Specifically, based on step S31, the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel are determined and sequentially input into the historical fault early warning model. After processing by multiple feature mapping layers and classification layers within the model, the fault risk probability of the unmanned vessel in the current and historical operating states is obtained.

[0081] Step S42: Analyze the differences between failure risk probabilities using transfer learning methods to determine the applicability of the historical failure early warning model in the current operating state of the unmanned vessel.

[0082] Specifically, the difference between the failure risk probability of the unmanned vessel in its current operating state and the failure risk probability in its historical operating state is calculated using the transfer learning method. This difference serves as an applicability evaluation index for the historical failure warning model in the current operating state, denoted as the applicability degree, to assess the predictive performance of the historical failure warning model in the current operating state of the unmanned vessel. Here, the transfer learning method refers to machine learning technology, which can effectively transfer and apply knowledge and models learned in one domain to another related but different domain, which is beneficial for subsequent analysis of the generalization performance and predictive accuracy of the historical failure warning model.

[0083] Further, step S5: Correct the fault warning threshold of the current historical fault warning model based on the characteristic patterns and applicability of the changes in operating status, and output the operating condition warning model, including:

[0084] Step S51: Determine the degree of change in the unmanned vessel's operating status based on the characteristic patterns of the changes.

[0085] Specifically, based on the determined characteristic patterns of unmanned surface vessel (USV) operational status changes, step S3 can specifically determine the deviation of the clustering positions of sensor drift features corresponding to the current operational status and historical operational status of the USV. The degree of change in the USV's operational status is obtained through the deviation, that is, the deviation is divided into different degree levels according to the magnitude of the deviation, such as low degree of change, medium degree of change, and high degree of change. Then, a preset threshold is set based on the corresponding degree level of change and compared with the deviation. For example, the preset threshold corresponding to low degree of change is the first predetermined threshold. If the deviation is less than the first predetermined threshold, the degree of change in the USV's operational status is determined to be low. The lower limit corresponding to medium degree of change is the first predetermined threshold, and the upper limit is the second predetermined threshold. If the deviation is between the first predetermined threshold and the second predetermined threshold, the degree of change in the USV's operational status is determined to be medium. If the deviation is greater than the second predetermined threshold, the degree of change in the USV's operational status is determined to be high.

[0086] Step S52: Evaluate the applicability of the model based on the applicability of the historical fault warning model under the current operating state of the unmanned vessel.

[0087] Specifically, before analyzing the applicability of the historical fault warning model under the current operating state of the unmanned vessel, step S4 determines the fault risk probability of the unmanned vessel under the current and historical operating states. The fault probability difference is obtained by comparing the fault risk probabilities under the two operating states. Based on the fault probability difference, the applicability of the historical fault warning model is classified into levels, such as high applicability, medium applicability, and low applicability. A difference threshold is preset for each level of model applicability and compared with the fault probability difference. For example, the difference threshold corresponding to high applicability is the low difference threshold. If the fault probability difference is less than the low difference threshold, the historical fault warning model is determined to be highly applicable under the current operating state of the unmanned vessel. The difference threshold corresponding to medium applicability has a lower limit of the low difference threshold and an upper limit of the high difference threshold. If the fault probability difference is between the low and high difference thresholds, the historical fault warning model is determined to be moderately applicable under the current operating state of the unmanned vessel. If the fault probability difference is greater than the high difference threshold, the historical fault warning model is determined to be low applicable under the current operating state of the unmanned vessel.

[0088] Step S53: Establish a correspondence between the degree of change in the unmanned vessel's operating status and the applicability of the model, correct the fault warning threshold in the historical fault warning model, and output a working condition warning model applicable to the current operating status of the unmanned vessel.

[0089] Specifically, a correspondence is established by creating a two-dimensional correspondence matrix with the degree of change in the unmanned vessel's operating state as the horizontal axis and the degree of model applicability as the vertical axis. Each position in the matrix represents a combination of the degree of change in the operating state and the degree of model applicability.

[0090] It can be explained that when the unmanned vessel's operating status changes at a high degree of change and the model's applicability is low, the combination position in the corresponding relationship matrix indicates that the historical fault warning model must be corrected for the fault warning threshold under the current operating status. When the unmanned vessel's operating status changes at a moderate degree of change and the model's applicability is moderate, the combination position in the corresponding relationship matrix indicates that the fault warning threshold in the historical fault warning model needs to be appropriately corrected. When the unmanned vessel's operating status changes at a low degree of change and the model's applicability is high, the combination position in the corresponding relationship matrix indicates that the fault warning threshold does not need to be corrected. For example, when the combination position in the corresponding relationship indicates that the fault warning threshold must be corrected, the original fault warning threshold in the historical fault warning model is increased or decreased. When the combination position in the corresponding relationship indicates that appropriate correction is needed, the fault warning threshold is moderately adjusted to adapt to the fault prediction of the current operating status. When the combination position in the corresponding relationship indicates that correction is not needed, the fault warning threshold is not adjusted.

[0091] Next, based on the historical fault warning model with the corrected fault warning threshold, that is, using the historical fault warning model as a foundation, the corresponding fault warning thresholds in the historical fault warning model are updated using the corrected fault warning thresholds to obtain a condition warning model applicable to the current operating state of the unmanned vessel. For example, when the unmanned vessel's speed parameters are high, the wave environment parameters are high, and the model's applicability is low under the current operating state, the original fault warning thresholds in the historical fault warning model are increased. After updating the original fault warning thresholds in the historical fault warning model, a new condition warning model is output. The condition warning model can reflect the actual fault risk probability of the unmanned vessel under the current operating conditions and is more suitable for judging the fault risk of the unmanned vessel under the current operating state.

[0092] Further, step S6: Analyze standardized monitoring data using an operational condition early warning model to generate early warning information for unmanned vessel operational malfunctions, including:

[0093] Step S61: Analyze standardized monitoring data based on the working condition early warning model to obtain the failure risk probability of the unmanned vessel under the current operating state.

[0094] Specifically, standardized monitoring data is input into the operating condition early warning model, which is processed through multiple feature mapping layers and classification prediction layers to output the failure risk probability corresponding to the current operating state of the unmanned vessel. The operating condition early warning model is based on the historical failure early warning model and includes multiple nonlinear mapping processing units and prediction units. The nonlinear mapping processing units perform layer-by-layer feature transformation processing on the standardized monitoring data, gradually mapping the input speed parameters, wave environment parameters, and load state parameters into high-dimensional feature representations. The prediction units then calculate the failure risk probability based on the high-dimensional feature representations.

[0095] Step S62: Compare the corrected fault warning threshold and fault risk probability to determine whether the unmanned vessel has reached the fault warning condition under the current operating state. If so, generate a warning message for the unmanned vessel's operating fault.

[0096] It can be explained that fault warning conditions refer to the specific rules and standards used to judge and determine whether to issue warning signals regarding potential or actual faults during the operation of unmanned surface vessels (USVs). These rules provide timely decision support for operations and maintenance personnel, ensuring navigation safety and mission reliability. Warning information includes, but is not limited to, a description of the fault warning level, the possible location of the fault, and the fault type. The fault warning level is typically divided into different levels based on severity and urgency, indicating the potential impact of the fault. The possible location of the fault directly identifies the specific area or component within the USV where the fault exists. The description of the fault type further details the nature, characteristics, and possible manifestations of the potential fault. These warning messages collectively constitute a comprehensive alert to potential problems, helping to take appropriate preventative or responsive measures in advance.

[0097] Specifically, by comparing the fault risk probability obtained from standardized monitoring data with the corrected fault warning threshold, it is determined whether the unmanned vessel has reached the fault warning condition under the current operating state. If the fault risk probability is less than the corrected fault warning threshold, it is determined that the unmanned vessel has not reached the fault warning condition under the current operating state. Conversely, if the fault risk probability is greater than or equal to the corrected fault warning threshold, it is determined that the unmanned vessel has reached or exceeded the fault warning condition under the current operating state, and a warning message of unmanned vessel operation failure is generated to alert maintenance personnel to carry out timely maintenance and adjustment.

[0098] For example, in actual operation, when the probability of failure risk of the unmanned vessel in its current operating state is greater than the corrected failure warning threshold, a warning message is output. The warning message is as follows: the failure warning level of the unmanned vessel's current operating state is high risk, the possible failure location is the shafting component of the unmanned vessel's propulsion system, and the failure type is that due to long-term operation in high-speed, high-wave environment and high-load conditions, the shafting component has a potential risk of wear or fatigue failure. It is recommended to immediately reduce the operating speed or take active maintenance measures.

[0099] Understandably, by analyzing the changing trends of potential coupling relationships between operational data during the operation of unmanned vessels, sensor drift characteristics can be dynamically identified, enhancing the ability to identify nonlinear interference in complex operating environments. Next, characteristic patterns of operational state changes can be identified to effectively address monitoring deviations caused by sudden changes in operational state. Then, the applicability of historical fault warning models can be evaluated based on the changing patterns of sensor drift characteristics, enabling these models to make adaptive judgments under new operating conditions. By adjusting the fault warning threshold using the characteristic patterns of operational state changes and the applicability evaluation index of historical fault warning models (i.e., the degree of applicability), a condition-based warning model can be generated, enhancing the model's generalization ability and response accuracy. Based on the condition-based warning model, fault risk assessment can be performed, enabling early identification and proactive warning of unmanned vessel operational faults, thereby improving the intelligence level and safety assurance capabilities of unmanned vessel operation.

[0100] The second embodiment of the present invention provides an electronic device, which includes a processor, a communication interface, a memory, and a communication bus. The processor, the communication interface, and the memory communicate with each other through the communication bus. The processor calls logical instructions in the memory to execute the unmanned vessel status monitoring and fault early warning system based on big data analysis described in any embodiment of the present invention.

[0101] During its operation, it requires the use of an unmanned vessel status monitoring and fault early warning system based on big data analysis. Therefore, whether the equipment and program data are integrated or different hardware is configured to produce functions with similar effects to those achieved by this invention, they all fall within the protection scope of this invention. This equipment has the same beneficial effects as the aforementioned unmanned vessel status monitoring and fault early warning system based on big data analysis, and will not be elaborated here.

[0102] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0103] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0104] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0105] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0106] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0107] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0108] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0109] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0111] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A system for monitoring the condition and providing early warning of faults of unmanned vessels based on big data analysis, characterized in that: The system includes: The parameter acquisition module is used to: collect operational data of the unmanned vessel through sensors and preprocess it to obtain standardized monitoring data; The coupling analysis module is used to: analyze the changing trends of potential coupling relationships in operational data based on standardized monitoring data, and determine the changing patterns of sensor drift characteristics; The status recognition module is used to: identify the characteristic pattern deviation between the current and historical operating states of the unmanned vessel based on the changing patterns of sensor drift characteristics, and to identify the characteristic patterns of changes in operating state; The model evaluation module is used to: acquire historical fault warning models and quantify the applicability of historical fault warning models in the current operating state of unmanned vessels based on the changing patterns of sensor drift characteristics; The threshold correction module is used to: comprehensively adjust the fault warning threshold of the current historical fault warning model based on the characteristic patterns and applicability of the changes in operating status, and output the operating condition warning model. The early warning output module is used to: analyze standardized monitoring data using an operational condition early warning model to generate early warning information for unmanned vessel operational failures.

2. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, The operational data includes speed parameters, wave environment parameters, and load status parameters.

3. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 2, characterized in that, The unmanned surface vessel (USV) collects operational data through sensors, and then preprocesses it to obtain standardized monitoring data, including: Based on the sensor data collected during the operation of the unmanned vessel, including speed parameters, wave environment parameters, and load status parameters; Abnormal data and / or noisy data in the operation data are removed by an abnormal data monitoring method, and the time series sequence of the operation data is unified by a time series interpolation method to obtain standardized monitoring data.

4. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, Based on standardized monitoring data, analyze the changing trends of potential coupling relationships in operational data, and determine the changing patterns of sensor drift characteristics, including: Based on standardized monitoring data, nonlinear correlation analysis was used to analyze the potential coupling relationship of operational data and extract the changing trend of operational data under different operating conditions of unmanned vessels. A nonlinear feature extraction algorithm is used to identify sensor drift characteristics corresponding to changing trends, and a model of the variation of sensor drift characteristics with operating data is established to obtain the variation law of sensor drift characteristics.

5. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, Based on the changing patterns of sensor drift characteristics, the system identifies the characteristic pattern deviation between the current and historical operating states of the unmanned surface vessel (USV), and identifies the characteristic patterns of changes in operating state, including: Based on the changing patterns of sensor drift characteristics, deep clustering analysis is used to perform feature pattern clustering analysis on the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel, and the clustering positions of the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel in the feature space are determined in turn. Analyze the deviation in cluster positions to identify characteristic patterns of changes in the operational status of unmanned vessels.

6. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, The applicability of historical fault early warning models to the current operating state of unmanned vessels is quantified based on the changing patterns of sensor drift characteristics, including: Based on the changing patterns of sensor drift characteristics, the sensor drift characteristics corresponding to the current and historical operating states of the unmanned vessel are obtained and input into the historical fault early warning model to obtain the fault risk probability of the unmanned vessel under the current and historical operating states. By analyzing the differences between failure risk probabilities using transfer learning methods, the applicability of historical failure early warning models to the current operating state of unmanned vessels can be determined.

7. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, The fault warning threshold of the current historical fault warning model is revised based on the characteristic patterns and applicability of the comprehensive operating status changes, and the operating condition warning model is output, including: The degree of change in the unmanned vessel's operational status is determined based on the characteristic patterns of the changes. The applicability of the model is evaluated based on its applicability to the current operating state of the unmanned vessel, according to the historical fault early warning model. By establishing a correlation between the degree of change in the unmanned vessel's operating status and the applicability of the model, the fault warning threshold in the historical fault warning model is corrected, and a working condition warning model applicable to the current operating status of the unmanned vessel is output.

8. The unmanned vessel status monitoring and fault early warning system based on big data analysis according to claim 1, characterized in that, Standardized monitoring data is analyzed using an operational condition early warning model to generate early warning information for unmanned surface vessel (USV) operational malfunctions, including: Based on the analysis of standardized monitoring data using the operating condition early warning model, the probability of failure risk of the unmanned vessel under the current operating condition is obtained; By comparing the corrected fault warning threshold and fault risk probability, it is determined whether the unmanned vessel has reached the fault warning condition under the current operating state. If so, a warning message for the unmanned vessel's operating fault is generated.

9. An electronic device, characterized in that, The device includes a processor, a communication interface, a memory, and a communication bus. The processor, the communication interface, and the memory communicate with each other through the communication bus. The processor calls logical instructions in the memory to execute the unmanned vessel status monitoring and fault early warning system based on big data analysis as described in any one of claims 1 to 8.

Images