A net cage boat failure early warning system

Abstract

The present application relates to the technical field of fault early warning, and particularly relates to a net cage ship fault early warning system, comprising: a data acquisition module, comprising an acquisition unit for acquiring operation data of each acquisition point of the net cage ship and a data transmission unit connected with the acquisition unit for converting the operation data into communication signals and transmitting the communication signals to a monitoring center; a data processing module connected with the data acquisition module, comprising a preprocessing unit for preprocessing the communication signals to output normalized data; an early warning module connected with the data processing module for issuing a fault early warning according to an analysis result; and a control module connected with the data acquisition module, the data processing module and the early warning module respectively for determining a change rate of adjacent acquisition points according to a false alarm rate of the fault early warning. The present application improves the accuracy of fault early warning.

Description

Technical Field

[0001] This invention relates to the field of fault early warning technology, and in particular to a fault early warning system for net cage ships. Background Technology

[0002] In the current technology, as aquaculture develops to the deep sea, the application of cage vessels is becoming more and more widespread. Cage vessels face complex marine environments and various potential risks during the aquaculture process, such as equipment failure and severe weather. These factors may lead to aquaculture losses or even casualties. In order to ensure the safe operation of cage vessels and the smooth progress of aquaculture production, improve the intelligence level of aquaculture platforms, and reduce the management difficulty of aquaculture platforms, an effective fault early warning system is needed to monitor the status of cage vessels in real time, detect potential faults in advance, and take timely measures.

[0003] Chinese Patent Publication No. CN118230517A discloses a ship fault early warning and positioning system, comprising: a data processing module, which is connected to a ship positioning module, a fault diagnosis module, a system management module, a backend database module, and an environmental monitoring module. The fault diagnosis module is also connected to a real-time monitoring module and an equipment positioning module. The system management module is also connected to a backend data module, an early warning alarm module, and a user operation module. This invention features a simple and easy-to-operate system that can send alarm signals to the user's terminal while responding to ship alarm devices, enabling rapid and timely responses to different alarm types. It rapidly feeds back positioning information via BeiDou satellite and ship locators. Furthermore, it can not only issue alarms for ship equipment malfunctions but also for deviations from the ship's navigation route and sudden emergencies. It can also monitor the ship's surrounding navigation environment in real time, allowing for timely avoidance of potentially hazardous routes.

[0004] It is evident that existing technologies have the following problems: when faced with a large amount of real-time data, the computing resources of the data processing module may be insufficient, resulting in untimely or incomplete data processing, which affects the accuracy and timeliness of fault warnings. Summary of the Invention

[0005] To address this issue, the present invention provides a fault early warning system for net cage vessels, which overcomes the problem in the prior art where, when faced with a large amount of real-time data, the computing resources of the data processing module may be insufficient, resulting in untimely or incomplete data processing, thus affecting the accuracy and timeliness of fault early warning.

[0006] To achieve the above objectives, the present invention provides a fault early warning system for net cage vessels, comprising:

[0007] The data acquisition module includes an acquisition unit for collecting operational data from various collection points of the net cage vessel and a data transmission unit connected to the acquisition unit for converting the operational data into communication signals and transmitting them to the monitoring center.

[0008] The data processing module, which is connected to the data acquisition module, includes a preprocessing unit for preprocessing the communication signal to output normalized data and a data analysis unit connected to the preprocessing unit for performing early warning analysis on the normalized data through a neural network model to output analysis results.

[0009] An early warning module, which is connected to the data processing module, is used to issue fault warnings based on the analysis results;

[0010] The control module, which is connected to the data acquisition module, the data processing module and the early warning module respectively, is used to determine the change rate of adjacent acquisition points based on the false alarm rate of fault early warning, or to determine the training enhancement ratio of missing data in the neural network model based on the byte missing rate of normalized data, and to determine the fusion threshold of running data of different acquisition points based on the signal-to-noise ratio of communication signals.

[0011] Furthermore, the control module is used to determine whether the accuracy of the fault warning meets the requirements based on the false alarm rate of the fault warning. If the false alarm rate of the fault warning is greater than the preset first false alarm rate, then it is determined that the accuracy of the fault warning does not meet the requirements.

[0012] Furthermore, the control module is used to preliminarily determine that the effectiveness of the early warning analysis does not meet the requirements when the false alarm rate of the fault warning is greater than the preset first false alarm rate and less than or equal to the preset second false alarm rate, and to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data.

[0013] Furthermore, the control module is used to increase the rate of change of adjacent acquisition points when the false alarm rate of the fault warning is greater than the preset second false alarm rate;

[0014] The increase in the rate of change of adjacent acquisition points is determined by the difference between the false alarm rate of the fault warning and the preset second false alarm rate.

[0015] Furthermore, the control module is used to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data. If the byte missing rate of the normalized data is greater than the preset first missing rate, then the effectiveness of the early warning analysis is determined to be unsatisfactory.

[0016] Furthermore, the control module is used to increase the training enhancement ratio of missing data in the neural network model when the byte missing rate of the normalized data is greater than the preset first missing rate and less than or equal to the preset second missing rate.

[0017] Furthermore, the control module is used to preliminarily determine that the transmission stability of the running data does not meet the requirements when the byte missing rate of the normalized data is greater than the preset second missing rate, and to determine whether the transmission stability of the running data meets the requirements based on the signal-to-noise ratio of the communication signal.

[0018] Furthermore, the increase in the training enhancement ratio of missing data in the neural network model is determined by the difference between the byte missing rate of the normalized data and the preset first missing rate.

[0019] Furthermore, the control module is used to determine whether the transmission stability of the operating data meets the requirements based on the signal-to-noise ratio of the communication signal. If the signal-to-noise ratio of the communication signal is less than the preset signal-to-noise ratio, it is determined that the transmission stability of the operating data does not meet the requirements, and the fusion threshold of the operating data from different collection points is increased.

[0020] Furthermore, the increase in the data fusion threshold of different collection points is determined by the difference between the preset signal-to-noise ratio and the signal-to-noise ratio of the communication signal.

[0021] Compared with existing technologies, the beneficial effects of this invention are as follows: The system of this invention, by setting up a data acquisition module, a data processing module, an early warning module, and a control module, adjusts the rate of change of adjacent acquisition points according to the false alarm rate of fault warning. Since equipment vibration may cause slight displacement of some components, changing parameters such as connections and gaps that were originally normal, sensors monitoring these parameters will detect numerical changes. The system cannot accurately distinguish between normal component wear and temporary displacement caused by vibration, and may mistakenly issue an early warning as a component failure. By increasing the rate of change of adjacent acquisition points, high rate of change signals in a short period can be ignored, and alarms can only be triggered for displacements that have been stably exceeding limits for a long time, thereby avoiding false judgments. Furthermore, the system adjusts the training enhancement ratio of missing data in the neural network model according to the byte missing rate of normalized data. Because data from some monitoring points in the dataset is not uploaded in time or data gaps occur during certain periods due to equipment failure, the analysis results may be biased and unable to fully and accurately reflect the actual situation. The accuracy of model training and results may be affected, potentially failing to accurately describe the underlying patterns in the data, leading to inaccurate early warning results. Increasing the training augmentation ratio of missing data in the neural network model allows the model more opportunities to learn from samples containing missing data, thus better adapting to situations with missing data, improving the model's ability to handle missing data, and reducing bias caused by missing data. Adjusting the data fusion threshold for different collection points based on the signal-to-noise ratio of the communication signal is also crucial. Since various electromagnetic interference sources may exist around the net cage vessel, such as generators installed too close to sensitive electronic equipment, the resulting electromagnetic fields can directly couple to these devices, affecting their performance and interfering with the communication link of the fault early warning system, leading to signal distortion and increased bit error rate. Increasing the data fusion threshold for different collection points can, to some extent, filter out unreliable data that may be affected by electromagnetic interference, thereby improving the accuracy and reliability of data fusion and reducing the risk of erroneous fusion due to interference.

[0022] Furthermore, the system described in this invention adjusts the rate of change of adjacent acquisition points by setting a first false alarm rate and a second false alarm rate. Since equipment vibration may cause some components to undergo slight displacement, causing changes in normally normal parameters such as connections and gaps, the sensors monitoring these parameters will detect the numerical changes. The system cannot accurately distinguish whether it is normal component wear or temporary displacement caused by vibration, and may mistakenly judge it as a component failure and issue an early warning. By increasing the rate of change of adjacent acquisition points, high rate of change signals in a short period of time can be ignored, and alarms can only be triggered for displacements that are stable and exceed limits for a long period of time, thereby avoiding misjudgment and improving the accuracy of fault early warning.

[0023] Furthermore, the system of the present invention adjusts the training augmentation ratio of missing data in the neural network model by setting a first missing rate and a second missing rate. Since some monitoring points in the dataset are not uploaded in time or data gaps occur in certain periods due to equipment failure, the analysis results may be biased and unable to fully and accurately reflect the actual situation, affecting the accuracy of model training and results. It may not be able to accurately describe the patterns behind the data, resulting in inaccurate early warning results. By increasing the training augmentation ratio of missing data in the neural network model, the model can have more opportunities to learn samples containing missing data, thereby better adapting to the situation of missing data, improving the model's ability to process missing data, reducing the bias caused by missing data, and further improving the accuracy of fault early warning.

[0024] Furthermore, the system described in this invention adjusts the data fusion threshold of different collection points by preset signal-to-noise ratio. Since there may be various electromagnetic interference sources around the net cage vessel, such as the generator on the vessel being installed too close to sensitive electronic equipment, the electromagnetic field generated by it will directly couple into these devices, affecting their performance and interfering with the communication link of the fault early warning system, leading to signal distortion and increased bit error rate. By increasing the data fusion threshold of different collection points, unreliable data that may be affected by electromagnetic interference can be filtered out to a certain extent, thereby improving the accuracy and reliability of data fusion, reducing the risk of erroneous fusion due to interference, and further improving the accuracy of fault early warning. Attached Figure Description

[0025] Figure 1 This is an overall structural block diagram of the cage ship fault early warning system according to an embodiment of the present invention;

[0026] Figure 2 This is a flowchart illustrating the process of determining the rate of change of adjacent data collection points in the cage ship fault early warning system of this invention.

[0027] Figure 3 This is a flowchart illustrating the process of determining the training enhancement ratio of missing data in the neural network model of the cage ship fault early warning system according to an embodiment of the present invention.

[0028] Figure 4 This is a flowchart illustrating the process of determining the data fusion threshold for different collection points in the cage vessel fault early warning system according to an embodiment of the present invention. Detailed Implementation

[0029] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0030] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0031] Please see Figure 1 , Figure 2 , Figure 3 as well as Figure 4 The diagrams shown are, respectively, the overall structural block diagram of the cage vessel fault early warning system according to an embodiment of the present invention, the logical flowchart of the process for determining the rate of change of adjacent collection points, the logical flowchart of the process for determining the training enhancement ratio of missing data in the neural network model, and the logical flowchart of the process for determining the fusion threshold of operating data from different collection points. The present invention provides a cage vessel fault early warning system, comprising:

[0032] The data acquisition module includes an acquisition unit for collecting operational data from various collection points of the net cage vessel and a data transmission unit connected to the acquisition unit for converting the operational data into communication signals and transmitting them to the monitoring center.

[0033] The data processing module, which is connected to the data acquisition module, includes a preprocessing unit for preprocessing the communication signal to output normalized data and a data analysis unit connected to the preprocessing unit for performing early warning analysis on the normalized data through a neural network model to output analysis results.

[0034] An early warning module, which is connected to the data processing module, is used to issue fault warnings based on the analysis results;

[0035] The control module, which is connected to the data acquisition module, the data processing module and the early warning module respectively, is used to determine the change rate of adjacent acquisition points based on the false alarm rate of fault early warning, or to determine the training enhancement ratio of missing data in the neural network model based on the byte missing rate of normalized data, and to determine the fusion threshold of running data of different acquisition points based on the signal-to-noise ratio of communication signals.

[0036] Specifically, the operating data includes engine speed, inverter temperature, and hull tilt angle.

[0037] Specifically, preprocessing includes converting the signal into data, removing noise, handling missing values, and normalization.

[0038] Specifically, neural network models include feedforward neural networks, convolutional neural networks, and recurrent neural networks.

[0039] Specifically, the normalized data includes noise-removed engine speeds, inverter temperatures after processing for missing values, and the ship's inclination angles after normalization.

[0040] Specifically, the analysis results include alarm time, alarm type, and fault frequency.

[0041] Specifically, the false alarm rate of fault warning is the ratio of the number of false warning signals issued to the total number of warning signals issued.

[0042] Specifically, erroneous warning signals are issued when the warning system misjudges normal equipment operating parameters as abnormal due to reasons such as sensor failure, signal transmission error, data processing algorithm failure, or environmental interference.

[0043] Specifically, the rate of change of adjacent sampling points is the ratio of the difference between the measurements of two time-adjacent sampling points (the value at the later time point minus the value at the previous time point) to the measurement of the previous sampling point.

[0044] Specifically, the byte missing rate of normalized data is the ratio of the number of missing bytes in the normalized data to the total number of bytes in the normalized data.

[0045] Specifically, the training augmentation ratio for missing data in a neural network model is the ratio of the number of data points with missing values ​​added during the training process to the original number of data points with missing values.

[0046] Specifically, the signal-to-noise ratio (SNR) of a communication signal is the ratio of the communication signal power to the additive white Gaussian noise power.

[0047] Specifically, the data fusion threshold for different collection points is a boundary value used to determine whether data from different collection points should be fused.

[0048] In implementation, the system of this invention, by setting up a data acquisition module, a data processing module, an early warning module, and a control module, adjusts the rate of change of adjacent acquisition points based on the false alarm rate of fault warnings. Since equipment vibration may cause minor displacements in some components, altering normally functioning parameters such as connections and gaps, sensors monitoring these parameters will detect these changes. However, the system cannot accurately distinguish between normal component wear and temporary displacement caused by vibration, potentially misjudging it as a component failure and issuing an early warning. By increasing the rate of change of adjacent acquisition points, high-rate-of-change signals in short periods can be ignored, and alarms will only be triggered for long-term stable out-of-limit displacements, thus avoiding false judgments. Furthermore, the system adjusts the training augmentation ratio for missing data in the neural network model based on the byte missing rate of normalized data. Because data from some monitoring points in the dataset is not uploaded in a timely manner or data gaps occur during certain periods due to equipment failure, the analysis results may be biased, failing to fully and accurately reflect the actual situation and affecting model training and results. While accurate data fusion might not accurately describe the underlying patterns, leading to inaccurate early warning results, increasing the training augmentation ratio of missing data in the neural network model allows the model more opportunities to learn from samples containing missing data. This improves the model's ability to handle missing data, reducing bias caused by data gaps. Adjusting the data fusion threshold for different collection points based on the signal-to-noise ratio of the communication signal is also crucial. Since various electromagnetic interference sources may exist around the net cage vessel, such as generators installed too close to sensitive electronic equipment, their electromagnetic fields can directly couple to these devices, affecting their performance and interfering with the communication link of the fault early warning system, leading to signal distortion and increased bit error rate. Increasing the data fusion threshold for different collection points can filter out unreliable data that may be affected by electromagnetic interference, thereby improving the accuracy and reliability of data fusion and reducing the risk of erroneous fusion due to interference.

[0049] Specifically, the control module is used to determine whether the accuracy of the fault warning meets the requirements based on the false alarm rate of the fault warning. If the false alarm rate of the fault warning is greater than the preset first false alarm rate, then it is determined that the accuracy of the fault warning does not meet the requirements.

[0050] Specifically, the control module is used to preliminarily determine that the effectiveness of the early warning analysis does not meet the requirements when the false alarm rate of the fault warning is greater than the preset first false alarm rate and less than or equal to the preset second false alarm rate, and to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data.

[0051] It is understandable that the three intervals defined by the preset first false positive rate and the preset second false positive rate correspond to three different scenarios:

[0052] The first interval is when the false alarm rate of the fault warning is less than or equal to the preset first false alarm rate, which corresponds to the situation where the accuracy of the fault warning meets the requirements.

[0053] The second interval is when the false alarm rate of the fault warning is greater than the preset first false alarm rate and less than or equal to the preset second false alarm rate. The corresponding situation is: due to the failure to upload data from some monitoring points in the dataset in a timely manner or due to equipment failure, data gaps may occur in certain periods, which may cause deviations in the analysis results. This may not fully and accurately reflect the actual situation, affecting model training and the accuracy of results. It may not be able to accurately describe the patterns behind the data, resulting in inaccurate warning results.

[0054] The third interval is when the false alarm rate of the fault warning is greater than the preset second false alarm rate. The corresponding situation is: due to equipment vibration, some components may undergo slight displacement, causing changes in parameters such as normal connections and gaps. The sensors monitoring these parameters will detect the changes in values. The system cannot accurately distinguish whether it is normal component wear or temporary displacement caused by vibration, and may mistakenly judge it as a component failure and issue a warning.

[0055] Understandably, the preset false alarm rate can be set based on historical data, aiming to ensure the accuracy and practicality of the test results. Optionally, the preset false alarm rate is determined by analyzing past operating data of the net cage vessel fault early warning system and statistically analyzing the ratio of actual false alarms to total alarms within a certain period, using this as a reference to determine the value range. For example, the preset first false alarm rate is generally selected in the range of [1%, 3%], and the preset second false alarm rate is generally selected in the range of [4%, 6%].

[0056] Preferably, the first false alarm rate is 2% in a preferred embodiment, and the second false alarm rate is 5% in a preferred embodiment.

[0057] Specifically, the control module is used to increase the rate of change of adjacent acquisition points when the false alarm rate of the fault warning is greater than the preset second false alarm rate;

[0058] The increase in the rate of change of adjacent acquisition points is determined by the difference between the false alarm rate of the fault warning and the preset second false alarm rate.

[0059] Specifically, when the difference between the false alarm rate of the fault warning and the preset second false alarm rate is within 1%, the change rate of adjacent data collection points increases to 1.2 times the original rate. When the difference between the false alarm rate of the fault warning and the preset second false alarm rate exceeds 1%, in addition to increasing to 1.2 times the original rate, for every 0.5% increase, the change rate of adjacent data collection points increases by 1%. For example, if the difference between the false alarm rate of the fault warning and the preset second false alarm rate is 2%, and the current change rate of adjacent data collection points is 5%, the increased change rate of adjacent data collection points will be 5 × 1.2 + 1 × 2 = 8%.

[0060] Specifically, in this scheme, the rate of change of adjacent sampling points is the rate of change of liquid tank pressure in the cage ship.

[0061] In practice, the system described in this invention adjusts the rate of change of adjacent acquisition points by setting a first false alarm rate and a second false alarm rate. Since equipment vibration may cause some components to undergo slight displacement, causing changes in parameters such as connections and gaps that were originally normal, the sensors monitoring these parameters will detect the changes in values. The system cannot accurately distinguish whether it is normal component wear or temporary displacement caused by vibration, and may mistakenly judge it as a component failure and issue an early warning. By increasing the rate of change of adjacent acquisition points, high rate of change signals in a short period of time can be ignored, and alarms can only be triggered for displacements that have been stably exceeding the limit for a long time, thereby avoiding misjudgment and improving the accuracy of fault early warning.

[0062] Specifically, the control module is used to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data. If the byte missing rate of the normalized data is greater than the preset first missing rate, then the effectiveness of the early warning analysis is determined to be unsatisfactory.

[0063] Specifically, the control module is used to increase the training enhancement ratio of missing data in the neural network model when the byte missing rate of the normalized data is greater than the preset first missing rate and less than or equal to the preset second missing rate.

[0064] Specifically, the control module is used to preliminarily determine that the transmission stability of the running data does not meet the requirements when the byte missing rate of the normalized data is greater than the preset second missing rate, and to determine whether the transmission stability of the running data meets the requirements based on the signal-to-noise ratio of the communication signal.

[0065] It is understandable that the three intervals defined by the preset first missing rate and the preset second missing rate correspond to three different scenarios:

[0066] The first interval is when the byte missing rate of the normalized data is less than or equal to the preset first missing rate, which corresponds to the situation where the effectiveness of the early warning analysis is confirmed to meet the requirements.

[0067] The second interval is when the byte missing rate of the normalized data is greater than the preset first missing rate and less than or equal to the preset second missing rate. The corresponding situation is: due to the failure to upload data from some monitoring points in the dataset in a timely manner or due to equipment failure, data gaps may occur in certain periods, which may cause deviations in the analysis results. This may not fully and accurately reflect the actual situation, affecting model training and the accuracy of results. It may not be able to accurately describe the patterns behind the data, resulting in inaccurate warning results.

[0068] The third interval is when the byte missing rate of normalized data is greater than the preset second missing rate. The corresponding situation is: due to the existence of various electromagnetic interference sources around the cage ship, such as the generator on the ship being installed too close to sensitive electronic equipment, the electromagnetic field generated by it will be directly coupled to these devices, affecting the performance of the devices, interfering with the communication link of the fault early warning system, resulting in signal distortion and increased bit error rate.

[0069] Understandably, the preset missing rate can be set based on system performance evaluation, aiming to ensure the accuracy and usability of test results. Optionally, the preset missing rate is determined by evaluating system performance to determine the tolerable range of missing data, taking into account the overall performance requirements of the net cage vessel fault early warning system, including data acquisition, transmission, storage, and processing, to ensure the system can operate normally and provide accurate early warning information. For example, the preset first missing rate is generally selected in the range of [0.3%, 0.5%], and the preset second missing rate is generally selected in the range of [0.6%, 0.8%].

[0070] Preferably, the first missing rate is 0.4% in a preferred embodiment, and the second missing rate is 0.7% in a preferred embodiment.

[0071] Specifically, the increase in the training enhancement ratio of missing data in the neural network model is determined by the difference between the byte missing rate of the normalized data and the preset first missing rate.

[0072] Specifically, when the difference between the byte missing rate of normalized data and the preset first missing rate is within 0.2%, the training augmentation ratio of missing data in the neural network model increases to 1.2 times the original. When the difference between the byte missing rate of normalized data and the preset first missing rate exceeds 0.2%, in addition to increasing to 1.2 times the original, for every 0.1% exceeding the original, the training augmentation ratio of missing data in the neural network model increases by 2%. For example, if the difference between the byte missing rate of normalized data and the preset first missing rate is 0.4%, the current training augmentation ratio of missing data in the neural network model is 15%, and the increased training augmentation ratio of missing data in the neural network model is 15×1.2+2×2=22%.

[0073] In implementation, the system of this invention adjusts the training enhancement ratio of missing data in the neural network model by setting a preset first missing rate and a preset second missing rate. Because some monitoring points in the dataset are not uploaded in time or data gaps occur in certain periods due to equipment failure, the analysis results may be biased and unable to fully and accurately reflect the actual situation, affecting the accuracy of model training and results. It may not be able to accurately describe the patterns behind the data, resulting in inaccurate early warning results. By increasing the training enhancement ratio of missing data in the neural network model, the model can have more opportunities to learn samples containing missing data, thereby better adapting to the situation of missing data, improving the model's ability to process missing data, reducing the bias caused by missing data, and further improving the accuracy of fault early warning.

[0074] Specifically, the control module is used to determine whether the transmission stability of the operating data meets the requirements based on the signal-to-noise ratio of the communication signal. If the signal-to-noise ratio of the communication signal is less than the preset signal-to-noise ratio, it is determined that the transmission stability of the operating data does not meet the requirements, and the fusion threshold of the operating data from different collection points is increased.

[0075] It is understandable that the two preset signal-to-noise ratio intervals correspond to two different scenarios:

[0076] The first interval is when the signal-to-noise ratio of the communication signal is less than the preset signal-to-noise ratio. The corresponding situation is: due to the existence of various electromagnetic interference sources around the net cage ship, such as the generator on the ship being installed too close to sensitive electronic equipment, the electromagnetic field generated by it will be directly coupled to these devices, affecting the performance of the devices, interfering with the communication link of the fault warning system, resulting in signal distortion and increased bit error rate.

[0077] The second interval is when the signal-to-noise ratio of the communication signal is greater than or equal to the preset signal-to-noise ratio, which corresponds to the situation where the transmission stability of the running data meets the requirements.

[0078] Understandably, the preset signal-to-noise ratio (SNR) can be set according to the sensor performance parameters, aiming to ensure the accuracy and usability of the test results. Optionally, the preset SNR is determined by collecting operational data from the cage vessel fault early warning system over a period of time, including sensor data and system logs, analyzing the frequency and distribution of missing values ​​in this data, and understanding the missing patterns of different types of data. For example, the preset SNR is typically selected within the range of [20dB, 25dB].

[0079] Preferably, the preset signal-to-noise ratio is 23dB in the preferred embodiment.

[0080] Specifically, the increase in the data fusion threshold of different collection points is determined by the difference between the signal-to-noise ratio of the communication signal and the preset signal-to-noise ratio.

[0081] Specifically, when the difference between the signal-to-noise ratio (SNR) of the communication signal and the preset SNR is within 2dB, the data fusion threshold for different acquisition points is increased to 1.5 times the original value. When the difference between the SNR of the communication signal and the preset SNR exceeds 2dB, in addition to increasing to 1.5 times the original value, the data fusion threshold for different acquisition points increases by 0.1 for every 1dB exceeding the original value. For example, when the difference between the SNR of the communication signal and the preset SNR is 3dB, the current data fusion threshold for different acquisition points is 0.4, and the increased data fusion threshold for different acquisition points is 0.4×1.5+1×0.1=0.7.

[0082] In practice, the system described in this invention adjusts the fusion threshold of operational data from different collection points by setting a preset signal-to-noise ratio. Since there may be various electromagnetic interference sources around the net cage vessel, such as the generator on the vessel being installed too close to sensitive electronic equipment, the electromagnetic field it generates will directly couple into these devices, affecting their performance and interfering with the communication link of the fault early warning system, leading to signal distortion and increased bit error rate. By increasing the fusion threshold of operational data from different collection points, unreliable data that may be affected by electromagnetic interference can be filtered out to a certain extent, thereby improving the accuracy and reliability of data fusion, reducing the risk of erroneous fusion due to interference, and further improving the accuracy of fault early warning.

[0083] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A fault early warning system for net cage vessels, characterized in that, include: The data acquisition module includes an acquisition unit for collecting operational data from various collection points of the net cage vessel and a data transmission unit connected to the acquisition unit for converting the operational data into communication signals and transmitting them to the monitoring center. The data processing module, which is connected to the data acquisition module, includes a preprocessing unit that preprocesses the communication signal to output normalized data and a data analysis unit connected to the preprocessing unit that performs early warning analysis on the normalized data through a neural network model to output analysis results. An early warning module, which is connected to the data processing module, is used to issue fault warnings based on the analysis results; The control module is connected to the data acquisition module, the data processing module and the early warning module respectively, and is used to determine the change rate of adjacent acquisition points based on the false alarm rate of the fault early warning, and to determine the training enhancement ratio of missing data in the neural network model based on the byte missing rate of normalized data, and to determine the fusion threshold of the running data of different acquisition points based on the signal-to-noise ratio of the communication signal. The rate of change of adjacent sampling points is the ratio of the difference between the measured values ​​of the next moment and the previous moment of two time-adjacent sampling points to the measured value of the previous sampling point. The data fusion threshold for different collection points is a boundary value used to determine whether data from different collection points should be fused. The increase in the rate of change of adjacent acquisition points is determined by the difference between the false alarm rate of the fault warning and the preset second false alarm rate. When the difference between the false alarm rate of the fault warning and the preset second false alarm rate is within 1%, the change rate of adjacent data collection points increases to 1.2 times the original rate; when the difference between the false alarm rate of the fault warning and the preset second false alarm rate exceeds 1%, in addition to increasing to 1.2 times the original rate, for every 0.5% exceeding the original rate, the change rate of adjacent data collection points increases by 1%. The control module is used to increase the training enhancement ratio of missing data in the neural network model when the byte missing rate of the normalized data is greater than a preset first missing rate and less than or equal to a preset second missing rate. When the difference between the byte missing rate of normalized data and the preset first missing rate is within 0.2%, the training augmentation ratio of missing data in the neural network model increases to 1.2 times the original value. When the difference between the byte missing rate of normalized data and the preset first missing rate exceeds 0.2%, in addition to increasing to 1.2 times the original value, for every 0.1% exceeding the original value, the training augmentation ratio of missing data in the neural network model increases by 2%. The control module is used to determine whether the transmission stability of the running data meets the requirements based on the signal-to-noise ratio of the communication signal. If the signal-to-noise ratio of the communication signal is less than the preset signal-to-noise ratio, it is determined that the transmission stability of the running data does not meet the requirements, and the fusion threshold of the running data from different collection points is increased. When the difference between the signal-to-noise ratio (SNR) of the communication signal and the preset SNR is within 2dB, the data fusion threshold of different acquisition points is increased to 1.5 times the original value. When the difference between the SNR of the communication signal and the preset SNR exceeds 2dB, in addition to increasing to 1.5 times the original value, the data fusion threshold of different acquisition points increases by 0.1 for every 1dB exceeding the original value.

2. The cage vessel fault early warning system according to claim 1, characterized in that, The control module is used to determine whether the accuracy of the fault warning meets the requirements based on the false alarm rate of the fault warning. If the false alarm rate of the fault warning is greater than the preset first false alarm rate, it is determined that the accuracy of the fault warning does not meet the requirements.

3. The cage vessel fault early warning system according to claim 2, characterized in that, The control module is used to preliminarily determine that the effectiveness of the early warning analysis does not meet the requirements when the false alarm rate of the fault warning is greater than the preset first false alarm rate and less than or equal to the preset second false alarm rate, and to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data.

4. The cage vessel fault early warning system according to claim 3, characterized in that, The control module is used to increase the rate of change of adjacent acquisition points when the false alarm rate of the fault warning is greater than the preset second false alarm rate.

5. The cage vessel fault early warning system according to claim 4, characterized in that, The control module is used to determine whether the effectiveness of the early warning analysis meets the requirements based on the byte missing rate of the normalized data. If the byte missing rate of the normalized data is greater than the preset first missing rate, the effectiveness of the early warning analysis is determined to be unsatisfactory.

6. The cage vessel fault early warning system according to claim 5, characterized in that, The control module is used to preliminarily determine that the transmission stability of the running data does not meet the requirements when the byte missing rate of the normalized data is greater than the preset second missing rate, and to determine whether the transmission stability of the running data meets the requirements based on the signal-to-noise ratio of the communication signal.

7. The cage vessel fault early warning system according to claim 6, characterized in that, The increase in the training enhancement ratio of missing data in the neural network model is determined by the difference between the byte missing rate of the normalized data and the preset first missing rate.

8. The cage vessel fault early warning system according to claim 7, characterized in that, The increase in the data fusion threshold of different collection points is determined by the difference between the preset signal-to-noise ratio and the signal-to-noise ratio of the communication signal.

Images