A method for predicting shipborne communication hardware failure based on multidimensional data analysis technology

Abstract

The present invention discloses a method for predicting shipboard communication hardware faults based on multidimensional data analysis technology, and the method includes data collection, data preprocessing, establishing a shipboard communication hardware fault prediction model, and shipboard communication hardware fault prediction based on multidimensional data analysis. The present invention belongs to the field of fault prediction technology, and specifically refers to a method for predicting shipboard communication hardware faults based on multidimensional data analysis technology. This scheme defines an insensitive loss function by introducing a weighted decreasing function, which has a higher penalty for large prediction errors, but is relatively tolerant for slight deviations, and better copes with small fluctuations in hardware data; adopts a regularization strategy based on a kernel function to improve the accuracy of fault prediction; tracks error changes during training through an adaptive learning rate adjustment mechanism; and controls model performance by designing an error threshold, thereby ensuring that the efficiency of the final fault prediction is high enough and the prediction result is more reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a method for predicting faults of shipborne communication hardware based on multidimensional data analysis technology. Background Art

[0002] The shipboard communication hardware fault prediction method is a technical means to predict the health status of the hardware equipment by collecting, analyzing and modeling various sensor data of shipboard communication equipment. However, the traditional shipboard communication hardware fault prediction method has the problem of improper handling of noise and outliers in the hardware data, poor effect in long-term monitoring, and low accuracy of hardware fault prediction; the traditional shipboard communication hardware fault prediction method has the problem of difficulty in handling complex multi-dimensional hardware data, improper design of the model establishment mechanism, and low efficiency of fault prediction. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method for predicting shipboard communication hardware faults based on multidimensional data analysis technology. The traditional shipboard communication hardware fault prediction method has the problem of improper handling of noise and outliers in hardware data, poor effect in long-term monitoring, and low accuracy of hardware fault prediction. This scheme defines an insensitive loss function by introducing a weighted decreasing function, which has a higher penalty for large prediction errors and is relatively tolerant to slight deviations. It can better deal with small fluctuations in hardware data and reduce the impact of noise. By dynamically adjusting the model parameters, the model can automatically predict the hardware faults according to the changes in the distribution of hardware data during the training process. Adjustment makes the long-term monitoring of hardware data more flexible and efficient, and the prediction accuracy is higher; the traditional shipborne communication hardware fault prediction method has the problem that it is difficult to handle complex multi-dimensional hardware data, and the improper design of the model establishment mechanism leads to low efficiency of fault prediction. This scheme adopts a regularization strategy based on kernel functions to improve the generalization ability of processing high-dimensional hardware data, thereby improving the accuracy of fault prediction; through the adaptive learning rate adjustment mechanism, the error changes in the training process are tracked, which is more in line with the changes in data characteristics in the shipborne communication environment; by designing the error threshold, the model performance is controlled, thereby ensuring that the efficiency of the final fault prediction is high enough and the prediction result is more reliable.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a method for predicting shipborne communication hardware failures based on multidimensional data analysis technology, the method comprising the following steps:

[0005] Step S1: data collection;

[0006] Step S2: data preprocessing;

[0007] Step S3: Establishing a shipborne communication hardware failure prediction model;

[0008] Step S4: Prediction of shipborne communication hardware failure based on multi-dimensional data analysis.

[0009] Furthermore, in step S1, the data collection is to collect historical shipboard communication hardware detection data; the historical shipboard communication hardware detection data includes signal strength, temperature, voltage, environmental conditions, equipment operation time, working current, communication delay, humidity, network traffic data and hardware health status at the next moment; the hardware health status includes normal and faulty; the hardware health status is used as a data label.

[0010] Furthermore, in step S2, the data preprocessing is to perform data cleaning, data conversion and standardization on the collected data; the data cleaning is to process missing values ​​and duplicate values; the data conversion is to convert the data into a vector form; the standardization is to standardize the data based on the maximum and minimum normalization method; and the data set D is obtained.

[0011] Furthermore, in step S3, the establishment of the shipborne communication hardware failure prediction model specifically includes the following steps:

[0012] Step S31: model design; select support vector machine as the basic structure of the shipborne communication hardware fault prediction model; initialize the support vector machine regularization parameters and kernel function parameters; divide the data set D into a training set and a test set;

[0013] Step S32: define a loss function; introduce a relaxation parameter and a weighted decreasing function into the loss function to cope with different levels of noise in the hardware data and process the error distribution in the hardware data;

[0014] Step S33: dynamically updating model parameters; initializing relaxation parameters and weighted decrement parameters, and dynamically updating relaxation parameters, weighted decrement parameters and regularization parameters of the support vector machine;

[0015] Step S34: Regularization strategy; minimizing the error to obtain the final shipboard communication hardware fault prediction model;

[0016] Step S35: Adaptive learning rate adjustment mechanism;

[0017] Step S36: define an error threshold; if the error threshold is not met during training or the maximum number of training times is reached, adjust the initial parameters, re-divide the data set, and train the model; when the maximum number of training times is not reached and the loss convergence is achieved under the condition of meeting the error threshold, the shipborne communication hardware fault prediction model is established.

[0018] Further, in step S4, the shipborne communication hardware fault prediction based on multidimensional data analysis is based on the established shipborne communication hardware fault prediction model, and real-time acquisition of signal strength, temperature, voltage, environmental conditions, equipment operating time, operating current, communication delay, humidity and network traffic data is input into the shipborne communication hardware fault prediction model after preprocessing, and the hardware health status output by the model is used as the prediction result. When the output hardware health status is a fault, an early warning reminder is issued to the management personnel.

[0019] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0020] (1) In view of the problem that the traditional shipborne communication hardware fault prediction method has improper handling of noise and outliers in the hardware data, which has poor effect in long-term monitoring and leads to low accuracy of hardware fault prediction, this scheme defines an insensitive loss function by introducing a weighted decreasing function, which has a higher penalty for large prediction errors and is relatively tolerant to slight deviations. It can better deal with small fluctuations in hardware data and reduce the impact of noise. By dynamically adjusting the model parameters, it can automatically adjust according to the changes in the distribution of hardware data during the training process, making the long-term monitoring of hardware data more flexible and efficient, and the prediction accuracy higher.

[0021] (2) In view of the fact that traditional shipborne communication hardware fault prediction methods have difficulty in processing complex multi-dimensional hardware data and the improper design of the model establishment mechanism leads to low efficiency of fault prediction, this scheme adopts a regularization strategy based on kernel functions to improve the generalization ability of processing high-dimensional hardware data, thereby improving the accuracy of fault prediction; through the adaptive learning rate adjustment mechanism, the error changes during the training process are tracked to better conform to the changes in data characteristics in the shipborne communication environment; by designing the error threshold, the model performance is controlled, thereby ensuring that the efficiency of the final fault prediction is high enough and the prediction results are more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic flow chart of a method for predicting shipborne communication hardware failures based on multidimensional data analysis technology provided by the present invention;

[0023] Figure 2 It is a schematic diagram of the process of step S3.

[0024] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0027] Example 1, see Figure 1 The present invention provides a method for predicting shipborne communication hardware failures based on multidimensional data analysis technology, the method comprising the following steps:

[0028] Step S1: data collection, collecting historical shipboard communication hardware detection data;

[0029] Step S2: Data preprocessing, data cleaning, data conversion and standardization of the collected data;

[0030] Step S3: Establish a shipborne communication hardware fault prediction model, introduce a weighted decreasing function to define an insensitive loss function, dynamically adjust model parameters, adopt a regularization strategy based on kernel functions, design an adaptive learning rate adjustment mechanism, and use an error threshold to implement the establishment of a shipborne communication hardware fault prediction model;

[0031] Step S4: Fault prediction of shipborne communication hardware based on multi-dimensional data analysis, and fault prediction based on the established shipborne communication hardware fault prediction model.

[0032] Example 2, see Figure 1 This embodiment is based on the above embodiment. In step S1, the historical shipborne communication hardware detection data includes signal strength, temperature, voltage, environmental conditions, equipment operating time, working current, communication delay, humidity, network traffic data and hardware health status at the next moment; the hardware health status includes normal and faulty; the hardware health status is used as a data label.

[0033] Example 3, see Figure 1This embodiment is based on the above embodiment. In step S2, data cleaning is to process missing values ​​and duplicate values; data conversion is to convert data into vector form; standardization is to standardize data based on the maximum and minimum normalization method; and a data set D is obtained, which is expressed as: ,in, is a multidimensional feature vector, It is the status label of the hardware.

[0034] Example 4, see Figure 1 and Figure 2 Based on the above embodiment, in step S3, establishing a shipborne communication hardware failure prediction model specifically includes the following steps:

[0035] Step S31: model design; select support vector machine as the basic structure of the shipborne communication hardware fault prediction model; initialize the support vector machine regularization parameters and kernel function parameters; divide the data set D into a training set and a test set;

[0036] Step S32: define a loss function;

[0037] Step S33: dynamically update model parameters;

[0038] Step S34: Regularization strategy: A regularization strategy based on kernel function is introduced to improve the performance of the model on hardware data. The generalized error of regularization is defined as: ; Minimize the error to obtain the final shipborne communication hardware failure prediction model, expressed as: ;in, is the generalized error of the model; is the regularization parameter; ρ is the data distribution; F is the decision function of SVM; is the squared norm in kernel space; It is the final shipboard communication hardware failure prediction model;

[0039] Step S35: Adaptive learning rate adjustment mechanism; expressed as: ;in, and are the learning rate and regularization factor at the t+1th training respectively; τ is a very small positive number; and are the generalized errors of the training set and the test set, respectively; is the gradient; α1 is the amplitude parameter; β1 is the rate coefficient; is the error coefficient;

[0040] Step S36: define an error threshold; the error threshold is expressed as: ;in, is the expectation of the t-th training of the dataset; is the generalized error of the model at the t+1th time; is the expected optimal generalized error; C is the data characteristic constant; is the regularization constant; α1 is the step size adjustment parameter; γ is the regularization adjustment parameter; if the error threshold is not met or the maximum number of training times is reached during the training process, the initial parameters are adjusted, the data set is re-divided, and the model is trained; when the maximum number of training times is not reached and the loss convergence is achieved under the condition of meeting the error threshold, the shipborne communication hardware fault prediction model is established.

[0041] By performing the above operations, in view of the problem that traditional shipborne communication hardware fault prediction methods have difficulty in processing complex multi-dimensional hardware data, and the improper design of the model establishment mechanism leads to low efficiency of fault prediction, this scheme adopts a regularization strategy based on kernel functions to improve the generalization ability of processing high-dimensional hardware data, thereby improving the accuracy of fault prediction; through the adaptive learning rate adjustment mechanism, the error changes during the training process are tracked, which is more in line with the changes in data characteristics in the shipborne communication environment; by designing the error threshold, the model performance is controlled, thereby ensuring that the efficiency of the final fault prediction is high enough and the prediction result is more reliable.

[0042] Example 5, see Figure 2 This embodiment is based on the above embodiment. In step S3, the loss function is defined as introducing a relaxation parameter and a weighted decreasing function into the loss function. To cope with different degrees of noise in the hardware data and error distribution in the processing hardware data, the loss function is defined as: ; ;in, is the model loss function; y is the predicted label; β is the weighted reduction parameter; is the relaxation parameter; u is the prediction loss;

[0043] Dynamically updating model parameters specifically involves initializing relaxation parameters and weighted decrement parameters, and dynamically updating relaxation parameters, weighted decrement parameters, and regularization parameters of the support vector machine, which can be expressed as: ; Among them, f t+1 and f t are the model parameters at the t+1th training and the tth training respectively; is the learning rate during the tth training; is the derivative of the loss function with respect to the model parameters; f t (x t ) is the model's response to the input data x t The predicted value of is the true label of the input data; is the Gaussian kernel function during the t-th training; is the regularization factor during the t-th training.

[0044] By performing the above operations, the traditional shipborne communication hardware fault prediction method has the problem of improper handling of noise and outliers in hardware data, poor effect in long-term monitoring, and low accuracy of hardware fault prediction. This scheme introduces a weighted decreasing function to define an insensitive loss function, which has a higher penalty for large prediction errors and is relatively tolerant to slight deviations. It can better deal with small fluctuations in hardware data and reduce the impact of noise. By dynamically adjusting the model parameters, it automatically adjusts according to the changes in the distribution of hardware data during the training process, making long-term monitoring of hardware data more flexible and efficient, and the prediction accuracy is higher.

[0045] Example 6, see Figure 1 This embodiment is based on the above embodiment. In step S4, the shipboard communication hardware fault prediction based on multidimensional data analysis is based on the established shipboard communication hardware fault prediction model. Signal strength, temperature, voltage, environmental conditions, equipment operating time, working current, communication delay, humidity and network traffic data are collected in real time, and input into the shipboard communication hardware fault prediction model after preprocessing. The hardware health status output by the model is used as the prediction result. When the output hardware health status is a fault, an early warning reminder is issued to the management personnel.

[0046] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0047] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0048] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. A method for predicting shipborne communication hardware failure based on multidimensional data analysis technology, characterized in that: The method comprises the following steps: Step S1: data collection, collecting historical shipboard communication hardware detection data; Step S2: data preprocessing; Step S3: Establish a shipborne communication hardware fault prediction model, introduce a weighted decreasing function to define an insensitive loss function, dynamically adjust model parameters, adopt a regularization strategy based on kernel functions, design an adaptive learning rate adjustment mechanism, and use an error threshold to implement the establishment of a shipborne communication hardware fault prediction model; Step S4: Fault prediction of shipborne communication hardware based on multi-dimensional data analysis, and fault prediction based on the established shipborne communication hardware fault prediction model; The establishment of the shipborne communication hardware failure prediction model specifically includes the following steps: Step S31: model design; select support vector machine as the basic structure of the shipborne communication hardware fault prediction model; initialize the support vector machine regularization parameters and kernel function parameters; divide the data set D into a training set and a test set; Step S32: define a loss function; introduce a relaxation parameter and a weighted decreasing function into the loss function to deal with different levels of noise in the hardware data and error distribution in the processing hardware data; the loss function is defined as: ; ;in, is the model loss function; y is the predicted label; β is the weighted reduction parameter; is the relaxation parameter; u is the prediction loss; Dynamically updating model parameters specifically involves initializing relaxation parameters and weighted decrement parameters, and dynamically updating relaxation parameters, weighted decrement parameters, and regularization parameters of the support vector machine, which can be expressed as: ; Among them, f t+1 and f t are the model parameters at the t+1th training and the tth training respectively; is the learning rate during the tth training; is the derivative of the loss function with respect to the model parameters; f t (x t ) is the model's response to the input data x t The predicted value of is the true label of the input data; is the Gaussian kernel function during the t-th training; is the regularization factor during the t-th training; Step S33: dynamically updating model parameters; initializing relaxation parameters and weighted decrement parameters, and dynamically updating relaxation parameters, weighted decrement parameters and regularization parameters of the support vector machine; Step S34: Regularization strategy: A regularization strategy based on kernel function is introduced, and the generalized error of regularization is defined as: ; Minimize the error to obtain the final shipborne communication hardware failure prediction model, expressed as: ;in, is the generalized error of the model; is the regularization parameter; ρ is the data distribution; F is the decision function of SVM; is the squared norm in the kernel space; It is the final shipboard communication hardware failure prediction model; Step S35: Adaptive learning rate adjustment mechanism; expressed as: ;in, and are the learning rate and regularization factor at the t+1th training respectively; τ is a very small positive number; and are the generalized errors of the training set and the test set, respectively; is the gradient; α1 is the amplitude parameter; β1 is the rate coefficient; is the error coefficient; Step S36: define an error threshold; the error threshold is expressed as: ;in, is the expectation of the t-th training of the dataset; is the generalized error of the model at the t+1th time; is the expected optimal generalized error; C is the data characteristic constant; is the regularization constant; α1 is the step size adjustment parameter; γ is the regularization adjustment parameter; if the error threshold is not met or the maximum number of training times is reached during the training process, the initial parameters are adjusted, and the data set is re-divided to train the model; when the maximum number of training times is not reached and the loss convergence is achieved under the condition of meeting the error threshold, the shipborne communication hardware fault prediction model is established; In step S1, the historical shipborne communication hardware detection data includes signal strength, temperature, voltage, environmental conditions, equipment running time, working current, communication delay, humidity, network traffic data and hardware health status at the next moment; the hardware health status includes normal and fault; the hardware health status is used as a data label; In step S4, the shipborne communication hardware fault prediction based on multidimensional data analysis is based on the established shipborne communication hardware fault prediction model, and real-time data of signal strength, temperature, voltage, environmental conditions, equipment operating time, working current, communication delay, humidity and network traffic are collected. After preprocessing, the data is input into the shipborne communication hardware fault prediction model, and the hardware health status output by the model is used as the prediction result. When the output hardware health status is a fault, an early warning reminder is issued to the management personnel.

2. The method for predicting shipborne communication hardware failure based on multidimensional data analysis technology according to claim 1, characterized in that: In step S2, the data preprocessing is to clean, convert and standardize the collected data; the data cleaning is to process missing values ​​and duplicate values; the data conversion is to convert the data into vector form; the standardization is to standardize the data based on the maximum and minimum normalization method; and the data set D is obtained.

Images