Network equipment fault prediction method and system based on multi-source data fusion

Abstract

The invention discloses a network equipment fault prediction method and system based on multi-source data fusion, and relates to the technical field of network equipment fault prediction, and the system comprises a disk vibration prediction module, a vibration transmission analysis module, and a fault disk positioning module. The disk vibration prediction module trains a vibration signal prediction neural network model by analyzing vibration signal historical data generated when each disk unit reads and writes data in a disk array, and predicts a vibration signal generated when each disk unit reads and writes data according to a data access request monitored in real time; the vibration transmission analysis module is used for analyzing the amplitude and phase change of each frequency component of a vibration signal transmitted to each vibration monitoring sensor from the orientation of each disk unit by constructing a three-dimensional material model for the disk array; and the fault disk positioning module constructs a fault positioning model by comparing and analyzing predicted vibration signal data and actually measured vibration signal data of the vibration monitoring sensor, and positions a fault disk unit.

Description

Technical Field

[0001] The present invention relates to the technical field of network equipment fault prediction, and in particular to a network equipment fault prediction method and system based on multi-source data fusion. Background Art

[0002] During long-term operation, disk units are prone to performance degradation or even failure due to mechanical failure, overload, aging and other problems.

[0003] However, traditional disk array fault detection methods have the following technical drawbacks: First, traditional disk array fault prediction is primarily based on SMART (Self-Monitoring, Analysis and Reporting Technology). During fault analysis, the data layer and physical structure layer parameters of the disk operation are relatively independent, lacking correlation analysis between read and write data and disk vibration, making it impossible to accurately determine whether the vibration anomaly is due to disk load characteristics or potential faults. Second, the use of multiple sensors to independently monitor and analyze faults for all disk arrays in the disk array also requires synchronous updating of fault analysis components during disk expansion, severely limiting the scalability of the disk array.

[0004] Therefore, a network equipment fault prediction method and system based on multi-source data fusion is needed to solve the above technical defects. Summary of the Invention

[0005] The purpose of the present invention is to provide a network equipment fault prediction method and system based on multi-source data fusion to solve the problems raised in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A network equipment fault prediction method based on multi-source data fusion, the method comprising the following steps:

[0008] Step S100: Obtain historical data of vibration signals generated when each disk unit in the disk array reads and writes data, use Fourier transform to obtain spectrum data of the vibration signals, and perform vibration signal characteristic parameter annotation to construct a vibration signal characteristic parameter set, and train a vibration signal prediction neural network model; the vibration signal characteristic parameters include the read and write data characteristic parameters of the disk units and the vibration signal characteristic parameters;

[0009] Step S200: Acquire physical structure data of the disk array, construct a three-dimensional material model, analyze the phase offset and amplitude attenuation coefficient of each frequency component of the vibration signal transmitted from the position of each disk unit to each vibration monitoring sensor, and obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve by fitting;

[0010] Step S300: monitoring data access requests in the server transmission link, extracting characteristic parameters of read and write data, predicting characteristic parameters of the read and write data of each disk unit, and using the vibration signal prediction neural network model to predict characteristic parameters of the vibration signal generated by the read and write data of each disk unit;

[0011] Step S400: Setting a frequency step size, dividing the frequency level intervals, dividing the frequency components of the vibration signal into frequency levels, and calculating the amplitude attenuation coefficient and phase offset of each frequency component of the disk unit vibration signal based on the frequency-amplitude attenuation coefficient curve and the frequency-phase offset curve transmitted from each disk unit vibration signal to each vibration monitoring sensor;

[0012] Step S500: Acquire vibration signal monitoring data from each vibration monitoring sensor in real time, compare and analyze the data with the predicted vibration signal characteristic parameters at each vibration monitoring sensor, build a fault location model, and locate the fault of each disk unit.

[0013] In the above technical solution, step S100 is divided into the following steps:

[0014] Step S101: Acquire historical vibration signal data generated when each disk unit in the disk array reads and writes data; the historical vibration signal data is vibration signal data generated when each disk unit reads and writes data independently;

[0015] Step S102: using Fourier transform to obtain spectrum data of the vibration signal and perform vibration signal feature annotation;

[0016] For any vibration signal history data s of any disk unit x in the disk array, the feature labeling result is x_s[D_x_s,S_x_s]; where D_x_s is the read and write data feature parameter set of the vibration signal history record s of disk unit x, and S_x_s is the vibration signal feature parameter of the vibration signal history record s of disk unit x;

[0017] Step S103: constructing a vibration signal feature data set and training a vibration signal prediction neural network model;

[0018] The read and write operation characteristics of the disk are combined with the spectral characteristics of the vibration signal to construct a multi-dimensional input neural network model. By monitoring data access requests in real time, the vibration characteristics under different loads are dynamically predicted, breaking through the limitations of isolated analysis of multi-source data in traditional methods.

[0019] In the above technical solution, step S200 is divided into the following steps:

[0020] Step S201: Setting a vibration monitoring sensor to monitor the disk array vibration signal in real time;

[0021] Step S202: obtaining physical structure data of the disk array, and constructing a three-dimensional material model for the physical structure of the disk array;

[0022] Step S203: respectively monitoring the amplitude and phase change data of vibration signals of different frequencies transmitted from each disk unit to each vibration monitoring sensor, and then calculating the phase offset and amplitude attenuation coefficient of the vibration signal transmitted by each disk unit;

[0023] The phase difference between the vibration signal at the disk unit and the vibration monitoring sensor is used as the phase offset of the vibration signal transmission, and the amplitude ratio of the vibration signal at the disk unit and the vibration monitoring sensor is used as the amplitude attenuation coefficient of the vibration signal transmission;

[0024] Step S204: using curve fitting to obtain frequency-amplitude attenuation coefficient curves and frequency-phase offset curves of vibration signals of each disk unit at different frequencies transmitted to each vibration monitoring sensor;

[0025] For any disk unit x, the frequency-amplitude attenuation coefficient curve of the vibration signal transmitted to the vibration monitoring sensor s is recorded as L_a(x,s), and the frequency-phase offset curve is recorded as

[0026] Build a three-dimensional material model of the disk array and analyze the phase offset and amplitude attenuation coefficient of the vibration signal to accurately calculate the propagation characteristics of the vibration signal and improve prediction accuracy.

[0027] In the above technical solution, step S400 is divided into the following steps:

[0028] Step S401: Setting a frequency step f_step for the vibration signal spectrum, dividing all frequency components in the vibration signal belonging to the frequency level interval [(n-1)×f_step,n×f_step] into frequency level n, and summing the amplitudes of all frequency components of the same frequency level and dividing by the frequency step f_step as the amplitude of the corresponding frequency level; where n is the frequency level number, which is a positive integer;

[0029] Step S402: For any disk unit x, the amplitude attenuation coefficient k(x, s, f) of any frequency f component of the vibration signal generated by reading and writing data and transmitted to the vibration monitoring sensor s is calculated as follows:

[0030] Extract the frequency level interval to which frequency f belongs, calculate the integrated average value of L_a(x,s) over the frequency in the interval, and divide the calculated result by the frequency step size to obtain k(x,s,f);

[0031] For any disk unit x, the phase offset of any frequency f component of the vibration signal generated by reading and writing data transmitted to the vibration monitoring sensor s The calculation method is:

[0032] Extract the frequency level interval to which the frequency f belongs, and calculate the frequency level interval For the integrated average value of the frequency, divide the calculated result by the frequency step size to get

[0033] The frequency step division method is adopted to divide the vibration signal into multiple frequency levels, and the amplitude attenuation and phase shift of different frequency components are calculated by integral averaging, which reduces the amount of calculation while improving the accuracy of the equalization algorithm analysis and the real-time performance.

[0034] In the above technical solution, step S500 is divided into the following steps:

[0035] Step S501: using the predicted characteristic parameters of the vibration signal generated by the read / write data of each disk unit, and generating the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal generated by the read / write data of the disk unit and transmitted to each vibration monitoring sensor, simulating the spectrum data of the vibration signal of each disk unit transmitted to the vibration monitoring sensor;

[0036] Step S502: performing frequency domain complex addition on the vibration signals transmitted from each disk unit to the vibration monitoring sensor to obtain the amplitude and phase of each frequency component of the vibration signal predicted by each vibration monitoring sensor;

[0037] Step S503: monitoring the vibration signal of the disk array in real time through a vibration monitoring sensor provided on the outer contour of the disk array, and obtaining the amplitude and phase of each frequency component of the measured vibration signal through Fourier transform;

[0038] Step S504: Calculating the deviation between the measured vibration signal and the predicted vibration signal of the disk array vibration signal, performing a weighted summation, and setting a vibration deviation threshold; when the weighted summation result is greater than or equal to the vibration deviation threshold, determining that a faulty disk unit exists in the disk array, further constructing a fault location model, and locating the faulty disk unit in the disk array;

[0039] A fault location model is constructed, and a sparse optimization method is used to iteratively solve the spectrum deviation of the disk unit. The failure probability of each disk unit is accurately calculated, and fault priority ranking is provided to help managers conduct accurate troubleshooting.

[0040] In the above technical solution, the step S504 of constructing a fault location model to locate a faulty disk unit in the disk array includes the following:

[0041] Construct a fault location model M, which is as follows:

[0042]

[0043] M=min_ΔS x (f){∑ s,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]|| 2 +λ×∑ x,f (ΔS x (f))};

[0044] Where, ΔS x (f) is the spectrum deviation of the frequency f component of the vibration signal of disk unit x, ΔS s (f) is the spectrum deviation of the frequency f component of the vibration signal of the vibration monitoring sensor s, k x,f is the transfer function of the vibration signal of frequency f from the disk unit x to the vibration monitoring sensor s, λ is the regularization parameter, j is a complex unit; ||p|| 2 is the L2 norm square of parameter p;

[0045] Initialize the spectrum deviation of the frequency f component of the vibration signal of each disk unit to 0, select the spectrum deviation ΔS of the frequency f component of the vibration signal of any disk unit x x (f) is used as a variable, and the spectrum deviation of the frequency f component of the vibration signal of the remaining disk units is fixed, and ΔS is solved to minimize the value of the fault location model M. x (f);

[0046] Set the threshold for the number of failed disks to be determined, using the above ΔS x (f) The solution method iterates the spectrum deviation of the vibration signal frequency f component of each disk unit multiple times until the number of disk units whose spectrum deviation of the vibration signal frequency f component is greater than the vibration abnormality threshold S' is less than or equal to the threshold of the number of faulty disks to be determined, and all disk units whose spectrum deviation of the vibration signal frequency f component is greater than the vibration abnormality threshold S' are marked as faulty disk units;

[0047] The vibration signal deviation of each disk unit is further calculated based on the iterative results. For any disk unit x, the vibration signal deviation Res(x) is calculated as follows:

[0048] Res(x)=∑ s,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]||2;

[0049] For any disk units a and b, if Res(a) > Res(b), it is determined that the failure probability of disk unit a is less than that of disk unit b; if Res(a) < Res(b), it is determined that the failure probability of disk unit a is greater than that of disk unit b.

[0050] Furthermore, each disk unit is sequentially fed back to the administrator for disk unit failure troubleshooting in ascending order of the vibration signal deviation.

[0051] A network device fault prediction system based on multi-source data fusion in the above technical solution, the system includes: a disk vibration prediction module, a vibration transfer analysis module, and a faulty disk positioning module;

[0052] The disk vibration prediction module analyzes the historical data of the vibration signals generated when each disk unit in the disk array reads and writes data, trains a vibration signal prediction neural network model, and predicts the vibration signals generated when each disk unit reads and writes data according to the real-time monitored data access requests; the vibration transfer analysis module analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the azimuth of each disk unit to each vibration monitoring sensor by constructing a three-dimensional material model of the disk array; the faulty disk positioning module constructs a fault positioning model by comparing and analyzing the predicted vibration signals and the measured vibration signal data of the vibration monitoring sensors, and locates the faults of each disk unit.

[0053] In the above technical solution, the disk vibration prediction module includes: a vibration data processing unit, a vibration feature analysis unit, and a vibration signal prediction unit;

[0054] The vibration data processing unit is used to process the historical data of the vibration signals generated when each disk unit in the disk array reads and writes data; the vibration feature analysis unit uses Fourier transform to obtain the spectrum data of the vibration signals and performs vibration signal feature parameter annotation; the vibration signal prediction unit trains a vibration signal prediction neural network model by constructing a vibration signal feature parameter set.

[0055] In the above technical solution, the vibration transfer analysis module includes: a material model construction unit and a signal transfer analysis unit;

[0056] The material model construction unit obtains the physical structure data of the disk array and constructs a three-dimensional material model; the signal transfer analysis unit analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the azimuth of each disk unit to each vibration monitoring sensor, and fits to obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve.

[0057] In the above technical solution, the faulty disk locating module includes: a vibration deviation analyzing unit, a positioning model building unit, and a faulty disk screening unit;

[0058] The vibration deviation analysis unit compares and analyzes the vibration signal predicted by the vibration monitoring sensor with the measured vibration signal data; the positioning model construction unit constructs a fault positioning model based on the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal; the faulty disk screening unit iterates the spectrum deviation of the frequency f component of the vibration signal of each disk unit multiple times, and screens the faulty disk units according to the threshold value of the number of faulty disks to be determined, and performs disk unit fault troubleshooting feedback in order from small to large according to the vibration signal deviation of each faulty disk unit.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] In the present invention, by integrating the disk read and write data characteristics with the vibration signal characteristics, a neural network model is trained, and the correlation between the disk data read and write operations and the generated vibration signals is fully considered, thereby enhancing the fault prediction capability and improving the accuracy of fault analysis.

[0061] In this invention, fault tracing is performed through the external vibration signal of the disk array, a sparse optimization method is used to build a fault location model, an L1 regularization term is introduced to force a sparse solution, and a non-negative amplitude constraint is added to ensure the rationality of fault location. At the same time, it also effectively solves the disk array expansion obstacle caused by monitoring each disk unit separately.

[0062] In addition, the present invention combines the characteristics of disk read and write data and the material characteristics of the disk array to comprehensively predict and analyze the effects of vibration data generated when each disk unit in the disk array reads and writes data after being collected by the sensor, so as to perform comparative analysis to locate the faulty disk unit, breaking through the isolation of multi-source data analysis in traditional methods and significantly improving the accuracy and reliability of disk fault analysis and positioning results. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a flow chart of the network equipment fault prediction method based on multi-source data fusion of the present invention;

[0064] Figure 2 This is the organizational structure diagram of the network equipment fault prediction system based on multi-source data fusion of the present invention. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0066] Example: See Figure 1-Figure 2 , the present invention provides the following technical solutions:

[0067] like Figure 1 As shown, the present invention provides a network equipment fault prediction method based on multi-source data fusion, the method comprising the following steps:

[0068] Step S100: Obtain historical data of vibration signals generated when each disk unit in the disk array reads and writes data, use Fourier transform to obtain spectrum data of the vibration signals, and perform vibration signal characteristic parameter annotation to construct a vibration signal characteristic parameter set, and train a vibration signal prediction neural network model; the vibration signal characteristic parameters include the read and write data characteristic parameters of the disk units and the vibration signal characteristic parameters;

[0069] Step S200: Acquire physical structure data of the disk array, construct a three-dimensional material model, analyze the phase offset and amplitude attenuation coefficient of each frequency component of the vibration signal transmitted from the position of each disk unit to each vibration monitoring sensor, and obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve by fitting;

[0070] Step S300: monitoring data access requests in the server transmission link, extracting characteristic parameters of read and write data, predicting characteristic parameters of the read and write data of each disk unit, and using the vibration signal prediction neural network model to predict characteristic parameters of the vibration signal generated by the read and write data of each disk unit;

[0071] Step S400: Setting a frequency step size, dividing the frequency level intervals, dividing the frequency components of the vibration signal into frequency levels, and calculating the amplitude attenuation coefficient and phase offset of each frequency component of the disk unit vibration signal based on the frequency-amplitude attenuation coefficient curve and the frequency-phase offset curve transmitted from each disk unit vibration signal to each vibration monitoring sensor;

[0072] Step S500: Acquire vibration signal monitoring data from each vibration monitoring sensor in real time, compare and analyze the data with the predicted vibration signal characteristic parameters at each vibration monitoring sensor, build a fault location model, and locate the fault of each disk unit.

[0073] The step S100 is divided into the following steps:

[0074] Step S101: Acquire historical vibration signal data generated when each disk unit in the disk array reads and writes data; the historical vibration signal data is vibration signal data generated when each disk unit reads and writes data independently;

[0075] Step S102: using Fourier transform to obtain spectrum data of the vibration signal and perform vibration signal feature annotation;

[0076] For any vibration signal history data s of any disk unit x in the disk array, the feature labeling result is x_s[D_x_s,S_x_s]; where D_x_s is the read and write data feature parameter set of the vibration signal history record s of disk unit x, and S_x_s is the vibration signal feature parameter of the vibration signal history record s of disk unit x;

[0077] Step S103: constructing a vibration signal feature data set and training a vibration signal prediction neural network model;

[0078] In specific implementation, as parameters such as the data access mode of disk read and write data, the data reading range of a single IO operation, and the concurrency of data access requests change, the vibration frequency and amplitude of the magnetic head inside the mechanical disk also vary greatly. In a disk array, each disk stores different data, and the vibration conditions generated by each disk during read and write operations are also different. Therefore, when analyzing disk failures, it is necessary to conduct a comprehensive analysis based on the data read and write operations being performed on the disk. For disk units with high concurrency and high operation rates, the vibration signal analysis basis should be more significantly different from that of disks with low data operation rates, so that the fault analysis results are reliable and reasonable.

[0079] Furthermore, in a specific implementation, an index is constructed for the data in each disk unit, and the index is used to analyze which data each disk unit needs to read and write in each data access request of the disk array, thereby determining the data reading and writing behavior characteristics of each disk unit, and using the vibration signal prediction neural network model to predict the vibration data generated by each disk unit due to data reading and writing.

[0080] The step S200 is divided into the following steps:

[0081] Step S201: Setting a vibration monitoring sensor to monitor the disk array vibration signal in real time;

[0082] Step S202: obtaining physical structure data of the disk array, and constructing a three-dimensional material model for the physical structure of the disk array;

[0083] Step S203: respectively monitoring the amplitude and phase change data of vibration signals of different frequencies transmitted from each disk unit to each vibration monitoring sensor, and then calculating the phase offset and amplitude attenuation coefficient of the vibration signal transmitted by each disk unit;

[0084] The phase difference between the vibration signal at the disk unit and the vibration monitoring sensor is used as the phase offset of the vibration signal transmission, and the amplitude ratio of the vibration signal at the disk unit and the vibration monitoring sensor is used as the amplitude attenuation coefficient of the vibration signal transmission;

[0085] Step S204: using curve fitting to obtain frequency-amplitude attenuation coefficient curves and frequency-phase offset curves of vibration signals of each disk unit at different frequencies transmitted to each vibration monitoring sensor;

[0086] For any disk unit x, the frequency-amplitude attenuation coefficient curve of the vibration signal transmitted to the vibration monitoring sensor s is recorded as L_a(x,s), and the frequency-phase offset curve is recorded as

[0087] In the specific implementation, a 3D scanner is used to obtain the physical structure data of the disk array, which is then imported into COMSOL Multiphysics to construct a three-dimensional material model, define material properties, and determine the position of each disk unit in the three-dimensional material model. Furthermore, an exciter is used to apply a 10Hz to 10kHz sinusoidal sweep signal to each disk unit, and the sensor response is recorded. Based on the sweep signal input and response, the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal of each disk unit are calculated, and the calculation results of different frequency components are curve fitted.

[0088] The step S400 is divided into the following steps:

[0089] Step S401: Setting a frequency step f_step for the vibration signal spectrum, dividing all frequency components in the vibration signal belonging to the frequency level interval [(n-1)×f_step,n×f_step] into frequency level n, and summing the amplitudes of all frequency components of the same frequency level and dividing by the frequency step f_step as the amplitude of the corresponding frequency level; where n is the frequency level number, which is a positive integer;

[0090] Step S402: For any disk unit x, the amplitude attenuation coefficient k(x, s, f) of any frequency f component of the vibration signal generated by reading and writing data and transmitted to the vibration monitoring sensor s is calculated as follows:

[0091] Extract the frequency level interval to which frequency f belongs, calculate the integrated average value of L_a(x,s) over the frequency in the interval, and divide the calculated result by the frequency step size to obtain k(x,s,f);

[0092] For any disk unit x, the phase offset of any frequency f component of the vibration signal generated by reading and writing data transmitted to the vibration monitoring sensor s The calculation method is:

[0093] Extract the frequency level interval to which the frequency f belongs, and calculate the frequency level interval For the integrated average value of the frequency, divide the calculated result by the frequency step size to get

[0094] In a specific implementation, the frequency step size f_step is set to 100 Hz, and the frequency level intervals are further divided into: level 1 (0-100 Hz), level 2 (100-200 Hz) ... level 100 (9.9-10 kHz). Considering that the high-frequency component of the vibration signal has a strong interference on the working condition and life of the disk, when the high frequency band is higher than 5 kHz, the frequency step size is reduced to 50 Hz, and the frequency level intervals are further corrected to level 1 (0-100 Hz), level 2 (100-200 Hz) ... level 50 (4.9 kHz-5 kHz), level 51 (5 kHz-5.05 kHz) ... level 150 (9.95-10 kHz);

[0095] Further, based on the amplitude attenuation coefficient and phase offset fitting curve, the amplitude attenuation coefficient k(x,s,f) and phase offset of each frequency level interval are calculated. For any frequency level n, calculate according to the formula:

[0096] k(x,s,f)=[1 / (f_step)]×∫ n [L_a(x,s)]df;

[0097]

[0098] Among them, ∫ n [L_a(x,s)] is the integral of the frequency-amplitude attenuation coefficient curve within the frequency level interval of frequency level n, is the integral of the frequency-phase offset curve within the frequency level interval of frequency level n;

[0099] In the specific implementation, considering that the amplitude attenuation coefficient and phase offset may be nonlinear with frequency, using the value of the center point of the interval or the average of the boundaries on both sides of the interval as the final calculation result cannot reflect the overall relationship between the amplitude attenuation coefficient and phase offset and frequency in each frequency level interval. Therefore, the integration and averaging method is adopted to ensure that the overall amplitude attenuation coefficient and phase offset in each frequency level interval are accurately reflected.

[0100] The step S500 is divided into the following steps:

[0101] Step S501: using the predicted characteristic parameters of the vibration signal generated by the read / write data of each disk unit, and generating the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal generated by the read / write data of the disk unit and transmitted to each vibration monitoring sensor, simulating the spectrum data of the vibration signal of each disk unit transmitted to the vibration monitoring sensor;

[0102] Step S502: performing frequency domain complex addition on the vibration signals transmitted from each disk unit to the vibration monitoring sensor to obtain the amplitude and phase of each frequency component of the vibration signal predicted by each vibration monitoring sensor;

[0103] Step S503: monitoring the vibration signal of the disk array in real time through a vibration monitoring sensor provided on the outer contour of the disk array, and obtaining the amplitude and phase of each frequency component of the measured vibration signal through Fourier transform;

[0104] Step S504: Calculating the deviation between the measured vibration signal and the predicted vibration signal of the disk array vibration signal, performing a weighted summation, and setting a vibration deviation threshold; when the weighted summation result is greater than or equal to the vibration deviation threshold, determining that a faulty disk unit exists in the disk array, further constructing a fault location model, and locating the faulty disk unit in the disk array;

[0105] The step S504 constructs a fault location model to locate the faulty disk unit in the disk array, including the following:

[0106] Construct a fault location model M, which is as follows:

[0107]

[0108] M=min_ΔS x (f){∑ s,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]|| 2 +λ×∑ x,f (ΔS x (f))};

[0109] Where, ΔS x (f) is the spectrum deviation of the frequency f component of the vibration signal of disk unit x, ΔS s (f) is the spectrum deviation of the frequency f component of the vibration signal of the vibration monitoring sensor s, k x,f is the transfer function of the vibration signal of frequency f from the disk unit x to the vibration monitoring sensor s, λ is the regularization parameter, j is a complex unit; ||p|| 2 is the L2 norm square of parameter p;

[0110] In the fault location model M formula, ∑ s,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]|| 2 It is a residual term used to ensure that the model prediction is as close as possible to the actual observed data. By minimizing the prediction error, the vibration anomaly ΔS output by the model iteration is guaranteed. x (f) can reasonably explain the deviation observed by the sensor, λ×∑ x,f (ΔS x (f)) is a regularization term that accurately locates disk units with a high probability of abnormal vibration signals through sparse constraints. It also filters out false abnormal signals caused by sensor noise or cross-interference by screening disk units, thus preventing small vibration deviations or noise of disk units from interfering with the location of the fault source.

[0111] Initialize the spectrum deviation of the frequency f component of the vibration signal of each disk unit to 0, select the spectrum deviation ΔS of the frequency f component of the vibration signal of any disk unit x x (f) is used as a variable, and the spectrum deviation of the frequency f component of the vibration signal of the remaining disk units is fixed, and ΔS is solved to minimize the value of the fault location model M. x (f);

[0112] In the specific implementation, each frequency component is iterated separately, the frequency component spectrum deviation of each disk unit is preset to 0, and stored in the deviation matrix of the corresponding frequency component, and the frequency f component deviation ΔS of the vibration signal of any disk unit x is selected. x (f) is calculated as a variable, and the fault location model is iteratively simplified to about ΔS x (f) One-dimensional optimization problem;

[0113] ΔS x (f) Take the derivative and set it to zero to calculate ΔS x The solution of (f) is ΔS x The iterative result of (f) is that the variable ΔS corresponding to the other x values with different values is used in the above iterative method. x (f) After iterative update, it is used as the variable ΔS corresponding to all x values x (f) initial value;

[0114] Set the threshold for the number of failed disks to be determined, using the above ΔS x(f) The solution method iterates multiple times on the spectral deviation of the frequency f component of the vibration signal of each disk unit until the number of disk units with the spectral deviation of the frequency f component of the vibration signal greater than the vibration anomaly threshold S' in all disk units is less than or equal to the threshold of the number of faulty disks to be determined, and marks the disk units with the spectral deviation of the frequency f component of all vibration signals greater than the vibration anomaly threshold S' as faulty disk units; when setting the vibration anomaly threshold S', a dynamic threshold can be set according to the analyzed frequency f to retain more high-frequency anomaly components, so as to ensure more sensitive detection and identification of the high-frequency abnormal vibration signals of each disk unit;

[0115] Further, in the subsequent iteration variable ΔS x (f), set the vibration anomaly threshold S'. When ΔS x (f) < S', correct the corresponding item value of ΔS x (f) in the deviation matrix to 0. When ΔS x (f) ≥ S', continue the iteration to achieve the selection and rejection of the abnormal amounts of different frequency components of each disk unit during the iteration process, and then correct most of the elements in the deviation matrix to 0 through multiple iterations;

[0116] In the iteration result of any frequency f component, the vibration signals of the disk units corresponding to the remaining elements all satisfy: If the abnormal vibration of the frequency f component of disk unit x is transmitted to vibration monitoring sensor s as a predicted vibration signal, and when the deviation between the predicted vibration signal of vibration monitoring sensor s and the measured vibration signal is the smallest, the abnormal vibration of the frequency f component of disk unit x is stronger and is recognized as abnormal vibration in actual operation;

[0117] Further calculate the vibration signal deviation of each disk unit according to the iteration result. For any disk unit x, the calculation formula of the vibration signal deviation Res(x) is as follows:

[0118] Res(x) = ∑ s,f ||ΔS s (f) - ∑ x [k x,f ×ΔS x (f)]||2;

[0119] For any disk units a and b, if Res(a) > Res(b), it is determined that the failure probability of disk unit a is less than that of disk unit b. If Res(a) < Res(b), it is determined that the failure probability of disk unit a is greater than that of disk unit b;

[0120] Further, arrange each disk unit in ascending order of vibration signal deviation and feedback the disk unit fault investigation to the management personnel in turn;

[0121] In specific implementation, when the frequency f component deviation ΔS x (f) When the number of disk units with vibrations greater than the abnormal vibration threshold is less than or equal to the threshold of the number of faulty disks to be determined, it is determined that the above iterative algorithm has converged. For any frequency f component, if the deviation of the vibration signal calculated by any disk unit x is small, it means in actual operation that the abnormal vibration of the frequency component is more consistent with the actual vibration signal of the vibration monitoring sensor s. Therefore, it can be considered that the probability of abnormal vibration of the frequency f component of the disk unit x is greater, and therefore each disk unit involved is regarded as a faulty disk unit.

[0122] Furthermore, the vibration signal deviation between the signal spectrum deviation predicted by the model and the actual signal spectrum deviation is calculated when a single disk failure is assumed. The smaller the vibration signal deviation, the more consistent the disk failure is with the analysis result, that is, the higher the possibility of disk failure. Therefore, each disk unit is sorted in ascending order according to the vibration signal deviation calculation results, as the order of the failure probability of each disk unit from large to small.

[0123] like Figure 2 As shown, the present invention also provides a network equipment fault prediction system based on multi-source data fusion, the system comprising: a disk vibration prediction module, a vibration transmission analysis module, and a fault disk location module;

[0124] The disk vibration prediction module trains a vibration signal prediction neural network model by analyzing the historical data of vibration signals generated when each disk unit in the disk array reads and writes data, and predicts the vibration signals generated by each disk unit reading and writing data based on real-time monitored data access requests; the vibration transmission analysis module constructs a three-dimensional material model for the disk array, and analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the position of each disk unit to each vibration monitoring sensor; the fault disk locating module constructs a fault locating model by comparing and analyzing the vibration signal predicted by the vibration monitoring sensor and the measured vibration signal data, and locates the fault of each disk unit.

[0125] The disk vibration prediction module includes: a vibration data processing unit, a vibration feature analysis unit, and a vibration signal prediction unit;

[0126] The vibration data processing unit is used to process the historical data of vibration signals generated when each disk unit in the disk array reads and writes data; the vibration feature analysis unit uses Fourier transform to obtain the spectrum data of the vibration signal and annotate the vibration signal feature parameters; the vibration signal prediction unit trains the vibration signal prediction neural network model by constructing a vibration signal feature parameter set.

[0127] The vibration transmission analysis module includes: a material model construction unit and a signal transmission analysis unit;

[0128] The material model construction unit obtains the physical structure data of the disk array and constructs a three-dimensional material model; the signal transmission analysis unit analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the position of each disk unit to each vibration monitoring sensor, and fits to obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve.

[0129] The faulty disk locating module includes: a vibration deviation analyzing unit, a positioning model building unit, and a faulty disk screening unit;

[0130] The vibration deviation analysis unit compares and analyzes the vibration signal predicted by the vibration monitoring sensor with the measured vibration signal data; the positioning model construction unit constructs a fault positioning model based on the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal; the faulty disk screening unit iterates the spectrum deviation of the frequency f component of the vibration signal of each disk unit multiple times, and screens the faulty disk units according to the threshold value of the number of faulty disks to be determined, and performs disk unit fault troubleshooting feedback in order from small to large according to the vibration signal deviation of each faulty disk unit.

[0131] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. Network equipment fault prediction method based on multi-source data fusion, characterized by , the method comprises the following steps: Step S100: Obtain historical data of vibration signals generated when each disk unit in the disk array reads and writes data, use Fourier transform to obtain spectrum data of the vibration signals, and perform vibration signal characteristic parameter annotation to construct a vibration signal characteristic parameter set, and train a vibration signal prediction neural network model; the vibration signal characteristic parameters include the read and write data characteristic parameters of the disk units and the vibration signal characteristic parameters; Step S200: Acquire physical structure data of the disk array, construct a three-dimensional material model, analyze the phase offset and amplitude attenuation coefficient of each frequency component of the vibration signal transmitted from the position of each disk unit to each vibration monitoring sensor, and obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve by fitting; Step S300: monitoring data access requests in the server transmission link, extracting characteristic parameters of read and write data, predicting characteristic parameters of the read and write data of each disk unit, and using the vibration signal prediction neural network model to predict characteristic parameters of the vibration signal generated by the read and write data of each disk unit; Step S400: Setting a frequency step size, dividing the frequency level intervals, dividing the frequency components of the vibration signal into frequency levels, and calculating the amplitude attenuation coefficient and phase offset of each frequency component of the disk unit vibration signal based on the frequency-amplitude attenuation coefficient curve and the frequency-phase offset curve transmitted from each disk unit vibration signal to each vibration monitoring sensor; Step S500: Acquire vibration signal monitoring data from each vibration monitoring sensor in real time, compare and analyze the data with the predicted vibration signal characteristic parameters at each vibration monitoring sensor, build a fault location model, and locate the fault of each disk unit.

2. The network equipment fault prediction method based on multi-source data fusion according to claim 1 is characterized in that: The step S100 is divided into the following steps: Step S101: Acquire historical vibration signal data generated when each disk unit in the disk array reads and writes data; the historical vibration signal data is vibration signal data generated when each disk unit reads and writes data independently; Step S102: using Fourier transform to obtain spectrum data of the vibration signal and perform vibration signal feature annotation; For any vibration signal history data s of any disk unit x in the disk array, the feature labeling result is x_s[D_x_s,S_x_s]; where D_x_s is the read and write data feature parameter set of the vibration signal history record s of disk unit x, and S_x_s is the vibration signal feature parameter of the vibration signal history record s of disk unit x; Step S103: constructing a vibration signal feature data set and training a vibration signal prediction neural network model.

3. The network equipment fault prediction method based on multi-source data fusion according to claim 1 is characterized in that: The step S200 is divided into the following steps: Step S201: Setting a vibration monitoring sensor to monitor the disk array vibration signal in real time; Step S202: obtaining physical structure data of the disk array, and constructing a three-dimensional material model for the physical structure of the disk array; Step S203: respectively monitoring the amplitude and phase change data of vibration signals of different frequencies transmitted from each disk unit to each vibration monitoring sensor, and then calculating the phase offset and amplitude attenuation coefficient of the vibration signal transmitted by each disk unit; The phase difference between the vibration signal at the disk unit and the vibration monitoring sensor is used as the phase offset of the vibration signal transmission, and the amplitude ratio of the vibration signal at the disk unit and the vibration monitoring sensor is used as the amplitude attenuation coefficient of the vibration signal transmission; Step S204: using curve fitting to obtain frequency-amplitude attenuation coefficient curves and frequency-phase offset curves of vibration signals of each disk unit at different frequencies transmitted to each vibration monitoring sensor; For any disk unit x, the frequency-amplitude attenuation coefficient curve of the vibration signal transmitted to the vibration monitoring sensor s is recorded as L_a(x,s), and the frequency-phase offset curve is recorded as 4. The network equipment fault prediction method based on multi-source data fusion according to claim 1 is characterized in that: The step S400 is divided into the following steps: Step S401: Setting a frequency step f_step for the vibration signal spectrum, dividing all frequency components in the vibration signal belonging to the frequency level interval [(n-1)×f_step,n×f_step] into frequency level n, and summing the amplitudes of all frequency components of the same frequency level and dividing by the frequency step f_step as the amplitude of the corresponding frequency level; where n is the frequency level number, which is a positive integer; Step S402: For any disk unit x, the amplitude attenuation coefficient k(x, s, f) of any frequency f component of the vibration signal generated by reading and writing data and transmitted to the vibration monitoring sensor s is calculated as follows: Extract the frequency level interval to which frequency f belongs, calculate the integrated average value of L_a(x,s) over the frequency in the interval, and divide the calculated result by the frequency step size to obtain k(x,s,f); For any disk unit x, the phase offset of any frequency f component of the vibration signal generated by reading and writing data transmitted to the vibration monitoring sensor s The calculation method is: Extract the frequency level interval to which the frequency f belongs, and calculate the frequency level interval For the integrated average value of the frequency, divide the calculated result by the frequency step size to get 5. The network equipment fault prediction method based on multi-source data fusion according to claim 4 is characterized in that: The step S500 is divided into the following steps: Step S501: using the predicted characteristic parameters of the vibration signal generated by the read / write data of each disk unit, and generating the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal generated by the read / write data of the disk unit and transmitted to each vibration monitoring sensor, simulating the spectrum data of the vibration signal of each disk unit transmitted to the vibration monitoring sensor; Step S502: performing frequency domain complex addition on the vibration signals transmitted from each disk unit to the vibration monitoring sensor to obtain the amplitude and phase of each frequency component of the vibration signal predicted by each vibration monitoring sensor; Step S503: monitoring the vibration signal of the disk array in real time through a vibration monitoring sensor provided on the outer contour of the disk array, and obtaining the amplitude and phase of each frequency component of the measured vibration signal through Fourier transform; Step S504: Calculate the deviation between the measured vibration signal and the predicted vibration signal of the disk array vibration signal, perform weighted summation, and set a vibration deviation threshold; when the weighted summation result is greater than or equal to the vibration deviation threshold, determine that there is a faulty disk unit in the disk array, further construct a fault location model, and locate the faulty disk unit in the disk array.

6. The network equipment fault prediction method based on multi-source data fusion according to claim 5 is characterized in that: The step S504 constructs a fault location model to locate the faulty disk unit in the disk array, including the following: Construct a fault location model M, which is as follows: M=min_ΔS x (f){∑ s,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]|| 2 +λ×∑ x,f (ΔS x (f))}; Where, ΔS x (f) is the spectrum deviation of the frequency f component of the vibration signal of disk unit x, ΔS s (f) is the spectrum deviation of the frequency f component of the vibration signal of the vibration monitoring sensor s, k x,f is the transfer function of the vibration signal of frequency f from the disk unit x to the vibration monitoring sensor s, λ is the regularization parameter, j is a complex unit; ||p|| 2 is the L2 norm square of parameter p; Initialize the spectrum deviation of the frequency f component of the vibration signal of each disk unit to 0, select the spectrum deviation ΔS of the frequency f component of the vibration signal of any disk unit x x (f) is used as a variable, and the spectrum deviation of the frequency f component of the vibration signal of the remaining disk units is fixed, and ΔS is solved to minimize the value of the fault location model M. x (f); Set the threshold for the number of failed disks to be determined, using the above ΔS x (f) The solution method iterates the spectrum deviation of the vibration signal frequency f component of each disk unit multiple times until the number of disk units whose spectrum deviation of the vibration signal frequency f component is greater than the vibration abnormality threshold S' is less than or equal to the threshold of the number of faulty disks to be determined, and all disk units whose spectrum deviation of the vibration signal frequency f component is greater than the vibration abnormality threshold S' are marked as faulty disk units; The vibration signal deviation of each disk unit is further calculated based on the iterative results. For any disk unit x, the vibration signal deviation Res(x) is calculated as follows: Res(x)=∑ x,f ||ΔS s (f)-∑ x [k x,f ×ΔS x (f)]||2; For any disk units a and b, if Res(a) > Res(b), it is determined that the failure probability of disk unit a is less than that of disk unit b; if Res(a) < Res(b), it is determined that the failure probability of disk unit a is greater than that of disk unit b. Furthermore, each disk unit is sequentially fed back to the administrator for disk unit fault troubleshooting in ascending order of the vibration signal deviation.

7. A network equipment fault prediction system based on multi-source data fusion using the network equipment fault prediction method based on multi-source data fusion according to any one of claims 1 to 6, characterized in that: The system includes: a disk vibration prediction module, a vibration transmission analysis module, and a faulty disk location module. The disk vibration prediction module analyzes the historical vibration signal data generated when each disk unit in the disk array reads and writes data, trains a vibration signal prediction neural network model, and predicts the vibration signals generated when each disk unit reads and writes data according to the real-time monitored data access requests; the vibration transmission analysis module constructs a three-dimensional material model for the disk array and analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the position where each disk unit is located to each vibration monitoring sensor; the faulty disk location module constructs a fault location model by comparing and analyzing the predicted vibration signals and the measured vibration signal data of the vibration monitoring sensors, and locates faults for each disk unit.

8. The network equipment fault prediction system based on multi-source data fusion according to claim 7 is characterized in that: The disk vibration prediction module includes: a vibration data processing unit, a vibration feature analysis unit, and a vibration signal prediction unit. The vibration data processing unit is used to process the historical vibration signal data generated when each disk unit in the disk array reads and writes data; the vibration feature analysis unit uses Fourier transform to obtain the spectral data of the vibration signal and performs vibration signal feature parameter annotation; the vibration signal prediction unit trains a vibration signal prediction neural network model by constructing a vibration signal feature parameter set.

9. The network equipment fault prediction system based on multi-source data fusion according to claim 7, characterized in that: The vibration transmission analysis module includes: a material model construction unit and a signal transmission analysis unit. The material model construction unit obtains the physical structure data of the disk array and constructs a three-dimensional material model; the signal transmission analysis unit analyzes the amplitude and phase changes of each frequency component of the vibration signal transmitted from the position where each disk unit is located to each vibration monitoring sensor, and fits to obtain a frequency-amplitude attenuation coefficient curve and a frequency-phase offset curve.

10. The network equipment fault prediction system based on multi-source data fusion according to claim 7, characterized in that: The faulty disk location module includes: a vibration deviation analysis unit, a location model construction unit, and a faulty disk screening unit. The vibration deviation analysis unit compares and analyzes the predicted vibration signals and the measured vibration signal data of the vibration monitoring sensors; the location model construction unit constructs a fault location model according to the amplitude attenuation coefficient and phase offset of each frequency component of the vibration signal; the faulty disk screening unit iterates the spectral deviation of the f component of the vibration signal of each disk unit multiple times, screens faulty disk units according to the threshold of the number of faulty disks to be determined, and sequentially feeds back the disk unit fault troubleshooting in ascending order of the vibration signal deviation of each faulty disk unit.

Images