Carrier module intelligent detection system based on artificial intelligence analysis

Abstract

The invention relates to the technical field of intelligent detection, in particular to a carrier module intelligent detection system based on artificial intelligence analysis, which comprises a fault data basic filling module, a fault correlation cross analysis module, a fault relation accurate analysis module and an accurate fault data recovery module, wherein the fault data basic filling module adopts a data filling method to calculate filling data and fill missing historical fault data, the fault correlation cross analysis module adopts a cross method to calculate a correlation coefficient between the historical fault data, and when correlation exists between two pieces of fault data, the fault correlation cross analysis module calculates the correlation coefficient between the historical fault data. If yes, the correlation relation is determined according to the timestamps of the fault data, and when correlation exists between the multiple pieces of fault data, the fault relation accurate analysis module calculates partial correlation coefficients between the two pieces of fault data one by one through a partial correlation analysis method; and the precise fault data recovery module adopts a fusion algorithm and partial correlation coefficients to establish a linear regression model to calculate supplementary data and replace filling data.

Description

Technical Field

[0001] The present invention relates to the field of intelligent detection technology, and in particular to an intelligent detection system for a carrier module based on artificial intelligence analysis. Background Art

[0002] In today's digital communication era, the stable operation of carrier communication systems plays a vital role in ensuring efficient and accurate information transmission. As the core component of carrier communication systems, the performance of carrier modules directly affects the reliability of the entire system.

[0003] The carrier module detection system is designed to ensure the stable operation of the carrier communication system. To this end, in actual applications, carrier module fault data is recorded over a long period of time to form a historical fault database containing multiple aspects of information such as fault occurrence time, fault characteristics, and fault content. Historical fault data is usually indexed by timestamps, and each timestamp contains detection data from multiple equidistant time nodes, presenting a complex structure with multiple dimensions and time series.

[0004] However, during the data collection and recording process, due to various reasons (such as sensor failure, communication interruption, data storage error, etc.), fault data will inevitably be missing. After the missing fault data is missing, the existing technology usually fills the missing data by analyzing the changing trend of the fault data, and then analyzes the correlation between different fault data. However, when filling the data, if the overall distribution law of the data, the complex correlation between multiple variables and other factors are not fully considered, the missing data will be filled inaccurately. When the staff views the filled historical fault data, it is difficult to accurately grasp the overall picture of the fault and judge the development trend of the fault due to inaccurate data. In view of this, we propose a carrier module intelligent detection system based on artificial intelligence analysis. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem of missing fault data due to sensor failure, communication interruption, etc. during data collection and recording. The existing technology mostly fills the missing data by analyzing the data change trend and then analyzes the correlation of the fault data. However, if the overall distribution of the data and the complex association of multiple variables are not fully considered during the filling, it is easy to cause inaccurate filling, making it difficult for staff to accurately grasp the overall picture of the fault and judge the development trend when viewing the historical fault data after filling.

[0006] To achieve the above objectives, the present invention provides an intelligent detection system for carrier modules based on artificial intelligence analysis, including a fault data basic filling module, a fault correlation cross-analysis module, a fault relationship precise analysis module, and a precise fault data repair module, wherein:

[0007] The fault data basic filling module adopts the data filling method to calculate the filling data to fill in the missing historical fault data. The fault correlation cross analysis module adopts the cross method to calculate the correlation coefficient between the historical fault data. When there is a correlation between two fault data, the correlation relationship is determined according to the timestamp of the fault data. When there is a correlation between multiple fault data, the fault relationship precise analysis module adopts the biased correlation analysis method to calculate the partial correlation coefficient between the two fault data one by one; the precise fault data repair module adopts the fusion algorithm and the partial correlation coefficient to establish a linear regression model to calculate the supplementary data and replace the filling data.

[0008] As a further improvement of the present technical solution, the fault data basic filling module senses the historical fault data of the carrier module, the historical fault data includes multiple fault occurrence timestamps, and each timestamp corresponds to fault data detected at multiple equidistant time nodes;

[0009] As a further improvement of the present technical solution, the data filling method in the fault data basic filling module calls out the timestamp where the missing data is located, perceives two other non-missing data points adjacent to the corresponding time node, determines a straight line through the two other non-missing data points, and uses the straight line trend to calculate and fill in the missing data.

[0010] The beneficial effect of adopting the above-mentioned further scheme is that the filling method in the present invention is relatively simple and direct, based on basic mathematical principles (two points determine a straight line), and does not require complex calculations and model construction. This enables missing data to be quickly filled in during data processing, reducing the time cost and computing resource consumption of data processing, improving the efficiency of data processing, and enabling the filled fault data to meet the basic requirements of subsequent analysis methods for data integrity, so that the fault correlation cross-analysis module and the fault relationship precise analysis module can proceed smoothly.

[0011] On the basis of the above technical solution, the present invention can also be improved as follows.

[0012] As a further improvement of this technical solution, the fault correlation cross analysis module senses the fault data Where fault data X and fault data Y are detection data at the same time node, the correlation coefficient is:

[0013]

[0014] in is the mean of X, =The mean value of Y; set the correlation coefficient threshold. If the absolute value of the correlation coefficient is less than or equal to the correlation coefficient threshold, it is determined that there is no correlation between the fault data X and the fault data Y; if the absolute value of the correlation coefficient is greater than the correlation coefficient threshold, it is determined that there is correlation between the fault data X and the fault data Y; call out the timestamp X corresponding to the fault data X and the fault data Y T 、Y T , if timestamp X T Earlier than timestamp Y T , then the correlation relationship between fault data X and fault data Y is fault data X→fault data Y.

[0015] The beneficial effect of adopting the above further solution is that the present invention calculates the correlation coefficient between different fault data at the same time node in order to identify the potential correlation between the fault data. During the operation of the carrier module, multiple faults may occur simultaneously or successively, and the fault data may be causally related or caused by common factors. The fault correlation cross-analysis module accurately identifies the correlation between the fault data in a quantitative manner, so as to gain a deeper understanding of the mechanism of fault occurrence and clarify the order of fault occurrence through correlation, facilitating the analysis of the propagation path and root cause of the fault data.

[0016] On the basis of the above technical solution, the present invention can also be improved as follows.

[0017] As a further improvement of this technical solution, the fault correlation cross analysis module calculates the correlation coefficient r between fault data X and fault data Y. XY Then, calculate the correlation coefficient r between the fault data X and the fault data Z XZ , if | r XY |>correlation coefficient threshold, and |r XZ |>correlation coefficient threshold, the correlation coefficient r between fault data Y and fault data Z is calculated at the first time YZ , if the correlation coefficient r is calculated YZ > correlation coefficient threshold, it is determined that there is correlation among X, Y, and Z, and the correlation relationship of the fault data is determined according to the corresponding timestamps.

[0018] The beneficial effect of adopting the above-mentioned further scheme is that the present invention takes into account that among a large number of fault data, if a comprehensive pairwise correlation coefficient calculation is performed on all fault data, the amount of calculation will increase exponentially with the increase in the number of fault data. The fault correlation cross-analysis module adopts a cross-calculation method to first determine the correlation between some fault data pairs, and then calculate the correlation coefficients of other fault data pairs in a targeted manner based on the results to avoid blind calculation.

[0019] On the basis of the above technical solution, the present invention can also be improved as follows.

[0020] As a further improvement of this technical solution, the partial correlation analysis method in the fault relationship precise analysis module is used to calculate the partial correlation coefficient r between fault data Y and fault data Z under the influence of control fault data X. YZ.X , and calculate the partial correlation coefficient r between fault data X and fault data Z under the influence of controlled fault data Y XZ.Y If r YZ.X >r XZ.Y When , it is judged that the fault data X has the greatest impact;

[0021] As a further improvement of this technical solution, the calculation formula of the partial correlation analysis method is as follows: Perceive fault data X, Y and Z, control the influence of fault data X, and the partial correlation coefficient r between fault data Y and fault data Z YZ.X for:

[0022] The beneficial effect of adopting the above-mentioned further scheme is that the partial correlation coefficient of the present invention can more accurately express the degree of linear relationship between two fault data after controlling other factors. Compared with the simple correlation coefficient, the partial correlation coefficient eliminates the interference of other factors and can better reflect the intrinsic connection between variables, further helping to analyze the cause-effect relationship between faults. By controlling different variables and observing the changes in the partial correlation coefficient, it is possible to determine which fault data is the cause and which is the result, or whether there are common influencing factors.

[0023] On the basis of the above technical solution, the present invention can also be improved as follows.

[0024] As a further improvement of the present technical solution, the fusion filling method in the precise fault data repair module senses multiple fault data that are correlated with the missing fault data, retrieves the partial correlation coefficients corresponding to the multiple fault data and the missing fault data, and establishes multiple linear regression models respectively: Y=a0+a1X; wherein Y is the supplementary data, X is the fault data that is correlated with the supplementary data, a1 is the regression coefficient, and a0 is the intercept.

[0025] As a further improvement of this technical solution, the regression coefficient a1 is calculated by the partial correlation coefficient: where r XY.Z is the partial correlation coefficient between fault data X and fault data Y, σ X and σ Y are the standard deviations of fault data X and fault data Y respectively;

[0026] intercept in and are the means of fault data X and fault data Y respectively.

[0027] As a further improvement of the present technical solution, after the precise fault data repair module calculates a plurality of supplementary data, it calls out the most supplementary data, selects the one with the highest frequency of occurrence as the final supplementary data, and replaces the filling data.

[0028] The beneficial effect of adopting the above-mentioned further scheme is that the data filling method adopted in the fault data basic filling module of the present invention may be inaccurate. The fusion filling method of the precise fault data repair module establishes a more accurate linear regression model by considering multiple fault data related to the missing fault data and the partial correlation coefficients between them, thereby improving the accuracy of missing data filling. When staff perform fault diagnosis, analyze the correlation between faults, and predict fault development trends, they can make more accurate judgments and decisions based on accurate fault data.

[0029] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is the overall module principle diagram of the present invention;

[0031] Figure 2 This is a flow chart of the working principle of the fault data basic filling module of the present invention;

[0032] Figure 3 This is a flow chart showing the working principle of the correlation coefficient in the fault correlation cross analysis module of the present invention;

[0033] Figure 4 This is a flow chart showing the working principle of the cross calculation method in the fault correlation cross analysis module of the present invention;

[0034] Figure 5 This is a flowchart of the working principle of the fault relationship accurate analysis module of the present invention;

[0035] Figure 6 This is a flow chart of the working principle of the precise fault data repair module of the present invention.

[0036] The meaning of each number in the figure is:

[0037] 100. Fault data basic filling module; 200. Fault correlation cross analysis module; 300. Fault relationship precise analysis module; 400. Precise fault data repair module. DETAILED DESCRIPTION

[0038] 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 any creative efforts shall fall within the scope of protection of the present invention.

[0039] Example 1, reference Figures 1-6 As shown, the carrier module intelligent detection system based on artificial intelligence analysis includes a fault data basic filling module 100, a fault correlation cross analysis module 200, a fault relationship precise analysis module 300 and a precise fault data repair module 400, wherein:

[0040] The fault data basic filling module 100 senses the historical fault data of the carrier module. The historical fault data includes data such as the fault occurrence time, fault characteristic data, and fault content. The historical fault data includes multiple fault occurrence timestamps, and each timestamp corresponds to fault data detected at multiple equidistant time nodes:

[0041] Fault occurrence timestamp Time Node Fault characteristic data 1 Fault characteristic data 2 ... Fault content ×year×month×day 1 Data value 1 Data value 2 ... Fault Description 1 ×year×month×day 2 Data value 3 Missing data value 4 ... Fault Description 2 ×year×month×day 3 Data value 5 Data value 6 ... Fault Description 3 ... ... ... ... ... ... ×year×month×day 1 Data value 7 Data value 8 ... Fault Description 4 ×year×month×day 2 Data value 9 Data value 10 ... Fault Description 5 ×year×month×day 3 Data value 11 Data value 12 ... Fault Description 6 ... ... ... ... ... ...

[0042] In the above table:

[0043] The "Fault Occurrence Timestamp" column records the specific time when the fault occurred;

[0044] The “Time Node” column indicates the equidistant time nodes under each fault occurrence timestamp;

[0045] The "Fault Feature Data" column contains various feature fault data related to the fault;

[0046] The "Fault Content" column describes in detail the fault conditions detected at each time node.

[0047] The fault data basic filling module 100 senses missing fault data in historical fault data and fills the missing fault data using a data filling method. The data filling method calls out two other non-missing data points adjacent to the corresponding time node in the timestamp where the missing data is located, uses the two other non-missing data points to determine a straight line, and uses the straight line trend to calculate and fill the missing data. The corresponding expression is as follows: the time node of the perceived missing data is t, the previous non-missing data point is (t1, x1), and the next non-missing data point is (t2, x2) (t1<t<t2). According to the principle of determining a straight line from two points, the slope k of the straight line is The filling value x at the missing data point t is calculated by the following formula: x=x1+k×(t-t1).

[0048] In the actual fault data recording process, data missing will inevitably occur due to various reasons such as sensor failure, communication interruption, etc. The fault data basic filling module 100 uses a data filling method to supplement these missing fault data to make the historical fault data complete, providing a more comprehensive data basis for subsequent fault analysis and processing, and providing preliminary data support for the fault correlation cross-analysis module 200 and the fault relationship precise analysis module 300 to calculate the correlation between fault data, determine the root cause of the fault, etc.

[0049] refer to Figure 3 As shown, the fault correlation cross analysis module 200 uses a cross method to calculate the correlation coefficient between multiple historical fault data within the same time stamp: the perceived fault data Among them, fault data X and fault data Y are detection data at the same time node, so the correlation coefficient

[0050] in is the mean of X, is the mean value of Y. The correlation coefficient is used to measure the degree of association between fault data. Its value range is usually between -1 and 1. When the correlation coefficient is positive, it means that there is a positive correlation between the two variables; when the correlation coefficient is negative, it means that there is a negative correlation. Set the correlation coefficient threshold. If the absolute value of the correlation coefficient is less than or equal to the correlation coefficient threshold, it is judged that there is no correlation between fault data X and fault data Y. If the absolute value of the correlation coefficient is greater than the correlation coefficient threshold, it is judged that there is a correlation between fault data X and fault data Y, and the timestamp X corresponding to fault data X and fault data Y is retrieved. T 、Y T , if timestamp X T Earlier than timestamp Y T , then the correlation relationship between fault data X and fault data Y is fault data X→fault data Y.

[0051] For example, if the correlation coefficient between the carrier signal strength fault data and the data transmission fault data is greater than the correlation coefficient threshold, it means that there is an intrinsic connection between the two faults, which are caused by the same potential cause, or the occurrence of one fault will trigger another fault. The relationship between the faults is determined, and the size of the correlation coefficient is used to further determine the closeness between the two fault data. The closer the correlation coefficient of two correlated fault data is to 1, the stronger the positive correlation between the two variables; the closer it is to -1, the stronger the negative correlation.

[0052] When fault data appears, the root cause of the fault can be located more quickly based on the correlation between the fault data. For example, if a certain fault is known to be related to several other faults, then during troubleshooting, the fault data that may appear at the next timestamp can be predicted based on the correlation coefficients and correlation relationships between different fault data.

[0053] In the historical fault data, there will be a large amount of fault data, and there may be intricate relationships between different fault data. If all fault data are randomly and comprehensively calculated for pairwise correlation coefficients, the amount of calculation will increase exponentially as the number of fault data increases. Therefore, in order to avoid blind calculation when calculating the correlation coefficient, reference is made to Figure 4 As shown, the fault correlation cross analysis module 200 uses the cross method to calculate the correlation coefficient. The specific working principle is: calculate the correlation coefficient r between fault data X and fault data Y XY Then, calculate the correlation coefficient r between the fault data X and the fault data Z XZ , if | r XY |>correlation coefficient threshold, and |r XZ |>correlation coefficient threshold, the correlation coefficient r between fault data Y and fault data Z is calculated at the first time YZ , if the correlation coefficient r is calculated YZ > correlation coefficient threshold, it is determined that there is correlation among X, Y, and Z, and the correlation relationship of the fault data is determined according to the corresponding timestamps;

[0054] Since the fault correlation cross analysis module 200 calculates that there is a correlation between fault data X and fault data Y, and between fault data X and fault data Z, in order to avoid multi-factor interaction between X, Y and Z, the fault correlation cross analysis module 200 calculates the correlation coefficient r between fault data Y and fault data Z. YZ When the correlation coefficient r YZ Affected by the fault data X, the fault relationship precise analysis module 300 uses the partial correlation analysis method to calculate the partial correlation coefficient r between the fault data Y and the fault data Z under the influence of the control fault data X. YZ.X , and calculate the partial correlation coefficient r between fault data X and fault data Z under the influence of controlled fault data Y XZ.Y ;

[0055] If r YZ.X >r XZ.Y When , it is determined that the fault data X has the greatest impact. When the fault correlation cross analysis module 200 outputs the correlation between the fault data, the impact degree of different fault data is output synchronously to assist the staff in determining the fault data to be repaired.

[0056] Partial correlation analysis is a method of analyzing the linear relationship between two fault data under the condition of controlling the linear influence of other fault data. The partial correlation coefficient is used to represent the closeness of this relationship, thereby eliminating the interference of other factors and calculating the correlation coefficient r between fault data Y and fault data Z. YZ , to avoid the occurrence of "pseudo-correlation", the specific calculation formula is as follows: sensing fault data X, Y and Z, controlling the influence of fault data X, the partial correlation coefficient r between fault data Y and fault data Z YZ.X for:

[0057] Calculate the partial correlation coefficient r between fault data Y and fault data Z YZ.X Finally, the key influencing factors between different fault data can be determined, which helps the staff to fundamentally solve the fault problem and reduce the probability of fault recurrence.

[0058] Example 2, based on Example 1, in order to avoid the situation where the data filling method adopted by the fault data basic filling module 100 is inaccurate when filling missing data due to missing fault data, and taking into account that the missing data filled by the fault data basic filling module 100 will not affect the calculation of the correlation coefficient of the fault correlation cross analysis module 200 (the correlation coefficient is a statistical indicator used to measure the strength and direction of the linear relationship between two variables. Its calculation is mainly based on the relative position and change trend between data points, rather than the specific value of a single data point. After the missing data filled by the fault data basic filling module 100 is filled, although the specific value may have errors, the relative position of the filled data in the data set and its relationship with other variables are determined according to the overall characteristics of the original data. Therefore, when calculating the correlation coefficient, the filled data can be integrated into the overall data relationship and will not have a significant impact on the calculation result of the correlation coefficient). However, inaccurate filled data may still interfere with the staff's review and in-depth analysis of historical fault data, thereby affecting the staff's review and analysis of historical fault data;

[0059] refer to Figure 4-Figure 6 As shown, this embodiment differs from the first embodiment in that the precise fault data repair module 400 comprehensively senses the historical fault data, precisely locates the missing fault data therein, and then processes the data using a fusion filling method. The fusion filling method senses multiple fault data that are correlated with the missing fault data, retrieves the partial correlation coefficients corresponding to the multiple fault data and the missing fault data, establishes multiple linear regression models, calculates multiple supplementary data, retrieves the most supplementary data, selects the data with the highest frequency of occurrence as the final supplementary data, and replaces the supplementary data. The calculation formula for calculating the supplementary data Y based on the fault data X by the fusion filling method is as follows:

[0060] Perceive the partial correlation coefficient between each known fault data and the missing data, and treat the fault data as the variable that affects the missing data. If the partial correlation coefficient is positive, it means that the two variables are positively correlated, that is, when one variable increases, the other variable also tends to increase; if it is negative, it is a negative correlation. Establish a linear regression model, Y = a0 + a1X; where Y is the supplementary data, X is the fault data that is correlated with the supplementary data, a1 is the regression coefficient, and a0 is the intercept. Calculate the partial correlation coefficient: where r XY.Z is the partial correlation coefficient between fault data X and fault data Y, reflecting the direction and degree of linear correlation between X and Y; σ X and σ Y are the standard deviations of fault data X and fault data Y, The correlation is scaled so that the regression coefficient a1 accurately represents the average change in Y for each unit change in X;

[0061] Calculate the intercept in and are the means of the fault data X and fault data Y respectively, and the intercept a0 is the estimated value of Y when X=0, which ensures that the regression line can pass through the mean point of the sample data.

[0062] The fusion filling method in the precise fault data repair module 400 takes the partial correlation coefficient into account during the calculation process, which can eliminate the interference of other factors and accurately reflect the relationship between the fault data. For example, when analyzing the fault of the carrier module in the communication system, accurate filling data can help staff accurately judge whether the relationship between different fault characteristics (such as signal strength, bit error rate, etc.) is a true causal relationship or a false correlation caused by inaccurate data, thereby providing a reliable basis for subsequent workers' fault diagnosis.

[0063] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. The carrier module intelligent detection system based on artificial intelligence analysis is characterized by: The system comprises a fault data basic filling module (100), a fault correlation cross analysis module (200), a fault relationship precise analysis module (300) and a precise fault data repair module (400), wherein: The fault data basic filling module (100) uses a data filling method to calculate filling data to fill in missing historical fault data. The fault correlation cross analysis module (200) uses a cross method to calculate the correlation coefficient between historical fault data. When there is a correlation between two fault data, the correlation relationship is determined according to the time stamp of the fault data. When there is a correlation between multiple fault data, the fault relationship precise analysis module (300) uses a partial correlation analysis method to calculate the partial correlation coefficient between the two fault data one by one. The precise fault data repair module (400) uses a fusion algorithm and partial correlation coefficients to establish a linear regression model to calculate supplementary data and replace the filling data.

2. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 1 is characterized in that: The fault data basic filling module (100) senses historical fault data of the carrier module, the historical fault data including a plurality of fault occurrence timestamps, each of which corresponds to fault data detected at a plurality of equidistant time nodes.

3. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 2 is characterized in that: The data filling method in the fault data basic filling module (100) retrieves the timestamp where the missing data is located, senses two other non-missing data points adjacent to the corresponding time node, determines a straight line through the two other non-missing data points, and calculates and fills the missing data using the straight line trend.

4. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 3 is characterized in that: The fault correlation cross analysis module (200) senses fault data Where fault data X and fault data Y are detection data at the same time node, the correlation coefficient is: in is the mean of X, is the mean of Y; Set the correlation coefficient threshold. If the absolute value of the correlation coefficient is less than or equal to the correlation coefficient threshold, it is determined that there is no correlation between the fault data X and the fault data Y. If the absolute value of the correlation coefficient is greater than the correlation coefficient threshold, it is determined that there is correlation between the fault data X and the fault data Y. The timestamp X corresponding to the fault data X and the fault data Y is retrieved. T 、Y T , if timestamp X T Earlier than timestamp Y T , then the correlation relationship between fault data X and fault data Y is fault data X→fault data Y.

5. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 4 is characterized in that: The fault correlation cross analysis module (200) calculates the correlation coefficient r between fault data X and fault data Y XY Then, calculate the correlation coefficient r between the fault data X and the fault data Z XZ , if | r XY |>correlation coefficient threshold, and |r XZ |>correlation coefficient threshold, the correlation coefficient r between fault data Y and fault data Z is calculated at the first time YZ , if the correlation coefficient r is calculated YZ > correlation coefficient threshold, it is determined that there is correlation among X, Y, and Z, and the correlation relationship of the fault data is determined according to the corresponding timestamps.

6. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 5 is characterized in that: The partial correlation analysis method in the fault relationship precise analysis module (300) is used to calculate the partial correlation coefficient r between fault data Y and fault data Z under the influence of control fault data X. YZ.X , and calculate the partial correlation coefficient r between fault data X and fault data Z under the influence of controlled fault data Y XZ.Y If r YZ.X >r XZ.Y When , it is judged that the fault data X has the greatest impact.

7. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 6 is characterized in that: The calculation formula of the partial correlation analysis method is as follows: Perceive fault data X, Y and Z, under the influence of control fault data X, the partial correlation coefficient r between fault data Y and fault data Z is YZ.X for:

8. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 1 is characterized in that: The fusion filling method in the precise fault data repair module (400) senses multiple fault data that are correlated with the missing fault data, retrieves partial correlation coefficients corresponding to the multiple fault data and the missing fault data, and respectively establishes multiple linear regression models: Y=a0+a1X; wherein Y is the supplementary data, X is the fault data that is correlated with the supplementary data, a1 is the regression coefficient, and a0 is the intercept.

9. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 8, characterized in that: The regression coefficient a1 is calculated by the partial correlation coefficient: where r XY.Z is the partial correlation coefficient between fault data X and fault data Y, σ X and σ Y are the standard deviations of fault data X and fault data Y respectively; intercept in and are the means of fault data X and fault data Y respectively.

10. The carrier module intelligent detection system based on artificial intelligence analysis according to claim 9, characterized in that: After calculating a plurality of supplementary data, the precise fault data repair module (400) calls out the most supplementary data, selects the data with the highest frequency of occurrence as the final supplementary data, and replaces the filling data.

Images