Railway facility damage assessment method based on multi-source data prediction

Abstract

The present invention relates to the technical field of fault prediction, and specifically to a method for assessing damage to railway facilities based on multi-source data prediction; obtaining speed difference characteristic values ​​according to the difference characteristics of speed sequences at different positions; obtaining vibration difference characteristic values ​​according to the difference characteristics of vibration sequences at different positions; obtaining vibration difference coefficients between any position and other positions according to speed difference characteristic values ​​and vibration difference characteristic values; obtaining vibration anomaly indexes according to distance characteristics and vibration difference coefficients at different positions; and obtaining reference characteristic values ​​according to the difference characteristics of vibration anomaly indexes between any position and neighboring positions. The present invention obtains a damage coefficient according to the difference characteristics of the vibration anomaly index at any position and the reference characteristic value, and the distribution characteristics of all vibration anomaly indices; constructs a damage coefficient sequence according to the damage coefficients of historical dates at any position and predicts the maintenance date, thereby improving the accuracy and timeliness of track damage assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction, and in particular to a railway facility damage assessment method based on multi-source data prediction. Background Art

[0002] With the continuous development of transportation technology, high-speed railways are becoming more and more popular. Under high loads and harsh environments, the structural stability and working performance of railways and other related facilities may be attenuated and damaged, seriously affecting the safety of train travel. Therefore, it is necessary to conduct damage status assessment of railway facilities to improve the safety of train travel. When testing and evaluating the status of railways, manual inspections and installation of sensors to collect relevant data for analysis and testing are generally used. Such methods have limited coverage and high costs. It is difficult to determine the damage location and degree of the railway in a timely and accurate manner, which affects the accuracy and timeliness of railway damage assessment. Summary of the invention

[0003] In order to solve the technical problems that the above-mentioned existing railway detection methods have limited coverage and high costs, which affect the accuracy and timeliness of railway damage assessment, the purpose of the present invention is to provide a railway facility damage assessment method based on multi-source data prediction. The technical solution adopted is as follows:

[0004] Obtain the speed and vibration intensity of the train when it is traveling at any position on the track;

[0005] Constructing a speed sequence according to the speeds of the arbitrary position and the neighboring position on the track; constructing a vibration sequence according to the vibration intensity of the arbitrary position and the neighboring position; obtaining a speed difference characteristic value according to the difference characteristics of the speed sequence of the arbitrary position and the other positions; obtaining a vibration difference characteristic value according to the difference characteristics of the vibration sequence of the arbitrary position and the other positions; obtaining a vibration difference coefficient between the arbitrary position and the other positions according to the difference characteristics of the speed difference characteristic value and the vibration difference characteristic value;

[0006] Obtaining the vibration anomaly index of the arbitrary position according to the distance characteristics between the arbitrary position and any other position and the vibration difference coefficient; obtaining the reference characteristic value of the vibration anomaly index according to the difference characteristics between the vibration anomaly index of the arbitrary position and the neighboring position; obtaining the damage coefficient according to the difference characteristics between the vibration anomaly index of the arbitrary position and the reference characteristic value and the distribution characteristics of all vibration anomaly indices;

[0007] A damage coefficient sequence is constructed according to the damage coefficients of the historical dates of the arbitrary location; and a maintenance date of the arbitrary location is predicted according to the damage coefficient sequence.

[0008] Furthermore, the step of obtaining a speed difference characteristic value according to the difference characteristics of the speed sequences of the arbitrary position and other positions includes:

[0009] The mean square error of the speed sequence between the arbitrary position and the other positions is calculated and normalized to the maximum and minimum values ​​to obtain the speed difference characteristic value between the arbitrary position and the other positions.

[0010] Furthermore, the step of obtaining a vibration difference characteristic value according to the difference characteristics of the vibration sequences between the arbitrary position and the other positions includes:

[0011] The mean square error of the vibration sequence between the arbitrary position and the other positions is calculated and normalized to the maximum and minimum values ​​to obtain the vibration difference characteristic value between the arbitrary position and the other positions.

[0012] Further, the step of obtaining the vibration difference coefficient between the arbitrary position and the other positions according to the difference characteristics between the speed difference characteristic value and the vibration difference characteristic value comprises:

[0013] The difference between the vibration difference characteristic value and the speed difference characteristic value is calculated and positively correlated with each other to obtain the vibration difference coefficient between the arbitrary position and the other positions.

[0014] Furthermore, the step of obtaining the vibration anomaly index of the arbitrary position according to the distance feature between the arbitrary position and any other position and the vibration difference coefficient includes:

[0015] Calculate the inverse of the interval distance between the arbitrary position and any other position to obtain the distance weight; calculate the product of the distance weight and the vibration difference coefficient to obtain the weighted value of the vibration difference between the arbitrary position and any other position; calculate the average value of the weighted value of the vibration difference between the arbitrary position and all other positions to obtain the vibration anomaly index of the arbitrary position.

[0016] Furthermore, the step of obtaining a reference characteristic value of the vibration anomaly index according to the difference characteristics of the vibration anomaly index between the arbitrary position and the neighboring position includes:

[0017] The average absolute value of the difference between the vibration anomaly index of the arbitrary position and all neighboring positions is calculated and negatively correlated to obtain the vibration confidence of the arbitrary position; the vibration anomaly index corresponding to the maximum value of the vibration confidence is used as the reference characteristic value.

[0018] Furthermore, the step of obtaining the damage coefficient according to the difference characteristics between the vibration anomaly index at the arbitrary position and the reference characteristic value and the distribution characteristics of all vibration anomaly indexes includes:

[0019] Calculate the standard deviation of the vibration anomaly index of all positions where the train passes to obtain a comprehensive vibration index; calculate the absolute value of the difference between the vibration anomaly index at any position and the reference characteristic value and perform maximum and minimum value normalization to obtain the damage characteristic value at any position; calculate the product of the damage characteristic value and the comprehensive vibration index to obtain the damage coefficient at any position.

[0020] Furthermore, the step of predicting the maintenance date of the arbitrary position according to the damage coefficient sequence includes:

[0021] The damage coefficient sequence is predicted by a local weighted regression algorithm to obtain a predicted damage coefficient for a future date at any position; and the date when the predicted damage coefficient exceeds a preset threshold is used as the maintenance date.

[0022] The present invention has the following beneficial effects:

[0023] In the present invention, constructing a speed sequence and a vibration sequence can compare the vibration degree of the train passing through different positions of the track; obtaining a speed difference characteristic value can characterize the speed difference of the train at different positions, and obtaining a vibration difference characteristic value can characterize the vibration intensity difference of the train at different positions. Obtaining a vibration difference coefficient can obtain the vibration difference of different positions based on the characteristics that the train speed is similar under normal circumstances and its vibration intensity should be similar, and then the position where the track damage may exist can be determined according to the vibration difference coefficient. Obtaining a vibration anomaly index can avoid the vibration difference coefficient from being interfered by external factors when the train is running, and further improve the accuracy of calculating the track damage degree at any position. Obtaining a reference characteristic value based on the characteristics of similar vibration conditions at adjacent positions of a normal track, and then comparing the vibration anomaly index at any position with the reference characteristic value to improve the accuracy of judging the track damage at the arbitrary position. Obtaining a damage coefficient can accurately characterize the damage degree at different positions in the track; constructing a damage coefficient sequence based on the damage coefficient and predicting the maintenance date can timely and accurately determine the position in the track that needs maintenance, thereby improving the safety and stability of train travel. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A flow chart of a railway facility damage assessment method based on multi-source data prediction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the railway facility damage assessment method based on multi-source data prediction proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0027] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0028] The following is a detailed description of a railway facility damage assessment method based on multi-source data prediction provided by the present invention in conjunction with the accompanying drawings.

[0029] See also Figure 1 , which shows a flow chart of a railway facility damage assessment method based on multi-source data prediction provided by an embodiment of the present invention, the method comprising the following steps:

[0030] Step S1, obtaining the speed and vibration intensity of the train when the train is traveling at any position on the track.

[0031] In the embodiment of the present invention, the implementation scenario is to conduct damage assessment on the railway to improve the accuracy and timeliness of damage assessment. When there are problems such as wear and looseness on the track, it will have a certain impact on the stability of the train during driving. Therefore, the degree of track damage can be analyzed through the vibration characteristics of the train; first, the speed and vibration intensity of the train when the train is traveling at any position on the track are obtained. The speed is obtained through the driving log, and the vibration intensity is characterized by the vertical acceleration collected by the acceleration sensor when the train passes through different positions. The greater the vertical acceleration, the greater the vibration intensity.

[0032] Step S2, constructing a speed sequence according to the speed of any position and its neighboring positions on the track; constructing a vibration sequence according to the vibration intensity of any position and its neighboring positions; obtaining a speed difference characteristic value according to the difference characteristics of the speed sequence between any position and other positions; obtaining a vibration difference characteristic value according to the difference characteristics of the vibration sequence between any position and other positions; obtaining a vibration difference coefficient between any position and other positions according to the difference characteristics of the speed difference characteristic value and the vibration difference characteristic value.

[0033] When the track is damaged such as loosening and wear, it will affect the stability of the train, causing vibration and noise in the body, and the vibration intensity will increase; at the same time, the higher the speed of the train, the more likely it is to vibrate. If the vibration intensity at different positions is different when the train is running at a constant speed, the greater the vibration intensity, the more likely the track damage problem will occur. When the track is damaged at a certain place, it will cause the train to vibrate for a short period of time. Therefore, the vibration difference can be analyzed according to the speed characteristics and vibration intensity characteristics of any position and the neighborhood. First, a speed sequence is constructed according to the speed of any position and the neighborhood position on the track. In the embodiment of the present invention, the neighborhood position is the position within five meters before and after the arbitrary position. The implementer can determine the interval between the analyzed positions on the track and the range of the neighborhood position according to the implementation scenario; the speed sequence reflects the speed of the arbitrary position and the neighborhood position. Further, a vibration sequence is constructed according to the vibration intensity of the arbitrary position and the neighborhood position. After obtaining the speed sequence and vibration sequence of the same length, the vibration difference at different positions can be analyzed; if the speed sequence of the train passing near two positions in the track is more similar, but there is a difference in the vibration sequence, that is, there may be damage at a certain position. Therefore, the speed difference characteristic value is obtained according to the difference characteristics of the speed sequence between any position and other positions, and the vibration difference characteristic value is obtained according to the difference characteristics of the vibration sequence between any position and other positions.

[0034] Preferably, in an embodiment of the present invention, the step of obtaining the speed difference characteristic value and the vibration difference characteristic value includes: calculating the mean square error of the speed sequence between any position and other positions and performing maximum and minimum normalization to obtain the speed difference characteristic value between the arbitrary position and other positions. Calculating the mean square error of the vibration sequence between the arbitrary position and other positions and performing maximum and minimum normalization to obtain the vibration difference characteristic value between the arbitrary position and other positions. It should be noted that the calculation of the mean square error belongs to the prior art, and the specific calculation steps will not be repeated; when the difference between the two speed sequences is smaller, the speed difference characteristic value is smaller; when the difference between the vibration sequences is larger, the vibration difference characteristic value is larger. Maximum and minimum normalization belongs to the prior art, and the specific steps will not be repeated.

[0035] Furthermore, when the speed difference characteristic value between the arbitrary position and other positions is smaller and the vibration difference characteristic value is larger, it means that the vibration conditions of the two positions are more likely to be different, so the vibration difference coefficient between the arbitrary position and other positions is obtained according to the difference characteristics of the speed difference characteristic value and the vibration difference characteristic value; preferably, in the embodiment of the present invention, the step of obtaining the vibration difference coefficient includes: calculating the difference between the vibration difference characteristic value and the speed difference characteristic value and positively mapping them to obtain the vibration difference coefficient between the arbitrary position and other positions. When the vibration difference coefficient is larger, it means that the difference in vibration intensity between the arbitrary position and other positions when the train is traveling at a similar speed is greater, and one of the positions is more likely to be damaged.

[0036] Step S3, obtaining the vibration anomaly index of any position based on the distance characteristics and vibration difference coefficient between any position and any other position; obtaining the reference characteristic value of the vibration anomaly index based on the difference characteristics between the vibration anomaly index of any position and the neighboring positions; obtaining the damage coefficient based on the difference characteristics between the vibration anomaly index of any position and the reference characteristic value and the distribution characteristics of all vibration anomaly indices.

[0037] Due to the different laying conditions of the tracks at different locations and the differences in the train running environment, the vibration difference coefficient between any two locations will also be large. However, the train running environment and laying conditions of trains at similar locations are the same, so the impact of external factors on the vibration conditions at similar locations is relatively small; if the vibration difference coefficient at similar locations is larger, it is more likely to indicate that there is abnormal damage to the track at one of the locations. Therefore, the vibration anomaly index of any location is obtained based on the distance characteristics and vibration difference coefficient between any location and any other location.

[0038] Preferably, in an embodiment of the present invention, the step of obtaining the vibration anomaly index includes: calculating the inverse of the interval distance between the arbitrary position and any other position to obtain the distance weight; the interval distance is the length of the interval between the two positions on the track. The larger the distance weight, the more similar the external factors of the two positions are, and the more the vibration difference coefficient can reflect the degree of track damage; conversely, when the distance weight is smaller, the smaller the confidence of the vibration difference coefficient between the two positions. Calculate the product of the distance weight and the vibration difference coefficient to obtain the weighted value of the vibration difference between the arbitrary position and any other position; when the vibration difference weighted value is larger, it means that the vibration characteristics between the two positions are more different. Calculate the average of the weighted values ​​of the vibration difference between the arbitrary position and all other positions to obtain the vibration anomaly index of the arbitrary position; if the weighted values ​​of the vibration difference between the arbitrary position and all other positions are larger, the larger the vibration anomaly index is, which means that the arbitrary position is more likely to have track damage, and the difference in vibration intensity with other positions is larger. The formula for obtaining the vibration anomaly index includes:

[0039]

[0040] In the formula, represents the vibration anomaly index at the i-th position in the track, M represents the number of other positions, represents the vibration difference coefficient between the ith position and the mth other position, Represents the distance between the i-th position and the m-th other position, represents the distance weight, Indicates the vibration difference weighted value.

[0041] Furthermore, the stability at the normally operating track is relatively stable, so the vibration difference weighted values ​​at the normally operating track are relatively similar, and then the more normal vibration difference weighted values ​​in the track can be determined based on this feature. Therefore, the reference characteristic value of the vibration anomaly index is obtained based on the difference characteristics of the vibration anomaly index between any position and the neighboring position; preferably, in an embodiment of the present invention, the step of obtaining the reference characteristic value includes: calculating the average value of the absolute value of the difference between the vibration anomaly index of the arbitrary position and all neighboring positions and negatively mapping them to obtain the vibration confidence of the arbitrary position; if the vibration anomaly index of the arbitrary position is closer to that of the neighboring position, it means that the section of track is more normal, the train is running more smoothly, and then the vibration confidence at the arbitrary position is greater. The vibration anomaly index corresponding to the maximum value of the vibration confidence is used as the reference characteristic value, and the reference characteristic value represents the value of the vibration anomaly index when the train passes through the normal track. The formula for obtaining the vibration confidence includes:

[0042]

[0043] Where P represents the vibration confidence at any position, N represents the number of neighboring positions, and T represents the vibration anomaly index at the arbitrary position. represents the vibration anomaly index of the nth neighborhood position, Represents an exponential function with a natural constant as its base.

[0044] After obtaining the reference characteristic value of the normal track, if the difference between the vibration anomaly index at a certain position and the reference characteristic value is greater, it means that the track at that position is more likely to be damaged; if the distribution of the vibration anomaly index at different positions in the track is more discrete, it means that there may be multiple damages in the track, and the damage problem is more common, which further indicates that the track at that position is more likely to be damaged. Therefore, the damage coefficient can be obtained based on the difference characteristics between the vibration anomaly index at any position and the reference characteristic value and the distribution characteristics of all vibration anomaly indexes.

[0045] Preferably, in an embodiment of the present invention, the step of obtaining the damage coefficient includes: calculating the standard deviation of the vibration anomaly index of all positions where the train passes to obtain a comprehensive vibration index; the larger the comprehensive vibration index, the more serious the overall damage to the track. Calculate the absolute value of the difference between the vibration anomaly index at the arbitrary position and the reference characteristic value and perform maximum and minimum value normalization to obtain the damage characteristic value at the arbitrary position; when the vibration anomaly index at the arbitrary position is greater than the reference characteristic value, the larger the damage characteristic value, and the more likely the track is to be damaged at the arbitrary position. Calculate the product of the damage characteristic value and the comprehensive vibration index to obtain the damage coefficient at the arbitrary position; when the damage coefficient is larger, it means that the degree of damage at the arbitrary position is more serious.

[0046] Step S4, constructing a damage coefficient sequence according to the damage coefficients of historical dates at any location; and predicting the maintenance date of any location according to the damage coefficient sequence.

[0047] After obtaining the damage coefficient of each position in the track, since the track is used every day under normal circumstances, its wear and looseness becomes more and more serious, and the daily damage coefficient will gradually increase, so it can be predicted based on the daily damage coefficient. The damage coefficient of any position of the track is obtained once a day, and then the damage coefficient sequence of the arbitrary position is constructed based on the damage coefficient of the historical date of the arbitrary position, and the damage coefficient of the future date is predicted by the damage coefficient sequence. Finally, the maintenance date of any position is predicted based on the damage coefficient sequence; preferably, in an embodiment of the present invention, the step of obtaining the maintenance date includes: predicting the damage coefficient sequence through a local weighted regression algorithm to obtain the predicted damage coefficient of the future date of the arbitrary position; and taking the date when the predicted damage coefficient exceeds the preset threshold as the maintenance date. It should be noted that the local weighted regression algorithm belongs to the prior art, and the specific prediction will not be repeated. In order to improve the prediction accuracy, the damage coefficient sequence can be filtered using the existing Gaussian filtering algorithm before the prediction; the implementer can determine it according to the implementation scenario. The preset threshold can be set by the implementer according to the implementation scenario. The preset threshold represents that the damage level at any position has reached the level that requires maintenance and inspection, so that inspection and maintenance can be carried out in advance before the inspection date to improve the stability and safety of train operation. So far, the damage coefficient at each position in the track can improve the accuracy of track damage assessment, which can not only quickly determine the damage location, but also predict the damage level in advance, greatly improving the timeliness and accuracy of track damage assessment.

[0048] In summary, the embodiments of the present invention provide a railway facility damage assessment method based on multi-source data prediction; obtain speed difference characteristic values ​​according to the difference characteristics of speed sequences at different positions; obtain vibration difference characteristic values ​​according to the difference characteristics of vibration sequences at different positions; obtain vibration difference coefficients between any position and other positions according to speed difference characteristic values ​​and vibration difference characteristic values; obtain vibration anomaly indexes according to distance characteristics and vibration difference coefficients at different positions; obtain reference characteristic values ​​according to the difference characteristics of vibration anomaly indexes between any position and neighboring positions. The present invention obtains damage coefficients according to the difference characteristics of the vibration anomaly index at any position and the reference characteristic value and the distribution characteristics of all vibration anomaly indices; constructs a damage coefficient sequence according to the damage coefficients of historical dates at any position and predicts the maintenance date, thereby improving the accuracy and timeliness of track damage assessment.

[0049] It should be noted that the sequence of the above embodiments of the present invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0050] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A railway facility damage assessment method based on multi-source data prediction, characterized in that: The method comprises the following steps: Obtaining the speed and vibration intensity of the train when the train is traveling at any position on the track, wherein the vibration intensity is the vertical acceleration collected by the acceleration sensor when the train passes through any position; Constructing a speed sequence according to the speeds of the arbitrary position and the neighboring position on the track; constructing a vibration sequence according to the vibration intensity of the arbitrary position and the neighboring position; obtaining a speed difference characteristic value according to the difference characteristics of the speed sequence of the arbitrary position and the other positions; obtaining a vibration difference characteristic value according to the difference characteristics of the vibration sequence of the arbitrary position and the other positions; obtaining a vibration difference coefficient between the arbitrary position and the other positions according to the difference characteristics of the speed difference characteristic value and the vibration difference characteristic value; Obtaining the vibration anomaly index of the arbitrary position according to the distance characteristics between the arbitrary position and any other position and the vibration difference coefficient; obtaining the reference characteristic value of the vibration anomaly index according to the difference characteristics between the vibration anomaly index of the arbitrary position and the neighboring position; obtaining the damage coefficient according to the difference characteristics between the vibration anomaly index of the arbitrary position and the reference characteristic value and the distribution characteristics of all vibration anomaly indices; A damage coefficient sequence is constructed according to the damage coefficients of the historical dates of the arbitrary location; and a maintenance date of the arbitrary location is predicted according to the damage coefficient sequence.

2. A railway facility damage assessment method based on multi-source data prediction according to claim 1, characterized in that: The step of obtaining a speed difference characteristic value according to the difference characteristics of the speed sequences of the arbitrary position and other positions comprises: The mean square error of the speed sequence between the arbitrary position and the other positions is calculated and normalized to the maximum and minimum values ​​to obtain the speed difference characteristic value between the arbitrary position and the other positions.

3. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of obtaining a vibration difference characteristic value according to the difference characteristics of the vibration sequence between the arbitrary position and the other positions comprises: The mean square error of the vibration sequence between the arbitrary position and the other positions is calculated and normalized to the maximum and minimum values ​​to obtain the vibration difference characteristic value between the arbitrary position and the other positions.

4. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of obtaining the vibration difference coefficient between the arbitrary position and the other positions according to the difference characteristics between the speed difference characteristic value and the vibration difference characteristic value comprises: The difference between the vibration difference characteristic value and the speed difference characteristic value is calculated and positively correlated with each other to obtain the vibration difference coefficient between the arbitrary position and the other positions.

5. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of obtaining the vibration anomaly index of the arbitrary position according to the distance feature between the arbitrary position and any other position and the vibration difference coefficient comprises: Calculate the inverse of the interval distance between the arbitrary position and any other position to obtain the distance weight; calculate the product of the distance weight and the vibration difference coefficient to obtain the weighted value of the vibration difference between the arbitrary position and any other position; calculate the average value of the weighted value of the vibration difference between the arbitrary position and all other positions to obtain the vibration anomaly index of the arbitrary position.

6. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of obtaining a reference characteristic value of the vibration anomaly index according to the difference characteristics of the vibration anomaly index between the arbitrary position and the neighboring position comprises: The average absolute value of the difference between the vibration anomaly index of the arbitrary position and all neighboring positions is calculated and negatively correlated to obtain the vibration confidence of the arbitrary position; the vibration anomaly index corresponding to the maximum value of the vibration confidence is used as the reference characteristic value.

7. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of obtaining the damage coefficient according to the difference characteristics between the vibration anomaly index at the arbitrary position and the reference characteristic value and the distribution characteristics of all vibration anomaly indexes comprises: Calculate the standard deviation of the vibration anomaly index of all positions where the train passes to obtain a comprehensive vibration index; calculate the absolute value of the difference between the vibration anomaly index at any position and the reference characteristic value and perform maximum and minimum value normalization to obtain the damage characteristic value at any position; calculate the product of the damage characteristic value and the comprehensive vibration index to obtain the damage coefficient at any position.

8. The railway facility damage assessment method based on multi-source data prediction according to claim 1 is characterized in that: The step of predicting the maintenance date of the arbitrary position according to the damage coefficient sequence comprises: The damage coefficient sequence is predicted by a local weighted regression algorithm to obtain a predicted damage coefficient for a future date at any position; and the date when the predicted damage coefficient exceeds a preset threshold is used as the maintenance date.

Images