Data processing system and method for trackless maintenance trolley for bridge

Abstract

The invention discloses a data processing system and method for a trackless maintenance trolley for a bridge, and belongs to the technical field of data processing. According to the dispersion amplitude of output power value dispersion, the quantity of a starting sub-category of a histogram is actively obtained, and the histogram is constructed; then, according to the difference between an output power value formed by increasing or reducing the vehicle speed and a non-low or non-high output power value formed by failure of the driving motor, the outlier probability of each numerical value in each sub-domain of the histogram is obtained, and then the unified dispersivity of each sub-domain of the histogram is obtained; then, splitting inevitability of each sub-domain of the histogram is obtained according to the uniform dispersivity, and the sub-domains are split, so that the accuracy of outlier identification is improved; a small sub-category formed by an output power value formed during the conversion period of the vehicle speed is identified as an outlier, the histograms with the small number of values are merged, the optimal histogram is obtained, outlier identification is executed, and the outlier identification accuracy is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and in particular relates to a data processing system and method for a trackless maintenance vehicle for a bridge. Background Art

[0002] Bridges generally refer to structures built across rivers, lakes, and seas to allow vehicles and pedestrians to pass smoothly. In order to adapt to the modern high-speed development of the transportation industry, bridges are also extended to buildings that are built across mountain streams, poor geology, or to meet other transportation needs to make travel more convenient. Bridges are generally composed of superstructures, substructures, supports, and auxiliary structures. The superstructure, also known as the span structure, is the main structure for crossing obstacles; the substructure includes abutments, piers, and foundations; the supports are force transmission devices installed at the supporting locations of the span structure and piers or abutments.

[0003] On the other hand, trackless maintenance vehicles for bridges are increasingly used. As mentioned in the prior art solution with patent publication number "CN212270675U", the power source of the trackless maintenance vehicle for bridges is a drive motor provided on the running frame, and the drive motor is used to drive the operation of the trackless maintenance vehicle for bridges.

[0004] In addition, when the drive motor fails during operation, the driving work of the drive motor will be damaged. Therefore, by collecting the output power value of the drive motor during operation and combining the collected torque value of the drive motor, the output power value is analyzed in real time and outlier analysis is performed. This can detect defects at the initial stage of drive motor failure, which is conducive to the maintainer to efficiently perform maintenance on the failure condition of the drive motor and improve the driving work of the drive motor.

[0005] At present, outlier analysis of output power values ​​is often performed in combination with a bar chart. When performing outlier analysis on output power values ​​in combination with a bar chart, the numerical range of the output power values ​​must be divided into several sub-ranges to construct a bar chart, and the probability measure of each output power value is evaluated based on the bar chart, and the outlier score of each output power value is obtained based on this to achieve the identification of the outliers. However, because the number of outliers in the collected output power values ​​is unclear, if the number of sub-ranges of the bar chart defined in advance is too high or too low, the evaluation of the probability measure of the output power value is inaccurate, which is not suitable for outlier identification. In addition, the output power value caused by the increase or decrease in the speed of the trackless maintenance vehicle for the bridge and the not low or not high output power value caused by the failure of the drive motor are similar in value, and are often cut into the same sub-range of the bar chart, resulting in the error of distinguishing reasonable values ​​and outliers as the same value. Summary of the invention

[0006] In order to solve the defects in the prior art, the present invention proposes a data processing system and method for a trackless bridge maintenance vehicle, which effectively avoids the defects in the prior art that the quantity of the sub-category of the pre-defined bar graph is too high or too low, resulting in inaccurate evaluation of the probability measurement of the output power value, unsuitable for outlier identification, and erroneous identification of reasonable values ​​and outliers as the same value.

[0007] The present invention utilizes the following technical solutions.

[0008] A data processing method for a trackless bridge maintenance vehicle, comprising:

[0009] The torque sensor collects the torque value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The power sensor collects the output power value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The PLC confirms whether the drive motor is in failure or not based on the collected output power value and torque value of the drive motor.

[0010] The method for confirming whether the drive motor is in a failure state includes:

[0011] Step 1, obtaining the output power value and the torque value of the driving motor transmitted by sampling, and forming an output power value queue and a torque value queue;

[0012] Step 2, according to the number of output power values ​​and the dispersion amplitude of the output power values, actively obtain the amount of the initial sub-category of the histogram, so as to allocate the output power values ​​to the corresponding sub-category of the histogram, and obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability;

[0013] Step 3, based on the distribution uniformity of each subcategory of the bar chart and the comparison range after each subcategory is split, the necessity of splitting each subcategory of the bar chart is obtained, and each subcategory is split to obtain a refreshed bar chart; based on the similarity between each subcategory of the refreshed bar chart and the next subcategory and the numerical difference between each subcategory and its similar subcategory, the mergeability of each subcategory of the refreshed bar chart is obtained, and merged to obtain an optimal bar chart; based on the optimal bar chart, the outlier score of each output power value is obtained, thereby obtaining the outlier value.

[0014] Furthermore, in step 1, the output power value and the torque value are collected synchronously and transmitted to the PLC respectively. The PLC arranges the output power value at each collection time point in the order of the collection time points to form an output power value queue, and arranges the torque value at each collection time point in the order of the collection time points to form a torque value queue.

[0015] Furthermore, step 2 specifically includes:

[0016] Get the amount of the starting sub-category of the histogram. The equation is:

[0017]

[0018] In the equation, The amount representing the starting sub-category of the histogram; Indicates the number of output power values ​​in the output power value queue; Indicates the dispersion of the output power values ​​in the output power value queue.

[0019] Furthermore, in step 2, the bar graph is constructed by taking the numerical range of the output power value as the X-coordinate and the number of the output power values ​​as the Y-coordinate in the rectangular coordinate system, and dividing the numerical range of the output power value into each sub-range according to the amount of the starting sub-range of the bar graph, so as to configure each output power value in the output power value queue into the corresponding sub-range of the bar graph to obtain the bar graph.

[0020] Furthermore, step 2 specifically includes:

[0021] The equation for obtaining the outlier probability of each value in each sub-category of the histogram is:

[0022]

[0023] In the equation, Indicates the first The first The outlier probability of a value; Indicates the first The first The value of a numeric value; Indicates the first The first The average of the neighboring values ​​of a value; Indicates the first The first The highest value among the neighboring values ​​of a value; Indicates the first The first The lowest value among the neighboring values ​​of a value; Indicates the first The first The number of peaks and valleys in the neighborhood of a value; Indicates the first The first The correlation coefficient between the neighboring values ​​of a value and the torque value at the corresponding acquisition time point, is to use the Z-score method to Perform normalization.

[0024] Further, in step 2, each value in each sub-category of the bar graph is placed before the value in the output power value queue. The value and the The values ​​are treated as neighboring values ​​of each value in each sub-category of the histogram.

[0025] Furthermore, step 2 specifically includes:

[0026] The equation for obtaining the uniformity of the distribution of each subcategory of the histogram is:

[0027]

[0028] In the equation, Indicates the first the uniformity of the distribution of subcategories; Indicates the first The first The outlier probability of a value; Indicates the first The number of values ​​within a subcategory; Indicates the first The mean outlier probability of all values ​​within a subcategory; Indicates the first The total range of outlier probabilities for all values ​​within a subcategory; is the Euler number.

[0029] Furthermore, step 3 specifically includes:

[0030] The first The outlier probability of each value in the sub-category is arranged from low to high, and the first The outlier queue of the subcategory is Each value of the outlier queue in each subcategory is used as a cutting point to obtain the first The cutting point of each sub-category, facing any cutting point, use the cutting point to change the first The outlier queue of a subcategory is cut into a pair of subqueues, and the means of all values ​​of the pair of subqueues are defined as mean 1 and mean 2 respectively, and the quantity obtained by subtracting mean 2 from mean 1 is defined as , This is the first The comparison range of each subcategory at this cutting point.

[0031] Furthermore, step 3 specifically includes:

[0032] The equation for obtaining the necessity of splitting each subcategory of the bar chart is:

[0033]

[0034] In the equation, Indicates the first The inevitability of the splitting of subcategories; Indicates the first the uniformity of the distribution of subcategories; Indicates the first The highest amount of comparison amplitude of each subcategory under the overall cut-off point; Indicates the first The number of cut-off points for each subcategory; Indicates the first The subcategory is The control amplitude under the cutting point; Indicates the first The subcategory is The control amplitude under the cutting point, is the Euler number, is to use the Z-score method to Perform normalization.

[0035] Furthermore, step 3 specifically includes:

[0036] Define the critical value of the necessity of splitting in advance , if the bar graph The necessity of splitting up the sub-category is higher than the critical value of the necessity of splitting up , get the first The output power value corresponding to the cut-off point corresponding to the maximum amount of the comparison amplitude of the sub-category under the overall cut-off point is defined as The split value of the sub-category is divided into The sub-category is cut into a pair of sub-category, and the updated histogram is obtained accordingly.

[0037] Furthermore, step 3 specifically includes:

[0038] Get the average of the outlier probability of all values ​​in each sub-category of the refresh bar chart, define it as the outlier amplitude of each sub-category of the refresh bar chart, and get the first The quantity obtained by subtracting the outlier magnitude of each subcategory from the outlier magnitude of each other subcategory is defined as ,Bundle Lower than The other sub-category of An approximate subcategory of a subcategory.

[0039] Furthermore, step 3 specifically includes:

[0040] Get the merged properties of each subcategory of the refreshed histogram:

[0041]

[0042] In the equation, Indicates refreshing the bar graph. The integrability of subcategories; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. sub-category and the similarity of the subcategories; Indicates refreshing the bar graph. The number of values ​​within a subcategory; Indicates refreshing the bar graph. The average number of values ​​in all approximate subcategories of the subcategories; is the Euler number, is to use the Z-score method to Perform normalization.

[0043] Furthermore, step 3 specifically includes:

[0044] Define the merge necessity threshold in advance , if the mergeability of the random subcategory of the refreshed histogram is higher than the merge necessity threshold When , the subcategory of the updated histogram is merged with the next subcategory of the subcategory, thereby obtaining the optimal histogram.

[0045] Furthermore, step 3 specifically includes:

[0046] Using the bar chart, the probability measure of each output power value is obtained according to the optimal bar chart. The probability measure of the output power value is the quotient obtained by dividing the outlier probability of the output power value by the span of the subcategory to which the output power value belongs on the optimal bar chart. The quotient obtained by dividing the probability measure of each output power value by 1 is used as the outlier score of each output power value, and the outlier score critical value is defined in advance. , if the outlier score of the output power value is higher than the outlier score threshold When the output power value is outlier, the drive motor is considered to be in failure state, and the PLC sends a message of drive motor failure to the display screen; if the outlier score of the output power value is higher than the outlier score critical value When , the output power values ​​are all reasonable values.

[0047] A data processing system for a trackless bridge maintenance vehicle, comprising:

[0048] Display screen, power sensor, torque sensor, PLC and driving motor of trackless maintenance trolley for bridges. Display screen, power sensor and torque sensor are all connected to PLC. Torque sensor is used to collect torque value of driving motor of trackless maintenance trolley for bridges and transmit it to PLC. Power sensor is used to collect output power value of driving motor of trackless maintenance trolley for bridges and transmit it to PLC. PLC is used to confirm whether driving motor is in failure state according to output power value and torque value of driving motor collected and transmitted.

[0049] The modules within the PLC include:

[0050] A queue module, which is used to obtain the output power value and the torque value of the driving motor transmitted by sampling, and form an output power value queue and a torque value queue;

[0051] A dispersion module is used to actively obtain the amount of the initial sub-category of the histogram according to the number of output power values ​​and the dispersion amplitude of the output power values, so as to allocate the output power values ​​to the corresponding sub-category of the histogram to obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability;

[0052] The outlier module is used to obtain the necessity of splitting each subcategory of the bar chart according to the distribution uniformity of each subcategory of the bar chart and the comparison range after splitting each subcategory, and perform splitting on each subcategory to obtain a refreshed bar chart; obtain the mergeability of each subcategory of the refreshed bar chart according to the similarity between each subcategory of the refreshed bar chart and the next subcategory and the numerical difference between each subcategory and its similar subcategory, and perform merging on them to obtain an optimal bar chart; obtain the outlier score of each output power value according to the optimal bar chart, thereby obtaining the outlier value.

[0053] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0054] Initially, based on the dispersion of the power value distribution, the quantity of the initial sub-category of the histogram is automatically obtained, and the histogram is constructed. The dispersion of the power value distribution is added so that the quantity of the initial sub-category of the histogram can be adapted to the power value distribution, thereby improving the accuracy of outlier identification. Then, based on the difference between the power value caused by the increase or decrease of the vehicle speed and the not low or not high power value caused by the failure of the drive motor, the outlier probability of each value in each sub-category of the histogram is obtained, and then the outlier probability of each sub-category of the histogram is obtained. The distribution is unified, and then the sub-category is split according to the necessity of splitting each sub-category of the bar graph obtained based on the unified distribution. The output power value caused by increasing or decreasing the vehicle speed and the high or low output power value caused by the drive motor failure can be distinguished, thereby improving the accuracy of outlier identification. In order to identify the small sub-category formed by the output power value generated during the change of vehicle speed as an outlier, the bar graphs with low number of values ​​are merged to obtain the optimal bar graph for outlier identification, thereby improving the accuracy of outlier identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a flow chart of the data processing method for the trackless bridge maintenance vehicle described in the present invention;

[0056] Figure 2 It is a partial structural diagram of the data processing system of the trackless bridge maintenance vehicle described in the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention clearer, the following will be combined with the drawings in the embodiments of the present invention to clearly and completely express the technical solution of the present invention. The embodiments expressed in this application are only partial embodiments of the present invention, not all embodiments. According to the spirit of the present invention, other embodiments obtained by technicians in this field without making creative work are all within the protection scope of the present invention.

[0058] like Figure 1 As shown, the data processing method for a trackless bridge maintenance vehicle according to the present invention comprises:

[0059] The torque sensor collects the torque value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The power sensor collects the output power value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The PLC confirms whether the drive motor is in failure or not based on the collected output power value and torque value of the drive motor.

[0060] The method for confirming the failure status of the drive motor, running on the PLC, includes:

[0061] Step 1, obtaining the output power value and the torque value of the driving motor transmitted by sampling, and forming an output power value queue and a torque value queue;

[0062] In a preferred but non-limiting embodiment of the present invention, in step 1, during the operation of the drive motor, the acquisition frequency of the torque sensor and the power sensor is pre-defined as 1 time / second, and the output power value and torque value of the pre-defined time length (the time length can be 30 seconds or 60 seconds) are synchronously acquired according to the acquisition frequency (synchronous acquisition means that the torque sensor and the power sensor acquire the same acquisition time points each time) and transmitted to the PLC, and the PLC arranges the output power values ​​of each acquisition time point according to the order of the acquisition time points to form an output power value queue, and arranges the torque values ​​of each acquisition time point according to the order of the acquisition time points to form a torque value queue. During the transmission of the output power value and the torque value to the PLC, the acquisition time points of the output power value and the torque value are also transmitted to the PLC.

[0063] Step 2, according to the number of output power values ​​and the dispersion amplitude of the output power values, actively obtain the amount of the initial sub-category of the histogram, so as to allocate the output power values ​​to the corresponding sub-category of the histogram, and obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability;

[0064] In a preferred but non-limiting embodiment of the present invention, step 2 specifically comprises:

[0065] Step 2-1, according to the number of output power values ​​and the dispersion amplitude of the output power values, actively obtain the quantity of the initial sub-range of the histogram, thereby allocating the output power values ​​to the corresponding sub-range of the histogram, and obtaining the histogram;

[0066] When using a bar graph to perform outlier identification on the output power value, the numerical range of the output power value must first be divided into various sub-ranges to construct a bar graph, and the probability measurement of each output power value is evaluated based on the bar graph, and the outlier score of each output power value is obtained based on this to achieve the identification of the outlier. However, because the collected output power values ​​contain outliers, it is inaccurate to manually define the amount of the sub-range of the bar graph directly. If the amount of the sub-range is too high, it often causes the subtle characteristics of the output power value distribution to be lost; if the amount of the sub-range is too low, it often causes the amount of the numerical points in each sub-range to be too low, resulting in inaccurate evaluation of the probability measurement of the output power value. Therefore, the present invention first actively obtains the amount of the initial sub-range of the bar graph based on the number of output power values, and then analyzes the dispersion amplitude of the output power value. If the dispersion amplitude of the output power value is larger, it means that the overall output power value is more volatile, and the amount of the initial sub-range of the bar graph must be increased a little based on the dispersion amplitude of the output power value distribution, thereby matching the fluctuation of the numerical value.

[0067] Get the amount of the starting sub-category of the histogram. The equation is:

[0068]

[0069] In the equation, The amount representing the starting sub-category of the histogram; Indicates the number of output power values ​​in the output power value queue; and the amount of the sub-range of the output power value is proportional to the number of the output power values, so firstly the amount of the starting sub-range of the bar graph is obtained according to the number of the output power values; It indicates the dispersion amplitude of the output power values ​​in the output power value queue (the dispersion amplitude can be the variance of the output power values ​​in the output power value queue); the more dispersed the output power values ​​in the output power value queue are (the larger the variance is), the more fluctuating the overall output power values ​​are. In this case, the amount of the starting sub-category of the histogram should be increased according to the dispersion amplitude of the output power value distribution; if the dispersion amplitude of the output power values ​​is smaller (the smaller the variance is), the overall output power values ​​are denser. In this case, the amount of the starting sub-category of the histogram should be increased according to the dispersion amplitude of the output power value distribution.

[0070] In a preferred but non-restrictive embodiment of the present invention, in step 2, the bar graph is constructed by taking the numerical range of the output power value as the X coordinate and the number of the output power values ​​as the Y coordinate in the rectangular coordinate system, and the numerical range of the output power value is evenly divided into each sub-range according to the amount of the starting sub-range of the bar graph, so as to configure each output power value in the output power value queue to the corresponding sub-range of the bar graph, and obtain the bar graph. The numerical range of the output power value is the minimum and maximum amount of the output power value in the output power value queue, which are respectively regarded as the minimum and maximum amount of the numerical range of the output power value. Here, the number of sub-ranges is the amount of the starting sub-range, and configuring each output power value in the output power value queue to the corresponding sub-range of the bar graph is to send each output power value in the output power value queue to the sub-range to which its value belongs. Each column of the bar graph corresponds to a sub-category, the width of each column (that is, the X-coordinate range) is the span of the corresponding sub-category (that is, the distance between the lowest and highest quantities in the sub-category), and the height of each column (that is, the Y-coordinate value) is the number of output power values ​​configured in the corresponding sub-category.

[0071] In a preferred but non-limiting embodiment of the present invention, step 2 specifically further comprises:

[0072] Step 2-2, obtain the outlier probability of each value in each subcategory of the histogram, and obtain the distribution uniformity of each subcategory of the histogram based on the outlier probability.

[0073] When the drive motor works reasonably (reasonable work means effective work, that is, the drive motor works without failure), its collected output power value should be stable accordingly. However, if the speed of the trackless maintenance vehicle driven by the drive motor is not stable or the drive motor fails, the output power value will show a large disordered fluctuation and the output power at this time will be significantly greater than or less than the output power value when working reasonably. The output power of the drive motor is closely related to the speed of the trackless maintenance vehicle. When the speed of the trackless maintenance vehicle driven by the drive motor increases, the drive motor must have a higher torque value to maintain the increase in speed. Then, according to the relationship between the torque value and the output power of the drive motor (directly proportional), it can be seen that when the speed of the trackless maintenance vehicle driven by the drive motor increases and does not exceed the limit, the output power value will gradually increase and then become stable. When the speed of the trackless maintenance vehicle driven by the drive motor gradually decreases, the output power value will gradually decrease and then become stable.

[0074] Therefore, since the output power value caused by the increase or decrease of the vehicle speed and the medium or medium output power value caused by the failure of the drive motor are similar in value, they are often cut into the same sub-range of the histogram. The use of the histogram is to evaluate the probability measurement of each output power value based on the histogram to perform outlier identification, so reasonable values ​​and outliers are often identified as the same value. Therefore, the present invention first analyzes the output power values ​​in each sub-range of the histogram to obtain the distribution uniformity of each sub-range of the histogram, and then splits the sub-range with two values ​​in the histogram according to the distribution uniformity.

[0075] In a preferred but non-limiting embodiment of the present invention, in step 2, each value in each sub-category of the bar graph (the value is the output power value) is placed before it (before it is before the value) in the output power value queue. The value and the value after it (after it is after this value) The values ​​are treated as neighboring values ​​of each value in each sub-category of the histogram. The value can be defined according to specific requirements.

[0076] The equation for obtaining the outlier probability of each value in each sub-category of the histogram is:

[0077]

[0078] In the equation, Indicates the first The first The outlier probability of a value; Indicates the first The first The value of a numeric value; Indicates the first The first The average of the neighboring values ​​of a value; Indicates the first The first The highest value among the neighboring values ​​of a value; Indicates the first The first The lowest value among the neighboring values ​​of a value; Indicates the first The first The number of peaks and valleys in the neighborhood of a value; Indicates the first The first The correlation coefficient between the neighboring values ​​of a value and the torque value at the corresponding acquisition time point, is to use the Z-score method to Perform normalization. The peak-valley value in the neighboring values ​​is the value in the neighboring values. If the two adjacent values ​​on both sides are larger or smaller than it, the value is the peak-valley value. The first The correlation coefficient between the neighboring values ​​of a value and the torque value at the corresponding acquisition time point is: The first The torque values ​​corresponding to each acquisition time point of the neighboring values ​​of the value form a group of values The first The correlation coefficient of another set of values ​​formed by the neighboring values ​​of a value.

[0079] In a preferred but non-limiting embodiment of the present invention, step 2 specifically further comprises:

[0080] The equation for obtaining the uniformity of the distribution of each subcategory of the histogram is:

[0081]

[0082] In the equation, Indicates the first the uniformity of the distribution of subcategories; Indicates the first The first The outlier probability of a value; Indicates the first The number of values ​​within a subcategory; Indicates the first The mean outlier probability of all values ​​within a subcategory; Indicates the first The total range of outlier probabilities for all values ​​within a subcategory; is the Euler number; The larger the value, the higher the There are usually two kinds of values ​​in each sub-range (reasonable value and outlier value, the reasonable value is the output power value when the drive motor works reasonably, and the outlier value is the output power value under the fault working condition). The lower the uniformity of the distribution of the subcategories; The larger the value, the There are often two kinds of values ​​(reasonable values ​​and outliers) in a sub-range. The lower the uniformity of the distribution of the subcategories.

[0083] Step 3, based on the distribution uniformity of each subcategory of the bar chart and the comparison range after each subcategory is split, the necessity of splitting each subcategory of the bar chart is obtained, and each subcategory is split to obtain a refreshed bar chart; based on the similarity between each subcategory of the refreshed bar chart and the next subcategory and the numerical difference between each subcategory and its similar subcategory, the mergeability of each subcategory of the refreshed bar chart is obtained, and merged to obtain an optimal bar chart; based on the optimal bar chart, the outlier score of each output power value is obtained, thereby obtaining the outlier value.

[0084] In a preferred but non-limiting embodiment of the present invention, step 3 specifically comprises:

[0085] Step 3-1, according to the distribution uniformity of each sub-category of the bar chart and the comparison range after each sub-category is split, the necessity of splitting each sub-category of the bar chart is obtained, and each sub-category is split to obtain a refreshed bar chart; according to the similarity between each sub-category of the refreshed bar chart and the next sub-category and the numerical difference between each sub-category and its similar sub-category, the merging of each sub-category of the refreshed bar chart is obtained, and it is merged to obtain the optimal bar chart;

[0086] The distribution uniformity of each subcategory of the bar chart is obtained. The lower the value, the more likely it is that the subcategory should be split. Then, each subcategory is divided into a pair of sub-queues based on the outlier probability of the values ​​in each subcategory. The value obtained by subtracting the mean of the outlier probability of the values ​​in a pair of sub-queues is used to determine whether the subcategory should be split.

[0087] The first The outlier probability of each value in the sub-category is arranged from low to high, and the first The outlier queue of the subcategory is Each value of the outlier queue in each subcategory is used as a cutting point to obtain the first The cutting point of each sub-category, facing any cutting point, use the cutting point to change the first The outlier queue of the subcategory is cut into a pair of subqueues (the pair of subqueues is the one that takes the cutting point as the demarcation point to separate the first The outlier queues in the subcategory are divided into a pair of subqueues, where one subqueue is defined as subqueue 1, which contains the values ​​of the demarcation point and the outlier queues to the left of the demarcation point, and the other subqueue is defined as subqueue 2, which contains the values ​​of the remaining outlier queues outside subqueue 1. If subqueue 1 or subqueue 2 is empty, its mean is defined as 0). The means of all values ​​of a pair of subqueues are defined as mean 1 and mean 2, respectively, and the quantity obtained by subtracting mean 2 from mean 1 is defined as , This is the first The comparison range of each subcategory at this cutting point.

[0088] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0089] The equation for obtaining the necessity of splitting each subcategory of the bar chart is:

[0090]

[0091] In the equation, Indicates the first The inevitability of the splitting of subcategories; Indicates the first the uniformity of the distribution of subcategories; Indicates the first The highest amount of comparison amplitude of each subcategory under the overall cut-off point; Indicates the first The number of cut-off points for each subcategory; Indicates the first The subcategory is The control amplitude under the cutting point; Indicates the first The subcategory is The control amplitude under the cutting point, is the Euler number, is to use the Z-score method to Perform normalization; the histogram The greater the separability of a subcategory, the more likely it is that the subcategory will contain both reasonable values ​​and outliers, so the subcategory should be split.

[0092] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0093] Define the critical value of the necessity of splitting in advance , if the bar graph The necessity of splitting up the sub-category is higher than the critical value of the necessity of splitting up , get the first The output power value corresponding to the cut-off point corresponding to the maximum amount of the comparison amplitude of the sub-category under the overall cut-off point is defined as The split value of the sub-category is divided into subcategory into a pair of subcategory (split the first A subcategory is cut into a pair of subcategories by taking the split value as the dividing point. The sub-category is divided into two parts, the left and right parts are the pair of sub-categories respectively), and the refresh bar chart is obtained accordingly. Each column of the refresh bar chart still corresponds to a sub-category respectively. The width of each column (that is, the X-coordinate range) is the span of the corresponding sub-category (that is, the distance between the minimum and maximum quantities of the sub-category), and the height of each column (that is, the value of the Y-coordinate) is the number of output power values ​​configured in the corresponding sub-category. The pre-defined split necessity critical value It can be seventy percent, and of course, other split necessity critical values ​​can be defined in advance according to specific requirements.

[0094] While the driving motor is in operation, the speed of the trackless maintenance vehicle often changes, so the collected power values ​​are often different. Therefore, the power values ​​formed during the speed change will form some small sub-ranges, and the number of power values ​​in each sub-range is small, which is not suitable for outlier identification of power values. Therefore, it is necessary to merge the sub-ranges of such bar charts to more clearly compare the power distribution under different speed conditions. After the above sub-range splitting, some small sub-ranges of reasonable values ​​will also be formed. In order to accurately identify outliers, they must also be merged.

[0095] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0096] Get the average of the outlier probability of all values ​​in each sub-category of the refresh bar chart, define it as the outlier amplitude of each sub-category of the refresh bar chart, and get the first The quantity obtained by subtracting the outlier magnitude of each subcategory from the outlier magnitude of each other subcategory is defined as , is the reduction of the other subcategory. Lower than Another subcategory of sub-category) of the refresh bar graph outside the sub-category is defined as the first Predefine the reduction threshold The value can be 20%, of course, you can also define another value in advance according to specific requirements The value of .

[0097] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0098] Get the merged properties of each subcategory of the refreshed histogram:

[0099]

[0100] In the equation, Indicates refreshing the bar graph. The integrability of subcategories; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. sub-category and the similarity of the subcategories; Indicates refreshing the bar graph. The number of values ​​within a subcategory; Indicates refreshing the bar graph. The average number of values ​​in all approximate subcategories of subcategories, Indicates The numerical difference between a subcategory and its approximate subcategory; is the Euler number, is to use the Z-score method to Perform normalization; if you refresh the histogram The lower the number of values ​​in a subcategory and the greater the difference between the average number of values ​​in its similar subcategory, the more likely the subcategory should be merged. Subcategories and their The more similar the subcategories are, the more they should be merged. The larger the value, the more the bar chart is refreshed. The more subcategories there are, the more they should be merged.

[0101] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0102] Define the merge necessity threshold in advance , if the mergeability of the random subcategory of the refreshed histogram is higher than the merge necessity threshold When the sub-category of the refresh bar graph is merged with the next sub-category of the sub-category (merging means combining the sub-category and the next sub-category of the sub-category into one sub-category), the optimal bar graph is obtained accordingly. Each column of the refresh bar graph still corresponds to a sub-category, and the width of each column (that is, the X-coordinate range) is the span of the corresponding sub-category (that is, the distance between the minimum and maximum quantities of the sub-category), and the height of each column (that is, the value of the Y-coordinate) is the number of output power values ​​configured in the corresponding sub-category. Define the merge necessity threshold in advance The value can be 70%, of course, other merging necessity thresholds can be defined in advance according to specific requirements The value of .

[0103] Step 3-2, according to the optimal histogram, obtain the outlier score of each output power value, thereby obtaining the outlier value.

[0104] In a preferred but non-limiting embodiment of the present invention, step 3 specifically further comprises:

[0105] Using a bar chart, the probability measure of each output power value is obtained according to the optimal bar chart. The probability measure of the output power value is the quotient obtained by dividing the outlier probability of the output power value by the span of the subcategory to which the output power value belongs on the optimal bar chart (that is, the distance between the minimum and maximum quantities of the subcategory). The quotient obtained by dividing the probability measure of each output power value by 1 is used as the outlier score of each output power value, and the outlier score critical value is defined in advance. , if the outlier score of the output power value is higher than the outlier score threshold When the output power value is outlier, the drive motor is considered to be in failure state, and the PLC sends a message of drive motor failure to the display screen to prompt the maintainer to repair the drive motor; if the outlier score of the output power value is higher than the outlier score critical value , the output power values ​​are all reasonable values, and the drive motor is then deemed to have no failure condition.

[0106] Pre-defined outlier score cutoffs The value can be 65%, of course, other outlier score critical values ​​can be defined in advance according to specific requirements The value of .

[0107] like Figure 2 As shown, the data processing system for a trackless bridge maintenance vehicle according to the present invention comprises:

[0108] Display screen, power sensor, torque sensor, PLC and driving motor of trackless maintenance trolley for bridges, the driving motor is arranged on the running upper frame of the trackless maintenance trolley for bridges, the display screen, power sensor and torque sensor are all connected with PLC, the torque sensor is used to collect the torque value of the driving motor of the trackless maintenance trolley for bridges and transmit it to PLC, the power sensor is used to collect the output power value of the driving motor of the trackless maintenance trolley for bridges and transmit it to PLC, the PLC is used to confirm whether the driving motor is in failure state according to the collected output power value and torque value of the driving motor;

[0109] The modules within the PLC include:

[0110] A queue module, which is used to obtain the output power value and the torque value of the driving motor transmitted by sampling, and form an output power value queue and a torque value queue;

[0111] A dispersion module is used to actively obtain the amount of the initial sub-category of the histogram according to the number of output power values ​​and the dispersion amplitude of the output power values, so as to allocate the output power values ​​to the corresponding sub-category of the histogram to obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability;

[0112] The outlier module is used to obtain the inevitability of the splitting of each subcategory of the histogram according to the uniformity of the distribution of each subcategory of the histogram and the comparison range after the splitting of each subcategory, and to split each subcategory to obtain the updated histogram; according to the similarity between each subcategory of the updated histogram and the next subcategory and the numerical difference between each subcategory and its similar subcategory, the mergeability of each subcategory of the updated histogram is obtained, and the merge is performed to obtain the optimal histogram; according to the optimal histogram, the outlier score of each output power value is obtained, thereby obtaining the outlier value. The trackless maintenance trolley for bridges can be a trackless rotating pier bridge inspection trolley.

[0113] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0114] Initially, based on the dispersion of the power value distribution, the quantity of the initial sub-category of the histogram is automatically obtained, and the histogram is constructed. The dispersion of the power value distribution is added so that the quantity of the initial sub-category of the histogram can be adapted to the power value distribution, thereby improving the accuracy of outlier identification. Then, based on the difference between the power value caused by the increase or decrease of the vehicle speed and the not low or not high power value caused by the failure of the drive motor, the outlier probability of each value in each sub-category of the histogram is obtained, and then the outlier probability of each sub-category of the histogram is obtained. The distribution is unified, and then the sub-category is split according to the necessity of splitting each sub-category of the bar graph obtained based on the unified distribution. The output power value caused by increasing or decreasing the vehicle speed and the high or low output power value caused by the drive motor failure can be distinguished, thereby improving the accuracy of outlier identification. In order to identify the small sub-category formed by the output power value generated during the change of vehicle speed as an outlier, the bar graphs with low number of values ​​are merged to obtain the optimal bar graph for outlier identification, thereby improving the accuracy of outlier identification.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not deviate from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A data processing method for a trackless bridge maintenance vehicle, characterized in that: include: The torque sensor collects the torque value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The power sensor collects the output power value of the drive motor of the trackless bridge maintenance vehicle and transmits it to the PLC. The PLC confirms whether the drive motor is in failure or not based on the collected output power value and torque value of the drive motor. The method for confirming whether the drive motor is in a failure state includes: Step 1, obtaining the output power value and the torque value of the driving motor transmitted by sampling, and forming an output power value queue and a torque value queue; Step 2, according to the number of output power values ​​and the dispersion amplitude of the output power values, actively obtain the amount of the initial sub-category of the histogram, so as to allocate the output power values ​​to the corresponding sub-category of the histogram, and obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability; Step 3, based on the distribution uniformity of each subcategory of the bar chart and the comparison range after each subcategory is split, the necessity of splitting each subcategory of the bar chart is obtained, and each subcategory is split to obtain a refreshed bar chart; based on the similarity between each subcategory of the refreshed bar chart and the next subcategory and the numerical difference between each subcategory and its similar subcategory, the mergeability of each subcategory of the refreshed bar chart is obtained, and merged to obtain an optimal bar chart; based on the optimal bar chart, the outlier score of each output power value is obtained, thereby obtaining the outlier value.

2. The data processing method for a trackless bridge maintenance vehicle according to claim 1 is characterized in that: In step 1, the output power value and torque value are collected synchronously and transmitted to the PLC. The PLC arranges the output power value at each collection point in the order of the collection time points to form an output power value queue, and arranges the torque value at each collection point in the order of the collection time points to form a torque value queue.

3. The data processing method for the trackless bridge maintenance vehicle according to claim 2 is characterized in that: Step 2 specifically includes: Get the amount of the starting sub-category of the histogram. The equation is: ； In the equation, The amount representing the starting sub-category of the histogram; Indicates the number of output power values ​​in the output power value queue; Indicates the dispersion of the output power values ​​in the output power value queue.

4. The data processing method for the trackless bridge maintenance vehicle according to claim 3 is characterized in that: In step 2, the bar graph is constructed by taking the numerical range of the output power value as the X-coordinate and the number of the output power values ​​as the Y-coordinate in the rectangular coordinate system, and dividing the numerical range of the output power value into each sub-range according to the amount of the starting sub-range of the bar graph, so as to configure each output power value in the output power value queue into the corresponding sub-range of the bar graph to obtain the bar graph.

5. The data processing method for the trackless bridge maintenance vehicle according to claim 4 is characterized in that: Step 2 specifically also includes: The equation for obtaining the outlier probability of each value in each sub-category of the histogram is: ； In the equation, Indicates the first The first The outlier probability of a value; Indicates the first The first The value of a numeric value; Indicates the first The first The average of the neighboring values ​​of a value; Indicates the first The first The highest value among the neighboring values ​​of a value; Indicates the first The first The lowest value among the neighboring values ​​of a value; Indicates the first The first The number of peaks and valleys in the neighborhood of a value; Indicates the first The first The correlation coefficient between the neighboring values ​​of a value and the torque value at the corresponding acquisition time point, is to use the Z-score method to Perform normalization; In step 2, each value in each sub-category of the bar graph is placed before the output power value queue. The value and the The values ​​are treated as neighboring values ​​of each value in each sub-category of the histogram.

6. The data processing method for the trackless bridge maintenance vehicle according to claim 5 is characterized in that: Step 2 specifically also includes: The equation for obtaining the uniformity of the distribution of each subcategory of the histogram is: ； In the equation, Indicates the first the uniformity of the distribution of subcategories; Indicates the first The first The outlier probability of a value; Indicates the first The number of values ​​within a subcategory; Indicates the first The mean outlier probability of all values ​​within a subcategory; Indicates the first The total range of outlier probabilities for all values ​​within a subcategory; is the Euler number.

7. The data processing method for the trackless bridge maintenance vehicle according to claim 6 is characterized in that: Step 3 specifically includes: The first The outlier probability of each value in the sub-category is arranged from low to high, and the first The outlier queue of the subcategory is Each value of the outlier queue in each subcategory is used as a cutting point to obtain the first The cutting point of each sub-category, facing any cutting point, use the cutting point to change the first The outlier queue of a subcategory is cut into a pair of subqueues, and the means of all values ​​of the pair of subqueues are defined as mean 1 and mean 2 respectively, and the quantity obtained by subtracting mean 2 from mean 1 is defined as , This is the first The comparison range of each subcategory at the cut-off point; Step 3 specifically also includes: The equation for obtaining the necessity of splitting each subcategory of the bar chart is: ； In the equation, Indicates the first The inevitability of the splitting of subcategories; Indicates the first the uniformity of the distribution of subcategories; Indicates the first The highest amount of comparison amplitude of each subcategory under the overall cut-off point; Indicates the first The number of cut-off points for each subcategory; Indicates the first The subcategory is The control amplitude under the cutting point; Indicates the first The subcategory is The control amplitude under the cutting point, is the Euler number, is to use the Z-score method to Perform normalization.

8. The data processing method for the trackless bridge maintenance vehicle according to claim 7 is characterized in that: Step 3 specifically also includes: Define the critical value of the necessity of splitting in advance , if the bar graph The necessity of splitting up the sub-category is higher than the critical value of the necessity of splitting up When , get the first The output power value corresponding to the cut-off point corresponding to the maximum amount of the comparison amplitude of the sub-category under the overall cut-off point is defined as The split value of the sub-category is divided into The sub-category is cut into a pair of sub-category, and the refresh bar chart is obtained accordingly; Step 3 specifically also includes: Get the average of the outlier probability of all values ​​in each sub-category of the refresh bar chart, define it as the outlier amplitude of each sub-category of the refresh bar chart, and get the first The quantity obtained by subtracting the outlier magnitude of each subcategory from the outlier magnitude of each other subcategory is defined as ,Bundle Lower than The other sub-category of a subcategory that is close to a subcategory; Step 3 specifically also includes: Get the merged properties of each subcategory of the refreshed histogram: ； In the equation, Indicates refreshing the bar graph. The integrability of subcategories; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. The magnitude of outliers in each subcategory; Indicates refreshing the bar graph. sub-category and the similarity of the subcategories; Indicates refreshing the bar graph. The number of values ​​within a subcategory; Indicates refreshing the bar graph. The average number of values ​​in all approximate subcategories of the subcategories; is the Euler number, is to use the Z-score method to Perform normalization.

9. The data processing method for the trackless bridge maintenance vehicle according to claim 8 is characterized in that: Step 3 specifically also includes: Define the merge necessity threshold in advance If the mergeability of the random subcategory of the refreshed histogram is higher than the merge necessity threshold , merge the subcategory of the updated histogram with the next subcategory of the subcategory, and obtain the optimal histogram accordingly; Step 3 specifically also includes: Using the bar chart, the probability measure of each output power value is obtained according to the optimal bar chart. The probability measure of the output power value is the quotient obtained by dividing the outlier probability of the output power value by the span of the subcategory to which the output power value belongs on the optimal bar chart. The quotient obtained by dividing the probability measure of each output power value by 1 is used as the outlier score of each output power value, and the outlier score critical value is defined in advance. , if the outlier score of the output power value is higher than the outlier score threshold When the output power value is outlier, the drive motor is considered to be in failure state, and the PLC sends a message of drive motor failure to the display screen; if the outlier score of the output power value is higher than the outlier score critical value When , the output power values ​​are all reasonable values.

10. A data processing system for a trackless bridge maintenance vehicle, characterized in that: include: Display screen, power sensor, torque sensor, PLC and driving motor of trackless maintenance trolley for bridges. Display screen, power sensor and torque sensor are all connected to PLC. Torque sensor is used to collect torque value of driving motor of trackless maintenance trolley for bridges and transmit it to PLC. Power sensor is used to collect output power value of driving motor of trackless maintenance trolley for bridges and transmit it to PLC. PLC is used to confirm whether driving motor is in failure state according to output power value and torque value of driving motor collected and transmitted. The modules within the PLC include: A queue module, which is used to obtain the output power value and the torque value of the driving motor transmitted by sampling, and form an output power value queue and a torque value queue; A dispersion module is used to actively obtain the amount of the initial sub-category of the histogram according to the number of output power values ​​and the dispersion amplitude of the output power values, so as to allocate the output power values ​​to the corresponding sub-category of the histogram to obtain the histogram; obtain the outlier probability of each value in each sub-category of the histogram, and obtain the distribution uniformity of each sub-category of the histogram according to the outlier probability; The outlier module is used to obtain the necessity of splitting each subcategory of the bar chart according to the distribution uniformity of each subcategory of the bar chart and the comparison range after splitting each subcategory, and perform splitting on each subcategory to obtain a refreshed bar chart; obtain the mergeability of each subcategory of the refreshed bar chart according to the similarity between each subcategory of the refreshed bar chart and the next subcategory and the numerical difference between each subcategory and its similar subcategory, and perform merging on them to obtain an optimal bar chart; obtain the outlier score of each output power value according to the optimal bar chart, thereby obtaining the outlier value.

Images