Method, apparatus, and storage medium for boom system operational data interpolation

Abstract

The application discloses a method, device and storage medium for operation data interpolation of an arm support system. The method comprises the following steps: acquiring a plurality of operation data of the arm support system, each operation data comprising a plurality of characteristic dimensions, and the plurality of operation data comprising first operation data and second operation data; acquiring a correlation coefficient between a first characteristic dimension and a second characteristic dimension, the second characteristic dimension being a characteristic dimension other than the first characteristic dimension in the plurality of characteristic dimensions; determining a data distance between the first operation data and the second operation data according to the correlation coefficient; and interpolating a characteristic value of the first characteristic dimension in the first operation data according to the data distance and a dimension value of the first characteristic dimension in the second operation data, to obtain target operation data. The application can improve the accuracy of sample data interpolation and the reliability of subsequent data analysis results.

Description

Technical Field

[0001] This invention relates to the field of engineering machinery technology, and more specifically to a method, apparatus and storage medium for interpolating boom system operation data. Background Technology

[0002] In the construction machinery sector, large and medium-sized machinery is trending towards higher automation and intelligence. Large-scale equipment clusters, numerous sensor points, and high-frequency data acquisition generate massive amounts of equipment monitoring data, making big data analytics a powerful tool for monitoring the condition of machinery. Real-time monitoring and anomaly detection enable machine self-diagnosis and predictive maintenance, preventing significant adverse effects such as personal injury and property damage caused by equipment failure. However, missing values ​​in the sample data can affect the reliability of subsequent data analysis results. To ensure the reliability of these results, current technologies typically impute missing values ​​based on the average of multiple nearest neighbor sample points, rarely considering the influence of factors such as the dimensionality of the missing values, resulting in low imputation accuracy. Summary of the Invention

[0003] The purpose of this invention is to provide a method, apparatus, and storage medium for interpolating operational data of a boom system, in order to solve the problem of low interpolation accuracy in the prior art due to insufficient consideration of factors such as missing value dimensions.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for interpolating operational data of a boom system, the method comprising:

[0005] Acquire multiple operational data of the boom system. Each operational data includes multiple feature dimensions. Among the multiple operational data, there are first operational data and second operational data. The first operational data is the operational data with missing dimension values ​​in the first feature dimension among the multiple feature dimensions, and the second operational data is the operational data with complete dimension values ​​in the multiple feature dimensions.

[0006] Obtain the correlation coefficient between the first feature dimension and the second feature dimension, where the second feature dimension is a feature dimension other than the first feature dimension among multiple feature dimensions;

[0007] The data distance between the first running data and the second running data is determined based on the correlation coefficient.

[0008] Based on the data distance and the dimension value of the first feature dimension in the second running data, the feature value of the first feature dimension in the first running data is interpolated to obtain the target running data.

[0009] In this embodiment of the invention, interpolating the feature values ​​of the first feature dimension in the first running data based on the data distance and the dimension values ​​of the first feature dimension in the second running data includes:

[0010] Based on the data distance, determine K minimum distance samples from the second running data, where K is an integer greater than 1;

[0011] The weight value corresponding to each minimum distance sample is determined based on the data distance corresponding to each minimum distance sample.

[0012] Based on the weight values, the feature values ​​of the first feature dimension in each minimum distance sample are weighted to obtain the weighted interpolation value of the feature values ​​of the first feature dimension in the first running data.

[0013] In this embodiment of the invention, the weighted interpolation value satisfies formula (1):

[0014]

[0015] Where x′ is the weighted interpolation value, ω i y represents the weight value corresponding to the i-th minimum distance sample, which is negatively correlated with the data distance. i Let K be the i-th minimum distance sample, and K be the number of minimum distance samples.

[0016] In this embodiment of the invention, the data distance d(x,y) between the first running data and the second running data satisfies formula (2):

[0017]

[0018] Where, x j y represents the dimension value of the first running data in the j-th second feature dimension. j Let p be the dimension value of the second running data in the j-th second feature dimension, m represents the number of second feature dimensions, and p is the dimension value of the second running data in the j-th second feature dimension. j This represents the correlation coefficient between the first feature dimension and the j-th second feature dimension.

[0019] In this embodiment of the invention, the target operating data includes operating condition data and boom status data, and the method further includes:

[0020] By combining operating data, threshold parameters are predicted using a target deep neural network;

[0021] Using boom status data as the object of anomaly detection, and based on threshold parameters, a clustering algorithm is used to identify abnormal data in the boom status data.

[0022] In this embodiment of the invention, the threshold parameter includes the radius of the neighborhood and the minimum number of points contained;

[0023] Using boom status data as the object of anomaly detection, and based on a threshold parameter, a clustering algorithm is used to identify abnormal data in the boom status data, including:

[0024] Select any sample point from the boom status data as the first sample point, and determine whether the first sample point is the core point;

[0025] With the first sample point as the core point, determine the local region centered on the first sample point and the local data set of sample points within the local region;

[0026] Select any sample point in the local dataset as the second sample point, and determine whether the second sample point is a core point;

[0027] With the second sample point as the core point, a cluster is established based on the second sample point;

[0028] Determine the set of representative points in the neighborhood of the second sample point, and expand the cluster based on the set of representative points;

[0029] After completing the expansion of the cluster, return to the step of selecting any sample point in the boom status data as the first sample point, until all sample points in the boom status data have been visited.

[0030] Sample points in the boom status data that do not belong to any cluster are identified as outliers.

[0031] In this embodiment of the invention, the key point is that the number of samples in the neighborhood is greater than or equal to the minimum number of contained points.

[0032] In this embodiment of the invention, determining the set of representative points within the neighborhood of the second sample point, and expanding the cluster based on the set of representative points includes:

[0033] Select any one representative point from the set of representative points and determine whether the representative point is a core point;

[0034] With the representative point as the core point, determine the neighborhood set of sample points within the neighborhood of the representative point, and classify the neighborhood set into a cluster.

[0035] When the representative point is not a core point, the representative point is classified into a cluster.

[0036] The representative point set is the set of representative points within a second preset range, where the second preset range is the range where the distance from the second sample point is greater than half the radius and less than or equal to the radius.

[0037] In this embodiment of the invention, the representative point includes a distance representative point and a feature representative point. The distance representative point is a sample point whose distance from the second sample point is greater than half the radius and less than or equal to the radius under the Euclidean distance scale. The feature representative point is the sample point with the largest distance from the second sample point under a single feature dimension in the neighborhood.

[0038] A second aspect of the present invention provides an apparatus for interpolating operational data of a boom system, comprising:

[0039] The memory is configured to store instructions; and

[0040] The processor is configured to retrieve instructions from memory and, when executing the instructions, to implement the aforementioned method for data interpolation for boom system operation.

[0041] A third aspect of the present invention provides a machine-readable storage medium storing instructions for causing a machine to perform the above-described method for interpolating data for boom system operation.

[0042] The above technical solution acquires multiple operational data points of the boom system, each including multiple feature dimensions. These multiple operational data points include first operational data and second operational data. The correlation coefficient between the first and second feature dimensions is then obtained, and the data distance between the first and second operational data points is determined based on the correlation coefficient. Based on the data distance and the dimension value of the first feature dimension in the second operational data, the feature values ​​of the first feature dimension in the first operational data are interpolated to obtain the target operational data. This invention determines the data distance between the first and second operational data points based on the correlation coefficient, and interpolates the feature values ​​of the first feature dimension in the first operational data based on the data distance and the dimension value of the first feature dimension in the second operational data. This improves the accuracy of interpolating sample data and helps enhance the reliability of subsequent data analysis results.

[0043] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0044] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0045] Figure 1 A flowchart illustrating a method for interpolating operational data of a boom system according to an embodiment of the present invention is shown schematically.

[0046] Figure 2The flowchart illustrates a method for interpolating operational data of a boom system according to a specific embodiment of the present invention.

[0047] Figure 3 A schematic diagram of a partial region according to an embodiment of the present invention is shown;

[0048] Figure 4 The diagram schematically illustrates a structural block diagram of an apparatus for interpolating operational data of a boom system according to an embodiment of the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustrating and explaining the embodiments of the present invention and are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0050] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0051] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0052] Figure 1 A flowchart illustrating a method for interpolating operational data of a boom system according to an embodiment of the present invention is shown. Figure 1 As shown, this embodiment of the invention provides a method for interpolating operational data of a boom system, which may include the following steps:

[0053] Step 101: Obtain multiple operational data of the boom system. Each operational data includes multiple feature dimensions. Among the multiple operational data, there are first operational data and second operational data. The first operational data is the operational data with missing dimension values ​​in the first feature dimension among the multiple feature dimensions, and the second operational data is the operational data with complete dimension values ​​in the multiple feature dimensions.

[0054] Step 102: Obtain the correlation coefficient between the first feature dimension and the second feature dimension, where the second feature dimension is a feature dimension other than the first feature dimension among multiple feature dimensions;

[0055] Step 103: Determine the data distance between the first running data and the second running data based on the correlation coefficient;

[0056] Step 104: Based on the data distance and the dimension value of the first feature dimension in the second running data, interpolate the feature value of the first feature dimension in the first running data to obtain the target running data.

[0057] The following description will primarily use a processor as the execution entity to illustrate the method for detecting boom system operating data provided in this embodiment of the invention.

[0058] Because existing technologies typically do not consider the impact of other factors on missing values, and only impute missing values ​​by determining the average of multiple nearest neighbor sample points, there is a problem of low accuracy in imputing sample data, leading to low reliability of subsequent data analysis results. Therefore, in this embodiment of the invention, the processor can acquire multiple operating data of the boom system and preprocess the operating data to obtain the target operating data of the boom system. The multiple operating data includes first operating data and second operating data. The first operating data is the operating data with missing dimension values ​​in the first feature dimension among multiple feature dimensions. The second operating data is the operating data with complete dimension values ​​in multiple feature dimensions. The multiple feature dimensions include the first feature dimension and the second feature dimension. The first feature dimension is the feature dimension with missing dimension values. The second feature dimension is the feature dimension other than the first feature dimension among multiple feature dimensions. The target operating data includes, but is not limited to, basic data, boom status data, and operating condition data.

[0059] The processor can obtain the correlation coefficient between the first feature dimension and the second feature dimension. The correlation coefficient can be the Pearson correlation coefficient. Based on the correlation coefficient, the processor can determine the data distance between the first and second running data. The data distance between the first and second running data can be Euclidean distance or Manhattan distance, or other distance metrics can be used; this is not limited here. Having determined the data distance between the first and second running data, the processor can determine K minimum distance samples from the second running data based on the data distance. K is an integer greater than 1 and can be determined according to the actual situation. Based on the data distance corresponding to each minimum distance sample, the processor can determine the weight value corresponding to each minimum distance sample, and weight the feature values ​​of the first feature dimension in each minimum distance sample according to the weight value, thereby obtaining the weighted interpolated value of the feature values ​​of the first feature dimension in the first running data. This weighted interpolated value is then used to fill in the feature values ​​of the first feature dimension in the first running data to complete the interpolation process of the running data, thus obtaining the target running data.

[0060] The above technical solution acquires multiple operational data points of the boom system, each including multiple feature dimensions. These multiple operational data points include first and second operational data. The correlation coefficient between the first and second feature dimensions is then obtained, and the data distance between the first and second operational data points is determined based on this correlation coefficient. Finally, based on the data distance and the dimension value of the first feature dimension in the second operational data, the feature values ​​of the first feature dimension in the first operational data are interpolated to obtain the target operational data. This invention determines the data distance between the first and second operational data points based on the correlation coefficient, and interpolates the feature values ​​of the first feature dimension in the first operational data based on the data distance and the dimension value of the first feature dimension in the second operational data. This improves the accuracy of interpolating sample data and helps enhance the reliability of subsequent data analysis results.

[0061] In this embodiment of the invention, interpolating the feature values ​​of the first feature dimension in the first running data based on the data distance and the dimension value of the first feature dimension in the second running data may include:

[0062] Based on the data distance, determine K minimum distance samples from the second running data, where K is an integer greater than 1;

[0063] The weight value corresponding to each minimum distance sample is determined based on the data distance corresponding to each minimum distance sample.

[0064] Based on the weight values, the feature values ​​of the first feature dimension in each minimum distance sample are weighted to obtain the weighted interpolation value of the feature values ​​of the first feature dimension in the first running data.

[0065] Specifically, given the data distance between the first running data and the second running data, the processor can determine K minimum distance samples in the second running data. A minimum distance sample is a running data point with a smaller data distance to the first running data compared to the remaining running data in the second running data. Based on the data distance corresponding to each minimum distance sample, the processor can obtain the weight value corresponding to each minimum distance sample. It should be noted that the weight value can be determined in various ways and is not limited to a specific method. In one example, the weight value can satisfy formula (3):

[0066]

[0067] Where, ω i Let d be the weight value corresponding to the i-th minimum distance sample, K be the number of minimum distance samples, and d be the weight value. i Let δ be the data distance corresponding to the i-th minimum distance sample, and let δ be a small positive perturbation much less than 1.

[0068] Furthermore, the processor can weight the feature values ​​of the first feature dimension in each minimum distance sample according to the weight values, thereby obtaining the weighted interpolated value of the feature values ​​of the first feature dimension in the first running data. In this way, the weighted interpolated value can be used to interpolate the feature values ​​of the first feature dimension in the first running data to obtain the target running data.

[0069] In this embodiment of the invention, the weighted interpolation value satisfies formula (1):

[0070]

[0071] Where x′ is the weighted interpolation value, ω i y represents the weight value corresponding to the i-th minimum distance sample, which is negatively correlated with the data distance. i Let K be the i-th minimum distance sample, and K be the number of minimum distance samples.

[0072] Specifically, the processor can determine the weight value corresponding to each minimum distance sample based on the data distance corresponding to each minimum distance sample, and further weight the feature values ​​of the first feature dimension in each minimum distance sample based on the weight values, thereby obtaining the weighted interpolation value of the feature values ​​of the first feature dimension in the first running data.

[0073] In this embodiment of the invention, the data distance d(x,y) between the first running data and the second running data satisfies formula (2):

[0074]

[0075] Where, x jy represents the dimension value of the first running data in the j-th second feature dimension. j Let p be the dimension value of the second running data in the j-th second feature dimension, m represents the number of second feature dimensions, and p is the dimension value of the second running data in the j-th second feature dimension. j This represents the correlation coefficient between the first feature dimension and the j-th second feature dimension.

[0076] Specifically, the processor can determine the correlation coefficient between the first feature dimension and the j-th second feature dimension. In this embodiment of the invention, the correlation coefficient can be the Pearson correlation coefficient. Based on the dimension value of the first running data in the j-th second feature dimension, the dimension value of the second running data in the j-th second feature dimension, and the correlation coefficient, the processor can obtain the data distance between the first running data and the second running data. It should be noted that the data distance between the first running data and the second running data is not limited to the above formula; other distance measurement methods can also be used to determine the data distance between the first running data and the second running data, which is not limited here.

[0077] Figure 2 The flowchart illustrating a method for interpolating operational data of a boom system according to a specific embodiment of the present invention is shown schematically. Figure 2 As shown, in a specific embodiment of the present invention, the method for interpolating boom system operating data can satisfy the following steps:

[0078] S201. Obtain historical data;

[0079] S202. Determine the correlation coefficient matrix p, and determine the number K of samples with minimum distance;

[0080] S203, Obtain multiple runtime data;

[0081] S204. Determine the data distance between the first running data and the second running data;

[0082] S205. Select the sample with the smallest distance and add it to the nearest neighbor set L;

[0083] S206. Determine whether the number of samples in the nearest neighbor set L is less than the number of samples with the minimum distance K. If yes, proceed to S205; otherwise, proceed to S207.

[0084] S207. Determine the weighted interpolation value and fill in the feature value of the first feature dimension in the first running data.

[0085] Specifically, the processor can acquire historical data and determine the correlation coefficient matrix p based on the historical data, as well as the number K of minimum distance samples. Simultaneously, the processor can acquire multiple running data sets, including first and second running data sets. Each running data set includes multiple feature dimensions, including a first feature dimension and a second feature dimension. Based on the correlation coefficients of the first and second feature dimensions, the data distance between the first and second running data sets is determined. Further, minimum distance samples are selected from the second running data sets and added to the nearest neighbor set L. Then, it is determined whether the number of samples in the nearest neighbor set L is less than the number K of minimum distance samples. If the number of samples in the nearest neighbor set L is less than K, the process returns to the step of selecting minimum distance samples from the second running data sets and adding them to the nearest neighbor set L. If the number of samples in the nearest neighbor set L is not less than K, the processor can determine a weighted interpolation value based on the data distance and corresponding weight value of the minimum distance sample, and use this weighted interpolation value to fill the feature values ​​of the first feature dimension in the first running data set, thereby obtaining the target running data.

[0086] In this embodiment of the invention, the target operating data includes operating condition data and boom status data, and the method further includes:

[0087] By combining operating data, threshold parameters are predicted using a target deep neural network;

[0088] Using boom status data as the object of anomaly detection, and based on threshold parameters, a clustering algorithm is used to identify abnormal data in the boom status data.

[0089] Specifically, in this embodiment of the invention, the target operating data includes boom status data and working condition data. Combining the working condition data, the processor can predict threshold parameters using a trained target deep neural network, thereby meeting the need to obtain threshold parameters during the complex operation of the engineering machinery boom. The target deep neural network refers to a deep neural network incorporating multiple pre-encoding trainers. The threshold parameters include the radius of the neighborhood and the minimum number of contained points.

[0090] To obtain the target deep neural network, an initial deep neural network can be pre-trained. Since the target pre-encoding trainer can simulate the extraction of hidden features from each first network layer, making the initial weights of the first network layers close to the optimal solution, and can quickly learn the distribution characteristics of the threshold parameters, this embodiment of the invention introduces a pre-encoding trainer during the training of the initial deep neural network. First, the processor can construct the initial deep neural network, which includes a first preset number of first network layers connected sequentially. The first preset number refers to the number of first network layers in the initial deep neural network, which can be determined according to the actual situation. Each first network layer has a corresponding number of nodes. For the first second preset number of first network layers in the initial neural network, the processor can construct corresponding initial pre-encoding trainers. The second preset number is the number of first network layers for which corresponding initial pre-encoding trainers need to be constructed, and its value can be equal to the first preset number minus one. Each initial pre-encoding trainer includes an input layer, hidden layers, and an output layer. The number of nodes in the input layer of the initial pre-encoding trainer is equal to the number of nodes in the corresponding first network layer. The number of nodes in the hidden layer of the previous initial pre-encoding trainer is equal to the number of nodes in the input layer of the next initial pre-encoding trainer. The number of nodes in the output layer of the initial pre-encoding trainer is equal to the number of nodes in the last first network layer of the initial deep neural network. Furthermore, the number of nodes in the hidden layer of the last pre-encoding trainer is set to 1. In one example, the initial deep neural network includes m sequentially connected first network layers. For the first m-1 first network layers, m-1 initial pre-encoding trainers corresponding to the first network layers can be constructed respectively. Furthermore, the number of nodes in the input layer of the i-th pre-encoding trainer is equal to the number of nodes in the i-th first network layer, the number of nodes in the hidden layer of the i-th pre-encoding trainer is equal to the number of nodes in the input layer of the (i+1)-th pre-encoding trainer, and the number of nodes in the hidden layer of the (m-1)-th initial pre-encoding trainer is 1. Here, i is a positive integer less than m.

[0091] After constructing the initial deep neural network and the initial pre-encoding trainer, the processor can train the initial pre-encoding trainer. The processor can train the first initial pre-encoding trainer based on training samples, thereby obtaining the first target pre-encoding trainer and the output values ​​of its hidden layers. During this process, noise needs to be added to the output values ​​of the hidden layers of the previous target pre-encoding trainer. This allows the target pre-encoding trainer to learn not only the original feature information of the data but also the degraded features destroyed by the noise. After adding noise to the output values ​​of the hidden layers of the previous target pre-encoding trainer, the processor can obtain the target hidden layer output value, and then use the target hidden layer output value as the input value of the input layer of the next initial pre-encoding trainer, thereby improving generalization performance. In this way, a second preset number of target pre-encoding trainers can be obtained.

[0092] After training the target precoder, the weights of the input layers of each target precoder are used as the initial weights of the first network layer corresponding to that target precoder. The initial deep neural network is then trained again until it converges to obtain the target deep neural network. During the training of the initial deep neural network, mean squared error can be used as the error function, gradient descent can be used for weight updates, and early stopping can be used to determine whether training is complete. Other methods can also be used, and no limitation is made here. This allows for the extraction of accurate input-output mapping relationships even with a small number of training samples, overcoming the high training difficulty of existing deep neural network technologies.

[0093] After obtaining the threshold parameters through target deep neural network prediction, the processor can use boom status data as anomaly detection data and, based on the threshold parameters, identify abnormal data in the boom status using a clustering algorithm. Unlike existing clustering algorithms, this embodiment of the invention expands the clusters by determining a representative point set and performing neighborhood scanning on the representative points in the representative point set, thereby significantly improving the efficiency of the clustering algorithm.

[0094] Figure 3 A schematic diagram of a partial region according to an embodiment of the present invention is shown. Figure 3 As shown, in this embodiment of the invention, the threshold parameter includes the radius of the neighborhood and the minimum number of points contained;

[0095] Using boom status data as the object of anomaly detection, and based on a threshold parameter, a clustering algorithm is used to identify abnormal data in the boom status data, including:

[0096] Select any sample point from the boom status data as the first sample point, and determine whether the first sample point is the core point;

[0097] With the first sample point as the core point, determine the local region centered on the first sample point and the local data set of sample points within the local region;

[0098] Select any sample point in the local dataset as the second sample point, and determine whether the second sample point is a core point;

[0099] With the second sample point as the core point, a cluster is established based on the second sample point;

[0100] Determine the set of representative points in the neighborhood of the second sample point, and expand the cluster based on the set of representative points;

[0101] After completing the expansion of the cluster, return to the step of selecting any sample point in the boom status data as the first sample point, until all sample points in the boom status data have been visited.

[0102] Sample points in the boom status data that do not belong to any cluster are identified as outliers.

[0103] Specifically, the processor can identify abnormal data in the boom status data through clustering algorithms to complete the detection of the boom system's operational data. Threshold parameters include the radius ε of the neighborhood and the minimum number of contained points (MinPts). Based on the radius ε, the range of the neighborhoods of the first sample point, the second sample point, and the representative point can be determined. A neighborhood refers to a circular area with a radius ε. The minimum number of contained points (MinPts) is the minimum number of samples required within the neighborhood to determine if a sample point is a core point. That is, if the minimum number of contained points (MinPts) equals 3, then if the number of samples in the neighborhood of a sample point is greater than or equal to 3, the sample point can be determined as a core point.

[0104] Therefore, all sample points in the boom status data can be initialized to an unvisited state, and after neighborhood scanning, they can be marked as visited. Neighborhood scanning refers to the process of determining the sample points and the number of samples within the neighborhood of a sample point. The processor can select any unvisited sample point in the boom status data as the first sample point and determine whether the first sample point is a core point. If the first sample point is a core point, a local region centered on the first sample point and a local data set of sample points within that local region are determined. The radius of the local region can be a preset multiple of the radius ε of the neighborhood, i.e., (n+1)ε. The preset multiple can be determined according to the actual situation. The local data set is the set of sample points within the local region.

[0105] Further, the processor selects any sample point in the local dataset as the second sample point and determines whether the second sample point is a core point. It should be noted that the distance between the second sample point and the first sample point should be less than or equal to nε. If the second sample point is a core point, the processor can establish a cluster based on the second sample point and determine a set of representative points within the neighborhood of the second sample point to expand the cluster. The set of representative points is a set of representative points within a preset range, which is a range whose distance from the second sample point is greater than half the radius and less than or equal to the radius. After expanding the cluster, the processor returns to the step of selecting any unvisited sample point in the boom status data as the first sample point, until all sample points in the boom status data have been visited. If all sample points in the boom status data have been visited, the processor can identify sample points in the boom status data that do not belong to any cluster as abnormal data, thus completing the identification of abnormal data in the boom system operation data.

[0106] In this embodiment of the invention, the key point is that the number of samples in the neighborhood is greater than or equal to the minimum number of contained points.

[0107] Specifically, a core point can be a sample point whose number of samples within its neighborhood is greater than or equal to the minimum number of contained points, MinPts. The minimum number of contained points, MinPts, refers to the minimum number of samples that must be contained within the neighborhood to determine if a sample point is a core point. The neighborhood is a circular region with a radius ε. By determining whether the number of samples within the neighborhood of a sample point is greater than or equal to the minimum number of contained points, MinPts, it can be determined whether the sample point is a core point.

[0108] In this embodiment of the invention, determining the set of representative points within the neighborhood of the second sample point, and expanding the cluster based on the set of representative points includes:

[0109] Select any one representative point from the set of representative points and determine whether the representative point is a core point;

[0110] With the representative point as the core point, determine the neighborhood set of sample points within the neighborhood of the representative point, and classify the neighborhood set into a cluster.

[0111] When the representative point is not a core point, the representative point is classified into a cluster.

[0112] In one example, the representative point set is the set of representative points within a second preset range, which is the range where the distance from the second sample point is greater than half the radius and less than or equal to the radius.

[0113] Specifically, to reduce unnecessary neighborhood scanning, in this embodiment of the invention, the cluster can be expanded using a set of representative points that are far from the second sample point. The set of representative points is a set of representative points within a preset range, where the distance from the second sample point is greater than half the radius ε and less than or equal to the radius ε. The processor can select any representative point from the set and determine whether it is a core point. Representative points include distance representative points and feature representative points. If the representative point is a core point, the neighborhood set of sample points within the representative point's neighborhood is determined, and this neighborhood set is classified into a cluster. Furthermore, the processor can incorporate representative points from the neighborhood set into the representative point set to update the representative point set. The neighborhood set refers to the set of sample points within the neighborhood of a representative point. If the representative point is not a core point, only that representative point is classified into a cluster. In this way, the cluster can be expanded based on the representative point set.

[0114] In this embodiment of the invention, the representative point includes at least one of the distance representative point and the feature representative point. The distance representative point is a sample point whose distance from the second sample point is greater than half the radius and less than or equal to the radius under the distance scale. The feature representative point is a sample point whose distance from the second sample point is the largest or smallest under a single feature dimension in the neighborhood.

[0115] Specifically, the representative point set includes a distance representative point set and a feature representative point set. Representative points include distance representative points and feature representative points. Distance representative points can be sample points whose distance to the second sample point is greater than half the radius and less than or equal to the radius, according to the Euclidean distance scale. Distance representative points can satisfy formula (4):

[0116]

[0117] Where n0 is the second sample point, N is the neighborhood set of the second sample point, q is any sample point in the neighborhood set N, Q1 is the distance representative point set, and ε is the radius of the neighborhood.

[0118] The representative feature point can be the sample point that is furthest from the second sample point in a single feature dimension within the neighborhood. The representative feature point satisfies formula (5):

[0119] Q2={q|q∈N,max(q j -n 0j )∪min(q j -n 0j (5)

[0120] Where Q2 is the set of feature representative points, and the subscript j represents the j-th feature dimension, (q j -n 0j Let be the distance between any sample point in the neighborhood set N and the second sample point on the j-th feature dimension.

[0121] The representative point set is the union of the distance representative point set and the feature representative point set. This allows us to determine the representative point set within the neighborhood of the second sample point, which can then be used to expand the cluster.

[0122] Figure 4 A schematic block diagram of a controller according to an embodiment of the present invention is shown. Figure 4 As shown, this embodiment of the invention provides a device for interpolating operational data of a boom system, which may include:

[0123] Memory 410 is configured to store instructions; and

[0124] The processor 420 is configured to retrieve instructions from the memory 410 and, when executing the instructions, to implement the aforementioned method for interpolating data for boom system operation.

[0125] Specifically, in this embodiment of the invention, the processor 420 can be configured to:

[0126] Acquire multiple operational data of the boom system. Each operational data includes multiple feature dimensions. Among the multiple operational data, there are first operational data and second operational data. The first operational data is the operational data with missing dimension values ​​in the first feature dimension among the multiple feature dimensions, and the second operational data is the operational data with complete dimension values ​​in the multiple feature dimensions.

[0127] Obtain the correlation coefficient between the first feature dimension and the second feature dimension, where the second feature dimension is a feature dimension other than the first feature dimension among multiple feature dimensions;

[0128] The data distance between the first running data and the second running data is determined based on the correlation coefficient.

[0129] Based on the data distance and the dimension value of the first feature dimension in the second running data, the feature value of the first feature dimension in the first running data is interpolated to obtain the target running data.

[0130] Furthermore, the processor 420 can also be configured as follows:

[0131] Based on the data distance, determine K minimum distance samples from the second running data, where K is an integer greater than 1;

[0132] The weight value corresponding to each minimum distance sample is determined based on the data distance corresponding to each minimum distance sample.

[0133] Based on the weight values, the feature values ​​of the first feature dimension in each minimum distance sample are weighted to obtain the weighted interpolation value of the feature values ​​of the first feature dimension in the first running data.

[0134] In this embodiment of the invention, the weighted interpolation value satisfies formula (1):

[0135]

[0136] Where x′ is the weighted interpolation value, ω i y represents the weight value corresponding to the i-th minimum distance sample, which is negatively correlated with the data distance. i Let K be the i-th minimum distance sample, and K be the number of minimum distance samples.

[0137] In this embodiment of the invention, the data distance d(x,y) between the first running data and the second running data satisfies formula (2):

[0138]

[0139] Where, x j y represents the dimension value of the first running data in the j-th second feature dimension. j Let p be the dimension value of the second running data in the j-th second feature dimension, m represents the number of second feature dimensions, and p is the dimension value of the second running data in the j-th second feature dimension. j This represents the correlation coefficient between the first feature dimension and the j-th second feature dimension.

[0140] Furthermore, the processor 420 can also be configured as follows:

[0141] By combining operating data, threshold parameters are predicted using a target deep neural network;

[0142] Using boom status data as the object of anomaly detection, and based on threshold parameters, a clustering algorithm is used to identify abnormal data in the boom status data.

[0143] Furthermore, the processor 420 can also be configured as follows:

[0144] Select any sample point from the boom status data as the first sample point, and determine whether the first sample point is the core point;

[0145] With the first sample point as the core point, determine the local region centered on the first sample point and the local data set of sample points within the local region;

[0146] Select any sample point in the local dataset as the second sample point, and determine whether the second sample point is a core point;

[0147] With the second sample point as the core point, a cluster is established based on the second sample point;

[0148] Determine the set of representative points in the neighborhood of the second sample point, and expand the cluster based on the set of representative points;

[0149] After completing the expansion of the cluster, return to the step of selecting any sample point in the boom status data as the first sample point, until all sample points in the boom status data have been visited.

[0150] Sample points in the boom status data that do not belong to any cluster are identified as outliers.

[0151] In this embodiment of the invention, the key point is that the number of samples in the neighborhood is greater than or equal to the minimum number of contained points.

[0152] Furthermore, the processor 420 can also be configured as follows:

[0153] Select any one representative point from the set of representative points and determine whether the representative point is a core point;

[0154] With the representative point as the core point, determine the neighborhood set of sample points within the neighborhood of the representative point, and classify the neighborhood set into a cluster.

[0155] When the representative point is not a core point, the representative point is classified into a cluster.

[0156] The representative point set is the set of representative points within a second preset range, where the second preset range is the range where the distance from the second sample point is greater than half the radius and less than or equal to the radius.

[0157] In this embodiment of the invention, the representative point includes a distance representative point and a feature representative point. The distance representative point is a sample point whose distance from the second sample point is greater than half the radius and less than or equal to the radius under the Euclidean distance scale. The feature representative point is the sample point with the largest distance from the second sample point under a single feature dimension in the neighborhood.

[0158] The above technical solution acquires multiple operational data points of the boom system, each including multiple feature dimensions. These multiple operational data points include first and second operational data. The correlation coefficient between the first and second feature dimensions is then obtained, and the data distance between the first and second operational data points is determined based on this correlation coefficient. Finally, based on the data distance and the dimension value of the first feature dimension in the second operational data, the feature values ​​of the first feature dimension in the first operational data are interpolated to obtain the target operational data. This invention determines the data distance between the first and second operational data points based on the correlation coefficient, and interpolates the feature values ​​of the first feature dimension in the first operational data based on the data distance and the dimension value of the first feature dimension in the second operational data. This improves the accuracy of filling in sample data and the reliability of subsequent data analysis results.

[0159] This invention also provides a machine-readable storage medium storing instructions that cause a machine to perform the above-described method for interpolating data for boom system operation.

[0160] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0161] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0162] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0163] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0164] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0165] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0166] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0167] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0168] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for interpolating operational data of a boom system, characterized in that, The method includes: Multiple operational data of the boom system are acquired. Each operational data includes multiple feature dimensions. The multiple operational data includes first operational data and second operational data. The first operational data is operational data in which dimension values ​​are missing in the first feature dimension among the multiple feature dimensions. The second operational data is operational data in which dimension values ​​are complete in the multiple feature dimensions. Obtain the correlation coefficient between the first feature dimension and the second feature dimension, wherein the second feature dimension is a feature dimension other than the first feature dimension among the plurality of feature dimensions; Based on the correlation coefficient, determine the data distance between the first running data and the second running data; Based on the data distance and the dimension value of the first feature dimension in the second running data, the feature value of the first feature dimension in the first running data is interpolated to obtain the target running data.

2. The method according to claim 1, characterized in that, The step of interpolating the feature values ​​of the first feature dimension in the first running data based on the data distance and the dimension values ​​of the first feature dimension in the second running data includes: Based on the data distance, determine K minimum distance samples from the second running data, where K is an integer greater than 1; Based on the data distance corresponding to each minimum distance sample, determine the weight value corresponding to each minimum distance sample; Based on the weight values, the feature values ​​of the first feature dimension in each of the minimum distance samples are weighted to obtain the weighted interpolation value of the feature values ​​of the first feature dimension in the first running data.

3. The method according to claim 2, characterized in that, The weighted interpolation value satisfies formula (1): Where, x ′ For the weighted interpolation value, ω i The weight value corresponding to the i-th minimum distance sample is negatively correlated with the data distance, y i Let K be the i-th minimum distance sample, and K be the number of minimum distance samples.

4. The method according to claim 1, characterized in that, The data distance d(x,y) between the first running data and the second running data satisfies formula (2): Where, x j y represents the dimension value of the first running data in the j-th second feature dimension. j Let p be the dimension value of the second running data in the j-th second feature dimension, m represent the number of the second feature dimensions, and p be the dimension value of the second running data in the j-th second feature dimension. j This represents the correlation coefficient between the first feature dimension and the j-th second feature dimension.

5. The method according to claim 1, characterized in that, The target operating data includes operating condition data and boom status data, and the method further includes: Based on the aforementioned operating data, threshold parameters are predicted using a target deep neural network; The boom status data is used as anomaly detection object data, and based on the threshold parameter, a clustering algorithm is used to identify abnormal data in the boom status data.

6. The method according to claim 5, characterized in that, The threshold parameters include the radius of the neighborhood and the minimum number of points contained; The step of using the boom status data as anomaly detection object data and identifying abnormal data in the boom status data based on the threshold parameter through a clustering algorithm includes: Select any sample point from the boom status data as the first sample point, and determine whether the first sample point is the core point; With the first sample point as the core point, determine the local region centered on the first sample point and the local data set of sample points within the local region; Select any sample point from the local dataset as the second sample point, and determine whether the second sample point is a core point; With the second sample point as the core point, a cluster is established based on the second sample point; Determine a set of representative points within the neighborhood of the second sample point, and expand the cluster based on the set of representative points; After expanding the cluster, return to the step of selecting any sample point in the boom status data as the first sample point, until all sample points in the boom status data have been visited. Sample points in the boom status data that do not belong to any cluster are identified as abnormal data.

7. The method according to claim 6, characterized in that, The core point is a sample point whose number of samples in the neighborhood is greater than or equal to the minimum number of contained points.

8. The method according to claim 6, characterized in that, Determining the set of representative points within the neighborhood of the second sample point, and expanding the cluster based on the set of representative points, includes: Select any one representative point from the set of representative points and determine whether the representative point is a core point; When the representative point is the core point, determine the neighborhood set of sample points in the neighborhood of the representative point, and classify the neighborhood set into the cluster. If the representative point is not a core point, the representative point is classified into the cluster. The representative point set is a set of representative points within a second preset range, where the second preset range is the range where the distance from the second sample point is greater than half the radius and less than or equal to the radius.

9. The method according to claim 8, characterized in that, The representative points include distance representative points and feature representative points. The distance representative points are sample points whose distance from the second sample point is greater than half the radius and less than or equal to the radius under the Euclidean distance scale. The feature representative points are sample points whose distance from the second sample point is the largest under a single feature dimension in the neighborhood.

10. A device for interpolating operational data of a boom system, characterized in that, include: The memory is configured to store instructions; as well as The processor is configured to retrieve the instructions from the memory and, when executing the instructions, to implement the method for data interpolation for boom system operation according to any one of claims 1 to 9.

11. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform a method for interpolating boom system operating data according to any one of claims 1 to 9.

Images