A method and device for monitoring the state of a construction machine and a construction machine

Abstract

The application discloses an engineering machinery state monitoring method and device and engineering machinery, and can improve the accuracy and efficiency of engineering machinery state monitoring. The engineering machinery state monitoring method comprises the following steps: acquiring sensor data of multiple dimensions collected by multiple sensors; wherein one sensor corresponds to sensor data of one dimension; determining the data correlation degree between the multiple sensors and the data correlation degree of sensor data of the same dimension in different lag times according to the sensor data of multiple dimensions; weighting the sensor data in the same dimension based on the data correlation degree of the sensor data of the same dimension in different lag times to obtain a weighted feature; and constructing a state monitoring feature according to the data correlation degree between the multiple sensors and the weighted feature; wherein the change trend of the state monitoring feature reflects the running state of the engineering machinery.

Description

Technical Field

[0001] This invention relates to the field of engineering machinery technology, specifically to a method, device, and engineering machinery for monitoring the condition of engineering machinery. Background Technology

[0002] In the field of construction machinery, with the rapid development of technology, various large, complex, and highly integrated mechanical equipment are widely used in infrastructure construction, mining, transportation, and other industries. The performance stability and operational efficiency of construction machinery are directly related to the progress, quality, and safety of projects. However, traditional methods for monitoring the condition of construction machinery often rely on manual inspections, periodic maintenance, and experience-based judgment, resulting in low efficiency. Current methods for monitoring the condition of construction machinery can be based on models or statistical reliability prediction techniques; however, both require significant time investment in initial training and statistical analysis, and subsequent iterative optimization is difficult, leading to continued low efficiency. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention is proposed. Embodiments of this invention provide a method, apparatus, and type of engineering machinery for condition monitoring, which can improve the accuracy and efficiency of engineering machinery condition monitoring.

[0004] According to a first aspect of this application, a method for monitoring the condition of construction machinery is provided, comprising: acquiring sensor data of multiple dimensions collected by multiple sensors; wherein, one sensor corresponds to sensor data of one dimension; determining the degree of data correlation between the multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times based on the sensor data of multiple dimensions; weighting the sensor data of the same dimension based on the degree of data correlation of the sensor data of the same dimension at different lag times to obtain weighted features; and constructing condition monitoring features based on the degree of data correlation between the multiple sensors and the weighted features; wherein, the changing trend of the condition monitoring features reflects the operating status of the construction machinery.

[0005] As one possible implementation, determining the degree of data correlation among the multiple sensors based on the multi-dimensional sensor data includes: determining a correlation matrix of sensor data among the multiple sensors based on the multi-dimensional sensor data and grey relational analysis; wherein the correlation matrix characterizes the visual correlation degree of multiple sensor data; and clustering the multiple sensors based on the correlation matrix; wherein each cluster includes a main sensor and at least one auxiliary sensor, and the auxiliary sensor is a sensor whose correlation degree with the main sensor reaches a preset value.

[0006] As one possible implementation, the step of weighting sensor data in the same dimension based on the data correlation of sensor data in different lag times to obtain weighted features includes: determining the autocorrelation coefficient based on the data correlation of sensor data in the same dimension in different lag times; determining the maximum lag order of the autocorrelation coefficient of each sensor based on a preset confidence interval; wherein the maximum lag order represents the time period with the largest time interval in the autocorrelation coefficient; and determining the weighted feature of the maximum lag order based on the sensor data in the same dimension and the corresponding maximum lag order.

[0007] As one possible implementation, constructing state monitoring features based on the data correlation between the multiple sensors and the weighted features includes: determining the main sensor parameters and auxiliary sensor parameters in each cluster based on the main sensor and at least one auxiliary sensor in each cluster; constructing a multi-dimensional sensor sequence for each cluster based on the main sensor parameters and the auxiliary sensor parameters; constructing a feature sequence based on the multi-dimensional sensor sequence of each cluster and the weighted features of the maximum hysteresis order corresponding to the main sensor parameters; training a first self-organizing map network with the total number of the main sensor parameters and the auxiliary sensor parameters as the dimension of the feature sequence; using the maximum value in the hidden matrix of each layer of the first self-organizing map network after training as the winning node value of the corresponding feature sequence; and weightedly fusing the feature sequence and the winning node value corresponding to each time node to obtain the state monitoring features.

[0008] As one possible implementation, determining the weighted feature of the maximum lag order based on sensor data of the same dimension and the corresponding maximum lag order includes: training a second self-organizing map network based on sensor data of the same dimension; inputting the sensor data of the same dimension into the trained second self-organizing map network to obtain network weights at different time points; the network weights represent weight vectors; determining the maximum weight value in the network weights as the winning node value at the corresponding time point; obtaining the winning node values ​​at all time points in a preset time series; using the maximum lag order as the time step and the winning node value in the time step as the weight, performing weighted calculation on the sensor data within the time step to obtain the weighted feature of the maximum lag order.

[0009] As one possible implementation, training a second self-organizing map network based on sensor data of the same dimension includes: extracting slope features from the sensor data in the same dimension to construct a two-dimensional time series; inputting the two-dimensional time series into the second self-organizing map network to train the network weights of the sensors; the second self-organizing map network clustering different development trends of the two-dimensional time series according to the network weights; wherein the development trends include gradual increase, gradual decrease, sharp increase, and sharp decrease; evaluating the clustering effect based on an internal metric for evaluating the clustering effect; and determining that the training of the second self-organizing map network is complete when the evaluation score of the clustering effect is greater than or equal to a preset score.

[0010] As one possible implementation, after constructing condition monitoring features based on the data correlation between the multiple sensors and the weighted features, the engineering machinery condition monitoring method further includes: converting the condition monitoring features into a visual graph; wherein the visual graph includes a line graph; when a raised portion deviating from a preset fluctuation range appears in the line graph of the condition monitoring features, it is determined that a fault occurred during the time period corresponding to the raised portion deviating from the preset fluctuation range.

[0011] As one possible implementation, after acquiring sensor data from multiple sensors across multiple dimensions, the engineering machinery condition monitoring method further includes: performing frequency histogram analysis on the sensor data of the target dimension to statistically analyze the frequency of sensor data occurrence in different value ranges; wherein, the target dimension is any dimension among the multiple dimensions; determining the upper and lower control limits of the value range distribution of sensor data in the target dimension based on the frequency of sensor data occurrence in different value ranges; normalizing the sensor data in the target dimension based on the upper and lower control limits of the value range distribution of sensor data in the target dimension to obtain normalized target dimension data; and determining the data correlation between the multiple sensors and the data correlation of sensor data of the same dimension at different lag times based on the sensor data of the multiple dimensions, including: determining the data correlation between the multiple sensors and the data correlation of sensor data of the same dimension at different lag times based on the normalized sensor data of the multiple dimensions.

[0012] According to a second aspect of this application, a condition monitoring device for construction machinery is provided, comprising: an acquisition module for acquiring sensor data of multiple dimensions collected by multiple sensors; wherein, one sensor corresponds to sensor data of one dimension; a determination module for determining, based on the sensor data of multiple dimensions, the degree of data correlation between the multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times; a weighting module for weighting the sensor data of the same dimension based on the degree of data correlation of the sensor data of the same dimension at different lag times to obtain weighted features; and a construction module for constructing condition monitoring features based on the degree of data correlation between the multiple sensors and the weighted features; wherein, the changing trend of the condition monitoring features reflects the operating status of the construction machinery.

[0013] According to a third aspect of this application, an engineering machinery is provided, comprising: a plurality of sensors for collecting sensor data of multiple dimensions in the engineering machinery; wherein, one sensor collects sensor data of one dimension; and a monitoring device communicatively connected to the plurality of sensors, the monitoring device being used to perform the engineering machinery condition monitoring method described in the first aspect or any implementation thereof.

[0014] The present invention provides a method, device, and machinery for monitoring the condition of construction machinery. This method can process multi-dimensional sensor data collected from complex system equipment. Based on the correlations between multi-dimensional sensors and between sensor data, condition monitoring features are constructed. These features can intuitively reflect the condition of the construction machinery, improving data representativeness. By comparing the changes in the condition of the construction machinery at different points in time, the development trend of the condition monitoring features over time can be obtained. The fluctuations in this trend can then be used to determine the operating status of the construction machinery, effectively improving the efficiency and accuracy of condition monitoring. Attached Figure Description

[0015] The above and other objects, features, and advantages of the present invention will become more apparent from the more detailed description of the embodiments of the invention in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same parts or steps.

[0016] Figure 1 This is a flowchart illustrating an exemplary embodiment of the engineering machinery condition monitoring method provided by the present invention.

[0017] Figure 2 This is a schematic diagram of the process for obtaining the weighted feature with the maximum lag order provided by an exemplary embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram illustrating the visualization of construction status monitoring features provided by an exemplary embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of the structure of an engineering machinery condition monitoring device provided in an exemplary embodiment of the present invention.

[0020] Figure 5 This is a structural diagram of an electronic device provided in an exemplary embodiment of the present invention. Detailed Implementation

[0021] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0022] The embodiments of the present invention can be applied to construction machinery, which includes: multiple sensors for collecting sensor data of multiple dimensions in the construction machinery; wherein, one sensor collects sensor data of one dimension; and a monitoring device that is communicatively connected to the multiple sensors, the monitoring device being used to execute the construction machinery condition monitoring method provided by the present invention.

[0023] In some embodiments, the construction machinery may be excavators, cranes, and transport vehicles, etc.

[0024] In some embodiments, the sensor may involve a sensor that collects engine speed, hydraulic oil temperature, coolant temperature, and pump pressure, wherein the collected engine speed can be regarded as sensor data in one dimension, and the hydraulic oil temperature can be regarded as sensor data in another dimension. The monitoring device acquires the sensor data collected by all sensors to construct condition monitoring features.

[0025] As an example, among the multiple sensors, there could be an engine speed sensor (such as a crankshaft position sensor) to detect engine speed. There could also be a hydraulic oil temperature sensor (such as a thermistor or thermocouple sensor) to detect the temperature of the hydraulic fluid in the hydraulic system. Finally, there could be a coolant temperature sensor (such as a thermistor or integrated temperature sensor) to detect engine coolant temperature. And finally, there could be a pump pressure sensor (such as a capacitive, resistive, vibration, or hydrostatic sensor) to detect the pump's operating pressure.

[0026] It should be understood that the above examples are only illustrative examples of sensors that collect engine speed, hydraulic oil temperature, coolant temperature and pump pressure. However, the embodiments of this application are not limited to this. Other data similar to engine speed, hydraulic oil temperature, coolant temperature and pump pressure can also be collected, such as fuel pressure, intake air temperature, exhaust air temperature, hydraulic oil pressure, vibration and noise levels, etc.

[0027] After collecting sensor data from multiple sensors in multiple dimensions, see [link to relevant documentation]. Figure 1 This invention can determine the degree of data correlation between sensors of different dimensions, and the degree of data correlation between sensor data of the same dimension at different lag times, based on sensor data of multiple dimensions (see [reference]). Figure 1 S100 to S200 (in the model), based on the correlation between sensor data of the same dimension at different lag times, weights are applied to the sensor data of the same dimension to obtain weighted features (see S100 to S200). Figure 1 Based on the correlation between data from multiple sensors and weighted features, state monitoring features are constructed (see S300 in the text). Figure 1 (S400 in the text). The changing trend of the condition monitoring characteristics is related to time. By observing the changing trend of the condition monitoring characteristics, it is possible to quickly determine whether there are any abnormal or prominent changes in the construction machinery. If the changing trend of the condition monitoring characteristics is stable and without large fluctuations, it can be said that the condition of the construction machinery is stable and without abnormalities. If there are abnormal or prominent changes in the changing trend of the condition monitoring characteristics, it can be said that a fault has occurred in the construction machinery during the corresponding time period.

[0028] The following text combines Figure 1 This application provides a more detailed description of the engineering machinery condition monitoring method provided in the embodiments.

[0029] In S100, sensor data from multiple sensors across various dimensions is acquired. Each sensor corresponds to one dimension of sensor data. The sensors collect data from various dimensions of the construction machinery; for example, the sensors could be engine speed sensors, hydraulic oil temperature sensors, coolant temperature sensors, and pump pressure sensors, and the corresponding collected sensor data could be engine speed, hydraulic oil temperature, coolant temperature, and pump pressure. Besides the sensors and sensor data exemplified above, other sensors and sensor data related to construction machinery can also be used.

[0030] In some implementations, one sensor corresponds to one dimension of sensor data. This can be understood as follows: when the sensor is an engine speed sensor, the collected sensor data is engine speed, which is one dimension of sensor data; when the sensor is a pump pressure sensor, the collected sensor data is pump pressure, which is another dimension of sensor data. In other words, data collected by one type of sensor represents one dimension, and data collected by another type of sensor represents another dimension; multiple sensors can collect sensor data across multiple dimensions.

[0031] Different types of sensors collect different sensor data; however, these data are complementary in reflecting the state of construction machinery, providing a more comprehensive description of its current condition. For example, multiple sensors may collect the same or similar data (such as temperature sensors at different locations), and these data can be cross-verified, improving data reliability and accuracy. Furthermore, state transitions in construction machinery trigger multiple sensors to collect data simultaneously. The sensor data collected by different sensors can be aligned along a timeline, thus establishing correlations between multi-dimensional sensor data. By performing correlation analysis on sensor data collected by multiple sensors, the inherent patterns and connections in the state of construction machinery can be revealed, thereby determining its condition. For example, analyzing the correlation between data such as engine speed, oil pressure, and oil temperature can determine whether the engine is operating normally.

[0032] As described in section S100, raw sensor data can be obtained. However, different sensors typically have their own unique dimensions and ranges. For example, temperature sensor data may be in degrees Celsius, ranging from -50°C to 100°C; while pressure sensor data may be in Pascals, ranging from 0 to 100,000 Pa. Due to these differences in dimensions and ranges, directly comparing data from different sensors may be meaningless or misleading.

[0033] To enable the fusion and comparison of all sensor data, in some embodiments, sensor data from multiple sensors across multiple dimensions is normalized. Specifically, frequency histogram analysis is performed on the sensor data in the target dimension to statistically analyze the frequency of sensor data occurrences in different value ranges; where the target dimension can be any dimension among multiple dimensions; based on the frequency of sensor data occurrences in different value ranges, upper and lower control limits for the value range distribution of sensor data in the target dimension are determined; based on the upper and lower control limits for the value range distribution of sensor data in the target dimension, the sensor data in the target dimension is normalized to obtain normalized target dimension data. Correspondingly, S200 can be adjusted to: based on the normalized sensor data across multiple dimensions, determine the data correlation between multiple sensors and the data correlation of sensor data in the same dimension at different lag times.

[0034] Normalization involves scaling data according to a certain ratio so that it falls within a specific interval, usually [0, 1], but it can also be other intervals, such as [-1, 1] or custom upper and lower bounds. The target dimension can represent the dimension collected by any one of the multiple sensors. In this invention, it is necessary to normalize the data from all sensors involved in state monitoring, but the specific normalization process is performed individually for each sensor.

[0035] As a normalization implementation method, for a single sensor, the steps can be as follows: 1. Perform frequency histogram analysis independently to statistically analyze the frequency of sensor data (sensor data collected by a single sensor) in different value ranges; 2. Use kernel density estimation to estimate the probability density of each sensor data with the kernel as the weight, and fit the distribution curve; 3. Use trapezoidal numerical integration to calculate the confidence interval of the kernel density curve (the confidence interval is obtained by subtracting twice the standard deviation from the sample mean), thereby determining the upper and lower control limits of the current sensor data value range distribution; 4. Normalize based on the upper and lower control limits of the sensor data. Each sensor's collected sensor data undergoes the normalization process described in steps 1-4 above, thus normalizing the sensor data from all sensors. Normalization can eliminate the influence of different sensor data due to differences in dimensions and value ranges, helping to accelerate the convergence speed of optimization algorithms such as gradient descent, reducing the imbalance of importance between different features, and improving the efficiency and accuracy of subsequent data utilization.

[0036] Kernel density estimation (KDE) is a nonparametric statistical method in probability theory used to estimate unknown density functions. KDE does not rely on prior knowledge of the data distribution and does not impose any assumptions on the data distribution. Instead, it starts directly from the data sample itself to study the characteristics of the data distribution, thus having strong applicability.

[0037] Continue to refer to Figure 1 In S200, based on sensor data from multiple dimensions, the correlation between multiple sensors and the correlation between sensor data of the same dimension at different lag times are determined.

[0038] In some embodiments, different types of sensors can collect sensor data from various aspects of the construction machinery, such as temperature, pressure, vibration, and rotational speed. This sensor data is complementary in reflecting the state of the construction machinery, providing a more comprehensive description of its operation. Therefore, there are often correlations between the sensor data collected. For example, changes in engine speed may affect the temperature and pressure of hydraulic oil, which in turn affects the performance and stability of the construction machinery. Combining sensor data from multiple dimensions in different ways can be used to determine different states of the construction machinery. Therefore, determining the degree of correlation between multiple sensors and selecting sensors with higher data correlation when monitoring the state of the construction machinery can improve the accuracy of monitoring and more accurately determine its operating status and potential faults.

[0039] As an example, when sensors include engine speed sensors, throttle position sensors, and torque sensors, combining data from these sensors can assess whether the power output of the construction machinery is stable and meets current operational requirements. When sensors include gearbox vibration sensors, oil temperature sensors, and oil pressure sensors, data from these sensors can determine the operating status of the construction machinery's transmission system, such as whether gear meshing is good and whether the fluid level is sufficient. Vibration and sound sensors can monitor the vibration and noise levels of mechanical components, thereby determining their wear level. For example, bearing wear leads to increased vibration and noise. Data from pressure and displacement sensors can determine whether the load conditions of the construction machinery are appropriate.

[0040] One possible implementation of S200 is as follows: Based on sensor data from multiple dimensions and grey relational analysis, determine the correlation matrix of sensor data among multiple sensors; wherein the correlation matrix represents the degree of visual correlation of multiple sensor data; based on the correlation matrix, cluster the multiple sensors; wherein each cluster includes a main sensor and at least one auxiliary sensor, and the auxiliary sensor is a sensor whose correlation with the main sensor reaches a preset value.

[0041] Grey Relational Analysis (GRA) is a multi-factor statistical analysis method. It can be understood as understanding the relative strength of the influence of other factors on a specific indicator within a grey system. Alternatively, in a given situation, assuming a certain indicator is likely related to several other factors, the goal is to determine the relative strength of this correlation. This process is repeated, ranking the factors to obtain an analytical result that reveals which factors are most relevant to the indicator. For example, analyzing the similarity or dissimilarity of trends in sensor data can be done by establishing a correlation matrix based on grey relational analysis. Based on this matrix, the elements and their relationships are visualized using a graph structure, revealing the spatial relationships between the elements. For each sensor, the top N (e.g., N=4) sensors with the highest correlation are clustered to optimize the visualized graph structure. Based on the spatial relationships visualized in the graph structure and monitoring requirements, the main and auxiliary sensors in each cluster are determined, with the number of auxiliary sensors equal to N. Clustering methods are not limited to grey relational analysis; they can also include hierarchical clustering, K-means clustering, and density clustering.

[0042] Clustering multiple sensors refers to dividing physical or abstract objects into several groups (clusters) such that objects within each group (cluster) have high similarity; that is, sensor types within the same cluster are highly correlated, while the similarity between objects in different groups (clusters) is relatively low. Based on the distribution of sensors between clusters, the main sensor type and auxiliary sensor types for constructing condition monitoring features can be determined. For example, rotational speed, oil temperature, and water temperature constitute a cluster; when constructing rotational speed condition monitoring features, rotational speed is selected as the main sensor parameter, and oil temperature, water temperature, etc., are selected as auxiliary sensor parameters.

[0043] To visually assess sensor clustering, a correlation heatmap can be used as an example. Within a first preset range (e.g., 0.80-1), values ​​closer to 1 are displayed as a darker first color (e.g., dark blue), indicating stronger correlations. Conversely, values ​​further from 1 are displayed as a lighter first color (e.g., light blue), showing weakening correlations. Correlation values ​​within a second preset range (e.g., 0.60-0.75) are all displayed in red. Within this second preset range, values ​​closer to 1 are displayed as a lighter second color (e.g., light red), and values ​​further from 1 are displayed as a darker second color (e.g., dark red), again indicating weakening correlations. All values ​​within the second preset range are less than the minimum value in the first preset range. Therefore, when selecting the primary and secondary sensors, after determining the primary sensor, the secondary sensor can be directly selected from the blue range.

[0044] In other embodiments, determining the correlation of sensor data of the same dimension at different lag times refers to the correlation between the values ​​of a signal at different points in time. This describes the strength and direction of the linear relationship between the current value and past or future values ​​of sensor data of the same dimension in a time series. By determining the correlation of sensor data of the same dimension at different lag times, the stability of a certain dimension of construction machinery can be monitored. For example, for sensor data of a specific dimension, if the correlation of the specified dimension remains within a relatively stable range, it indicates that the performance of the specified dimension of the construction machinery is consistent at different points in time, and it may be in a stable operating state. Conversely, if the correlation of the specified dimension fluctuates significantly, it may mean that there are unstable factors in that dimension during a certain period of time.

[0045] Therefore, after determining the degree of correlation between data from multiple sensors, more suitable sensor data can be selected when judging the state. The degree of correlation between sensor data of the same dimension at different lag times can be determined. When a fault occurs, the time point of the anomaly can be found, and combined with other sensors with high data correlation, the cause and time point of the fault can be determined together.

[0046] Continue to refer to Figure 1 In S300, sensor data in the same dimension are weighted based on the correlation between the data in the same dimension at different lag times to obtain weighted features.

[0047] For a given fault phenomenon or a state that needs to be monitored in a specific dimension, the importance of each single-dimensional sensor data point varies. Therefore, weighting sensor data from different time points according to their importance can highlight data points within the same dimension that have a greater impact on overall performance or prediction, thereby improving the representativeness of the data. Weighted features can help identify outliers or abrupt changes in sensor data. Since outliers typically have significantly different weights than normal data points, they will be more prominent in weighted features. In other words, when a fault phenomenon exists, weighted features can more clearly determine the time point when the fault occurred.

[0048] One possible implementation of S300 is as follows: determine the autocorrelation coefficient based on the correlation between sensor data of the same dimension at different lag times; determine the maximum lag order of the autocorrelation coefficient of each sensor based on a preset confidence interval; wherein the maximum lag order represents the time period with the largest time interval in the autocorrelation coefficient; and determine the weighted feature of the maximum lag order based on the sensor data of the same dimension and the corresponding maximum lag order.

[0049] Autocorrelation, also known as serial correlation, is the cross-correlation of a signal at different time points. It measures how the similarity between observations of a sequence at different time points changes with the time interval (or lag) between them. Simply put, it's a mathematical tool for assessing the similarity between current values ​​and past or future values ​​in a sequence, depending on the time difference. The autocorrelation coefficient is a statistic that measures the correlation between values ​​of the same time series at different time points, describing the strength and direction of the linear relationship between current values ​​and their past or future values. Autocorrelation coefficients are used to analyze the correlation of data from various sensors at different lag times. This process is specific to single-dimensional sensors, analyzing the correlation of data from each sensor at different lag times. For example, the autocorrelation coefficient ranges from -1 to 1. When the autocorrelation coefficient is close to 1, it indicates a highly positive correlation in the time series with a lag of k periods; when it is close to -1, it indicates a highly negative correlation; and when it is close to 0, it indicates almost no linear relationship in the time series with a lag of k periods. In autocorrelation, the correlation of sensor data gradually decreases as the lag order increases. The lag in autocorrelation refers to the number of time intervals that separate the time series. For example, for a series with fewer than 240 observations, the maximum lag is approximately n / 4; for a series with more than 240 observations, the maximum lag is... Where n is the number of observations, for sequences with a large number of observations n, the maximum lag number is usually expressed as similar to... The maximum lag order is determined by the formula, but a constant term is added to adjust the range of lags. This adjustment is to account for potential long-term correlations in the sequence while maintaining computational efficiency. The maximum lag order can be determined empirically or using software.

[0050] As an example of a possible implementation of S300, the maximum lag order for which the autocorrelation coefficients of each sensor data are significantly non-zero can be determined based on the confidence interval. The maximum lag order refers to the autocorrelation coefficient with the largest time interval (or lag) among all significantly non-zero autocorrelation coefficients. In other words, determining the maximum lag order for which the autocorrelation coefficients of each sensor data are significantly non-zero indicates the extent to which the similarity (i.e., the autocorrelation coefficient) between the observations at a given time point of sensor data and a past or future time point (this future time point is the lag order) can be considered significantly non-zero. Significantly non-zero means that, with a certain probability (determined by the confidence interval), the autocorrelation coefficient is considered to be real and not due to random error.

[0051] The maximum lag order reflects the largest time interval at which significant correlations exist between observations in a data sequence. Within this interval, subsequent observations may be influenced by those of preceding observations. Therefore, weighted features may adjust the original features based on this lag effect to better capture dynamic changes and trends in the data. When constructing weighted features, weights can be assigned based on the autocorrelation coefficient corresponding to the maximum lag order. For example, if the autocorrelation coefficient for a certain lag order is high, it indicates a high similarity between the observation at that lag order and the current observation, thus warranting a higher weight in the weighted features. Conversely, if the autocorrelation coefficient is low or close to 0, a lower weight can be assigned, or the effect of that lag order can be disregarded.

[0052] One possible approach to determining the weighted feature of the maximum lag order is as follows: A second self-organizing map network is trained based on sensor data of the same dimension; the sensor data of the same dimension is input into the trained second self-organizing map network to obtain network weights at different time points; the network weights represent weight vectors; the maximum weight value in the network weights is determined to be the winning node value at the corresponding time point; the winning node values ​​at all time points in a preset time series are obtained; the sensor data within the time step is weighted using the maximum lag order as the time step and the winning node value within the time step as the weight, to obtain the weighted feature of the maximum lag order.

[0053] Self-Organizing Maps (SOMs) are unsupervised learning neural networks whose main characteristic is their ability to map high-dimensional input data to a low-dimensional space (usually a two- or three-dimensional grid structure) while preserving the original structure and relationships between the data. SOMs can map complex high-dimensional data to a low-dimensional space, making data processing and analysis simpler and more efficient. By comparing network weights at different time points, the dynamic changes of data in the low-dimensional space can be observed, revealing hidden patterns and trends. Using the network weights generated by SOMs, distribution maps or heatmaps can be created, visually demonstrating the clustering and outlier distribution of data at different time points, providing decision-makers with intuitive visual references. Furthermore, by comparing network weights at different time points, subtle changes in the data can be detected. The clustering properties of SOMs enable them to classify input data and identify outliers or exceptions. Comparing network weights at different time points makes it easier to detect abnormal changes in the data. The winning node value refers to the influence weight of the data that has the greatest impact on the fault or state at the current time point. Since determining the winning node value is a necessary step for each single-dimensional sensor, the maximum lag order for each sensor is taken as the time step. The winning node value at each time point within the time step is then used as a weight, and a weighted calculation is performed to obtain the weighted feature of the maximum lag order of the single-dimensional data. By weighting data from different time steps according to their importance (i.e., the winning node value), data points with a greater impact on overall performance or prediction can be highlighted, thereby improving the representativeness of the data. In time series analysis, weighted features can be used to predict future trends or behaviors. By focusing on past time steps that have a significant impact on the future state of the system, the prediction of future trends can be improved.

[0054] As a more concrete example of a possible implementation of S300, Figure 2 This is a schematic flowchart illustrating the process of obtaining the maximum lag order weighted feature according to an exemplary embodiment of the present invention. See also... Figure 2 The time series signal is divided into multiple segments according to the maximum lag order k (divided into two segments in the figure), X1, X2, ... X k The time points are represented by matrices W, where each matrix W represents the weight of each sensor data point at that time point. For example, the weights W1 for X1 are 0.962, 0.682, 0.037, and 0.998. The maximum value in the hidden layer matrix of each layer of the SOM network is taken as the winning node value of the corresponding feature sequence. Therefore, the winning node value for X1 is 0.998. k weight W k The values ​​are -0.419, 0.737, 0.906, and 0.938, where X kThe winning node value is 0.938. The weight W of each time point in the first segment with the maximum lag order k is calculated using this method. k Then, the weighted features are calculated. Y1 represents the sensor data at time node X1, Y... k Represents time node X k Sensor data, in time node X k Sensor data multiplied by X k The weights of the winning node values ​​are calculated, and the results over all time points are summed to obtain the weighted features. The formula for calculating the weighted features can be: In the formula, Z1 represents the weighted feature with the maximum lag order k, the subscript i=1 indicates that the starting index of the summation is i=1, used to traverse each item in the sequence during the summation process. The superscript k indicates that the ending index of the summation is k, and the summation process will continue until i reaches k (inclusive). W i Let W represent the i-th term in the sequence being summed. For each value of i (from 1 to k), its corresponding W is... i The value is added to the sum, Y k Represents time node X k Sensor data.

[0055] As an example, a method for training a second self-organizing map network based on sensor data of the same dimension can be as follows: extract slope features from the sensor data in the same dimension to construct a two-dimensional time series; input the two-dimensional time series into the second self-organizing map network to train the network weights of the sensors; the second self-organizing map network clusters different development trends of the two-dimensional time series according to the network weights; wherein the development trends include gradual increase, gradual decrease, sharp increase, and sharp decrease; evaluate the clustering effect based on an internal metric for evaluating the clustering effect; when the evaluation score of the clustering effect is greater than or equal to a preset score, the training of the second self-organizing map network is considered complete.

[0056] As a more concrete example, for a single-dimensional sensor, slope features are extracted independently to construct a two-dimensional time series. A time series is a sequence formed by arranging the values ​​of a statistical indicator at different times under a certain working condition in chronological order. A 2×2×2 SOM network with hidden layers is initialized, and the two-dimensional time series is input into the SOM network to independently train the network weights for each sensor dimension.

[0057] As a more concrete example, visualize the SOM training results and cluster different development trends of the time series according to the SOM weights, including gradual rise, gradual fall, sharp rise, and sharp fall.

[0058] As a more concrete example, an internal metric could be the Calinski-Harabasz (CH) score, which evaluates clustering performance. The CH score is calculated as the ratio of the sum of squared distances between each point within a cluster and its cluster center (representing intra-cluster density) to the sum of squared distances between each cluster center and the centers of the entire dataset (representing dataset separation). A lower CH score indicates poor clustering. If the clustering result is poor (below a preset score), the network weights for each dimension of the sensor are retrained. A higher CH score (greater than or equal to a preset score) indicates better clustering, meaning the clusters are more compact and more dispersed, confirming the completion of the second self-organizing map (SOM) network training. If the clustering result is good, the normalized one-dimensional data is sequentially input into the trained SOM network to obtain the network weights at the current time point. The maximum value among the network weights is selected as the winning node value for the current network. This process of inputting normalized one-dimensional data into the trained SOM network and searching for the maximum value among the obtained network weights continues until the winning node values ​​for all time points in the current time series are obtained.

[0059] Continue to refer to Figure 1 In the S400, condition monitoring features are constructed based on the correlation and weighted characteristics of data from multiple sensors. The changing trends of these condition monitoring features reflect the operating status of the construction machinery.

[0060] Weighted features refer to calculating a weighted feature for values ​​with prominent characteristics in a single dimension. Constructing condition monitoring features is essentially the process of integrating data from multiple sources (multiple sensors) and time points to form more useful and comprehensive information. For example, by analyzing the correlation between data from multiple sensors, it can be found that oil temperature and water temperature have the greatest impact on speed when monitoring rotational speed. Therefore, when constructing speed condition monitoring features, speed, oil temperature, and water temperature can be considered together. Furthermore, the changes in speed, oil temperature, and water temperature at different time points must be considered. When constructing fused features, temporal correlations are also taken into account. This allows for rapid location of the fault time point and the faulty sensor data when abnormal trends appear. A flat trend in condition monitoring features indicates that the construction machinery is operating normally without faults or abnormalities. Sharp increases or decreases in the trend indicate faults or abnormalities in the construction machinery. Therefore, the condition monitoring features extracted in this invention can effectively and intuitively identify abnormal states of construction machinery compared to the original sensor data. Moreover, the fused condition monitoring features can more comprehensively evaluate the abnormal states of construction machinery, improving the accuracy and efficiency of condition monitoring.

[0061] One possible implementation of S400 is as follows: Based on the main sensor and at least one auxiliary sensor in each cluster, determine the main sensor parameters and auxiliary sensor parameters in each cluster; construct a multi-dimensional sensor sequence for each cluster based on the main sensor parameters and auxiliary sensor parameters; construct a feature sequence based on the multi-dimensional sensor sequence of each cluster and the weighted features of the maximum hysteresis order corresponding to the main sensor parameters; train a first self-organizing map network with the total number of main sensor parameters and auxiliary sensor parameters as the dimension of the feature sequence; use the maximum value in the hidden matrix of each layer of the trained first self-organizing map network as the winning node value of the corresponding feature sequence; and perform weighted fusion of the feature sequence and the winning node value corresponding to each time node to obtain the state monitoring features.

[0062] As an example of a possible implementation of S400, based on the sensor correlation clusters identified above, a multi-dimensional sensor sequence is constructed using the main sensor parameters and auxiliary sensor parameters. For example, when constructing a speed state monitoring feature, speed is the main sensor parameter, and oil temperature, water temperature, etc., are auxiliary sensor parameters. The multi-dimensional sensor sequence is used to construct a feature sequence with the maximum hysteresis order of the main parameters as a reference, resulting in a multi-dimensional feature sequence. This can be understood as fusing multi-dimensional data based on time series. A Self-Organizing Map (SOM) network (first self-organizing map network) with a hidden layer dimension of 15×15×K is initialized, where K is the dimension of the feature sequence, i.e., the total number of main and auxiliary sensor parameters. The multi-dimensional feature sequence is input into the initialized SOM network for iterative training. The maximum value in the hidden layer matrix of each layer of the SOM network is taken as the winning node value of the corresponding feature sequence. At each time point, the corresponding feature sequence and the winning node value are weighted and fused to construct the state monitoring feature.

[0063] In some embodiments, to facilitate intuitive viewing of the status of construction machinery and to identify faults, the status monitoring features can be visualized. Therefore, after S400, the construction machinery status monitoring method may further include: converting the status monitoring features into a visualized graph; wherein the visualized graph includes a line graph; when a raised portion deviating from a preset fluctuation range appears in the line graph of the status monitoring features, determining that a fault occurred during the time period corresponding to the raised portion deviating from the preset fluctuation range.

[0064] Taking the construction of speed condition monitoring features as an example, Figure 3 This is a schematic diagram illustrating the visualization of build status monitoring features provided in an exemplary embodiment of the present invention. See also: Figure 3The visualized line chart shows that the main sensor parameter is engine speed, while oil temperature and coolant temperature are auxiliary sensor parameters. The time period 8834 in the chart is the abnormal period, during which the condition monitoring characteristics exhibit a clear abnormal trend. Furthermore, in other time periods besides 8834, the engine speed characteristic shows an abnormal trend, but the coolant temperature and oil temperature characteristics remain stable elsewhere; therefore, the overall engine speed condition monitoring characteristics do not show an abnormal trend. Between 8900 and 8950, the coolant temperature characteristic shows an abnormal trend, but during this period, both the engine speed and oil temperature characteristics are relatively stable; therefore, the engine speed condition monitoring characteristics do not show an abnormal trend either. Therefore, it can be seen that when both engine speed and coolant temperature characteristics show abnormalities, the engine speed condition monitoring characteristics will also show abnormalities, and the fault can be investigated by checking the engine speed and coolant temperature separately.

[0065] Figure 4 This is a schematic diagram of the structure of an engineering machinery condition monitoring device provided in an exemplary embodiment of the present invention, as shown below. Figure 4 As shown, the construction machinery condition monitoring device 4 includes: an acquisition module 41, used to acquire sensor data of multiple dimensions collected by multiple sensors; wherein, one sensor corresponds to sensor data of one dimension; a determination module 42, used to determine the degree of data correlation between multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times based on the sensor data of multiple dimensions; a weighting module 43, used to weight the sensor data in the same dimension based on the degree of data correlation of sensor data of the same dimension at different lag times to obtain weighted features; and a construction module 44, used to construct condition monitoring features based on the degree of data correlation between multiple sensors and the weighted features; wherein, the changing trend of the condition monitoring features reflects the operating status of the construction machinery.

[0066] The engineering machinery condition monitoring device provided by this invention can process multi-dimensional sensor data collected from complex system equipment. Based on the correlation between multi-dimensional sensors and the correlation between sensor data, condition monitoring features are constructed. These features can intuitively reflect the condition of the engineering machinery, improving data representativeness. By comparing the changes in the condition of the engineering machinery at different time points, the development trend of the condition monitoring features over time can be obtained. Therefore, the operating status of the engineering machinery can be judged by the fluctuations in the development trend, effectively improving the efficiency and accuracy of engineering machinery condition monitoring.

[0067] In one embodiment, the determining module 42 can be configured to: determine the correlation matrix of sensor data among multiple sensors based on sensor data of multiple dimensions and grey relational analysis; wherein the correlation matrix represents the visual correlation degree of multiple sensor data; and cluster the multiple sensors based on the correlation matrix; wherein each cluster includes a main sensor and at least one auxiliary sensor, and the auxiliary sensor is a sensor whose correlation degree with the main sensor reaches a preset value.

[0068] In one embodiment, the weighting module 43 can be configured to: determine the autocorrelation coefficient based on the degree of data correlation of sensor data of the same dimension at different lag times; determine the maximum lag order of the autocorrelation coefficient of each sensor based on a preset confidence interval; wherein the maximum lag order represents the time period with the largest time interval in the autocorrelation coefficient; and determine the weighting feature of the maximum lag order based on the sensor data of the same dimension and the corresponding maximum lag order.

[0069] In one embodiment, the construction module 44 can be configured to: determine the main sensor parameters and auxiliary sensor parameters in each cluster based on the main sensor and at least one auxiliary sensor in each cluster; construct a multi-dimensional sensor sequence for each cluster based on the main sensor parameters and auxiliary sensor parameters of each cluster; construct a feature sequence based on the multi-dimensional sensor sequence of each cluster and the weighted features of the maximum hysteresis order corresponding to the main sensor parameters; train a first self-organizing map network with the total number of main sensor parameters and auxiliary sensor parameters as the dimension of the feature sequence; use the maximum value in the hidden matrix of each layer of the trained first self-organizing map network as the winning node value of the corresponding feature sequence; and perform weighted fusion of the feature sequence and the winning node value corresponding to each time node to obtain the state monitoring features.

[0070] In one embodiment, the weighting module 43 can also be configured to: train a second self-organizing map network based on sensor data of the same dimension; input sensor data of the same dimension into the trained second self-organizing map network to obtain network weights at different time points; the network weights represent weight vectors; determine the maximum weight value in the network weights as the winning node value at the corresponding time point; obtain the winning node values ​​at all time points in a preset time series; and perform weighted calculations on the sensor data within the time step using the maximum lag order as the time step and the winning node value in the time step as the weight to obtain the weighted feature of the maximum lag order.

[0071] In one embodiment, the weighting module 43 can also be configured to: extract slope features of sensor data in the same dimension to construct a two-dimensional time series; input the two-dimensional time series into a second self-organizing map network to train the network weights of the sensors; the second self-organizing map network clusters different development trends of the two-dimensional time series according to the network weights; wherein the development trends include gradual rise, gradual fall, sharp rise and sharp fall; evaluate the clustering effect based on an internal metric for evaluating the clustering effect; when the evaluation score of the clustering effect is greater than or equal to a preset score, it is determined that the training of the second self-organizing map network is complete.

[0072] In one embodiment, the engineering machinery condition monitoring device 4 can also be configured to: convert condition monitoring features into visual graphics; wherein the visual graphics include line graphs; when a protruding part deviating from a preset fluctuation range appears in the line graph of the condition monitoring features, it is determined that a fault occurred in the time period corresponding to the protruding part deviating from the preset fluctuation range.

[0073] In one embodiment, the engineering machinery condition monitoring device 4 can also be configured to: perform frequency histogram analysis on sensor data of the target dimension, and statistically analyze the frequency of sensor data occurrence in different value ranges; wherein, the target dimension is any dimension among multiple dimensions; based on the frequency of sensor data occurrence in different value ranges, determine the upper and lower control limits of the value range distribution of sensor data in the target dimension; based on the upper and lower control limits of the value range distribution of sensor data in the target dimension, normalize the sensor data in the target dimension to obtain normalized target dimension data; wherein, based on sensor data of multiple dimensions, determining the data correlation between multiple sensors and the data correlation of sensor data of the same dimension in different lag times includes: based on the normalized sensor data of multiple dimensions, determining the data correlation between multiple sensors and the data correlation of sensor data of the same dimension in different lag times.

[0074] An electronic device includes: a processor; a memory for storing processor-executable instructions; and the processor for executing the engineering machinery condition monitoring method described in the embodiments of the present invention.

[0075] Below, for reference Figure 5 This describes an electronic device according to embodiments of the present invention. The electronic device may be either or both of a first device and a second device, or a standalone device independent of them, which may communicate with the first device and the second device to receive acquired input signals from them.

[0076] Figure 5 A block diagram of an electronic device according to an embodiment of the present invention is shown.

[0077] like Figure 5 As shown, the electronic device 10 includes one or more processors 11 and memory 12.

[0078] The processor 11 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0079] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the engineering machinery condition monitoring method of the various embodiments of the present invention described above, and / or other desired functions. Various contents such as input signals, signal components, and noise components may also be stored in the computer-readable storage medium.

[0080] In one example, the electronic device 10 may also include an input device 13 and an output device 14, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0081] When the electronic device is a standalone device, the input device 13 can be a communication network connector for receiving the collected input signals from the first device and the second device.

[0082] In addition, the input device 13 may also include, for example, a keyboard, a mouse, etc.

[0083] The output device 14 can output various information to the outside, including determined distance information, direction information, etc. The output device 14 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0084] Of course, for the sake of simplicity, Figure 5 Only some of the components of the electronic device 10 relevant to the present invention are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device 10 may include any other suitable components depending on the specific application.

[0085] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of the present invention. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0086] A computer-readable storage medium stores a computer program for executing the engineering machinery condition monitoring method described in the embodiments of the present invention.

[0087] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0088] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the invention to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for monitoring the condition of engineering machinery, characterized in that, include: Acquire sensor data from multiple sensors in multiple dimensions; where each sensor corresponds to one dimension of sensor data. Based on the sensor data from the multiple dimensions, determine the degree of data correlation between the multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times; Based on the multi-dimensional sensor data and grey relational analysis, a correlation matrix of sensor data among the multiple sensors is determined; wherein, the correlation matrix represents the visual correlation degree of multiple sensor data; based on the correlation matrix, the multiple sensors are clustered; wherein, each cluster includes a main sensor and at least one auxiliary sensor, wherein the auxiliary sensor is a sensor whose correlation degree with the main sensor reaches a preset value; Based on the correlation between sensor data of the same dimension at different lag times, the autocorrelation coefficient is determined; based on a preset confidence interval, the maximum lag order of the autocorrelation coefficient of each sensor is determined; where the maximum lag order represents the time period with the largest time interval in the autocorrelation coefficient. A second self-organizing map network is trained based on sensor data of the same dimension. Sensor data of the same dimension is input into the trained second self-organizing map network to obtain network weights at different time points. Each network weight represents a weight vector. The maximum weight value in the network weights is determined to be the winning node value at the corresponding time point. The winning node values ​​at all time points in a preset time series are obtained. Using the maximum lag order as the time step and the winning node value within the time step as the weight, the sensor data within the time step is weighted to obtain the weighted feature of the maximum lag order. Based on the correlation between the data from the multiple sensors and the weighted characteristics of the maximum hysteresis order, a condition monitoring feature is constructed; wherein, the changing trend of the condition monitoring feature reflects the operating status of the construction machinery. Based on the main sensor and at least one auxiliary sensor in each cluster, the main sensor parameters and auxiliary sensor parameters in each cluster are determined; a multi-dimensional sensor sequence for each cluster is constructed according to the main sensor parameters and the auxiliary sensor parameters; a feature sequence is constructed based on the multi-dimensional sensor sequence of each cluster and the weighted feature of the maximum hysteresis order corresponding to the main sensor parameters; a first self-organizing map network is trained with the total number of the main sensor parameters and the auxiliary sensor parameters as the dimension of the feature sequence; the maximum value in the hidden matrix of each layer of the first self-organizing map network after training is used as the winning node value of the corresponding feature sequence; the feature sequence corresponding to each time node and the winning node value are weighted and fused to obtain the state monitoring features.

2. The method for monitoring the condition of engineering machinery according to claim 1, characterized in that, The training of the second self-organizing map network based on sensor data of the same dimension includes: Slope features of sensor data in the same dimension are extracted to construct a two-dimensional time series. The two-dimensional time series is input into a second self-organizing map network to train the network weights of the sensor; The second self-organizing map network clusters different development trends of the two-dimensional time series based on the network weights; wherein the development trends include gradual increase, gradual decrease, sharp increase, and sharp decrease; The clustering effect is evaluated based on internal metrics used to assess the clustering performance. When the evaluation score of the clustering effect is greater than or equal to the preset score, the training of the second self-organizing map network is determined to be complete.

3. The method for monitoring the condition of engineering machinery according to claim 1, characterized in that, After constructing the condition monitoring features based on the data correlation between the multiple sensors and the weighted features of the maximum hysteresis order, the engineering machinery condition monitoring method further includes: The status monitoring features are converted into visual graphics; wherein, the visual graphics include line graphs; When a raised portion deviating from the preset fluctuation range appears in the line graph of the status monitoring feature, it is determined that a fault occurred during the time period corresponding to the raised portion deviating from the preset fluctuation range.

4. The method for monitoring the condition of engineering machinery according to claim 1, characterized in that, After acquiring multi-dimensional sensor data from multiple sensors, the condition monitoring method for construction machinery also includes: Frequency histogram analysis is performed on sensor data of the target dimension to count the frequency of sensor data occurrence in different value ranges; wherein, the target dimension can be any dimension from multiple dimensions. Based on the frequency of sensor data occurrence in different value ranges, determine the upper and lower control limits of the value range distribution of sensor data in the target dimension. Based on the upper and lower control limits of the value range distribution of sensor data in the target dimension, the sensor data in the target dimension is normalized to obtain normalized target dimension data. The step of determining the correlation between the multiple sensors and the correlation of sensor data of the same dimension at different lag times based on the multi-dimensional sensor data includes: Based on the normalized multi-dimensional sensor data, the degree of data correlation among the multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times are determined.

5. A condition monitoring device for construction machinery, applied to the condition monitoring method for construction machinery according to any one of claims 1-4, characterized in that, include: The acquisition module is used to acquire sensor data of multiple dimensions collected by multiple sensors; where each sensor corresponds to sensor data of one dimension. The determination module is used to determine the degree of data correlation between the multiple sensors and the degree of data correlation of sensor data of the same dimension at different lag times based on the sensor data of the multiple dimensions. The weighting module is used to weight the sensor data in the same dimension based on the degree of data correlation at different lag times to obtain weighted features; A construction module is used to construct status monitoring features based on the correlation between data from the multiple sensors and the weighted features; wherein the changing trend of the status monitoring features reflects the operating status of the construction machinery.

6. An engineering machinery, characterized in that, include: Multiple sensors are used to collect sensor data in multiple dimensions in engineering machinery; wherein, each sensor collects sensor data in one dimension. A monitoring device, which is communicatively connected to the plurality of sensors, is used to perform the engineering machinery condition monitoring method according to any one of claims 1-4.

Images