An engineering machinery device remote monitoring system and method based on an internet of things

Abstract

The application provides an engineering mechanical equipment remote monitoring system and method based on an Internet of Things, relates to the technical field of equipment operation monitoring, and obtains frequency domain migration characteristics and energy attenuation characteristics by performing cross-scale characteristic response on the dynamic impact signal of the engineering mechanical equipment through an intelligent computing node, performing gradient operation on the operation thermal field data of the engineering mechanical equipment, and then obtaining thermal flow evolution characteristics; the dynamic interaction atlas is obtained by using a coupling analysis model; the node state and edge relationship of the dynamic interaction atlas are predicted by using space-time resonance convolution based on displacement trajectories, and the state prediction curve of the engineering mechanical equipment is obtained; the adaptive feedback network is constructed at the edge computing node of the Internet of Things according to the state prediction curve and the preset scheduling instruction, and then the engineering mechanical equipment is remotely monitored and scheduled based on the adaptive feedback network, so that the operation state prediction and scheduling closed-loop control of the engineering mechanical equipment can be realized, and the stability of the remote monitoring of the engineering mechanical equipment is improved.

Description

Technical Field

[0001] This application relates to the field of equipment operation monitoring technology, and more specifically, to a remote monitoring system and method for engineering machinery equipment based on the Internet of Things. Background Technology

[0002] Equipment operation monitoring technology refers to the real-time perception, analysis and management of equipment operation status through sensors, communication and data processing to ensure the safe, stable and efficient operation of equipment. It can be used in fields such as power, transportation, manufacturing and engineering machinery. The development trend of equipment operation monitoring is gradually evolving from traditional passive data collection and manual inspection to intelligent monitoring and predictive maintenance.

[0003] With the development of IoT technology, remote monitoring methods for construction machinery based on IoT are gradually emerging. By deploying multimodal sensors on construction machinery and combining edge computing nodes with cloud platforms, it is possible to remotely collect and transmit operating parameters such as equipment displacement, vibration, temperature, and pressure. These parameters can then be comprehensively analyzed and visualized in the cloud, enabling centralized management and remote maintenance of distributed construction machinery and improving equipment operation and maintenance efficiency. However, existing IoT-based remote monitoring of construction machinery still has some technical shortcomings. Most IoT remote monitoring technologies remain at the stage of static monitoring of single physical quantities, and the edge terminals are mostly used for simple data collection and transmission, lacking real-time computing and intelligent prediction capabilities, resulting in delayed anomaly detection and unstable monitoring. Furthermore, existing remote monitoring is mostly one-way information transmission, unable to achieve dynamic adaptive scheduling of equipment under complex working conditions. Therefore, how to achieve predictive and closed-loop control of the operating status of construction machinery to improve the stability of remote monitoring is a challenge facing the industry. Summary of the Invention

[0004] This application provides a remote monitoring system and method for construction machinery equipment based on the Internet of Things, which can realize the prediction of the operating status and closed-loop control of the scheduling of construction machinery equipment, so as to improve the stability of remote monitoring of construction machinery equipment.

[0005] In a first aspect, this application provides a method for remote monitoring of engineering machinery equipment based on the Internet of Things, the remote monitoring method comprising the following steps:

[0006] Collect displacement trajectory, dynamic impact signals, and operating thermal field data of engineering machinery and equipment;

[0007] By using IoT-based intelligent computing nodes to perform cross-scale characteristic response analysis on the dynamic impact signal, frequency domain migration characteristics and energy attenuation characteristics are obtained. Gradient calculations are then performed on the operating thermal field data to obtain thermal flow evolution characteristics.

[0008] A dynamic interaction map of engineering machinery is constructed using a coupled analysis model based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy decay characteristics.

[0009] Based on the displacement trajectory, spatiotemporal resonance convolution prediction is performed on the node states and edge relationships of the dynamic interaction graph to obtain the state prediction curve of the engineering machinery equipment.

[0010] An adaptive feedback network is constructed at the edge computing node of the Internet of Things based on the state prediction curve and the preset scheduling instructions, and then the engineering machinery equipment is remotely monitored and scheduled based on the adaptive feedback network.

[0011] In this embodiment, displacement trajectory, dynamic impact signal and operating thermal field data of engineering machinery are collected by multimodal sensing equipment.

[0012] In this embodiment, the frequency domain migration characteristics and energy attenuation characteristics obtained by performing cross-scale characteristic response analysis on the dynamic impact signal using IoT-based intelligent computing nodes specifically include:

[0013] The dynamic impact signal is segmented in the time domain and decomposed by wavelet packet to obtain sub-signal sequences at different scales, and then the main frequency component of each sub-signal sequence is extracted.

[0014] The dominant frequency offset and frequency domain shift characteristics are determined by using all dominant frequency components.

[0015] Based on the main frequency offset, energy integration is performed on all sub-signal sequences to obtain an energy distribution sequence;

[0016] The energy decay rate of the energy distribution sequence is calculated by using a time sliding window to obtain the energy decay characteristics.

[0017] In this embodiment, performing gradient calculations on the operating thermal field data to obtain the heat flow evolution characteristics specifically includes:

[0018] The operating thermal field data is subjected to two-dimensional rasterization to obtain a discrete raster point temperature matrix;

[0019] The temperature gradient field is obtained by performing spatial difference calculation on the grid point temperature matrix using a first-order difference algorithm.

[0020] The temperature gradient field is tracked over time to obtain the heat flow evolution characteristics.

[0021] In this embodiment, the construction of a dynamic interaction map of engineering machinery based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy decay characteristics using a coupled analysis model specifically includes:

[0022] The heat flow evolution characterization, the frequency domain migration characteristic, and the energy decay characteristic are subjected to feature normalization processing to obtain a multi-source feature vector set;

[0023] The correlation of the multi-source feature vector set is calculated to obtain the coupling coefficient matrix;

[0024] The coupling coefficient matrix is ​​clustered using a coupling analysis model to obtain clusters of different dynamic behaviors;

[0025] All dynamic behavior clusters are mapped to graph structure nodes, and the non-zero coefficients in the coupling coefficient matrix are mapped to graph structure edge relationships, thereby obtaining the dynamic interaction graph of engineering machinery equipment.

[0026] In this embodiment, the spatiotemporal resonance convolution prediction of the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory to obtain the state prediction curve of the engineering machinery equipment specifically includes:

[0027] The displacement trajectory is subjected to temporal segmentation and interpolation resampling to obtain a spatiotemporal displacement sequence;

[0028] Based on the node states and edge relationships of the dynamic interaction graph, the spatiotemporal displacement sequence and the dynamic interaction graph are aligned and matched to obtain the graph convolution input tensor;

[0029] The graph convolution input tensor is subjected to temporal convolution, frequency domain transformation, and resonance detection to obtain a frequency resonance feature set.

[0030] The frequency resonance feature set is fitted and calculated based on the predictive regression model to obtain the state prediction curve of the engineering machinery equipment.

[0031] In this embodiment, remote monitoring and scheduling of engineering machinery based on the adaptive feedback network specifically includes:

[0032] The adaptive feedback network uses a forward propagation mechanism to schedule the execution status of the engineering machinery equipment in real time and obtain the first execution command.

[0033] The reverse link of the adaptive feedback network predicts and corrects the first execution instruction based on the real-time scheduling result to obtain a remote monitoring and scheduling instruction, which is then used for remote monitoring and scheduling.

[0034] In this embodiment, the IoT-based intelligent computing node refers to an edge computing node built based on deep learning inference algorithms and multi-threaded parallel scheduling algorithms.

[0035] In this embodiment, the coupling analysis model is a multimodal coupling model based on graph neural networks.

[0036] Secondly, this application provides an Internet of Things (IoT)-based remote monitoring system for construction machinery equipment, used to execute an IoT-based remote monitoring method for construction machinery equipment, the remote monitoring system comprising:

[0037] The data acquisition module is used to collect displacement trajectory, dynamic impact signals, and operating thermal field data of engineering machinery and equipment;

[0038] The multi-source feature analysis module is used to perform cross-scale feature response on the dynamic impact signal through IoT-based intelligent computing nodes to obtain frequency domain migration features and energy decay features, and to perform gradient calculation on the operating thermal field data to obtain thermal flow evolution features.

[0039] The graph construction module is used to construct a dynamic interaction graph of engineering machinery equipment based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy decay characteristics using a coupled analysis model;

[0040] The predictive analysis module is used to perform spatiotemporal resonance convolution prediction on the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory, so as to obtain the state prediction curve of the engineering machinery equipment.

[0041] The collaborative scheduling module is used to construct an adaptive feedback network on the edge computing node of the Internet of Things based on the state prediction curve and preset scheduling instructions, and then to remotely monitor and schedule engineering machinery equipment based on the adaptive feedback network.

[0042] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0043] By collecting displacement trajectories, dynamic impact signals, and operating thermal field data of engineering machinery, and then using IoT-based intelligent computing nodes to perform cross-scale characteristic responses on the dynamic impact signals to obtain frequency domain migration characteristics and energy attenuation characteristics, and performing gradient calculations on the operating thermal field data to obtain heat flow evolution characteristics, a dynamic interaction map of the engineering machinery is constructed using a coupled analysis model based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy attenuation characteristics. Based on the displacement trajectory, spatiotemporal resonance convolution prediction is performed on the node states and edge relationships of the dynamic interaction map to obtain the state prediction curve of the engineering machinery. Based on the state prediction curve and preset scheduling instructions, an adaptive feedback network is constructed on the edge computing nodes of the IoT, and then the engineering machinery is remotely monitored and scheduled based on the adaptive feedback network.

[0044] Therefore, this application demonstrates that the operational status prediction and closed-loop control of engineering machinery equipment can be achieved. Firstly, by collecting displacement trajectories, dynamic impact signals, and operational thermal field data of the engineering machinery equipment, a comprehensive input condition covering structural deformation, dynamic response, and heat conduction characteristics of the equipment can be formed. Furthermore, cross-scale characteristic responses of the dynamic impact signals are performed to obtain frequency domain migration and energy attenuation characteristics, and gradient calculations are performed on the operational thermal field data to obtain heat flow evolution characteristics. This allows for a dual analysis of the energy dissipation law of the impact signal and the heat flow conduction process, effectively characterizing the dynamic loss mechanism of the equipment under complex operating conditions. This facilitates the transformation of previously fragmented sensor data into a calculable and trackable state representation. Secondly, constructing a dynamic interaction map of the engineering machinery equipment can integrate multi-source features... The nonlinear coupling relationship is expressed in the form of a graph structure, which is beneficial for capturing potential coupling effects that are difficult to identify by traditional methods, and provides higher accuracy and robustness for predictive analysis. Then, the spatiotemporal resonance convolution prediction of the node state and edge relationship of the dynamic interaction graph is performed to obtain the state prediction curve of the engineering machinery equipment. This realizes a deep prediction process that combines the temporal characteristics of displacement trajectory with dynamic interaction relationship, which is beneficial for identifying potential abnormal frequency patterns and operational trend drift, thereby discovering possible failure signs in advance. Finally, by constructing an adaptive feedback network and performing remote monitoring and scheduling of engineering machinery equipment based on this adaptive feedback network, the combination of forward command propagation and reverse prediction correction can enable engineering machinery equipment to maintain the stability of operation and scheduling under complex working conditions and sudden disturbances.

[0045] In summary, the technical solution adopted in this application can realize the prediction of the operating status and closed-loop control of engineering machinery equipment, so as to improve the stability of remote monitoring of engineering machinery equipment. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this embodiment of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of a remote monitoring method for engineering machinery equipment based on the Internet of Things, according to this application;

[0048] Figure 2 This is an exemplary flowchart for determining heat flow evolution characteristics according to the present application;

[0049] Figure 3 This is an exemplary flowchart for determining the dynamic interaction diagram of engineering machinery equipment according to the present application;

[0050] Figure 4 This is a module structure diagram of a remote monitoring system for engineering machinery equipment based on the Internet of Things, provided in this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0052] This application provides an IoT-based remote monitoring system and method for construction machinery. The core of this system involves collecting displacement trajectories, dynamic impact signals, and operational thermal field data of the construction machinery. Then, using IoT-based intelligent computing nodes, cross-scale characteristic responses are performed on the dynamic impact signals to obtain frequency domain migration characteristics and energy attenuation characteristics. Gradient calculations are performed on the operational thermal field data to obtain heat flow evolution characteristics. A coupled analysis model is used to construct a dynamic interaction graph of the construction machinery based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy attenuation characteristics. Based on the displacement trajectory, spatiotemporal resonance convolution prediction is performed on the node states and edge relationships of the dynamic interaction graph to obtain a state prediction curve for the construction machinery. An adaptive feedback network is constructed on IoT edge computing nodes based on the state prediction curve and preset scheduling instructions, and then the construction machinery is remotely monitored and scheduled based on the adaptive feedback network.

[0053] Example 1: To better understand the above technical solution, the following will provide a detailed description of the technical solution in conjunction with the accompanying drawings and specific implementation methods. (Refer to...) Figure 1 As shown in the figure, this is an exemplary flowchart of a remote monitoring method for construction machinery equipment based on the Internet of Things according to this embodiment of the present application. The remote monitoring method includes the following steps:

[0054] In step S1, the displacement trajectory, dynamic impact signal, and operating thermal field data of the engineering machinery equipment are collected.

[0055] In practice, multimodal sensing devices are used to collect displacement trajectories, dynamic impact signals, and operating thermal field data of construction machinery. These devices include a Global Positioning System (GPS) positioning module, an accelerometer, and an infrared thermal imaging sensor. The GPS positioning module is used in conjunction with an inertial navigation system to collect the displacement trajectory of the construction machinery. The inertial navigation system is used to measure the acceleration and angular velocity information of the equipment during operation. The accelerometer is used to collect the dynamic impact signals of the construction machinery, and the infrared thermal imaging sensor is used to collect the operating thermal field data of the construction machinery. Through the collaborative work of these multimodal sensing devices, the motion path, dynamic state, and thermal effects of the construction machinery can be collected synchronously.

[0056] It should be noted that the displacement trajectory in this application refers to the motion path data of the engineering machinery equipment, which includes timestamps, three-dimensional position coordinates and velocity information; the dynamic impact signal is a dynamic signal describing the engineering machinery equipment when subjected to force or collision; the operating thermal field data refers to the temperature distribution data of the engineering machinery equipment, which is used to characterize the heat conduction and energy dissipation process of the engineering machinery equipment.

[0057] In step S2, the dynamic impact signal is subjected to cross-scale characteristic response through an IoT-based intelligent computing node to obtain frequency domain migration characteristics and energy attenuation characteristics. Gradient calculation is performed on the operating thermal field data to obtain thermal flow evolution characteristics.

[0058] It should be noted that, in this application, the IoT-based intelligent computing node refers to an edge computing node built based on deep learning inference algorithms and multi-threaded parallel scheduling algorithms. The deep learning inference algorithm is used to quickly extract features and recognize patterns from the collected dynamic impact signals and operating thermal field data, enabling real-time computation under low power consumption and ensuring the accuracy of feature extraction and pattern recognition. The multi-threaded parallel scheduling algorithm is used to group and align the data streams uploaded from different sensors, ensuring data is transmitted to the edge computing node without loss and with low latency under complex operating conditions, thereby achieving efficient fusion and reliable processing of multi-source data. In this application, deploying intelligent computing nodes helps improve the real-time performance and stability of the adaptive feedback network, thereby enhancing the closed-loop control capability of remote monitoring and scheduling of engineering machinery equipment.

[0059] In this embodiment, the frequency domain migration characteristics and energy attenuation characteristics obtained by performing cross-scale characteristic response analysis on the dynamic impact signal using IoT-based intelligent computing nodes specifically include:

[0060] The dynamic impact signal is segmented in the time domain and decomposed by wavelet packet to obtain sub-signal sequences at different scales, and then the main frequency component of each sub-signal sequence is extracted.

[0061] The dominant frequency offset and frequency domain shift characteristics are determined by using all dominant frequency components.

[0062] Based on the main frequency offset, energy integration is performed on all sub-signal sequences to obtain an energy distribution sequence;

[0063] The energy decay rate of the energy distribution sequence is calculated by using a time sliding window to obtain the energy decay characteristics.

[0064] In specific implementation, firstly, a fixed-length time window can be used to divide the continuous dynamic impact signal into several time-segment signals. Then, common wavelet basis functions are used to perform multi-level decomposition on all time-segment signals to obtain sub-signal sequences at different scales. The spectral distribution of each sub-signal sequence is calculated using Fast Fourier Transform, and the peak energy frequency corresponding to each sub-signal sequence is selected to obtain the dominant frequency component of each sub-signal sequence. Secondly, the dominant frequency components of each scale sub-signal are arranged in scale order and the difference is calculated with the dominant frequency components of the original signal to obtain the dominant frequency offset. The dominant frequency offsets at all scales are statistically analyzed and normalized to obtain the frequency domain shift characteristics. Then, based on the mean of the dominant frequency offset, energy integration is performed on all sub-signal sequences, that is, the energy values ​​of each sub-signal in the sub-signal sequence are accumulated in the time series, and the accumulated sequence is used as the energy distribution sequence. Finally, the energy attenuation rate of the energy distribution sequence is calculated using a time sliding window, and the calculation result is used as the energy attenuation characteristic.

[0065] It should be noted that the energy attenuation feature in this application is a physical quantity that characterizes the attenuation trend of the impact signal energy over time during the operation of the process machinery equipment, and can be calculated by the time sliding window algorithm; in addition, the dominant frequency offset in this embodiment refers to the numerical quantity obtained by comparing the dominant frequency difference between the sub-signals of different scales and the reference scale signal, which is used to characterize the relative change of the signal frequency; the frequency domain migration feature is a feature set formed based on the statistical results and changing trends of all dominant frequency offsets, which is used to represent the overall migration law of the dynamic impact signal in the frequency domain with scale evolution.

[0066] Preferred, Reference Figure 2 As shown, this figure is an exemplary flowchart for determining heat flow evolution characteristics according to the present application. In this embodiment, the gradient calculation of the operating thermal field data to obtain the heat flow evolution characteristics can be achieved by the following steps:

[0067] In step S21, the operating thermal field data is subjected to two-dimensional rasterization to obtain a discrete raster point temperature matrix;

[0068] In step S22, the temperature gradient field is obtained by performing spatial difference calculation on the grid point temperature matrix using a first-order difference algorithm.

[0069] In step S23, the temperature gradient field is tracked over time to obtain the heat flow evolution characteristics.

[0070] In specific implementation, firstly, the operating thermal field data is processed into a two-dimensional rasterization, that is, the continuous temperature distribution area of ​​the operating thermal field data is divided into uniform squares according to rows and columns, and the center position of each square is taken as a grid point, and the temperature value of the grid point is recorded, thereby obtaining a discrete grid point temperature matrix; then, a first-order difference algorithm is used to perform spatial difference operations on the grid point temperature matrix, that is, the temperature difference between adjacent grid points in the horizontal and vertical directions is calculated respectively, and the ratio of the temperature difference to the distance between adjacent grid points is taken as the temperature gradient vector, and the matrix composed of all temperature gradient vectors is taken as the temperature gradient field; finally, the temperature gradient field at different times can be stored sequentially in chronological order using queue storage technology, and the trajectory of the gradient vector at the same location over time can be calculated, and the trajectory can be taken as the heat flow evolution feature, which can be used to represent the dynamic change law of heat conduction and dissipation during the operation of engineering machinery equipment.

[0071] It should be noted that the heat flow evolution characteristics in this application describe the changing patterns of heat conduction and dissipation during the operation of engineering machinery and equipment; the two-dimensional gridding processing in this embodiment is a data discretization technique, which divides the continuous temperature field into a finite number of grid points to facilitate subsequent numerical calculations and matrix operations; the first-order difference algorithm is a numerical differential calculation algorithm that approximates the rate of temperature change by calculating the difference between adjacent points; the queue storage technology is a simple data structure that can be created using Python.

[0072] In step S3, a dynamic interaction map of the engineering machinery equipment is constructed using a coupled analysis model based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy attenuation characteristics.

[0073] It should be noted that the coupling analysis model in this application is a multimodal coupling model based on graph neural networks. This multimodal coupling model can uniformly map multi-source heterogeneous data such as heat flow evolution characteristics, frequency domain migration characteristics, and energy decay characteristics to the same embedding space, so that data from different sources can be jointly modeled in the form of nodes and edges in the graph structure. Graph neural networks can capture the nonlinear interaction relationship between displacement trajectory, heat flow, and energy characteristics by iteratively updating node states and edge relationships, thereby revealing the coupled dynamic characteristics of engineering machinery during operation. The multimodal coupling model is used to perform feature alignment and unified expression of multi-source features in the embedding space, which is beneficial to improving the adaptability of the modeling process to complex working conditions and its robustness to abnormal disturbances.

[0074] Preferred, Reference Figure 3 As shown, this figure is an exemplary flowchart for determining the dynamic interaction map of engineering machinery equipment according to the present application. In this embodiment, the construction of the dynamic interaction map of engineering machinery equipment based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy attenuation characteristics using the coupling analysis model can be achieved through the following steps:

[0075] In step S31, the heat flow evolution characterization, the frequency domain migration feature, and the energy decay feature are subjected to feature normalization processing to obtain a multi-source feature vector set;

[0076] In step S32, the correlation of the multi-source feature vector set is calculated to obtain the coupling coefficient matrix;

[0077] In step S33, the coupling coefficient matrix is ​​clustered using a coupling analysis model to obtain clusters of different dynamic behaviors;

[0078] In step S34, all dynamic behavior clusters are mapped to graph structure nodes, and the non-zero coefficients in the coupling coefficient matrix are mapped to graph structure edge relationships, thereby obtaining the dynamic interaction graph of the engineering machinery equipment.

[0079] In specific implementation, firstly, the thermal evolution characteristics, frequency domain migration characteristics, and energy decay characteristics are numerically standardized, that is, the feature values ​​of different dimensions are scaled to the same numerical range by the maximum and minimum value normalization method, thereby obtaining a set of multi-source feature vectors. Secondly, the correlation of the set of multi-source feature vectors is calculated, that is, the Pearson correlation coefficient between each pair of feature vectors is calculated, and then the matrix composed of all Pearson correlation coefficients is used as the coupling coefficient matrix. Then, the coupling coefficient matrix is ​​clustered using a coupling analysis model, and data with high coupling relationships between features are grouped into the same class, thereby obtaining multiple dynamic behavior clusters. Finally, a graph data structure is created using Python, and all dynamic behavior clusters are mapped to nodes in the graph data structure, and the non-zero correlation coefficients in the coupling coefficient matrix are mapped to edge relationships in the graph data structure. Thus, the graph data structure mapped with nodes and edges serves as a dynamic interaction graph for characterizing the interaction relationships between the operating features of engineering machinery equipment.

[0080] It should be noted that the dynamic interaction graph in this application refers to a graph data structure composed of nodes and edge relationships. The nodes of the dynamic interaction graph are used to represent feature patterns, and the edge relationships are used to represent the coupling relationships between features. The dynamic interaction graph can intuitively represent the dynamic interaction between multiple source features during the operation of engineering machinery equipment. In this embodiment, the coupling coefficient matrix refers to a matrix structure composed of correlation indicators between features, which is used to reflect the coupling strength relationship between multiple source features. The dynamic behavior cluster is a feature set obtained by clustering, which is used to characterize the feature patterns of engineering machinery equipment under similar dynamic conditions.

[0081] In step S4, spatiotemporal resonance convolution prediction is performed on the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory to obtain the state prediction curve of the engineering machinery equipment.

[0082] In this embodiment, the spatiotemporal resonance convolution prediction of the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory to obtain the state prediction curve of the engineering machinery equipment specifically includes:

[0083] The displacement trajectory is subjected to temporal segmentation and interpolation resampling to obtain a spatiotemporal displacement sequence;

[0084] Based on the node states and edge relationships of the dynamic interaction graph, the spatiotemporal displacement sequence and the dynamic interaction graph are aligned and matched to obtain the graph convolution input tensor;

[0085] The graph convolution input tensor is subjected to temporal convolution, frequency domain transformation, and resonance detection to obtain a frequency resonance feature set.

[0086] The frequency resonance feature set is fitted and calculated based on the predictive regression model to obtain the state prediction curve of the engineering machinery equipment.

[0087] In practical implementation, firstly, the window length is 1 second with 50% overlap, and the sampling points within each segment are divided into sliding segments. Linear interpolation is used to fill in the unequal intervals of the sampling points within each segment, resulting in equal intervals. Zero-mean and unit-variance normalization is applied to each trajectory to obtain a spatiotemporal displacement sequence arranged by time step and spatial measurement points. Secondly, Euclidean distance nearest neighbor matching is used to map each measurement point to a graph node. Nodes not directly covered are filled in using the weighted average displacement of their neighboring nodes. The Laplacian normalization matrix required for standard graph convolution input is constructed using the adjacency matrix and degree matrix of the graph. Finally, a graph convolution input for computation is formed. The input tensor is then used; next, a one-dimensional time-domain convolution is performed on the input tensor along the time dimension, and then a fast Fourier transform is performed on the convolution output node by node to obtain the amplitude spectrum. Subsequently, the main peak frequency and bandwidth are identified by thresholding peak detection on the amplitude spectrum, thus obtaining the frequency resonance feature set for each time window. Finally, a predictive regression model is used to fit and calculate the frequency resonance feature set. Specifically, ridge regression can be used as the basic model, with input features including the main peak frequency, peak amplitude, and bandwidth, and the output being a time series of the operating status score of engineering machinery equipment. Then, a three-point moving average is used to smooth the time series to obtain the state prediction curve.

[0088] It should be noted that the state prediction curve in this application refers to the operating status score curve of engineering machinery equipment given along the time axis, which is used to reflect the health level or risk trend, and facilitates subsequent scheduling and control decisions. Among them, peak detection, fast Fourier transform, Euclidean distance matching, and ridge regression are all common signal processing and machine learning technical terms, which are used for main peak identification, spectrum calculation, spatial mapping and regression fitting, respectively. The spatiotemporal resonance convolution prediction in this embodiment refers to the prediction process of first extracting short-time dynamic features in the time domain, then identifying resonance peaks in the frequency domain and combining them with graph structure coherence for joint discrimination, which is used to amplify the energy accumulation phenomenon related to structural dynamics. The graph convolution input tensor refers to a three-dimensional data object organized according to time steps, graph nodes and physical features, which is used to carry basic quantities such as displacement and velocity and enters the convolution and frequency domain analysis stage together with the adjacency relationship of the graph. The frequency resonance feature set refers to the feature vector set composed of main peak frequency, peak amplitude and bandwidth, which is used to characterize the resonance strength and coupling degree of the structure within a specific time window.

[0089] In step S5, an adaptive feedback network is constructed at the edge computing node of the Internet of Things based on the state prediction curve and the preset scheduling instructions, and then the engineering machinery equipment is remotely monitored and scheduled based on the adaptive feedback network.

[0090] In this embodiment, constructing an adaptive feedback network at the edge computing node of the Internet of Things based on the state prediction curve and preset scheduling instructions specifically includes:

[0091] The state prediction curve is dynamically decomposed to obtain the prediction trend component and the abnormal deviation component.

[0092] In the edge computing node of the Internet of Things, a dual-channel input is formed by the characteristics of the predicted trend component and the abnormal deviation component, thereby obtaining the feedback coupling matrix;

[0093] An adaptive feedback network, comprising forward propagation links and backward correction links, is established based on the feedback coupling matrix.

[0094] In specific implementation, firstly, linear interpolation is used to unify the sampling rate and impute missing values ​​on the state prediction curve. Then, the predicted trend component is extracted using a sliding window smoothing method, and the moving average is subtracted from the original curve to obtain the residual sequence. Next, anomaly detection is performed on the residual sequence using a Hampel filter to identify the abnormal deviation component. Then, in the edge computing node of the Internet of Things, the predicted trend component and the abnormal deviation component are used as dual-channel inputs and time-aligned with a preset scheduling command to obtain the feedback coupling matrix. That is, firstly, the three time series (predicted trend, abnormal deviation, and scheduling command) are segmented according to a unified time window (e.g., 1s or 5s as the window) to calculate the in-window statistics. The mean, variance, and first difference are calculated. Then, the Pearson correlation coefficient or windowed covariance is calculated for each pair of statistics within the window to obtain the original coupling matrix. Row normalization is used to retain the coupling terms of the original coupling matrix to obtain the feedback coupling matrix. Finally, the feedback coupling matrix is ​​used as the weighted adjacency matrix to initialize the graph structure, and forward inference units (e.g., linear regression units) and backward correction units (e.g., adaptive gain modules for online gradient descent) are constructed. Forward inference is performed at the edge nodes at fixed intervals, including closed-loop iterations of simulation (excluding physical delivery), error calculation, and weight update, until the weights converge to a smooth horizontal line. The weight matrix structure at this point is then used as the adaptive feedback network.

[0095] It should be noted that in this application, the adaptive feedback network refers to an edge-end closed-loop mapping structure composed of forward inference units and backward correction units, based on the feedback coupling matrix as the topology. Its function is to realize dynamic mapping between prediction and scheduling and online adaptive weight update at the edge nodes. The backward correction unit preferably adopts small-step online gradient update with regularization terms, and sets upper and lower bounds on the weights and a learning rate decay strategy to avoid divergence. The prediction trend component refers to the long-term change component extracted from the state prediction curve through moving average or low-pass filtering, which is used to reflect the baseline evolution of the system. The abnormal deviation component refers to the short-term mutation identified from the residual through threshold filtering, which is used to characterize sudden disturbances or abnormal events. The feedback coupling matrix refers to the channel statistical coupling representation matrix calculated according to the time window and weighted, which quantifies the mutual influence strength between the prediction trend, abnormal deviation and scheduling instructions.

[0096] In this embodiment, remote monitoring and scheduling of engineering machinery based on the adaptive feedback network specifically includes:

[0097] The adaptive feedback network uses a forward propagation mechanism to schedule the execution status of the engineering machinery equipment in real time and obtain the first execution command.

[0098] The reverse link of the adaptive feedback network predicts and corrects the first execution instruction based on the real-time scheduling result to obtain a remote monitoring and scheduling instruction, which is then used for remote monitoring and scheduling.

[0099] In practice, firstly, the forward propagation unit of the adaptive feedback network performs time alignment and normalization processing on the state prediction curve, real-time sensor data, and preset scheduling instructions, and inputs them into the forward inference unit to obtain preliminary control suggestions. Then, the preliminary control suggestions are checked for upper and lower limits and safety constraints to generate the first execution instruction. Finally, the field controller returns the actual execution feedback, and the edge computing node compares the feedback with the predicted value to obtain the error sequence. The first execution instruction is then corrected using common online correction algorithms (such as Kalman filtering or recursive least squares), which can generate remote monitoring and scheduling instructions and issue them for execution.

[0100] It should be noted that the first execution instruction in this embodiment is the initial control command obtained by forward inference, which is used to quickly drive the equipment; predictive correction (i.e., reverse correction) refers to online correction based on the error sequence to improve control accuracy; the remote monitoring and scheduling instruction is the final corrected control command issued, which can ensure the stable operation of the engineering machinery equipment.

[0101] In summary, the technical solution adopted in this application can realize the prediction of the operating status and closed-loop control of engineering machinery equipment, so as to improve the stability of remote monitoring of engineering machinery equipment.

[0102] Example 2: This application provides a remote monitoring system for construction machinery equipment based on the Internet of Things (IoT), referencing... Figure 4 As shown, this figure is a modular structure diagram of a remote monitoring system for construction machinery equipment based on the Internet of Things (IoT) according to this application. The remote monitoring system includes:

[0103] Data acquisition module 100 is used to acquire displacement trajectory, dynamic impact signal and operating thermal field data of engineering machinery and equipment;

[0104] The multi-source feature analysis module 200 is used to perform cross-scale feature response on the dynamic impact signal through IoT-based intelligent computing nodes to obtain frequency domain migration features and energy decay features, and to perform gradient calculation on the operating thermal field data to obtain thermal flow evolution features.

[0105] The graph construction module 300 is used to construct a dynamic interaction graph of engineering machinery equipment based on the heat flow evolution characterization, the frequency domain migration characteristics, and the energy decay characteristics using a coupled analysis model;

[0106] The prediction and analysis module 400 is used to perform spatiotemporal resonance convolution prediction on the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory, so as to obtain the state prediction curve of the engineering machinery equipment.

[0107] The collaborative scheduling module 500 is used to construct an adaptive feedback network on the edge computing node of the Internet of Things based on the state prediction curve and the preset scheduling instructions, and then to remotely monitor and schedule engineering machinery equipment based on the adaptive feedback network.

[0108] This application is described with reference to flowchart illustrations and / or block diagrams of methods, medical aesthetic devices (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as 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 medical aesthetic device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing medical aesthetic device, generate instructions for implementing the process... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0109] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0110] 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, product, or medical aesthetic device 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, product, or medical aesthetic device. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, product, or medical aesthetic device that includes that element.

Claims

1. A method for remote monitoring of engineering machinery equipment based on the Internet of Things, characterized in that, The remote monitoring method includes the following steps: Collect displacement trajectory, dynamic impact signals, and operating thermal field data of engineering machinery and equipment; By using IoT-based intelligent computing nodes to perform cross-scale characteristic response analysis on the dynamic impact signal, frequency domain migration characteristics and energy attenuation characteristics are obtained. Gradient calculations are then performed on the operating thermal field data to obtain thermal flow evolution characteristics. A dynamic interaction map of engineering machinery equipment is constructed using a coupled analysis model based on the heat flow evolution characteristics, the frequency domain migration characteristics, and the energy decay characteristics. Based on the displacement trajectory, spatiotemporal resonance convolution prediction is performed on the node states and edge relationships of the dynamic interaction graph to obtain the state prediction curve of the engineering machinery equipment. An adaptive feedback network is constructed on the edge computing node of the Internet of Things based on the state prediction curve and the preset scheduling instructions, and then the engineering machinery equipment is remotely monitored and scheduled based on the adaptive feedback network. Specifically, the frequency domain migration characteristics and energy attenuation characteristics obtained by performing cross-scale characteristic response analysis on the dynamic impact signal using IoT-based intelligent computing nodes include: The dynamic impact signal is segmented in the time domain and decomposed by wavelet packet to obtain sub-signal sequences at different scales, and then the main frequency components of each sub-signal sequence are extracted. The dominant frequency offset and frequency domain shift characteristics are determined by using all dominant frequency components. Based on the main frequency offset, energy integration is performed on all sub-signal sequences to obtain an energy distribution sequence; The energy decay rate of the energy distribution sequence is calculated using a time sliding window to obtain the energy decay characteristics. The frequency domain migration feature is a feature set formed based on the statistical results and changing trends of all major frequency offsets, which is used to represent the overall migration law of the dynamic impact signal in the frequency domain as it evolves with scale. Specifically, the spatiotemporal resonance convolution prediction of the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory to obtain the state prediction curve of the engineering machinery equipment includes: The displacement trajectory is subjected to temporal segmentation and interpolation resampling to obtain a spatiotemporal displacement sequence; Based on the node states and edge relationships of the dynamic interaction graph, the spatiotemporal displacement sequence and the dynamic interaction graph are aligned and matched to obtain the graph convolution input tensor; The graph convolution input tensor is subjected to temporal convolution, frequency domain transformation, and resonance detection to obtain a frequency resonance feature set. Based on the prediction regression model, the frequency resonance feature set is fitted and calculated to obtain the state prediction curve of the engineering machinery equipment. The spatiotemporal resonance convolution prediction refers to a prediction process that first extracts short-time dynamic features in the time domain, then identifies resonance peaks in the frequency domain, and combines graph structure coherence for joint discrimination, which is used to amplify energy accumulation phenomena related to structural dynamics.

2. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, The displacement trajectory, dynamic impact signal, and operating thermal field data of engineering machinery are collected through multimodal sensing devices.

3. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, The gradient calculation of the operating thermal field data to obtain the heat flow evolution characteristics specifically includes: The operating thermal field data is subjected to two-dimensional rasterization to obtain a discrete raster point temperature matrix; The temperature gradient field is obtained by performing spatial difference calculation on the grid point temperature matrix using a first-order difference algorithm. The temperature gradient field is tracked over time to obtain the heat flow evolution characteristics.

4. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, Constructing a dynamic interaction map of engineering machinery based on the heat flow evolution characteristics, frequency domain migration characteristics, and energy decay characteristics using a coupled analysis model specifically includes: The heat flow evolution characteristics, the frequency domain migration characteristics, and the energy decay characteristics are normalized to obtain a set of multi-source feature vectors. The correlation of the multi-source feature vector set is calculated to obtain the coupling coefficient matrix; The coupling coefficient matrix is ​​clustered using a coupling analysis model to obtain clusters of different dynamic behaviors; All dynamic behavior clusters are mapped to graph structure nodes, and the non-zero coefficients in the coupling coefficient matrix are mapped to graph structure edge relationships, thereby obtaining the dynamic interaction graph of engineering machinery equipment.

5. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, Remote monitoring and scheduling of engineering machinery based on the aforementioned adaptive feedback network specifically includes: The adaptive feedback network uses a forward propagation mechanism to schedule the execution status of the engineering machinery equipment in real time and obtain the first execution command. The reverse link of the adaptive feedback network predicts and corrects the first execution instruction based on the real-time scheduling result to obtain a remote monitoring and scheduling instruction, which is then used for remote monitoring and scheduling.

6. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, The IoT-based intelligent computing node refers to an edge computing node built based on deep learning inference algorithms and multi-threaded parallel scheduling algorithms.

7. The method for remote monitoring of engineering machinery equipment based on the Internet of Things as described in claim 1, characterized in that, The coupling analysis model is a multimodal coupling model based on graph neural networks.

8. A remote monitoring system for construction machinery based on the Internet of Things (IoT), used to execute the remote monitoring method for construction machinery based on the IoT as described in any one of claims 1 to 7, characterized in that, The remote monitoring system includes: The data acquisition module is used to collect displacement trajectory, dynamic impact signals, and operating thermal field data of engineering machinery and equipment; The multi-source feature analysis module is used to perform cross-scale feature response on the dynamic impact signal through IoT-based intelligent computing nodes to obtain frequency domain migration features and energy decay features, and to perform gradient calculation on the operating thermal field data to obtain thermal flow evolution features. The graph construction module is used to construct a dynamic interaction graph of engineering machinery equipment based on the heat flow evolution characteristics, the frequency domain migration characteristics, and the energy decay characteristics using a coupled analysis model; The predictive analysis module is used to perform spatiotemporal resonance convolution prediction on the node states and edge relationships of the dynamic interaction graph based on the displacement trajectory, so as to obtain the state prediction curve of the engineering machinery equipment. The collaborative scheduling module is used to construct an adaptive feedback network on the edge computing node of the Internet of Things based on the state prediction curve and preset scheduling instructions, and then to remotely monitor and schedule engineering machinery equipment based on the adaptive feedback network.

Images