Multi-sensor fusion equipment operation under the health of the dual-mode data acquisition system

Abstract

This invention provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, comprising: a miniaturized data acquisition terminal, an edge computing module, and a low-power Bluetooth transmission module. The miniaturized data acquisition terminal is fixedly installed on the operating mechanism of the monitored equipment and internally includes an integrated sensor module. The integrated sensor module is used to collect operational behavior modal data and health modal data in real time during equipment operation. The low-power Bluetooth transmission module is used to transmit the operational behavior modal data and health modal data to the edge computing module. The edge computing module is used to perform time alignment and feature extraction on the operational behavior modal data and health modal data, and determine the dynamic correlation between equipment operational behavior parameters and equipment wear parameters based on the time alignment and feature extraction results. This effectively solves the problems of data isolation, delayed analysis, and disconnect between training and maintenance in traditional methods.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a dual-modal health data acquisition system for equipment operation using multi-sensor fusion. Background Technology

[0002] Currently, in modern equipment training, operational evaluation, and health management, it is crucial to comprehensively and accurately acquire equipment status data during real or simulated operations. However, existing technological solutions have significant shortcomings in terms of the comprehensiveness, real-time nature, and integration of data acquisition. (1) Isolated data acquisition and lack of correlation: Existing systems usually collect and process operational behavior data (such as the operation of the control handle and aiming posture) and equipment health data (such as mechanical vibration and temperature rise of key parts) separately. This isolated data flow makes it impossible to analyze the immediate damage or potential impact of specific operational behaviors on equipment in real time during operation, and it is difficult to establish a dynamic causal relationship between operation mode and equipment health status; (2) Complex sensor deployment affects operational realism: Traditional sensors and their cables used to collect equipment health data (such as vibration and temperature) are usually large in size and cumbersome to deploy. They need to be fixed to key parts of the equipment and connected to an external data logger. This deployment method may interfere with the normal operation of the equipment and affect the operator's real experience, especially in training scenarios involving precision operations or simulations. (3) Data processing delay and inability to provide real-time feedback: Existing solutions mostly adopt an offline mode of first collecting and storing data and then processing and analyzing it centrally. This mode has a significant time delay, making it impossible to evaluate the standardization of operational behavior or provide early warning of potential equipment anomalies in real time during training or operation, and it is difficult to establish a dynamic correlation between operational behavior and equipment wear and tear. It is also difficult to use it for immediate training optimization or preventive maintenance decision support. Therefore, in order to overcome the above-mentioned defects, the present invention provides a dual-modal health data acquisition system for equipment operation based on multi-sensor fusion. Summary of the Invention

[0003] This invention provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation. It integrates and synchronously collects operational actions and equipment status information, and analyzes the intrinsic relationship between the two in real time. This enables precise quantification of the human-machine interaction effect during equipment use, directly coupling discrete operational events with continuous equipment wear in the time domain. This allows training and evaluation to shift from qualitative observation to quantitative evaluation based on data correlation. At the same time, it provides a dynamic model derived from actual operational scenarios for predictive maintenance, effectively solving the problems of data isolation, analysis lag, and training-maintenance disconnect in traditional methods.

[0004] This invention provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, comprising: a miniaturized data acquisition terminal, an edge computing module, and a low-power Bluetooth transmission module; Among them, the miniaturized data acquisition terminal is fixedly installed on the operating mechanism of the monitored equipment, and contains an integrated sensor module. Integrated sensor modules are used to collect operational behavior modal data and health modal data in real time during equipment operation. A low-power Bluetooth transmission module is used to transmit operational behavior modal data and health modal data to the edge computing module; The edge computing module is used to perform time alignment and feature extraction on operational behavior modal data and health modal data, and to determine the dynamic correlation between equipment operational behavior parameters and equipment loss parameters based on the time alignment and feature extraction results.

[0005] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes an integrated sensor module comprising: The sensor synchronization start-up unit is used to synchronously start the internally integrated inertial measurement unit, vibration sensor and temperature sensor after the integrated sensor module is powered on and initialized. The data acquisition unit is used for: Based on the first preset sampling frequency, the inertial measurement device is controlled to collect the original angular velocity and original acceleration data of the operating mechanism in three-dimensional space in real time, and the attitude angle data of the operating mechanism in pitch, roll and yaw directions are determined based on the original angular velocity and original acceleration data. The attitude angle data constitutes the attitude time sequence in the operating behavior modal data. Based on the second preset sampling frequency, the vibration sensor collects high-frequency vibration signals of the operating mechanism under mechanical transmission and load change states, and performs time-domain and frequency-domain analysis on the high-frequency vibration signals to extract vibration characteristic parameters. The vibration characteristic parameters constitute the mechanical state indicators in the equipment health modal data. Based on the third preset sampling frequency, the temperature sensor monitors the surface temperature change of the operating mechanism in real time and compares the surface temperature change with the preset temperature baseline to obtain the real-time temperature rise rate and the duration of temperature exceeding the limit. The temperature rise rate and the duration constitute the thermal state indicators in the equipment health modal data.

[0006] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes a data acquisition unit comprising: The data preprocessing subunit is used for: The attitude time sequence, vibration characteristic parameters and thermal state index are obtained, and the basic value range of attitude time sequence, vibration characteristic parameters and thermal state index are obtained from the management terminal respectively. Based on the value range, the attitude time sequence, vibration characteristic parameters and thermal state index are filtered, and the filtering result is converted from analog to digital based on a preset analog-to-digital converter to obtain the corresponding standard digital signal. The time stamping subunit is used to obtain a unified time scale based on the management terminal and add timestamps to standard digital signals based on the unified time scale. The data output subunit is used to encapsulate the standard digital signal after adding a timestamp to obtain operational behavior modal data and device health modal data.

[0007] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes a low-power Bluetooth transmission module, comprising: The transmission preparation unit is used to encapsulate the obtained operation behavior modal data and health modal data according to the communication protocol to obtain the data frame to be transmitted, extract the sensor terminal identifier corresponding to the data frame to be transmitted, and add the sensor terminal identifier to the frame header of the data frame to be transmitted. Link building unit, used for: Obtain the access control address, and initiate a connection request based on the access control address to establish a point-to-point wireless communication link; Based on the current wireless environment noise intensity and link quality indicators, the communication channel and transmit power level are adaptively selected. The data frames to be transmitted are loaded into the transmission buffer in sequence, and the transmission priority and time slot allocation of the data frames to be transmitted are dynamically scheduled according to the data generation rate of the operation behavior mode data and the health mode data. Based on the transmission priority, the cyclic redundancy check code of each data frame to be transmitted is determined and appended to the end of the data frame to be transmitted. The data transmission unit is used to transmit the data frame to be transmitted to the edge computing module based on a determined communication channel and transmission power level.

[0008] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes a data transmission unit comprising: The monitoring subunit is used to start a preset response timer based on the transmission result of the data frame to be transmitted, and to monitor the reception confirmation frame fed back by the edge computing module in real time based on the start result. Data transmission subunit, used for: If a receive acknowledgment frame is received before the acknowledgment timer expires, and the acknowledgment frame indicates that the data frame to be transmitted is verified correctly, then the current data frame to be transmitted is removed from the transmit buffer, and the next data frame to be transmitted is sent. If the acknowledgment timer expires without receiving an acknowledgment frame or the received acknowledgment frame indicates a verification error in the data frame to be transmitted, the strategy of retransmitting the data frame to be transmitted or triggering a link reconnection is determined based on the relative size of the current number of consecutive retransmissions and the preset maximum retransmission threshold. Based on the time slot allocation results, the system dynamically switches to low-power monitoring or sleep mode according to a preset sleep strategy during the intervals between continuous data transmissions, until the next batch of data frames to be transmitted is ready.

[0009] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes a link building unit comprising: The signal measurement subunit is used for: After establishing a wireless communication link, a channel measurement request is sent to the edge computing module; Based on the channel measurement request, the signal strength and noise strength are measured on multiple predefined channels to obtain the received signal strength indication and ambient noise strength value of each channel; The channel determination subunit is used for: The signal-to-noise ratio (SNR) estimate for each channel is determined based on the received signal strength indication and the ambient noise intensity value, and selectable channels with an SNR higher than a preset threshold are selected. Select the channel with the fewest historical handovers and currently no interference flag from the available channels as the target communication channel; Transmit power determination subunit, used for: The initial transmit power level is determined by querying a pre-stored power lookup table based on the average received signal strength of the current wireless communication link. The target communication channel and the initial transmit power level are encapsulated as negotiation parameters and sent to the edge computing module. After the edge computing module confirms the negotiation parameters, it replies with a negotiation confirmation frame. After the negotiation confirmation frame is received, the communication link is switched to the target communication channel and the transmit power is adjusted to the initial transmit power level.

[0010] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes an edge computing module, comprising: Receive operational behavior modality data and health modality data, and extract the corresponding timestamps from the operational behavior modality data and health modality data; Based on timestamps, nonlinear alignment is performed between operational behavior modal data and equipment health modal data, and attitude time series is extracted from operational behavior modal data based on the nonlinear alignment results. The attitude time series includes discrete data points of pitch angle, roll angle and yaw angle as they change over time. Based on the attitude time series, the first-order difference of each attitude angle between adjacent sampling points is determined to obtain the attitude angle change rate sequence. The attitude angle change rate sequence is then divided into sliding windows, and the mean and standard deviation of the attitude angle change rate are determined in each window to obtain the operation smoothness characteristics and operation volatility characteristics. A short-time Fourier transform is performed on the attitude time sequence, and the energy distribution ratio in the preset low-frequency band and high-frequency band is determined based on the transform result to obtain the operating frequency domain characteristics. Zero-crossing points of the attitude time sequence are detected based on the attitude angle change rate sequence, and the number of zero-crossing points per unit time is counted to obtain the operation frequency characteristics. Simultaneously, the vibration signal time sequence and temperature signal time sequence were extracted from the equipment health modal data; By analyzing the time series of vibration signals, the root mean square value and peak value of the vibration signal time series within a complete analysis window are determined, and the vibration intensity characteristics and impact characteristics are obtained. Bandpass filtering is performed on the vibration signal time sequence to retain a preset frequency band related to the natural frequency of the equipment's mechanical structure, and the power spectral density integral of the vibration signal time sequence within the preset frequency band is determined to obtain the resonant energy characteristics; By analyzing the time series of temperature signals, the first derivative of the time series of temperature signals with time is determined, the temperature change gradient sequence is obtained, and the number of consecutive occurrences of positive values ​​and the average slope are determined based on the temperature change gradient sequence to obtain the temperature rise activity characteristics. Based on the time series of temperature signals and the corresponding time series of vibration signals, the peak value of the cross-correlation function of the time series of temperature signals and vibration signals within the same time window is determined, and the thermal-vibration coupling characteristics are obtained. By combining the characteristics of smooth operation, volatile operation, frequency operation, frequency operation, vibration intensity, impact, resonant energy, temperature rise activity, and thermal-vibration coupling, we obtain the operation behavior feature vector and the equipment health feature vector.

[0011] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes an edge computing module, comprising: Feature processing unit, used for: Obtain the operation behavior feature vector and the equipment health feature vector based on feature extraction, and input the operation behavior feature vector and the equipment health feature vector into the pre-built correlation analysis model; Based on the correlation analysis model, the operational behavior feature vector and the equipment health feature vector are analyzed to determine the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector, and output the dynamic correlation strength matrix. The correlation analysis unit is used for: Based on the dynamic correlation strength matrix, a combination of correlation features that matches the preset equipment wear mode is determined, and a mapping function from operation behavior parameters to equipment wear parameters is constructed in real time based on the combination of correlation features. Based on the mapping function, the operation behavior feature vector and the equipment health feature vector of real-time input are forward extrapolated to generate the predicted value and confidence interval of the equipment loss parameter. At the same time, the historical operation sequence is obtained and the operation chain pattern that leads to abnormal loss is determined by combining the historical operation sequence. The dynamic correlation strength matrix, mapping function, predicted value, and operation chain pattern are encapsulated to form an equipment operation-health correlation analysis report.

[0012] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes a feature processing unit comprising: Feature processing subunit, used for: Based on the correlation analysis model, a sliding window with a fixed time length is configured, and the operation behavior feature vector aligned with the current time and the device health feature vector are filled into the sliding window in chronological order; Within the sliding window, the operation behavior feature vector and the device health feature vector are standardized to obtain a standardized feature matrix. Based on the standardized feature matrix, the mutual information value between each operation behavior feature and each device health feature within the sliding window is determined. For feature pairs whose mutual information values ​​exceed a preset threshold, a state space model is constructed with the current operation behavior feature as input and the current device health feature as output, and the parameters of the state space model are predicted based on the operation behavior feature vector and the device health feature vector within the sliding window. Based on the parameters of the predicted state-space model, the conditional causal index of operational behavior characteristics on equipment health characteristics is determined, and the mutual information value and the conditional causal index are weighted and fused to obtain the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector. By iterating through all feature pairs, the mutual information value, state space model parameters, and conditional dependency coefficients corresponding to each feature pair are summarized to obtain the dynamic correlation strength matrix.

[0013] Preferably, a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation includes an edge computing module, comprising: The data packet construction unit is used to encapsulate the obtained dynamic correlation, original operational behavior modal data and health modal data, as well as the corresponding operational behavior feature vectors and device health feature vectors, to obtain the data packet to be uploaded; The result output unit is used to upload the data packet to be uploaded to the upper-level management terminal based on the communication interface.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: By integrating and synchronously collecting operational actions and equipment status information, and analyzing the intrinsic relationship between the two in real time, the system achieves precise quantification of the "human-machine" interaction effect during equipment use. It directly couples discrete operational events with continuous equipment wear in the time domain, enabling training evaluation to shift from qualitative observation to quantitative evaluation based on data correlation. At the same time, it provides a dynamic model derived from actual operational scenarios for predictive maintenance, effectively solving the problems of data isolation, analysis lag, and training-maintenance disconnect in traditional methods.

[0015] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a structural diagram of a multi-sensor fusion-based dual-modal health data acquisition system under equipment operation, as described in an embodiment of the present invention. Figure 2 This is a structural diagram of a low-power Bluetooth transmission module in a dual-modal health data acquisition system for equipment operation under multi-sensor fusion, as described in an embodiment of the present invention. Detailed Implementation

[0018] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0019] Example 1: This example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, such as... Figure 1 As shown, it includes: a miniaturized data acquisition terminal, an edge computing module, and a low-power Bluetooth transmission module; Among them, the miniaturized data acquisition terminal is fixedly installed on the operating mechanism of the monitored equipment, and contains an integrated sensor module. Integrated sensor modules are used to collect operational behavior modal data and health modal data in real time during equipment operation. A low-power Bluetooth transmission module is used to transmit operational behavior modal data and health modal data to the edge computing module; The edge computing module is used to perform time alignment and feature extraction on operational behavior modal data and health modal data, and to determine the dynamic correlation between equipment operational behavior parameters and equipment loss parameters based on the time alignment and feature extraction results.

[0020] In this embodiment, operational behavior modal data refers to the data sequence collected by an inertial measurement device to describe the action characteristics of personnel operating equipment. It mainly includes the three-dimensional attitude angles and angular velocities of the operating mechanism, as well as the derived action sequences and frequency information.

[0021] In this embodiment, health modal data refers to the data sequence collected by vibration and temperature sensors to reflect the mechanical and thermal state of equipment, mainly including characteristic parameters of mechanical vibration signals (such as amplitude and spectral energy) and temperature change indicators of key parts.

[0022] In this embodiment, the equipment operation behavior parameters refer to the quantitative indicators obtained by feature extraction from the operation behavior modal data, which are used to characterize the intensity, speed, stability, frequency and specific operation modes of the operation.

[0023] In this embodiment, the equipment loss parameter refers to the quantitative index obtained by feature extraction from health modal data, which is used to characterize the vibration level of the equipment's mechanical structure, thermal load status, and the wear, fatigue, or performance degradation trend inferred from it.

[0024] In this embodiment, dynamic correlation refers to the statistical dependency or causal relationship between time-varying operational behavior parameters and equipment loss parameters calculated by algorithmic models (such as mutual information analysis and state space modeling), which is used to quantify how a specific operation affects the equipment status immediately or in a short period of time.

[0025] The beneficial effects of the above technical solution are as follows: by collecting operational actions and equipment status information in an integrated and synchronous manner, and analyzing the intrinsic relationship between the two in real time on-site, the "human-machine" interaction effect during equipment use is accurately quantified. Discrete operational events are directly coupled with continuous equipment wear in the time domain, enabling training evaluation to be transformed from qualitative observation to quantitative evaluation based on data correlation. At the same time, it provides a dynamic model derived from actual operational scenarios for predictive maintenance, effectively solving the problems of data isolation, analysis lag, and training-maintenance disconnect in traditional methods.

[0026] Example 2: Based on Example 1, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, comprising an integrated sensor module, including: The sensor synchronization start-up unit is used to synchronously start the internally integrated inertial measurement unit, vibration sensor and temperature sensor after the integrated sensor module is powered on and initialized. The data acquisition unit is used for: Based on the first preset sampling frequency, the inertial measurement device is controlled to collect the original angular velocity and original acceleration data of the operating mechanism in three-dimensional space in real time, and the attitude angle data of the operating mechanism in pitch, roll and yaw directions are determined based on the original angular velocity and original acceleration data. The attitude angle data constitutes the attitude time sequence in the operating behavior modal data. Based on the second preset sampling frequency, the vibration sensor collects high-frequency vibration signals of the operating mechanism under mechanical transmission and load change states, and performs time-domain and frequency-domain analysis on the high-frequency vibration signals to extract vibration characteristic parameters. The vibration characteristic parameters constitute the mechanical state indicators in the equipment health modal data. Based on the third preset sampling frequency, the temperature sensor monitors the surface temperature change of the operating mechanism in real time and compares the surface temperature change with the preset temperature baseline to obtain the real-time temperature rise rate and the duration of temperature exceeding the limit. The temperature rise rate and the duration constitute the thermal state indicators in the equipment health modal data.

[0027] In this embodiment, the vibration characteristic parameters include the effective value of vibration, peak factor, and characteristic frequency band energy.

[0028] In this embodiment, the sensor synchronization start unit refers to a functional module inside the integrated sensor module, which is used to simultaneously trigger the inertial measurement device, vibration sensor and temperature sensor to start working after the module is powered on and initialized, so as to ensure that the start time of data acquisition is consistent.

[0029] In this embodiment, the data acquisition unit refers to a functional module inside the integrated sensor module, which is used to control each sensor to acquire data at a specific sampling frequency and to perform preliminary processing on the acquired raw data to generate modal data.

[0030] In this embodiment, the inertial measurement device refers to a combination of sensors used to measure the angular velocity and acceleration of an object in three-dimensional space, typically including a gyroscope and an accelerometer, to provide attitude information of the operating mechanism.

[0031] In this embodiment, the attitude timing sequence refers to the pitch angle, roll angle, and yaw angle data sequence that is collected and calculated by the inertial measurement device and arranged in chronological order, used to describe the dynamic attitude changes of the operating mechanism.

[0032] In this embodiment, vibration characteristic parameters refer to quantitative indicators such as amplitude and frequency components extracted by performing time-domain and frequency-domain analysis on the high-frequency vibration signals collected by the vibration sensor, which are used to reflect the mechanical state.

[0033] In this embodiment, the mechanical condition index refers to a set of quantitative data composed of vibration characteristic parameters used to assess the health status of the mechanical parts of equipment.

[0034] In this embodiment, the temperature baseline refers to a preset temperature reference value or normal operating temperature range, which is used to compare real-time temperature changes to detect anomalies.

[0035] In this embodiment, the real-time temperature rise rate refers to the slope of the surface temperature change over time as monitored by the temperature sensor, indicating how fast the temperature rises.

[0036] In this embodiment, the duration of temperature over-limit refers to the cumulative time during which the surface temperature exceeds a preset temperature baseline, and is used to assess the duration of the overheating state.

[0037] In this embodiment, the thermal status index refers to a set of quantitative data, consisting of real-time temperature rise rate and duration of temperature exceeding limits, used to assess the thermal health status of equipment.

[0038] The beneficial effects of the above technical solution are as follows: by synchronously activating multiple sensors and collecting data at customized frequencies, high-precision alignment of operational behavior and equipment health data in time is ensured, thereby improving the consistency of dual-modal data. By performing time-frequency analysis on vibration signals and real-time comparison of temperature changes with the baseline, more detailed mechanical and thermal state characteristics are extracted, providing richer and more accurate inputs for subsequent correlation analysis, and enhancing the system's ability to perceive changes in equipment status and the reliability of analysis.

[0039] Example 3: Based on Example 2, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation. The data acquisition unit includes: The data preprocessing subunit is used for: The attitude time sequence, vibration characteristic parameters and thermal state index are obtained, and the basic value range of attitude time sequence, vibration characteristic parameters and thermal state index are obtained from the management terminal respectively. Based on the value range, the attitude time sequence, vibration characteristic parameters and thermal state index are filtered, and the filtering result is converted from analog to digital based on a preset analog-to-digital converter to obtain the corresponding standard digital signal. The time stamping subunit is used to obtain a unified time scale based on the management terminal and add timestamps to standard digital signals based on the unified time scale. The data output subunit is used to encapsulate the standard digital signal after adding a timestamp to obtain operational behavior modal data and device health modal data.

[0040] In this embodiment, the management terminal refers to an external configuration and management device used to provide various reference parameters to the data acquisition unit, such as the value range and time reference.

[0041] In this embodiment, the unified time scale refers to a high-precision time source provided by the management terminal or generated internally by the system, which serves as the time synchronization benchmark for the entire system.

[0042] In this embodiment, the standard digital signal refers to the digitized data generated from the original analog signal after filtering, range adjustment, and analog-to-digital conversion, which conforms to the internal definition specifications of the system.

[0043] The beneficial effects of the above technical solution are: by filtering, digitizing and standardizing the time scale of the raw sensor data, noise and dimensional differences are effectively eliminated, ensuring accurate time synchronization and format standardization of multi-source data, and providing high-quality and highly consistent standardized data input for subsequent correlation analysis.

[0044] Example 4: Based on Example 1, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, such as... Figure 2 As shown, the low-power Bluetooth transmission module includes: The transmission preparation unit is used to encapsulate the obtained operation behavior modal data and health modal data according to the communication protocol to obtain the data frame to be transmitted, extract the sensor terminal identifier corresponding to the data frame to be transmitted, and add the sensor terminal identifier to the frame header of the data frame to be transmitted. Link building unit, used for: Obtain the access control address, and initiate a connection request based on the access control address to establish a point-to-point wireless communication link; Based on the current wireless environment noise intensity and link quality indicators, the communication channel and transmit power level are adaptively selected. The data frames to be transmitted are loaded into the transmission buffer in sequence, and the transmission priority and time slot allocation of the data frames to be transmitted are dynamically scheduled according to the data generation rate of the operation behavior mode data and the health mode data. Based on the transmission priority, the cyclic redundancy check code of each data frame to be transmitted is determined and appended to the end of the data frame to be transmitted. The data transmission unit is used to transmit the data frame to be transmitted to the edge computing module based on a determined communication channel and transmission power level.

[0045] In this embodiment, the data frame to be transmitted refers to a data packet that can be wirelessly transmitted after packaging operation behavior modal data and health modal data according to a specific communication protocol format.

[0046] In this embodiment, the sensor terminal identifier refers to a code or number that uniquely identifies a miniaturized data acquisition terminal and is used to distinguish different data sources in the system.

[0047] In this embodiment, the access control address refers to the hardware address used in Bluetooth Low Energy communication to uniquely identify and address the peer device (such as an edge computing module).

[0048] In this embodiment, the transmit buffer refers to a memory area inside the Bluetooth Low Energy transmission module used for temporary storage and sorting of data frames to be transmitted.

[0049] In this embodiment, the transmission priority refers to the level assigned to different data frames to be transmitted based on factors such as the data generation rate, which is used to determine their transmission order.

[0050] In this embodiment, time slot allocation refers to a specific time segment allocated for transmitting data during wireless communication.

[0051] In this embodiment, the Cyclic Redundancy Check (CRC) code refers to a check value appended to the end of a data frame to verify at the receiving end whether an error has occurred during data transmission.

[0052] The beneficial effects of the above technical solution are: by adaptively selecting channels and power, and combining with a dynamic scheduling mechanism, it can effectively cope with complex wireless environments, significantly improve the stability and real-time performance of data transmission, and ensure the integrity of data and terminal identifiability through protocol encapsulation and verification mechanisms, providing a reliable and orderly data stream for subsequent edge analysis.

[0053] Example 5: Based on Example 4, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including a data transmission unit comprising: The monitoring subunit is used to start a preset response timer based on the transmission result of the data frame to be transmitted, and to monitor the reception confirmation frame fed back by the edge computing module in real time based on the start result. Data transmission subunit, used for: If a receive acknowledgment frame is received before the acknowledgment timer expires, and the acknowledgment frame indicates that the data frame to be transmitted is verified correctly, then the current data frame to be transmitted is removed from the transmit buffer, and the next data frame to be transmitted is sent. If the acknowledgment timer expires without receiving an acknowledgment frame or the received acknowledgment frame indicates a verification error in the data frame to be transmitted, the strategy of retransmitting the data frame to be transmitted or triggering a link reconnection is determined based on the relative size of the current number of consecutive retransmissions and the preset maximum retransmission threshold. Based on the time slot allocation results, the system dynamically switches to low-power monitoring or sleep mode according to a preset sleep strategy during the intervals between continuous data transmissions, until the next batch of data frames to be transmitted is ready.

[0054] In this embodiment, the acknowledgment timer refers to a timer that is started after a data frame is sent and is used to wait for acknowledgment feedback. If no acknowledgment is received within the timeout period, the corresponding processing logic is triggered.

[0055] In this embodiment, the receiving confirmation frame refers to a data packet indicating the receiving status that is fed back to the sending end by the edge computing module after successfully receiving and verifying the data.

[0056] In this embodiment, the number of consecutive retransmissions refers to the cumulative number of attempts to repeatedly send the same data frame.

[0057] In this embodiment, the maximum retransmission threshold refers to the preset maximum number of times the same data frame can be retransmitted.

[0058] In this embodiment, the link reconnection strategy refers to a set of predefined rules that interrupt the current communication link and attempt to re-establish the connection when data transmission failure reaches certain conditions.

[0059] In this embodiment, the sleep strategy refers to a set of rules that control the wireless transmission module to enter a low-power state during data transmission gaps, in order to balance real-time performance and power consumption.

[0060] In this embodiment, low-power listening refers to a working state that maintains wireless listening but with significantly reduced power consumption.

[0061] In this embodiment, the sleep state refers to an extremely low-power state in which most of the wireless transceiver circuits are turned off and only basic timing functions are maintained.

[0062] The beneficial effects of the above technical solution are: through the response confirmation and intelligent retransmission mechanism, reliable delivery of wireless data in complex environments is ensured, data loss is effectively avoided, and combined with the on-demand sleep strategy, the overall power consumption is significantly reduced while ensuring real-time transmission, thereby extending the terminal's battery life and enhancing the system's applicability in field or mobile scenarios.

[0063] Example 6: Based on Example 4, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including a link construction unit comprising: The signal measurement subunit is used for: After establishing a wireless communication link, a channel measurement request is sent to the edge computing module; Based on the channel measurement request, the signal strength and noise strength are measured on multiple predefined channels to obtain the received signal strength indication and ambient noise strength value of each channel; The channel determination subunit is used for: The signal-to-noise ratio (SNR) estimate for each channel is determined based on the received signal strength indication and the ambient noise intensity value, and selectable channels with an SNR higher than a preset threshold are selected. Select the channel with the fewest historical handovers and currently no interference flag from the available channels as the target communication channel; Transmit power determination subunit, used for: The initial transmit power level is determined by querying a pre-stored power lookup table based on the average received signal strength of the current wireless communication link. The target communication channel and the initial transmit power level are encapsulated as negotiation parameters and sent to the edge computing module. After the edge computing module confirms the negotiation parameters, it replies with a negotiation confirmation frame. After the negotiation confirmation frame is received, the communication link is switched to the target communication channel and the transmit power is adjusted to the initial transmit power level.

[0064] In this embodiment, a channel measurement request refers to a control command frame sent to the communication peer to initiate a multi-channel measurement process.

[0065] In this embodiment, a predefined channel refers to a set of radio frequency channels for measurement and communication that are pre-defined in the communication protocol standard or system configuration.

[0066] In this embodiment, the selectable channel refers to the set of channels that have undergone preliminary screening and whose signal-to-noise ratio meets the preset quality requirements.

[0067] In this embodiment, the target communication channel refers to the specific channel that is ultimately selected from the available channels for subsequent data communication.

[0068] In this embodiment, the power lookup table refers to a pre-stored lookup table that describes the recommended transmit power levels under different received signal strength conditions.

[0069] In this embodiment, the negotiation parameters refer to data packets containing the target communication channel and initial transmit power level, used for configuration synchronization with the peer device.

[0070] In this embodiment, the negotiation confirmation frame refers to the response frame sent by the edge computing module to confirm acceptance of the negotiation parameters and complete the link configuration.

[0071] The beneficial effects of the above technical solution are: by dynamically measuring the signal and noise quality of multiple channels and automatically selecting the best communication channel based on the signal-to-noise ratio and historical stability, and matching the appropriate transmission power according to the real-time link status, the adaptive capability and anti-interference capability of wireless connection in complex electromagnetic environments are significantly enhanced, ensuring the continuous stability and reliability of data transmission.

[0072] Example 7: Based on Example 1, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including an edge computing module, comprising: Receive operational behavior modality data and health modality data, and extract the corresponding timestamps from the operational behavior modality data and health modality data; Based on timestamps, nonlinear alignment is performed between operational behavior modal data and equipment health modal data, and attitude time series is extracted from operational behavior modal data based on the nonlinear alignment results. The attitude time series includes discrete data points of pitch angle, roll angle and yaw angle as they change over time. Based on the attitude time series, the first-order difference of each attitude angle between adjacent sampling points is determined to obtain the attitude angle change rate sequence. The attitude angle change rate sequence is then divided into sliding windows, and the mean and standard deviation of the attitude angle change rate are determined in each window to obtain the operation smoothness characteristics and operation volatility characteristics. A short-time Fourier transform is performed on the attitude time sequence, and the energy distribution ratio in the preset low-frequency band and high-frequency band is determined based on the transform result to obtain the operating frequency domain characteristics. Zero-crossing points of the attitude time sequence are detected based on the attitude angle change rate sequence, and the number of zero-crossing points per unit time is counted to obtain the operation frequency characteristics. Simultaneously, the vibration signal time sequence and temperature signal time sequence were extracted from the equipment health modal data; By analyzing the time series of vibration signals, the root mean square value and peak value of the vibration signal time series within a complete analysis window are determined, and the vibration intensity characteristics and impact characteristics are obtained. Bandpass filtering is performed on the vibration signal time sequence to retain a preset frequency band related to the natural frequency of the equipment's mechanical structure, and the power spectral density integral of the vibration signal time sequence within the preset frequency band is determined to obtain the resonant energy characteristics; By analyzing the time series of temperature signals, the first derivative of the time series of temperature signals with time is determined, the temperature change gradient sequence is obtained, and the number of consecutive occurrences of positive values ​​and the average slope are determined based on the temperature change gradient sequence to obtain the temperature rise activity characteristics. Based on the time series of temperature signals and the corresponding time series of vibration signals, the peak value of the cross-correlation function of the time series of temperature signals and vibration signals within the same time window is determined, and the thermal-vibration coupling characteristics are obtained. By combining the characteristics of smooth operation, volatile operation, frequency operation, frequency operation, vibration intensity, impact, resonant energy, temperature rise activity, and thermal-vibration coupling, we obtain the operation behavior feature vector and the equipment health feature vector.

[0073] In this embodiment, nonlinear alignment refers to the time-series matching process that uses algorithms such as dynamic time warping to non-uniformly scale or shift two time series in order to compensate for possible inconsistencies in time scales or local deformations between them.

[0074] In this embodiment, the attitude angle change rate sequence refers to a new time sequence that reflects the rate of change of attitude angle, obtained by differential calculation of adjacent data points in the attitude time sequence.

[0075] In this embodiment, the smoothness feature refers to an index used to quantify the smoothness of the operation, which is calculated by the window mean of the attitude angle change rate sequence.

[0076] In this embodiment, the operational volatility characteristic refers to an indicator used to quantify the magnitude of operational volatility, which is calculated using the window standard deviation of the attitude angle change rate sequence.

[0077] In this embodiment, the operation frequency domain feature refers to the index that characterizes the frequency characteristics of the operation action by analyzing the energy distribution of the attitude time sequence at different frequency components through frequency domain transformation.

[0078] In this embodiment, the temperature change gradient sequence refers to a new time sequence that reflects the rate of temperature change, obtained by performing a first-order difference on the time sequence of the temperature signal.

[0079] In this embodiment, the thermal-vibration coupling characteristic refers to an index used to quantify the degree of correlation between temperature change and mechanical vibration in the time domain, which is obtained by calculating the peak value of the cross-correlation function of the two signals.

[0080] The beneficial effects of the above technical solution are: by performing nonlinear time alignment on the dual-modal data and extracting multi-dimensional time domain, frequency domain and time series statistical features, a refined characterization of the operation behavior mode and equipment state changes is achieved, providing a high-information feature input for accurately establishing the dynamic correlation between the two in the future.

[0081] Example 8: Based on Example 1, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including an edge computing module, comprising: Feature processing unit, used for: Obtain the operation behavior feature vector and the equipment health feature vector based on feature extraction, and input the operation behavior feature vector and the equipment health feature vector into the pre-built correlation analysis model; Based on the correlation analysis model, the operational behavior feature vector and the equipment health feature vector are analyzed to determine the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector, and output the dynamic correlation strength matrix. The correlation analysis unit is used for: Based on the dynamic correlation strength matrix, a combination of correlation features that matches the preset equipment wear mode is determined, and a mapping function from operation behavior parameters to equipment wear parameters is constructed in real time based on the combination of correlation features. Based on the mapping function, the operation behavior feature vector and the equipment health feature vector of real-time input are forward extrapolated to generate the predicted value and confidence interval of the equipment loss parameter. At the same time, the historical operation sequence is obtained and the operation chain pattern that leads to abnormal loss is determined by combining the historical operation sequence. The dynamic correlation strength matrix, mapping function, predicted value, and operation chain pattern are encapsulated to form an equipment operation-health correlation analysis report.

[0082] In this embodiment, the mapping function is updated online using the recursive least squares method.

[0083] In this embodiment, the conditional dependency coefficient refers to a numerical index calculated by statistical or modeling methods to quantify the strength of the correlation between a certain operational behavior feature and a certain device health feature under given other characteristics.

[0084] In this embodiment, the dynamic association strength matrix refers to a two-dimensional matrix with operational behavior features as rows, device health features as columns, and matrix elements as corresponding conditional dependency coefficients, which changes with the time window update. It is used to represent the dynamic association pattern between the two modal features as a whole.

[0085] In this embodiment, the associated feature combination refers to a set of paired operational behavior features and equipment health features identified from the dynamic association strength matrix that have a high conditional dependency coefficient and match a certain known equipment wear pattern.

[0086] In this embodiment, the mapping function refers to a mathematical model (such as a linear or nonlinear function) that takes operational behavior parameters as input and equipment loss parameters as output, and its parameters are constructed and updated in real time based on the data of associated feature combinations.

[0087] In this embodiment, a forward extrapolation is performed on the real-time input operation behavior feature vector and the device health feature vector based on a mapping function to generate predicted values ​​and confidence intervals for device loss parameters, including: The system acquires real-time input operation behavior feature vectors and uses them as input variables for a mapping function. Based on the predetermined weight matrix and bias vector in the mapping function, it performs a linear transformation on the input variables to obtain intermediate hidden state vectors. These intermediate hidden state vectors are then activated through the output layer of the mapping function to generate initial predicted values ​​for equipment loss parameters. Based on the parameter covariance matrix recursively estimated in the model training and evolution units, and combined with the values ​​of the input variables, the prediction variance of the initial predicted values ​​is calculated. Using the initial predicted values ​​as the expected values, and multiplying the arithmetic square root of the prediction variance by a pre-set confidence coefficient as the interval half-width, the predicted values ​​and confidence intervals for the equipment loss parameters are constructed.

[0088] In this embodiment, the confidence interval refers to a range of values ​​given when predicting equipment loss parameters, used to represent the uncertainty or reliability of the predicted value.

[0089] In this embodiment, the operation chain pattern refers to a typical operation sequence or pattern extracted from historical operation sequences, consisting of a series of specific operations arranged in chronological order, and significantly related to abnormal loss results.

[0090] In this embodiment, the equipment operation-health correlation analysis report refers to a structured data output that integrates a dynamic correlation strength matrix, a mapping function, equipment loss prediction values ​​and their confidence intervals, and identified operation chain patterns to comprehensively reflect the correlation analysis results between operation and health status.

[0091] The beneficial effects of the above technical solution are as follows: by constructing and applying the correlation analysis model, the dynamic dependence between operational behavior and equipment wear can be quantified from the bimodal features, and the correlation strength matrix can be output in real time. Furthermore, based on the matrix, key feature combinations can be identified, parameter mapping functions can be constructed, and wear trends can be predicted. This transforms discrete operational data into interpretable and decision-supporting correlation analysis reports, realizing a closed loop from data collection to causal insight, and providing direct and quantitative basis for operation optimization and preventive maintenance.

[0092] Example 9: Based on Example 8, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including a feature processing unit comprising: Feature processing subunit, used for: Based on the correlation analysis model, a sliding window with a fixed time length is configured, and the operation behavior feature vector aligned with the current time and the device health feature vector are filled into the sliding window in chronological order; Within the sliding window, the operation behavior feature vector and the device health feature vector are standardized to obtain a standardized feature matrix. Based on the standardized feature matrix, the mutual information value between each operation behavior feature and each device health feature within the sliding window is determined. For feature pairs whose mutual information values ​​exceed a preset threshold, a state space model is constructed with the current operation behavior feature as input and the current device health feature as output, and the parameters of the state space model are predicted based on the operation behavior feature vector and the device health feature vector within the sliding window. Based on the parameters of the predicted state-space model, the conditional causal index of operational behavior characteristics on equipment health characteristics is determined, and the mutual information value and the conditional causal index are weighted and fused to obtain the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector. By iterating through all feature pairs, the mutual information value, state space model parameters, and conditional dependency coefficients corresponding to each feature pair are summarized to obtain the dynamic correlation strength matrix.

[0093] In this embodiment, the rows and columns of the dynamic association strength matrix correspond to the operation behavior features and the device health features, respectively, and the elements are the corresponding conditional dependency coefficients. The current matrix is ​​the dynamic association strength matrix of the current window.

[0094] In this embodiment, a sliding window refers to a data interval with a fixed time length that slides continuously on the time axis, used to contain feature vector data from the current moment and a period of time before in chronological order.

[0095] In this embodiment, the standardized feature matrix refers to the matrix formed by arranging the operation behavior feature vector and the device health feature vector within the sliding window according to the feature dimensions after standardization processing.

[0096] In this embodiment, the mutual information value refers to a dimensionless numerical value calculated within a sliding window to measure the degree of statistical dependence between an operational behavior feature and a device health feature.

[0097] In this embodiment, the state-space model refers to a mathematical model that includes state equations and output equations, constructed to describe the dynamic influence of an operational behavior feature on a device health feature.

[0098] In this embodiment, the conditional causality index refers to an indicator calculated based on the parameters of the state-space model, used to quantify the causal influence of the operational behavior feature on the health feature of the device within the sliding window.

[0099] The beneficial effects of the above technical solution are: by adopting a sliding window mechanism combined with mutual information filtering and state space modeling, it is possible to dynamically capture the statistical dependence and causal relationship between operational behavior and equipment health characteristics over time, thereby quantifying the dynamic relationship between the two more precisely, enhancing the timeliness and adaptability of correlation analysis, and providing a more reliable and sensitive quantitative basis for subsequent loss prediction and pattern recognition.

[0100] Example 10: Based on Example 1, this example provides a multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, including an edge computing module, comprising: The data packet construction unit is used to encapsulate the obtained dynamic correlation, original operational behavior modal data and health modal data, as well as the corresponding operational behavior feature vectors and device health feature vectors, to obtain the data packet to be uploaded; The result output unit is used to upload the data packet to be uploaded to the upper-level management terminal based on the communication interface.

[0101] In this embodiment, the data packet to be uploaded refers to a structured data set generated by the data packet construction unit, which contains multi-layer analysis results and raw data and is ready to be sent to the upper-layer system.

[0102] In this embodiment, the upper-level management terminal refers to a higher-level computing device or software platform located above this system, used to receive, store, display, or further process the data uploaded by this system.

[0103] The beneficial effects of the above technical solution are: by encapsulating and uploading the analysis results and raw data in a structured manner, multi-level aggregation and standardized output of information are achieved, providing a complete and traceable data foundation for upper-level systems to make decisions, archive and conduct in-depth analysis.

[0104] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-sensor fusion-based dual-modal health data acquisition system for equipment operation, characterized in that, include: Miniaturized data acquisition terminal, edge computing module and low-power Bluetooth transmission module; Among them, the miniaturized data acquisition terminal is fixedly installed on the operating mechanism of the monitored equipment, and contains an integrated sensor module. Integrated sensor modules are used to collect operational behavior modal data and health modal data in real time during equipment operation. A low-power Bluetooth transmission module is used to transmit operational behavior modal data and health modal data to the edge computing module; The edge computing module is used to perform time alignment and feature extraction on operational behavior modal data and health modal data, and to determine the dynamic correlation between equipment operational behavior parameters and equipment loss parameters based on the time alignment and feature extraction results.

2. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 1, characterized in that, Integrated sensor module, including: The sensor synchronization start-up unit is used to synchronously start the internally integrated inertial measurement unit, vibration sensor and temperature sensor after the integrated sensor module is powered on and initialized. The data acquisition unit is used for: Based on the first preset sampling frequency, the inertial measurement device is controlled to collect the original angular velocity and original acceleration data of the operating mechanism in three-dimensional space in real time, and the attitude angle data of the operating mechanism in pitch, roll and yaw directions are determined based on the original angular velocity and original acceleration data. The attitude angle data constitutes the attitude time sequence in the operating behavior modal data. Based on the second preset sampling frequency, the vibration sensor collects high-frequency vibration signals of the operating mechanism under mechanical transmission and load change states, and performs time-domain and frequency-domain analysis on the high-frequency vibration signals to extract vibration characteristic parameters. The vibration characteristic parameters constitute the mechanical state indicators in the equipment health modal data. Based on a third preset sampling frequency, the temperature sensor is controlled to monitor the surface temperature of the operating mechanism in real time. The surface temperature change is compared with a preset temperature baseline to obtain the real-time temperature rise rate and the duration of temperature exceeding the limit. The temperature rise rate and the duration of temperature exceeding the limit constitute the thermal state indicators in the equipment health modal data.

3. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 2, characterized in that, The data acquisition unit includes: The data preprocessing subunit is used for: The attitude time sequence, vibration characteristic parameters and thermal state index are obtained, and the basic value range of attitude time sequence, vibration characteristic parameters and thermal state index are obtained from the management terminal respectively. Based on the value range, the attitude time sequence, vibration characteristic parameters and thermal state index are filtered, and the filtering result is converted from analog to digital based on a preset analog-to-digital converter to obtain the corresponding standard digital signal. The time stamping subunit is used to obtain a unified time scale based on the management terminal and add timestamps to standard digital signals based on the unified time scale. The data output subunit is used to encapsulate the standard digital signal after adding a timestamp to obtain operational behavior modal data and device health modal data.

4. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 1, characterized in that, The low-power Bluetooth transmission module includes: The transmission preparation unit is used to encapsulate the obtained operation behavior modal data and health modal data according to the communication protocol to obtain the data frame to be transmitted, extract the sensor terminal identifier corresponding to the data frame to be transmitted, and add the sensor terminal identifier to the frame header of the data frame to be transmitted. Link building unit, used for: Obtain the access control address, and initiate a connection request based on the access control address to establish a point-to-point wireless communication link; Based on the current wireless environment noise intensity and link quality indicators, the communication channel and transmit power level are adaptively selected. The data frames to be transmitted are loaded into the transmission buffer in sequence, and the transmission priority and time slot allocation of the data frames to be transmitted are dynamically scheduled according to the data generation rate of the operation behavior mode data and the health mode data. Based on the transmission priority, the cyclic redundancy check code of each data frame to be transmitted is determined and appended to the end of the data frame to be transmitted. The data transmission unit is used to transmit the data frame to be transmitted to the edge computing module based on a determined communication channel and transmission power level.

5. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 4, characterized in that, The data transmission unit includes: The monitoring subunit is used to start a preset response timer based on the transmission result of the data frame to be transmitted, and to monitor the reception confirmation frame fed back by the edge computing module in real time based on the start result. Data transmission subunit, used for: If a receive acknowledgment frame is received before the acknowledgment timer expires, and the acknowledgment frame indicates that the data frame to be transmitted is verified correctly, then the current data frame to be transmitted is removed from the transmit buffer, and the next data frame to be transmitted is sent. If the acknowledgment timer expires without receiving an acknowledgment frame or the received acknowledgment frame indicates a verification error in the data frame to be transmitted, the strategy of retransmitting the data frame to be transmitted or triggering a link reconnection is determined based on the relative size of the current number of consecutive retransmissions and the preset maximum retransmission threshold. Based on the time slot allocation results, the system dynamically switches to low-power monitoring or sleep mode according to a preset sleep strategy during the intervals between continuous data transmissions, until the next batch of data frames to be transmitted is ready.

6. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 4, characterized in that, Link building units include: The signal measurement subunit is used for: After establishing a wireless communication link, a channel measurement request is sent to the edge computing module; Based on the channel measurement request, the signal strength and noise strength are measured on multiple predefined channels to obtain the received signal strength indication and ambient noise strength value of each channel; The channel determination subunit is used for: The signal-to-noise ratio (SNR) estimate for each channel is determined based on the received signal strength indication and the ambient noise intensity value, and selectable channels with an SNR higher than a preset threshold are selected. Select the channel with the fewest historical handovers and currently no interference flag from the available channels as the target communication channel; Transmit power determination subunit, used for: The initial transmit power level is determined by querying a pre-stored power lookup table based on the average received signal strength of the current wireless communication link. The target communication channel and the initial transmit power level are encapsulated as negotiation parameters and sent to the edge computing module. After the edge computing module confirms the negotiation parameters, it replies with a negotiation confirmation frame. After the negotiation confirmation frame is received, the communication link is switched to the target communication channel and the transmit power is adjusted to the initial transmit power level.

7. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 1, characterized in that, The edge computing module includes: Receive operational behavior modality data and health modality data, and extract the corresponding timestamps from the operational behavior modality data and health modality data; Based on timestamps, nonlinear alignment is performed between operational behavior modal data and equipment health modal data, and attitude time series is extracted from operational behavior modal data based on the nonlinear alignment results. The attitude time series includes discrete data points of pitch angle, roll angle and yaw angle as they change over time. Based on the attitude time series, the first-order difference of each attitude angle between adjacent sampling points is determined to obtain the attitude angle change rate sequence. The attitude angle change rate sequence is then divided into sliding windows, and the mean and standard deviation of the attitude angle change rate are determined in each window to obtain the operation smoothness characteristics and operation volatility characteristics. A short-time Fourier transform is performed on the attitude time sequence, and the energy distribution ratio in the preset low-frequency band and high-frequency band is determined based on the transform result to obtain the operating frequency domain characteristics. Zero-crossing points of the attitude time sequence are detected based on the attitude angle change rate sequence, and the number of zero-crossing points per unit time is counted to obtain the operation frequency characteristics. Simultaneously, the vibration signal time sequence and temperature signal time sequence were extracted from the equipment health modal data; By analyzing the time series of vibration signals, the root mean square value and peak value of the vibration signal time series within a complete analysis window are determined, and the vibration intensity characteristics and impact characteristics are obtained. Bandpass filtering is performed on the vibration signal time sequence to retain a preset frequency band related to the natural frequency of the equipment's mechanical structure, and the power spectral density integral of the vibration signal time sequence within the preset frequency band is determined to obtain the resonant energy characteristics; By analyzing the time series of temperature signals, the first derivative of the time series of temperature signals with time is determined, the temperature change gradient sequence is obtained, and the number of consecutive occurrences of positive values ​​and the average slope are determined based on the temperature change gradient sequence to obtain the temperature rise activity characteristics. Based on the time series of temperature signals and the corresponding time series of vibration signals, the peak value of the cross-correlation function of the time series of temperature signals and vibration signals within the same time window is determined, and the thermal-vibration coupling characteristics are obtained. By combining the characteristics of smooth operation, volatile operation, frequency operation, frequency operation, vibration intensity, impact, resonant energy, temperature rise activity, and thermal-vibration coupling, we obtain the operation behavior feature vector and the equipment health feature vector.

8. The multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 1, characterized in that, The edge computing module includes: Feature processing unit, used for: Obtain the operation behavior feature vector and the equipment health feature vector based on feature extraction, and input the operation behavior feature vector and the equipment health feature vector into the pre-built correlation analysis model; Based on the correlation analysis model, the operational behavior feature vector and the equipment health feature vector are analyzed to determine the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector, and output the dynamic correlation strength matrix. The correlation analysis unit is used for: Based on the dynamic correlation strength matrix, a combination of correlation features that matches the preset equipment wear mode is determined, and a mapping function from operation behavior parameters to equipment wear parameters is constructed in real time based on the combination of correlation features. Based on the mapping function, the operation behavior feature vector and the equipment health feature vector of real-time input are forward extrapolated to generate the predicted value and confidence interval of the equipment loss parameter. At the same time, the historical operation sequence is obtained and the operation chain pattern that leads to abnormal loss is determined by combining the historical operation sequence. The dynamic correlation strength matrix, mapping function, predicted value, and operation chain pattern are encapsulated to form an equipment operation-health correlation analysis report.

9. A multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 8, characterized in that, Feature processing unit, including: Feature processing subunit, used for: Based on the correlation analysis model, a sliding window with a fixed time length is configured, and the operation behavior feature vector aligned with the current time and the device health feature vector are filled into the sliding window in chronological order; Within the sliding window, the operation behavior feature vector and the device health feature vector are standardized to obtain a standardized feature matrix. Based on the standardized feature matrix, the mutual information value between each operation behavior feature and each device health feature within the sliding window is determined. For feature pairs whose mutual information values ​​exceed a preset threshold, a state space model is constructed with the current operation behavior feature as input and the current device health feature as output, and the parameters of the state space model are predicted based on the operation behavior feature vector and the device health feature vector within the sliding window. Based on the parameters of the predicted state-space model, the conditional causal index of operational behavior characteristics on equipment health characteristics is determined, and the mutual information value and the conditional causal index are weighted and fused to obtain the conditional dependency coefficient between the operational behavior feature vector and the equipment health feature vector. By iterating through all feature pairs, the mutual information value, state space model parameters, and conditional dependency coefficients corresponding to each feature pair are summarized to obtain the dynamic correlation strength matrix.

10. A multi-sensor fusion-based dual-modal health data acquisition system for equipment operation according to claim 1, characterized in that, The edge computing module includes: The data packet construction unit is used to encapsulate the obtained dynamic correlation, original operational behavior modal data and health modal data, as well as the corresponding operational behavior feature vectors and device health feature vectors, to obtain the data packet to be uploaded; The result output unit is used to upload the data packet to be uploaded to the upper-level management terminal based on the communication interface.

Images