CN-121985358-A - Real-time monitoring and dynamic analysis system supporting hot plug of multiple Bluetooth sensors

CN121985358ACN 121985358 ACN121985358 ACN 121985358ACN-121985358-A

Abstract

The invention discloses a real-time monitoring and dynamic analysis system supporting hot plug of a plurality of Bluetooth sensors, which is used for solving the problems of unstable connection, incapability of dynamic maintenance and high-load data processing delay during collaborative monitoring of a plurality of wireless sensors. The system distributes independent data channels for each triaxial acceleration Bluetooth sensor through a dynamic resource isolation module, realizes disconnection automatic reconnection through a connection state management module, ensures continuous monitoring, decouples data acquisition and calculation analysis through a logic decoupling processing pipeline module and a unified calculation resource scheduling module, performs frequency domain and time domain feature extraction of multiple devices in parallel based on a thread pool, and realizes dynamic binding and contrast display between any on-line sensor data and display views in operation through a dynamic visual mapping module. The invention realizes multi-equipment hot plug management, continuous non-loss acquisition under high sampling rate and efficient parallel analysis, and improves the reliability, flexibility and real-time performance of the monitoring system.

Inventors

Wu Ruyue
TANG XIAOLI
Yao Longda
WANG MU
XU YUANDONG
GU FENGSHOU

Assignees

湖南科技大学

Dates

Publication Date: 20260505
Application Date: 20260122

Claims (10)

1. A real-time monitoring and dynamic analysis system supporting hot plug of multiple bluetooth sensors, comprising: The dynamic resource isolation module is configured to dynamically allocate and maintain a set of mutually independent data receiving threads and a multi-level buffer structure for each successfully accessed triaxial acceleration Bluetooth sensor so as to realize the physical isolation of data flows among different sensors, and automatically destroy all corresponding exclusive resources when the sensors are removed; A connection state management module configured to automatically add the connected sensor to a retry queue and perform a limited number of delayed reconnection attempts to achieve thermal recovery of the connection when an abnormal disconnection of the connected sensor is detected; A logic decoupling processing pipeline module connected with the connection state management module and configured to decouple the data sampling receiving process from each sensor from the subsequent computation intensive analysis process so that the data sampling continuity is not affected by the instantaneous fluctuation of the back-end computation load; The unified computing resource scheduling module is connected with the logic decoupling processing pipeline module and is configured to maintain a shared thread pool and used for receiving and executing parallel computing tasks from each sensor, wherein the computing tasks comprise fast Fourier transformation and time domain feature extraction of peak-to-peak value, standard deviation and skewness; The dynamic visualization mapping module is configured to provide a plurality of independent display views and allow the dynamic binding or unbinding of the data stream of any online sensor to a designated display view during the operation of the system so as to realize real-time contrast display.
2. The system of claim 1, wherein the multi-level buffer structure in the dynamic resource isolation module comprises an independent raw data buffer, an X/Y/Z three-way data buffer, and an independent spectrum calculation buffer, which are independent of each other and are sequentially arranged according to data processing stages; the independent original data buffer zone is configured as a first-in first-out byte stream buffer zone and is used for receiving data frames which are not analyzed and come from a Bluetooth communication link so as to absorb wireless link jitter and avoid data loss caused by upper layer processing delay; The X/Y/Z three-channel data buffer area is configured into a plurality of mutually independent double-end queues corresponding to the physical channels of the triaxial acceleration Bluetooth sensor one by one, and is used for analyzing the load of the data frame into time domain signal data of a plurality of physical channels after the data frame passes the verification, and respectively storing the time domain signal data after the conversion of the measuring range and the sensitivity is completed, so as to be used as a data source for time domain waveform display and time domain characteristic calculation; The independent spectrum calculation buffer is configured to buffer time domain data blocks to be subjected to frequency domain analysis and to be continuously filled in a sliding update or cyclic coverage manner so as to provide data input with consistent length and continuous time to a frequency domain analysis task, thereby supporting stable execution of real-time spectrum calculation.
3. The system of claim 1, wherein the connection state management module employs a progressive delay strategy when performing delayed reconnection attempts, and maintains a separate retry counter for each device, stopping reconnection and triggering device removal procedures after a maximum number of attempts is reached.
4. The system of claim 1, wherein the logic decoupling processing pipeline module, upon decoupling the data sample receiving process from the subsequent analysis processing process, is further configured to: The occupation state of each data buffer structure is monitored in real time, and the trigger time or the concurrency quantity of the calculation tasks is dynamically adjusted according to the occupation degree of the buffer structure, so that the concurrency strength of the calculation tasks is reduced when the load of the calculation resources is high or the buffer occupation is close to a threshold value, and the continuity of the data sampling process and the stability of the whole operation of the system are ensured.
5. The system of claim 1, wherein the uniform computing resource scheduler module is further configured to dynamically adjust the allocation of computing tasks based on the implicit state control signal of the user interface to a particular data channel, and to suspend or cancel execution of computing tasks to the channel in the implicit state.
6. The system of claim 1, wherein the dynamic visualization mapping module responds to user operations via an exclusive selection control in the user interface by immediately establishing an exclusive data path from a sensor to a view when the user selects to bind the sensor to the view and automatically disconnecting the view from any previously bound sensors.
7. The system of any one of claims 1 to 6, further comprising a non-blocking advanced analytical processing module configured to: Starting a background analysis thread in a window independent of a system main interface aiming at a sensor currently selected by the dynamic visual mapping module; In the background analysis thread, performing Hilbert transform on the time domain signal of the sensor to obtain an envelope signal and an envelope spectrum; and displaying the analysis result in the independent window in real time, and providing an interaction control for a user to adjust the display frequency range of the frequency spectrum.
8. A method of real-time monitoring and dynamic analysis based on the system of any one of claims 1 to 7, comprising the steps of: step S1, equipment dynamic access and resource isolation, namely, responding to a Bluetooth sensor connection request, and dynamically creating an independent data receiving thread and a multistage buffer structure for a sensor after connection establishment is successful; Step S2, connection state maintenance and heat recovery, namely automatically adding the sensor into a retry queue and executing delayed reconnection attempt if abnormal disconnection of the sensor is detected during system operation; Step S3, continuously collecting and checking high sampling rate data, namely continuously receiving data at a set sampling rate through the independent data receiving thread and the buffer structure, checking a frame structure of each received data packet, discarding invalid data packets, absorbing instantaneous delay caused by wireless link jitter through the buffer structure, and ensuring that effective data points are not lost; step S4, decoupling and parallel computing, namely packaging the ready data acquired and checked in the step S3 into a computing task, wherein the computing task comprises fast Fourier transform and calculation of peak-to-peak value, standard deviation and skewness, and submitting the computing task to a shared thread pool for parallel computing to realize decoupling of a data acquisition flow and a computing analysis flow; s5, extracting and dynamically displaying temperature data, namely extracting a temperature value from a data frame which passes verification, converting the temperature value, and dynamically updating the temperature value to a designated position of a main interface according to a sensor access sequence; Step 6, visual dynamic mapping and display, namely establishing or updating a dynamic binding relation between the sensor and the display view according to a user interaction instruction, and pushing data or a calculation result of the corresponding sensor to the bound view in real time for display; Step S7, equipment thermal removal, namely responding to a sensor removal instruction or a final failure of reconnection, sequentially executing and terminating a data receiving thread, destroying a buffer structure, releasing a visual binding relation and releasing all associated resources; step S8, on-demand advanced analysis, wherein envelope analysis is performed on currently selected sensors in an independent analysis environment in response to a user request, and the results are presented in an independent window.
9. The method according to claim 8, wherein in the step S3, the system can realize the high-frequency non-drop acquisition of 16kHz of single device and 4kHz of double devices, and the key is that the data competition between the devices is isolated by the independent buffer structure established for each sensor, and meanwhile, the calculation task is unloaded to the thread pool by the step S4, so that the calculation delay is avoided from blocking the data acquisition process.
10. The method of claim 8, further comprising the step of data persistence: creating a separate data storage file for each connected sensor; writing the original data received and checked in the step S3, the processing results generated in the steps S4 and S5 and the corresponding time stamps into corresponding files according to a preset format; and when the device heat removal is executed in the step S7, safely closing and storing the data file corresponding to the sensor.

Description

Real-time monitoring and dynamic analysis system supporting hot plug of multiple Bluetooth sensors Technical Field The invention belongs to the technical field of state monitoring of industrial equipment, and particularly relates to a real-time monitoring and dynamic analysis system supporting hot plug of a multi-Bluetooth sensor, which is suitable for state monitoring of rotary machinery. Background The state monitoring of the wind turbine generator gearbox is a key link for guaranteeing the safe and stable operation of wind power equipment. The traditional monitoring scheme is mainly divided into a wired mode and a wireless mode, and has obvious defects. The deployment process of the wired sensor is complex, cables are required to be laid in the gear box, the unit is required to be stopped for hours or even days for transformation, and the cost is high. However, the existing wireless sensor scheme avoids physical wiring, but faces serious challenges in the scenes with complex electromagnetic environments such as wind power towers. The dense metal structure in the tower barrel can generate strong multipath interference and shielding effect on wireless signals such as Bluetooth, so that the connection is unstable when the multiple sensors work concurrently, and abnormal disconnection frequently occurs. The existing system lacks an effective connection state maintenance and automatic recovery mechanism, and once the connection is broken, manual intervention is needed to restart, so that the continuity of state monitoring is seriously damaged. In addition, as the monitoring demand is developed to the direction of multiple measuring points and high precision, the deep contradiction of the traditional system architecture is increasingly highlighted. The solidified equipment management logic cannot support the dynamic addition and deletion of the sensor in the running state of the system, namely, hot plug is not supported, and the additionally arranged measuring points must be stopped and reset. At the data processing level, the system cannot efficiently cooperate with data streams generated by a plurality of high sampling rate sensors in parallel, and data loss or update delay is often caused by uneven distribution of computing resources or link blockage. The multi-channel data are mixed and displayed, the key fault characteristic frequency is easy to be submerged, and when the computation intensive analysis such as frequency spectrum, envelope and the like is carried out, the main thread is easy to be blocked, and interface blocking is caused. These defects not only affect the timely capture and accurate diagnosis of early weak fault characteristics of the gearbox, but also restrict the construction process of an intelligent and predictive maintenance system of the wind farm. Therefore, a new monitoring system is needed that can adapt to complex interference environments, support hot plug and connection self-recovery of devices, and realize parallel and reliable processing and collaborative analysis of multi-device data. Disclosure of Invention In order to solve the technical problems, the invention provides a real-time monitoring and dynamic analysis system supporting hot plug of multiple Bluetooth sensors. The technical scheme for solving the technical problems is that the system for supporting the real-time monitoring and dynamic analysis of the hot plug of the multi-Bluetooth sensor comprises: The dynamic resource isolation module is configured to dynamically allocate and maintain a set of mutually independent data receiving threads and a multi-level buffer structure for each successfully accessed triaxial acceleration Bluetooth sensor so as to realize the physical isolation of data flows among different sensors, and automatically destroy all corresponding exclusive resources when the sensors are removed; A connection state management module configured to automatically add the connected sensor to a retry queue and perform a limited number of delayed reconnection attempts to achieve thermal recovery of the connection when an abnormal disconnection of the connected sensor is detected; A logic decoupling processing pipeline module connected with the connection state management module and configured to decouple the data sampling receiving process from each sensor from the subsequent computation intensive analysis process so that the data sampling continuity is not affected by the instantaneous fluctuation of the back-end computation load; The unified computing resource scheduling module is connected with the logic decoupling processing pipeline module and is configured to maintain a shared thread pool and used for receiving and executing parallel computing tasks from each sensor, wherein the computing tasks comprise fast Fourier transformation and time domain feature extraction of peak-to-peak value, standard deviation and skewness; The dynamic visualization mapping module is configured to provide a plurality