【Intelligent Monitoring System for Motor Equipment】Work Submission

极致创客

【Intelligent Monitoring System for Motor Equipment】Work Submission [Copy link]

 

Intelligent monitoring system for motor equipment

Author: Extreme Maker

At present, the faults of motor equipment are mainly discovered through manual inspection or during major and medium repairs, which has certain errors and lags. In view of the above problems, this work designs a set of intelligent monitoring system for motor equipment.

This system is mainly composed of a data acquisition board and an intelligent monitoring center. The data acquisition board consists of modules such as temperature and humidity, vibration signal acquisition, status monitoring, power metering, communication, and display, which realize functions such as remote control of equipment, signal processing, data display, and parameter upload. As the center of intelligent monitoring, the STM32H7B3-DK development board receives data transmitted by the data acquisition board. The deep learning model deployed to the hardware platform can independently judge the received data, intelligently analyze the health status of the equipment, diagnose abnormal motor faults, and display them in different forms such as charts and curves.

The STM32H7B3-DK is an ARM-based microcontroller that provides 2M bytes of FLASH memory and 1.4MB RAM, 512-Mbit Octo-SPI NOR flash memory, and 128-Mbit SDRAM, which is fully capable of processing the microcontroller's algorithm for sensor data. The algorithm deployment platform is shown in Figure 1.

Figure 1 Algorithm deployment platform STM32H7B3I-DK

The design work is shown in Figure 2.

Figure 2 Design drawing of the work

The algorithm deployment function is shown in Figure 3.

Figure 3 Temperature rise trend prediction

2 System Block Diagram
1 Data Acquisition Board Hardware Block Diagram

2 Data acquisition board hardware design diagram

3 PCB manufacturing of data acquisition board

3. Functional description of each part

1 Electrical parameter measurement

The single-phase AC/DC adaptive energy metering module IM1266 is a module developed by Shenzhen Airuida Optoelectronics Co., Ltd. for product power consumption monitoring. It can measure electrical data such as voltage, current, power, power factor, frequency, etc., and communicate with the microcontroller through the serial port to realize automatic data collection and monitoring functions.

2 Temperature and humidity measurement

TH10S-B is a high-precision temperature and humidity transmitter with beautiful and compact appearance. It has super stability and anti-interference ability, strong product protection performance, and first-level lightning protection. RS485 interface allows multiple modules to be connected to the bus network to monitor the environment of multiple sites in real time.

Command Examples

Receive data code

3 Data acquisition board GUI display

4 Upper test interface

5. Wiring diagram for vibration signal collection

6. Introduction to Fault Diagnosis Model

STM32Cube.AI is an advanced toolkit launched by ST that can interoperate with popular deep learning libraries to convert and apply any artificial neural network to STM32 microcontrollers (MCUs). With STM32Cube.AI, edge IoT devices based on STM32 MCUs can now run neural networks directly, perform real-time AI calculations at the edge and respond instantly, thereby protecting privacy, reducing network bandwidth usage and consuming a lot of computer power.

STM32Cube.AI enables embedded systems to build a neural network in just five steps.

STM32Cube.AI supports fast and automatic import of neural networks trained by popular design frameworks such as Keras, TensorFlow-Lite, Caffe, Lasagne, and ConvnetJS.

Adding a neural network model

Generate an STM32 project, which contains the library file of the neural network, as well as: network_data.c: the weight information of the entire neural network, which represents the size and computational complexity of the entire network, network.c: this file is a callable API, which includes the initialization, creation, and operation of the neural network.

The verification results are shown in the figure, including the structure of the neural network, MACC, memory usage, etc.

Network model parameters

4. Source Code

The code is in the following link:

https://download.eeworld.com.cn/detail/eew_02EYVp/625208

https://download.eeworld.com.cn/detail/eew_02EYVp/625207

5. Demonstration video of the work’s functions

Demonstration of the upper test process: https://www.bilibili.com/video/BV1d84y1z7bR/

GUI display demonstration: https://www.bilibili.com/video/BV1q14y1L7RV/

6. Project Summary

This design has basically achieved the expected goal. Thanks to the strong support of Digi-Key Electronics and EEWORLD, the work was completed. Through this design, I have a deeper understanding of embedded development, completed hardware production and software testing, and realized the overall function. Due to the time constraints, there is still a lot of work to be done in the algorithm, and we will continue to work hard to realize all the functions of fault diagnosis.

The shared post link is as follows:

【2022 Digi-Key Innovation Design Competition】Material Unboxing STM32H7B3-DK - 2022 Digi-Key Electronics Innovation Design Competition - Electronic Engineering World - Forum (eeworld.com.cn)

[Design of online monitoring system] TouchGFX touch point based on STM32H7B3 -DK ... - 2022 Digi-Key Innovation Design Competition - Electronic Engineering World - Forum (eeworld.com.cn)

7. Innovation

1 For the traditional cloud-based mechanical fault monitoring system, edge computing is used to deploy the fault diagnosis model to the edge to achieve lower power consumption and minimal latency.

2. Give full play to the advantages of edge computing and deep migration models, establish an efficient online monitoring system, and achieve high diagnostic accuracy.