Self-Powered Wireless Sensor Platform for Online Motor Condition Monitoring at 5G Era

Nowadays, for protecting the environment, targets of automotive CO2 emission reduction and fuel efficiency are set by many countries. Because of these aggressive CO2 emission reduction plans and customers’ preference, automotive manufactures spend more resources into vehicle electrification. Consequently, the market of Battery Electric Vehicles (BEV), Hybrid Electric Vehicles (HEV), and Plug-in Hybrid Electric Vehicles (PHEV) as well as the respective semiconductor market grows. These facts indicate more and more resources will be invested into research in EV industry. Safety is one of the most important aspect in a vehicle. Besides driver safety, vehicle conditions monitoring functions are important as well. In order to prevent severe accidents caused by motor fault, monitoring of the power train, especially electric motor itself, is necessary. In addition, technologies on level 5 autonomous drive are significantly developed. In order to increase the reliability of self-driving vehicles and to reduce maintenance costs, online motor condition monitoring and fault diagnosis are also desired here. Methods for motor monitoring and fault diagnosis are developed extensively in the literature since the inception of electric machinery. Traditionally, these methods have been proven to be partially effective and accurate. However, these methods usually have many drawbacks. In order to remove the limitations, many efforts have been spent to apply machine learning and artificial-intelligence (AI) technologies to motor fault prediction and condition monitoring. However, these methods require strong computational platforms and large storage capability, as well as data-intensive communication capability. Powerful on-board microcontroller units (MCUs) are required for these applications. However, this introduces additional costs for MCUs, data storages, and communications equipment. Recently, the internet of things (IoT) introduces methods that can improve these drawbacks of AIbased motor condition monitoring. To implement these methods, 5G communication, which provides high data-rates, wide bandwidth, and URLLC, can support the advanced motor condition monitoring methods. In these motor condition monitoring methods, only the current spectrum data can be directly measured by the motor drive system, and the remaining methods require additional sensors. Therefore, self-powered multi-signal wireless sensors are desired. For these monitoring methods, the data sources can also produce power ranging from micro-watts to milli-watts through transducers. In this work, a general model for self-powered sensor platform is proposed and studied. By combining information theory, circuit theory, and energy model, the physical limitation and relationship among charge-discharge ratio, energy input, and information output is illustrated. An example of self-powered multi-signal sensing system is built for concept verification and model explanation. The present research shows that self powered sensing platforms can not only support the conditioning monitoring of the motors but also can provide a scalable solution for many other sensor intensive problems that are used in IoT systems.