State based monitoring (CbM) helps prevent device failures through predictive maintenance, but designing an effective system typically requires optimized integration of precision sensing, low-noise signal chains, power management, and wireless connectivity. These are complex features that may delay the deployment of CbM and increase costs. Designers also recognize the advantages of edge artificial intelligence (AI) analysis, but this also makes CbM more complex. We need to find a more direct and effective solution.
This article first briefly introduces proximity sensors, and then introduces Analog Devices' plug and play solutions. This solution enables immediate deployment of wireless CbM with edge AI capabilities.
The Importance of State Monitoring
Unplanned downtime remains a major challenge in maintaining high operational efficiency of equipment. Once an unexpected failure occurs in critical equipment, it may lead to the paralysis of the entire production line, interruption of the supply chain, and expensive maintenance services. Traditional maintenance methods include passive repair after a failure or strict periodic maintenance, but these methods have their drawbacks: passive maintenance can lead to costly downtime, while periodic maintenance can increase resource costs by unnecessarily replacing still running components.
Adopting CbM enables the implementation of more cost-effective predictive maintenance methods. By monitoring vibration, temperature, current, or other performance indicators, equipment operators can identify warning signals of component performance degradation before faults occur. This data-driven approach can reduce unplanned downtime, extend equipment lifespan, and lower total cost of ownership.
Despite the numerous advantages of CbM, its deployment may come to a standstill due to its complex requirements and the need for interdisciplinary expertise. For the industrial and automotive fields, overcoming these challenges is a major challenge in successfully applying CbM based predictive maintenance.
Challenges and requirements brought by state based monitoring
To fully leverage the potential advantages of CbM, CbM solutions must operate reliably in harsh industrial and automotive environments, while conducting timely analysis based on accurate measurement data. However, even during the normal operation of the monitored equipment, these specific operating conditions can subject the measuring equipment to enormous environmental and mechanical pressures. Industrial motors, transmission systems, and heavy rotating equipment can continuously expose monitoring devices to vibration, shock, extreme temperatures, and high levels of electromagnetic interference (EMI).
In order to achieve reliable predictive maintenance, vibration sensors in CbM devices must be able to detect finer changes, which are often the earliest clues of shaft imbalance, misalignment, or bearing wear. To ensure high-precision vibration measurement under harsh environmental conditions, a high bandwidth, low-noise sensor signal acquisition subsystem is required, which can maintain stable performance in harsh working environments.
As the core of the CbM method, vibration analysis lays the foundation for identifying patterns that can distinguish between normal operation and early signs of failure. In the past, vibration sensor systems transmitted measurement results to a central host or cloud resources for analysis. However, advanced CbM solutions have begun to increasingly shift analysis capabilities to the edge. By analyzing data within or near the sensor system, results can be obtained in the shortest possible time and reduce traffic in time sensitive industrial and automotive networks.
Specifically, edge AI inference based on convolutional neural network (CNN) models can provide real-time interpretation of vibration changes. However, using CNN for inference requires a significant amount of computation, making it more complex to achieve CbM goals without exceeding system power, size, or cost limitations.
With the increasing use of CbM in rotating devices, remote or mobile devices, and the impracticality of wired connections, minimizing power consumption has become more urgent. To meet the wireless connection requirements in these situations, Low Energy Bluetooth (BLE) can achieve the required combination of transmission distance, power, and reliability compared to other optional connection technologies (Table 1).
However, like edge AI processing, the challenge we face is finding a BLE connectivity solution that can operate normally within the power limitations of wireless sensor systems. In fact, ensuring extended battery life remains a challenge for any wireless sensor system designer. However, this is particularly important in industrial and automotive applications, where sensors may be difficult to reach. In CbM systems that require CNN inference, battery and power management are becoming increasingly important. The challenge in this regard is how to coordinate multiple voltage regulators, sequencers, and charging systems to reduce power consumption while ensuring stable operation.
The evaluation kit provides an embedded wireless CbM solution with edge AI functionality
The EV-CBM-VOYAGER4-1Z Voyager 4 kit from Analog Devices provides a complete battery powered vibration monitoring platform for continuous evaluation of CbM technology or immediate deployment in predictive maintenance applications, addressing various challenges faced when deploying wireless CbM with edge AI capabilities. This kit adopts a vertical support (Figure 1, top), firmly fixing the main printed circuit board (PC board) on one side and the battery on the other side to eliminate the impact of harsh environments. The power circuit board and sensor are located at the bottom of the support, close to the vibration source to be monitored. For ease of deployment, the vertical support components are placed inside an aluminum protective cover with a diameter of 46 mm and a height of 77 mm (Figure 1, bottom). The top of the protective cover is equipped with an ABS acrylic cover, which can be used for BLE connection.