Condition Monitoring Sensor and System Failure Considerations

　　First, before implementing condition monitoring, it is necessary to determine which indicators are most critical for evaluating the condition of the equipment. This usually includes but is not limited to physical quantities such as temperature, pressure, and vibration level. After the monitoring items are selected, sensors or other types of data acquisition devices will be installed at the corresponding locations. These devices can capture the required information with high accuracy and convert it into digital signals for further processing.

　　Next, the collected data is sent to the central control system via wired or wireless means. Advanced communication protocols are used here to ensure the security and accuracy of data transmission. Once the data reaches the server, a series of complex algorithms are activated to analyze it. The purpose of data analysis is to identify any trends or patterns that may indicate an impending failure. For example, if the vibration amplitude of a machine suddenly increases, it may be a sign of severe wear of internal parts; abnormally high temperatures often mean that there is a problem with the cooling system.

　　When a potential problem is found, the condition monitoring system will immediately sound an alarm to remind relevant personnel to intervene in the investigation and solve the problem in a timely manner. In addition, based on historical records and current conditions, the system can also predict the problems that may occur in the future and their likelihood, helping companies make more reasonable maintenance plans. This proactive maintenance strategy based on condition monitoring has obvious advantages over traditional regular maintenance methods: it can not only reduce unnecessary downtime, but also effectively extend the service life of equipment.

　　In short, condition monitoring technology provides users with a scientific and effective health management solution by continuously tracking the changes in important performance parameters. With the continuous improvement of IoT technology and big data analysis capabilities, the application scope of condition monitoring will be further expanded, becoming an important force in promoting the intelligent transformation of various industries.

　　‌Data processing and analysis technology‌: The acquired data needs to be processed and analyzed to extract useful information and features. Common data processing and analysis methods include filtering, noise reduction, feature extraction, fault diagnosis, etc. Filtering can remove noise and interference in the data and retain valid information. Noise reduction is to further reduce useless information in the data and improve the reliability and accuracy of the signal through signal processing methods. Feature extraction is to extract useful features from the original data. Commonly used methods include Fourier transform, wavelet transform, time-frequency analysis, etc. Fault diagnosis is to determine whether the object or system has a fault and locate the cause of the fault by analyzing the anomalies in the data‌1.

　　‌Data display and monitoring technology‌: Data display and monitoring is the visualization and management of condition monitoring results. Through appropriate charts and interfaces, the status information and change trends of objects or systems can be intuitively displayed. Data monitoring is to analyze and judge the monitoring data in real time, and when abnormal conditions are found, timely alarms or corresponding measures can be taken. Data display and monitoring can not only facilitate users to understand and manage data, but also help them quickly respond to and handle changes in the status of objects or systems‌1.

　　‌Communication and Data Analysis‌: Before implementing condition monitoring, it is necessary to determine which indicators are most critical for evaluating the condition of the equipment. These indicators usually include physical quantities such as temperature, pressure, and vibration level. After the monitoring project is selected, sensors or other types of data acquisition devices are installed at the corresponding locations. These devices can capture the required information with high accuracy and convert it into digital signals for further processing. The collected data is sent to the central control system via wired or wireless means, using advanced communication protocols to ensure the security and accuracy of the data transmission process. Once the data arrives at the server, a series of complex algorithms are launched to analyze it and identify any trends or patterns that may indicate an impending failure‌2.

‌Artificial Intelligence　　and Machine Learning‌: With the development of technology, artificial intelligence and machine learning have been introduced into condition monitoring, enabling the system to autonomously learn and identify the normal working mode of the equipment, and promptly detect signals that deviate from the normal mode, thereby predicting and warning of potential equipment failures. This intelligent development makes condition monitoring more accurate and efficient‌3.

　　In vibration detection, both bearing defects and gear defects require the vibration sensor to have noise controlled below 100 µg/√Hz, and the bandwidth needs to be above 5kHz, otherwise the faults on the bearings and gears cannot be sensed, and bearing defect detection requires an additional higher sensor g value range. Because this type of fault is not very obvious at the beginning, especially in the early stage, it is difficult to identify simply by increasing the vibration frequency. It is necessary to pair a vibration sensor with low noise and wide bandwidth with a high-performance signal chain, processing, transceiver, and post-processor for complete monitoring.

　　Moderate noise requirements above 100 µg/√Hz apply to unbalance and motor misalignment detection, which require bandwidths of 5× to 10× the fundamental frequency. Unbalance and motor misalignment faults require additional multi-axis detection, where unbalance faults require low-frequency response for slow-moving machines.

　　How to design a small PoDL circuit?

　　Coupled inductors and series capacitors are the basic components required for PoDL to work. In addition, other components are required to ensure robustness and fault tolerance.

　　Because the PoDL coupled inductor is a non-ideal component, some degree of differential-to-common-mode conversion occurs. This common-mode noise can degrade the signal quality. Connecting a common-mode choke close to the cable connector can help mitigate this non-ideality and protect the design from the cable common-mode noise. It is also necessary to check the common-mode choke ampacity and DCR to ensure it can provide enough power to the sensor.

　　To prevent incorrect polarity cable installation (for example, connecting the 24 VDC of the PSE PHY to the 0 VDC of the PD PHY), it is recommended to use a bridge rectifier diode. To ensure EMC robustness, use a TVS diode with a working voltage greater than 24 VDC. If the sensor hardware design is larger, other EMC components can also be used (for example, high-voltage capacitors on the signal line).

　　The best PdM strategy is to effectively use as many technologies and sensors as possible to detect faults early with high confidence, without the need for a one-size-fits-all sensor solution. This article attempts to explain why predictive maintenance sensors are critical for early detection of faults in PdM applications and what their advantages and disadvantages are.

　　Predictive maintenance sensors

　　Some sensors are able to detect certain faults earlier than others, such as bearing damage, as shown in Figure 1. In this section, we will discuss the sensors that are commonly used to detect faults as early as possible, typically accelerometers and microphones. Table 1 shows a list of sensor specifications and some of the faults they can detect. Most PdM systems only use some of these sensors, so it is important to ensure that in addition to using the right sensors to detect these potentially critical faults, you also have a deep understanding of these faults.

　　Sensor and System Failure Considerations

　　More than 90% of rotating machines in industrial and commercial applications use rolling element bearings. 3 The distribution of faulty components in an electric motor is shown in Figure 2, from which it is clear that special attention needs to be paid to bearing monitoring when selecting a PdM sensor. In order to detect, diagnose and predict potential failures, vibration sensors must have low noise and wide bandwidth.