In deep learning-based fault diagnosis of the wind turbine gearbox, a commonly faced challenge is the domain shift caused by differing operational conditions. Traditional domain adaptation methods aim to learn transferable features from the source domain and apply them to the target data. However, such methods still require access to target domain data during the training process, which limits their applicability in real-time fault diagnosis. To address this issue, we introduce an adversarial single-domain generalization network (ASDGN). It relies solely on data from a single length of data acquisition in wind turbine fault diagnosis. This novel approach introduces a more flexible and efficient solution to the field of real-time fault diagnosis for wind turbines.

Most of the sensors currently used for wind turbine generator system oil detection can only perform measurements of a single physical quantity, and some sensors are capable of measuring multiple contaminants but require switching modes to achieve this, making it impossible to detect various oil contaminants simultaneously. These methods not only require the addition of numerous sensors and burden the wind turbine generator system piping, but also increase the time and cost of maintenance and inspection. In addition, it may be unable to respond quickly to specific emergencies. To address these limitations, this study proposes an LC-based passive dual solenoid coil multi-contaminant oil detection sensor. An analytical expression for the magnetic field strength of the dual solenoidal coil is derived, and a corresponding mathematical model is developed. The validity of the model is verified compared it with the finite element analysis results. An experimental verification was carried out. The experimental results show that the sensor can recognize multiple oil pollutants simultaneously according to different signal curves without switching modes. The sensor can detect iron particles with diameters larger than 38 μm, copper particles with diameters larger than 100 μm, liquid droplets with diameters larger than 100-110 μm, and air bubbles with diameters larger than 180-190 μm. By comparing with the existing sensors, the LC-based passive dual-coil multi-pollutant oil detection sensor proposed in this paper has a broad application prospect in real-time monitoring of wind turbine oil conditions.

Hydraulic oil is the key medium for the normal operation of hydraulic machinery, which carries various wear debris. The information reflected by the wear debris can be used to predict the early failure of equipment and achieve predictive maintenance. In order to realize the real-time condition monitoring of hydraulic oil, an impedance debris sensor that can detect inductance and resistance parameters is designed and studied in this paper. The material and size of wear debris can be discriminated based on inductance-resistance detection method. Silicon steel strips and two rectangular channels are designed in the sensor. The silicon steel strips are used to enhance the magnetic field strength, and the double rectangular detection channels can make full use of the magnetic field distribution region, thereby improving the detection sensitivity and throughput of the sensor. The comparison experiment shows that the coils in series are more suitable for the monitoring of wear debris. By comparing and analyzing the direction and the presence or absence of the signal pulses, the debris sensor can detect and distinguish 46 µm iron particles and 110 µm copper particles. This impedance detection method provides a new technical support for the high-precision distinguishing measurement of metal debris. The sensor can not only be used for oil detection in the laboratory, but also can be made into portable oil detection device for machinery health monitoring.

Wear debris in the oil contains a wealth of information about the friction pairs of the mechanical equipment. By analyzing the size and type of wear debris through oil detection technology, condition monitoring and fault diagnosis of mechanical systems can be realized. This paper presents an inductive sensor based on differential detection and its signal conditioning circuit, which can detect metal wear debris in the oil. The sensor adopts the structure of two induction coils embedded in one excitation coil. The differential signal is obtained by reverse connecting two induction coils with the same parameters, which can suppress the common mode interference and eliminate the influence of ambient noise so that the sensor has extremely low noise. Through the designed signal conditioning circuit, the detection signal is phase-sensitive detected, and the information of wear debris is extracted by amplification and filtering. In this paper, the sensing principle of the sensor is derived, the spacing between the two induction coils is optimized using finite element simulation, and the optimal excitation frequency, detection limit, and detection error of the sensor are investigated through experiments. The experiment results show that the sensor can detect 20 μm iron particles and 130 μm copper particles in a 2 mm flow channel, and the detection error of the sensor is less than 22%. The sensor has the advantages of simple structure and high sensitivity and can be applied to detect metal wear debris in hydraulic oil.

Oil is rich in information about the operating status of machinery and equipment, and oil condition monitoring is essential for maintaining the safety and reliability of machinery and equipment. Most sensors not only have the problem of small detection throughput and low detection efficiency, but also large-sized contaminants can easily block or even scratch the detection channel. A sensor with high sensitivity and high throughput for detecting multiple oil contamination is designed in this article. The sensor comprises four glass capillary detection channels and four detection units, including a planar coil, solenoid coil, cylindrical parallel plate capacitor, and rectangular parallel plate capacitor. Detection flux of the sensor is increased by 12 times its initial value, which significantly improves the detection efficiency. Through experimental research, in the inductance mode, the best excitation frequency is 2 MHz, and the voltage is 2 V. In the capacitance mode, the best excitation frequency is 1.8 MHz, and the voltage is 2 V. In addition, the sensor has a higher detection capability for Fe than Cu in the inductance mode, and a higher detection capability for water droplets than air bubbles in the capacitance mode. By analyzing the signal characteristics of each sensor unit, the sensor can distinguish between Fe of 38– $900 ~\mu \text{m}$ and Cu of 88– $900 ~\mu \text{m}$ , as well as water droplets greater than $130 ~\mu \text{m}$ and air bubbles greater than $100 ~\mu \text{m}$ . This research provides a new idea for online oil monitoring.

Oil is crucial for the efficient operation of mechanical equipment as it transfers energy, lubricates, and cools the system. Real-time monitoring of the mechanical system’s condition requires analyzing abrasive particles in the oil through oil detection technology. However, during detection, these particles tend to adhere together, leading to signal overlap and misjudgments resulting in losses. To address this issue, this paper proposes a resistive oil detection sensor and a wireless signal transmission circuit. The sensor detects agglomerated metal abrasive particles in the oil, while the wireless circuit transmits resistance signals from the detection coil to the receiving coil, enabling the identification of metal particles in the oil. The study explores the superposition characteristics of mixed signals generated by metal particles, examining resistance signals of individual metal particles and mixed signals produced by particles of different sizes, magnetism, and adhesion. The results demonstrate that the sensor can detect 55 μm iron particles and 138 μm copper particles in a microfluidic channel measuring 1 mm. The amplitude of the superposition signal from two particles with the same magnetism exceeds that of a single particle of the same size, and the superposition resistance signal increases with particle size. In the case of two particles with different magnetisms, the superposition resistance signal exhibits a double-peak pattern. The sensor distinguishes various types of adhered magnetic metal particles by recognizing distinct resistance signal patterns, providing valuable insights into identifying the signal characteristics of overlapping particles.