Prediction of remaining useful life (RUL) is greatly significant for improving the safety and reliability of manufacturing equipment. However, in real industry, it is difficult for RUL prediction models trained on a small sample of faults to obtain satisfactory accuracy. To overcome this drawback, this paper presents a long short-term memory (LSTM) neural network with transfer learning and ensemble learning and combines it with an unsupervised health indicator (HI) construction method for remaining-useful-life prediction. This study consists of the following parts: (1) utilizing the characteristics of deep belief networks and self-organizing map networks to translate raw sensor data to a synthetic HI that can effectively reflect system health; and (2) introducing transfer learning and ensemble learning to provide the required degradation mechanism for the RUL prediction model based on LSTM to improve the performance of the model. The performance of the proposed method is verified by two bearing datasets collected from experimental data, and the results show that the proposed method obtains better performance than comparable methods.

Ocean temperature monitoring is of great significance to marine fishing, aquaculture, and marine operations. Traditional electric sensors lack the potential to multiplex several sensors, and may suffer from electromagnetic interference. Meanwhile, fiber Bragg grating-based sensors have the advantages of high sensitivity, possibility for large-scale multiplexing, and immunity to electromagnetic interference. In this paper, we propose a Fabry–Pérot (FP) interferometer based on the draw tower grating array and combine it with the phase measurement method for demonstration and testing. In the sensor system, two adjacent fiber Bragg gratings (FBGs) are used as mirrors and an optical fiber connects them, forming a sensor unit. The signal was detected through the compensation of the optical path difference via two-arm path differences in an unbalanced interferometer. The sensor is calibrated in the range of 36.00–36.50 °C, and back to 36.00 °C, in steps of 0.10 °C. A thermocouple (DW1222) is used as a reference. Experimental testing demonstrates that under the thermal loop, the temperature and phase can be approximated as a linear relationship, the Pearson square correlation coefficient is 0.9996, and the temperature sensitivity is −9846 rad/°C. To prove that our experimental device can achieve a higher temperature resolution, we measured the background noise of the system. The experimental results indicate that the order of magnitude of our system temperature resolution can reach 10⁻⁵ °C. Thus, we believe that the sensor system is promising for the application of ocean temperature detection, and owing to the ultraweak reflection characteristics of the FBG, this method provides the possibility for large-scale multiplexing of the system.

The demodulation algorithm based on 3 × 3 coupler in a fiber-optic hydrophone array has gained extensive attention in recent years. The traditional method uses a circulator to construct the normal path-match interferometry; however, the problem of increasing the asymmetry of the three-way signal to be demodulated is easily overlooked. To provide a solution, we report a pretreatment method for hydrophone array based on 3 × 3 coupler demodulation. We use cubic spline interpolation to perform nonlinear fitting to the reflected pulse train and calculate the peak-to-peak values of the single pulse to determine the light intensity compensation coefficient of the interference signal, so as to demodulate the corrected three-way interference signal. For experimental verification, ultra-weak fiber Bragg gratings (uwFBGs) with reflectivity of −50 dB are applied to construct a hydrophone array with 800 sensors, and a vibratory liquid column method is set up to generate a low-frequency hydroacoustic signal. Compared to the traditional demodulation algorithm based on a 3 × 3 coupler, the pretreatment method can improve the consistency of interference signals. The Lissajous figures show that cubic spline interpolation can improve the accuracy of monopulse peak seeking results by about 1 dB, and intensity compensation can further lead to a much lower noise density level for the interference pulse amplitude—specifically, more than 7 dB at 5~50 Hz—and the signal-to-noise ratio is improved by approximately 10 dB at 10 Hz. The distinct advantages of the proposed pretreatment method make it an excellent candidate for a hydrophone array system based on path-match interferometry.

In this paper, a fiber optic sensor with draw-tower-grating (DTG)-based Fabry-Perot interferometers (FPIs) are proposed to measure the temperature and salinity of seawater simultaneously. The sensing structure utilizes DTG as the reflector, and two adjacent gratings can be cascaded to compose an FPI. By employing moisture-sensitive materials to the surface of the optical fiber, the variation in salinity can be measured by the axial strain applied by the coating on the sensing optical fiber. The salinity altering causes swelling or shrinking actions of the coating through absorbing or releasing water. Two different moisture-sensitive materials are chosen to eliminate the cross-sensitivity of temperature and salinity. The experimental results show that the sensor exhibits linear response to temperature and salinity changes with good repeatability and stability, and the temperature and salinity sensitivities are 12308.65 rad/°C and 95.02 rad/‰, respectively. The optical path configuration of the DTG sensing array matching compensation interferometer provides the possibility for efficient distributed solutions, which has promising potential for application in marine engineering.

In this paper, a dual-axis Fabry–Pérot (FP) accelerometer assembled on single optical fiber is proposed. The sensor is equipped with a special beam-splitting prism to split the light into two perpendicular directions (the X- and Y-axes); the prism surface coated with semi-permeable film and the reflective sheet on the corresponding Be-Cu vibration-sensitive spring form two sets of FP cavities of different sizes. When the Be-Cu spring with a proof mass (PM) is subjected to the vibration signal, the cavity length of the corresponding FP cavity is changed and the interference signal returns to the collimator through the original path of the prism. After bandpass filtering and demodulation, the two cavity lengths are obtained, and the acceleration measurement in dual-axis directions is completed. The resonant frequency of the proposed dual-axis fiber optic accelerometer is around 280 Hz. The results of the spectral measurements show 3.93 $\mathsf{\mu }\mathrm{m}/\mathrm{g}$ (g = 9.8 m/s²: gravity constant) and 4.19 $\mathsf{\mu }\mathrm{m}/\mathrm{g}$ for the applied acceleration along the X- and Y-axes, respectively, and the cross-axis sensitivity is below 5.1%. Within the angle range of 180°, the maximum error of measured acceleration is less than 3.77%. The proposed fiber optic dual-axis FP accelerometer has high sensitivity and strong immunity to electromagnetic interference. The size of the sensor mainly depends on the size of the prism, which is easy to reduce and mass produce. Moreover, this FP construction method has high flexibility and development potential.

The temperature resolution of Fiber Optic Sensor is greatly limited by background noise. In this paper, we propose a noise suppression method, based on variational mode decomposition (VMD), to enhance the temperature resolution. The VMD algorithm is employed to decompose the original signal into several intrinsic mode function (IMF) components, each limited by a certain bandwidth. We calculate the correlation coefficient between each IMF component and the original sensing signal, and filter out IMFs with correlation coefficients less than the average, which primarily contain noise signal. The IMFs with correlation coefficients greater than the average are retained for signal reconstruction. The experimental and comparative results demonstrate the exceptional performance of the proposed algorithm in denoising temperature signals, effectively removing background and environmental noise while preserving temperature information. Furthermore, by analyzing the power spectrum of the temperature signal before and after noise suppression, we observe that the proposed algorithm can achieve a temperature resolution of 10 ⁻⁶ °C, which is two orders of magnitude higher than that before noise suppression (10 ⁻⁴ °C). These results validate the efficacy of our method in reducing system noise and improving the accuracy and reliability of interferometric temperature sensors.