An integrated fiber Michelson interferometer (MI) based on an asymmetrical twin-core fiber (TCF) cascaded with a side-hole fiber (SHF) for multi-parameter sensing is proposed and demonstrated. The asymmetrical TCF consists of a centric core and an eccentric core with same refractive index. The centric core of the TCF is aligned with the core of the SHF and the eccentric core is aligned with one air hole of the SHF. Because the eccentric core of the TCF is off the axis of the fiber, a bend or twist will cause a variation of the optical path difference (OPD) of the MI, which makes the MI structure be suitable for twist and vector curvature measurements. Experimental results show that the bending sensitivities of the sensor are -6.968 nm/m-1 and 6.978 nm/m-1 at the bending directions of 0° and 180° in the curvature ranges of 0 m-1 to 10.708 m-1, respectively, and the twist sensitivity of the sensor is 0.639 nm/(rad/m) in the twist ranges of 0 rad/m to 7.44 rad/m. The structure is also sensitive to strain and temperature, and the strain and temperature sensitivities of the interferometer are 1.36 pm/ $\mu\epsilon$ and 10.37 pm/°C, respectively.

We report on current theoretical and experimental results of hydroacoustic sensing array based on ultra-weak Fiber Bragg Gratings (FBG), using a modified phase generated carrier (PGC) demodulation method with a self-referencing signal. The self-referencing signal is obtained by a sensor isolated from acoustic signals and other environmental disturbances. We report improvements over the conventional PGC methods. Using our demodulation method and with non-sensitized bare fiber (reference sensor0) keeping separate from water, experiment testing demonstrates a minimum detectable hydroacoustic pressure of 2239 μPa/sqrtHz. The properties of our demodulation method are also compared with those of the conventional PGC algorithms. Both simulation and experiments indicate that our demodulation method is immune to the drifts of modulation depth C and of carrier frequency, and the detectable frequency range can be increased by 5 times compared to using conventional PGC methods with the same carrier frequency. The sensor array exhibits a particularly large response in the very low frequency (VLF) region, which is of great importance for underwater seismic detection and submarine applications.

We demonstrate interrogation of a large-capacity sensor array with nearly identical weak fiber Bragg gratings (FBGs) based on frequency-shifted interferometry (FSI). In contrast to time-division multiplexing, FSI uses continuous-wave light and therefore requires no pulse modulation or high-speed detection/acquisition. FSI utilizes a frequency shifter in the Sagnac interferometer to encode sensor location information into the relative phase between the clock-wise and counter-clockwise propagating lightwaves. Sixty-five weak FBGs with reflectivities in the range of −31 ~-34 dB and with near identical peak reflection wavelengths around 1555 nm at room temperature were interrogated simultaneously. Temperature sensing was conducted and the average measurement accuracy of the peak wavelengths was ± 3.9 pm, corresponding to a temperature resolution of ± 0.4 °C. Our theoretical analysis taking into account of detector noise, fiber loss, and sensor cross-talk noise shows that there exists an optimal reflectivity that maximizes multiplexing capacity. The multiplexing capacity can reach 3000 with the corresponding sensing range of 30 km, when the peak reflectivity of each grating is −40 dB, the sensor separation 10 m and the source power 14 mW. Experimental results and theoretical analysis reveal that FSI has distinct cost and speed advantages in multiplexing large-scale FBG networks.

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.

We combine wavelength-division multiplexing (WDM) and frequency shifted interferometry (FSI) to interrogate a large-scale ultra-weak fiber Bragg grating (FBG) array. Based on Sagnac interference, FSI can location resolve sensors without using pulses or fast detection, therefore considerably lowers the system cost. By combining FSI with WDM, higher spatial resolution can be achieved. We demonstrated simultaneous interrogation of 363 FBG sensors, grouped into 121 identical units of 3 FBGs of different central wavelengths and spaced two meters apart. Based on the performance of the 363 grating system, we show the potential for interrogating 3207 sensors with good signal-to-noise ratio. Stability test and temperature sensing were carried out, and the obtained temperature resolution was ± 0.4 ° C. The results indicate the proposed scheme can greatly enhance the multiplexing capacity and meet the requirements of large-scale optical fiber networks.