This article is devoted to the development of a classification method based on an artificial neural network architecture to solve the problem of recognizing the sources of acoustic influences recorded by a phase-sensitive OTDR. At the initial stage of signal processing, we propose the use of a band-pass filter to collect data sets with an increased signal-to-noise ratio. When solving the classification problem, we study three widely used convolutional neural network architectures: AlexNet, ResNet50, and DenseNet169. As a result of computational experiments, it is shown that the AlexNet and DenseNet169 architectures can obtain accuracies above 90%. In addition, we propose a novel CNN architecture based on AlexNet, which obtains the best results; in particular, its accuracy is above 98%. The advantages of the proposed model include low power consumption (400 mW) and high speed (0.032 s per net evaluation). In further studies, in order to increase the accuracy, reliability, and data invariance, the use of new algorithms for the filtering and extraction of acoustic signals recorded by a phase-sensitive reflectometer will be considered.

We demonstrated a fiber optic distributed acoustic sensor based on a double Sagnac interferometer, using two wavelengths separated by CWDM modules. A mathematical model of signal formation principle, based on a shift in two signals analysis, was described and substantiated mathematically. The dependence of the sensor sensitivity on a disturbance coordinate and frequency was found and simulated, and helped determine a low sensitivity zone length and provided sensor scheme optimization. A data processing algorithm without filtering, appropriate even in case of a high system noise level, was described. An experimental study of the distributed fiber optic sensor based on a Sagnac interferometer with two wavelengths divided countering loops was carried out. An accuracy of 24 m was achieved for 25.4 km SMF sensing fiber without phase unwrapping.

We present a theoretical and experimental study in which we increased the sensitivity of a phase-sensitive optical time-domain reflectometer (phi-OTDR). This was achieved by constructing coils in the sensor cable, which increased the total amplitude of the impact on the fiber. We demonstrate this theoretically using the example of a phase-sensitive reflectometer model and practically in testing grounds with a buried nearby conventional sensor and a sensor with coils. The sensitivity increased 2.2 times. We detected 95% of events when using coils, versus 20% when using a straight cable. The suggested method does not require any modifications to the device.

In this study, an experimental study of the burning rate of solid fuel in a model solid propellant rocket motor (SRM) E-5-0 was conducted using a non-invasive control method with fiber-optic sensors (FOSs). Three sensors based on the Mach–Zehnder interferometer (MZI), fixed on the SRM E-5-0, recorded the vibration signal during the entire cycle of solid fuel burning. The results showed that, when using MZI sensors, the non-invasive control of solid fuel burnout is made possible both by recording the time of arrival of the combustion front to the sensor and by analyzing the peaks on the spectrogram of the recorded FOS signal. The main mode of acoustic vibrations of the chamber of the model SRM is longitudinal, and it changes with time, depending on the chamber length. Longitudinal modes of the combustion chamber were detected by MZI only after the combustion front passed its fixing point, and the microphone was unable to register them at all. The results showed that the combustion rate was practically constant after the first second, which was confirmed by the graph of the pressure versus time at the nozzle exit.

The paper presents the application of a phase-sensitive optical time-domain reflectometer (phi-OTDR) in the field of urban infrastructure monitoring. In particular, the branched structure of the urban network of telecommunication wells. The encountered tasks and difficulties are described. The possibilities of usage are substantiated, and the numerical values of the event quality classification algorithms applied to experimental data are calculated using machine learning methods. Among the considered methods, the best results were shown by convolutional neural networks, with a probability of correct classification as high as 98.55%.

In recent years, attention to the realization of a distributed fiber-optic microphone for the detection and recognition of the human voice has increased, whereby the most popular schemes are based on φ-OTDR. Many issues related to the selection of optimal system parameters and the recognition of registered signals, however, are still unresolved. In this research, we conducted theoretical studies of these issues based on the φ-OTDR mathematical model and verified them with experiments. We designed an algorithm for fiber sensor signal processing, applied a testing kit, and designed a method for the quantitative evaluation of our obtained results. We also proposed a new setup model for lab tests of φ-OTDR single coordinate sensors, which allows for the quick variation of their parameters. As a result, it was possible to define requirements for the best quality of speech recognition; estimation using the percentage of recognized words yielded a value of 96.3%, and estimation with Levenshtein distance provided a value of 15.