Investigation On Refractive Index Sensing Of New Multi-core Asymmetric Tapered Fiber

Refractive index has always been an important parameter in the biochemical industry,so how to measure the refractive index easily has always been a concern.In the actual measurement process,the temperature of the solution cannot be kept constant.So in order to improve the measurement accuracy,a method to eliminate the influence of temperature should be found.In addition,sensitivity is also an important parameter to evaluate the performance of the sensor.Generally,the higher the sensitivity,the better the sensor.Based on the above needs and facts,this paper mainly carried out the following research.The paper mainly has three parts.The first part introduces the three-core asymmetric tapered fiber Mach-Zehnder interferometer sensor we designed and manufactured using the cascade and tapering fiber method.The optimal parameter of the interferometer length is 8.31 mm.The distance between the first tapering and the second tapering is 3 mm.Its purpose is to improve the sensitivity to a certain extent and produce different coupling parameters,to further ensure that the temperature and the concentration of glucose solution can be measured simultaneously,thereby decreasing the impact of temperature to measurement of glucose solution concentration.Since the length of the cascade structure sensor is only increased by 1.31 mm after the tapering is drawn,the smaller taper‘s waist diameter of the manufactured sensor is 74.06 um,which means that the sensor is durable.The three-core asymmetric double-taper fiber sensor made by this method conforms to Mazer interference sensing,so the principle of Mazer interference sensing of multi-core fiber is directly used.Then,the response of the three-core asymmetric biconical sensor to changes on temperature and glucose solution concentration was measured.This paper not only gives the demodulation matrix when temperature and glucose solution concentration are measured simultaneously,but also gives the theoretical formula.The main content of the second part is to study the sensor characteristics of the four-core cascade fiber sensor and compare it with the sensor characteristics of the three-core fiber sensor of the first part.First,the manufacturing method of the new four-core tapered fiber sensor is introduced,and then the transmission spectrum at different temperatures and glucose solution concentrations is measured at small intervals.Extracting the interference peak wavelength and the relative power of light transmission,We studied relationships between the temperature and the glucose solution concentrations.The third part establishes a three-layer BP neural network to post-process theexperimental data of the three-core optical fiber sensor,and can predict the output of input values that are not in the sample.The paper's innovations:(1)Innovatively proposed the introduction of asymmetric double tapers on the three-core tapered fiber to form different coupling parameters,thereby realizing multi-parameter simultaneous sensing.(2)As long as the straight state of the multi-core fiber sensor is ensured,the characteristics of the sensor can have good repeatability.It is easier to manufacture and not easy to break than sensors that need to maintain a special shape.(3)Using the BP neural network to process the sensor data can accurately predict the corresponding solution concentration and temperature of data points that has not been seen by the network during training.The sensitivity of the three-core asymmetric tapered fiber sensor proposed in this paper is-31.01 pm /?,-22.12 pm / ?and 195.67 pm /%,162.31 pm /%.In contrast,the temperature sensitivity of the four-core tapered fiber sensor is at least 4.7 times the sensitivity of the three-core sensor.But at the same time,the new three-core fiber sensor has a large measurement range,the structure is more difficult to break.Futhermore,the new three-core fiber sensor can also measure simultaneously two parameters,which can overcome cross-sensitivity among multiple parameters in fibre optic interferometer sensing,so we chose to apply the three-core asymmetric tapered fiber for practical application.In order to obtain more accurate measurement results,the experimental data of the three-core asymmetric tapered fiber was processed by BP.Compared with the known samples obtained by the measurement,the prediction results show that the accuracy is higher than the demodulation matrix,which further shows that using neural networks for data reprocessing has certain practical significance.