DESIGN OF MEASURING INSTRUMENT OF GLUCOSE CONCENTRATION IN SOLUTION USING NEAR-INFRARED SPECTROSCOPY (NIRS) METHOD

• Main Author: Sanjaya, Christian
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/68530
• Institution: Institut Teknologi Bandung

Diabetes mellitus is a long-term disease resulting from disturbances in insulin secretion. Based on data from the Ministry of Health, Indonesia is ranked seventh out of ten countries with the highest number of people with diabetes mellitus worldwide. These data show that one of the leading causes of the increasing number of people with diabetes mellitus in Indonesia is the consumption pattern of people who like to consume sweet drinks. As many as 61.3% of respondents stated that the frequency of consuming sweet beverages in one day was more than once. Through the Ministry of Health, the Indonesian government recommended a daily sugar intake limit of 50 g/ml. In contrast(Johnson dkk., 2009), the average sugar content in packaged drinks is 20 grams per serving. The consumption pattern of sugary beverages needs to be controlled to reduce risk factors for diabetes by limiting daily sugar intake. Therefore we need a hardware device to measure glucose levels in the solution. One method to measure the level of glucose concentration in a solution is by using the near-infrared spectroscopy (NIRS) method. The simple working principle of NIRS is to measure the absorption rate of the sample. The NIRS hardware design uses two primary circuits: signal acquisition and signal conditioning. The signal acquisition circuit consists of two circuits: the transmitter and the receiver. The transmitter circuit uses three infrared light-emitting diodes (LEDs) at 940 nm, 1200 nm, and 1450 nm. The receiver circuit used three photodiodes with a wavelength range of 800-1700 nm and an additional circuit that can convert current into voltage. The signal conditioning circuit consists of an active filter circuit with a cut-off frequency of 50 Hz. Signal conditioning data will be hardware-read data. NIRS hardware testing is tested by analyzing the reading data on the concentration of the solution. The analysis was carried out by least square analysis and the correlation coefficient calculation. The results of least squares show that the normalized root means square (RMS) error data is 0.2023 (940 nm), 0.1162 (1200 nm), and 0.0828 (1450 nm). The results of the data correlation coefficient calculation are 0.6567 (940 nm), 0.9057 (1200 nm), and 0.9480 (1450 nm), and the results of the analysis of the coefficient of determination were 43.1% (940 nm), 82% (1200 nm), and 89.9% (1450 nm). The NIRS hardware test data analysis results show a sufficient correlation at the 940 nm wavelength, and a strong relationship occurs at the 1200 nm and 1450 nm wavelengths. The analysis of test data using 29 sample solutions found that among the three wavelengths used in NIRS hardware, the wavelength of 1450 nm is the best wavelength used to detect glucose levels in the solution.