Document Type : Research Paper
Authors
Mechanical Engineering of Biosystems Department, Faculty of Agriculture, Shahrekord University, Shahrekord, Iran
Abstract
Keywords
Main Subjects
EXTENDED ABSTRACT
Checking food quality is an important issue in today's food industry, and various researches have been conducted in this regard. Due to the fact that food products are of general use and have many customers, this importance is doubled. The machine olfaction is one of these tools in checking the quality of food. Many of the devices and methods currently in use are located at a fixed location, meaning that the sample must be specifically transported from the production environment or factory to the laboratory location. Therefore, it is necessary to build a portable device that can transmit information wirelessly to overcome these problems and eliminate the unnecessary costs of testing. These items can help monitor food quality more closely.
The printed circuit board was designed in Proteus software. In this board, a 5V line was used to feed the sensors' heater and a 3.3V line was used to obtain the signals. In this device, two ADS1115 modules and one sht20 temperature and humidity sensor that use the I2C protocol are used at the same time. For this purpose, each of the two adc modules must have a specific address to communicate with the main board. In the written program, receiving data from sensors is done with the I2C communication protocol. The voltage of the sensors is calculated and converted into digital values by the ADS1115 analog-to-digital converter. These converters send data to the ESP32 board through two SCL and SDA lines, and the central board receives the data and processes it to be sent by Wi-Fi.
The process of measuring the odor and receiving the signal of the gas sensors was timed in three stages: baseline correction, gas injection of the space above the milk sample into the sensor chamber, and cleaning the sensor chamber. At first, oxygen gas was transferred to the sensor chamber for 125 seconds, and at the end of this time, the output signal of all sensors was considered as the base signal. Linear discriminant analysis was used in order to investigate the differentiation between pure milk and formalin adulterated groups. This method is similar to PCA as a feature reduction method that determines the upper plane with a smaller dimension and on which points with higher dimensions are imaged.
For LDA methods, two main components and second order method were used. The LDA method has the ability to distinguish pure milk from adulterated formalin samples with different percentages with an accuracy of 83%. The tests performed on milk samples containing adulterated formalin showed the quality of the system's information transmission and the positive performance of the device. The evaluation of the wireless smell machine device for food products will be further elaborated in future research.
Mahdi Ghasemi-Varnamkhasti: Writing – Supervision, Original draft, Methodology.
Zahra Izadi: Methodology, Experiments design, and Resources.
Data available on request from the authors. All the data used in this original research are presented throughout the text and in the form of Tables and Figures.
The authors avoided data fabrication, falsification, plagiarism, and misconduct.
This research has received the financial supports from the Institute of Standards and Industrial Research of Iran (Chaharmahal and Bakhtiari branch).
The authors extend their sincere appreciation to the Institute of Standards and Industrial Research of Iran (Chaharmahal and Bakhtiari branch) and Shahrekord University. Also, the authors thank to Mr. Mohammad Hossein Shams, Daryush Valipour, Amir Hossein Mohammadi, Mohammad Rasoul Amini and Mrs. Mahsa Edris for the cooperation in this project.
The author declares no conflict of interest.
)