ANALYSIS OF VISUAL CHARACTERISTICS OF A GAS FLAME AND CONSTRUCTION OF AN EMPIRICAL MODEL FOR REVERSE CALCULATION OF PARAMETERS TO IDENTIFY COMBUSTION MODES

Abstract

This paper presents the development and research of software for the automated assessment of combustion quality in a gas flame based on computer vision methods. The proposed system is designed for real-time video processing, extracting key visual parameters – flame area and its average brightness. The software complex is implemented in Python using the OpenCV library for video stream processing and image analysis and the NumPy library for mathematical calculations.The main scientific novelty lies in the use of an approximated mathematical model, derived from empirical data, which describes the functional relationship between the flame area and its brightness for the nominal operating mode. This model enables the reverse calculation of the expected flame area. By comparing this calculated value with the actual measured area, the system detects a transition to the undesirable mode of chemical underburning (incomplete combustion). A special logical indicator is introduced for this purpose. The research results confirm the effectiveness of the proposed approach for creating low-cost, fast, and reliable tools for visual combustion monitoring in industrial conditions. The main advantage of the method is its simplicity and high processing speed. The key limitation is its sensitivity to changes in the system configuration (camera position, optics), which requires model recalibration for each new setup. The developed software can be integrated into automatic control loops of boiler units via an OPC UA server for prompt elimination of combustion anomalies.