Madagascar

Energy production by methanization or gasification of biomass is dependant on the chemical composition of the gas generated. The resistive sensors based on semiconductor metal oxides like the MQ series sensors are inexpensive and frequently used in gas detection. These sensors initially dedicated to detecting gas leaks in safety systems have relatively small measurement ranges (i.e. limited to concentrations below 10000 ppm). It is therefore necessary to find solutions to adapt these categories of sensors for gas measurements in the energy sector where the gas concentration is much more significant. In this article we propose a protocol using an adaptable capsule for MQ-4 and MQ-8 sensors to measure high concentrations of CH4 and H2 respectively. The technique consists of diluting the gas to be studied in a known volume of air. Three methods are proposed and compared regarding the linearity and the repeatability of the measurements. The first method was done in an airtight enclosed chamber the second method consists of directly injecting the gas on the sensor placed in an open environment and the final method was accomplished by direct injection of the gas on the sensor placed in a partially closed capsule. Comparisons show that the first technique provides the best repeatability with a maximum standard deviation of 13.88% for CH4 measurement and 5.1% for H2. However its linearity is weak (i.e. R2 ¼ 0.8637 for CH4 and R2 ¼ 0.5756 for H2). The second technique has better linearity but bad repeatability. The third technique presents the best results with R2 values of 0.9973 for the CH4 measurement and 0.9472 for H2. The use of the partially closed capsule resulted in an acceptable linear response of the sensors by up to 20% concentration of CH4 and until 13.33% concentration of H2 in the studied gas. The use of this simple and low-cost technique facilitates the characterization of combustible gases in isolated areas. It allows local operators of biomass valorization systems to control and improve their installations while avoiding the high costs of conventional measurement devices. This study hence contributes to the development of rural electrification projects in remote areas.