Analysis of Stability Degree for Combustion of Waste-to-Energy Plants with Fuzzy Relational Map of Sensors — Construction of Maps Using Data from Different Timestamps

Abstract

In Waste-to-Energy plants, signs of unstable combustion states appear in other locations. We have proposed a fuzzy relational map of sensors (FuRMS) to represent relationships between many sensors as relations between fuzzy sets of sensor values. For all pairs of sensors in a specific combustion state, we express their relationships by quantifying distribution patterns of many sensor values as a few degrees of importance of the relationships between fuzzy sets. For data with unknown states, we calculate an evaluation value as a weighted sum of importance degrees for all sensor pairs, where each weight is given by the product of the degrees of the sensor values to the fuzzy sets. This value indicates how closely the data matches the state by the map. When only a few sensors are associated with a state, it is possible to identify those sensors and to understand the relationships between sensors. We have used sensor values at the same timestamp for constructing maps. However, effects from some sensors appear in others after some delay. In this paper, we propose construction of maps using sensor values from the current and previous timestamps. This is expected to more accurately capture the relationships between sensors. We applied this map to data from an actual plant. It was found that maps from different timestamps and from only the current timestamp enabled earlier detection of changes in the combustion state than only maps from the current timestamp.