TY - GEN
T1 - Early detection of lean blow out (LBO) via generalized D-Markov machine construction
AU - Sarkar, Soumalya
AU - Ray, Asok
AU - Mukhopadhyay, Achintya
AU - Chaudhari, Rajendra R.
AU - Sen, Swarnendu
PY - 2014
Y1 - 2014
N2 - This paper develops a method for early detection of lean-blow-out (LBO) in combustion systems by extracting low-dimensional features from chemiluminescence time series of optical sensor data. In the proposed method, symbol strings are generated by partitioning the (finite-length) time series to construct a special class of probabilistic finite state automata (PFSA), called D-Markov machines. These PFSA have a deterministic algebraic structure and their states are represented by symbol blocks of length D or less. The states of D-Markov machines are constructed in two steps: (i) state splitting, i.e., the states are split based on their information contents, and (ii) state merging, i.e., two or more states (of possibly different lengths) are merged together to form a new state without any significant loss of their embedded information. The modeling complexity (i.e., the number of states) of a D-Markov machine is observed to be drastically reduced as the combustion system approaches LBO. The prediction of LBO is posed as a pattern classification problem based on different ranges of equivalence ratio of the flame. It is shown that, over a wide range of air-fuel premixing, a generalized D-Markov machine (i.e., with D > 1) performs better than a D-Markov machine with D = 1 as a predictor of LBO.
AB - This paper develops a method for early detection of lean-blow-out (LBO) in combustion systems by extracting low-dimensional features from chemiluminescence time series of optical sensor data. In the proposed method, symbol strings are generated by partitioning the (finite-length) time series to construct a special class of probabilistic finite state automata (PFSA), called D-Markov machines. These PFSA have a deterministic algebraic structure and their states are represented by symbol blocks of length D or less. The states of D-Markov machines are constructed in two steps: (i) state splitting, i.e., the states are split based on their information contents, and (ii) state merging, i.e., two or more states (of possibly different lengths) are merged together to form a new state without any significant loss of their embedded information. The modeling complexity (i.e., the number of states) of a D-Markov machine is observed to be drastically reduced as the combustion system approaches LBO. The prediction of LBO is posed as a pattern classification problem based on different ranges of equivalence ratio of the flame. It is shown that, over a wide range of air-fuel premixing, a generalized D-Markov machine (i.e., with D > 1) performs better than a D-Markov machine with D = 1 as a predictor of LBO.
UR - http://www.scopus.com/inward/record.url?scp=84905695671&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84905695671&partnerID=8YFLogxK
U2 - 10.1109/ACC.2014.6859048
DO - 10.1109/ACC.2014.6859048
M3 - Conference contribution
AN - SCOPUS:84905695671
SN - 9781479932726
T3 - Proceedings of the American Control Conference
SP - 3041
EP - 3046
BT - 2014 American Control Conference, ACC 2014
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2014 American Control Conference, ACC 2014
Y2 - 4 June 2014 through 6 June 2014
ER -