You need JavaScript to view this

The application of system identification techniques to performance monitoring of four stroke turbocharged diesel engines

Abstract

The thesis has two main objectives; firstly to mathematically model, simulate the performance and carry out a theoretical sensitivity analysis, and secondly to evaluate a new method for condition monitoring of four stroke turbocharged diesel engines based on system identification techniques and parameter estimation. Fault detection and diagnosis are illustrated by application of off-line parameter estimation techniques to a non-linear model of diesel engines. A recursive prediction error method was used to estimate parameters, corresponding to faults in a lumped parameter system. The results clearly indicate the feasibility of fault detection as well as diagnosis, that is, discrimination among faults occuring approximately at the same time. Among the objectives of early detection of malfunctions are the prevention of sudden failure of equipment, the improved planning of maintenance and performance. Measurements of dynamic cylinder pressures and mean values of compressor outlet temperature, air receiver pressure and temperature, exhaust receiver pressure and temperature, engine speed, turbine outlet temperature and turbocharger speed are used as input to the program. Bearing in mind the simple mathematical model of the diesel engine, the predicted variables show quite good agreement with measurements. A number of faults were investigated experimentally both individually and as multiple faults in  More>>
Authors:
Publication Date:
Nov 01, 1991
Product Type:
Thesis/Dissertation
Report Number:
NEI-NO-300
Reference Number:
SCA: 330102; PA: NW-92:005248; SN: 93000918308
Resource Relation:
Other Information: TH: Thesis (Dr.ing.).; PBD: Nov 1991
Subject:
33 ADVANCED PROPULSION SYSTEMS; DIESEL ENGINES; PERFORMANCE; MATHEMATICAL MODELS; MONITORING; CONTROL THEORY; COMPUTERIZED SIMULATION; SENSITIVITY ANALYSIS; 330102; DIESEL
OSTI ID:
10113256
Research Organizations:
Trondheim Univ. (Norway). Norges Tekniske Hoegskole
Country of Origin:
Norway
Language:
English
Other Identifying Numbers:
Other: ON: DE93752997; ISBN 82-7119-344-9; TRN: NO9205248
Availability:
OSTI; NTIS
Submitting Site:
NW
Size:
187 p.
Announcement Date:
Jun 30, 2005

Citation Formats

Molteberg, G A. The application of system identification techniques to performance monitoring of four stroke turbocharged diesel engines. Norway: N. p., 1991. Web.
Molteberg, G A. The application of system identification techniques to performance monitoring of four stroke turbocharged diesel engines. Norway.
Molteberg, G A. 1991. "The application of system identification techniques to performance monitoring of four stroke turbocharged diesel engines." Norway.
@misc{etde_10113256,
title = {The application of system identification techniques to performance monitoring of four stroke turbocharged diesel engines}
author = {Molteberg, G A}
abstractNote = {The thesis has two main objectives; firstly to mathematically model, simulate the performance and carry out a theoretical sensitivity analysis, and secondly to evaluate a new method for condition monitoring of four stroke turbocharged diesel engines based on system identification techniques and parameter estimation. Fault detection and diagnosis are illustrated by application of off-line parameter estimation techniques to a non-linear model of diesel engines. A recursive prediction error method was used to estimate parameters, corresponding to faults in a lumped parameter system. The results clearly indicate the feasibility of fault detection as well as diagnosis, that is, discrimination among faults occuring approximately at the same time. Among the objectives of early detection of malfunctions are the prevention of sudden failure of equipment, the improved planning of maintenance and performance. Measurements of dynamic cylinder pressures and mean values of compressor outlet temperature, air receiver pressure and temperature, exhaust receiver pressure and temperature, engine speed, turbine outlet temperature and turbocharger speed are used as input to the program. Bearing in mind the simple mathematical model of the diesel engine, the predicted variables show quite good agreement with measurements. A number of faults were investigated experimentally both individually and as multiple faults in order to test the method. Results showed that some of the parameters were more sensitive for detecting faults than the more traditionally used methods of trending measured variables. 111 refs., 59 figs., 36 tabs.}
place = {Norway}
year = {1991}
month = {Nov}
}