The aim of the diagnosis techniques usually applied in energy systems is the detection of signs of possible malfunctions, in particular those malfunctions that can cause failures, if not removed. The thermoeconomic diagnosis is applied to detect and locate possible inefficiencies of the plant components that provoke variations of the plant efficiency.
Thermoeconomic diagnosis procedures are based on the comparison between the real operation conditions and carefully selected reference conditions, usually characterized by the same qualitative and quantitative plant productions. In this way a difference in the overall fuel consumption (fuel impact) is an index of the presence of malfunctions due to the component inefficiencies. The use of the thermoeconomic theories for the location of anomalies presents the problem of the induced malfunctions , In fact, when an anomaly takes place in a component, not only its efficiency decreases, but also other components change their behaviour, due to the modified working conditions provoked by the anomaly. The induced malfunctions have two different causes: 1) the malfunctioning component requires a larger amount of fuel to maintain constant its production, so the other components must increase their production too, thus varying their efficiency; 2) as the values of some set-point variables can have been modified, the regulation system must operate to restore its set-point value. Its action modifies the plant working conditions, inducing new malfunctions.
In this paper a procedure for filtering and canceling the induced malfunctions is proposed. Hence, the devices suffering the real inefficiencies can be identified and located. The proposed procedure is based on the knowledge of the system behaviour when the regulation parameters vary. A first fictitious working condition, called free condition , is mathematically built, starting from the operation conditions and restoring the regulation parameters corresponding to the reference operating conditions. In this way the free condition does not include the regulation system contribution. A second state is then built for every single component in order to take into account the malfunctions induced by the degradation of the components behavior. The procedure is particularly useful when the number of malfunctions is unknown, as it happens in the usual operation of an actual plant. If the induced malfunctions are filtered, all the component inefficiencies can be identified at the same time.
The proposed procedure is applied to a 33 MW gas turbine plant, also providing 63 MW of thermal power to a district-heating network. A mathematical model of the plant has been used to simulate the behavior degradation of the gas turbine devices, when several components inefficiencies occur.