Method for the estimation of combustion parameters
Method for combustion engine control comprising an engine, at least one sensor and at least one engine control unit ( 200 ) arranged to control the engine ( 100 ) by the use of real time sensor data S(k) characterized in that a prior (P 1 (k, j; z), P 2 (k; X, Y) is used to derive at least one combustion parameter estimate (X(k)) and/or at least one combustion parameter measurement (Y(k)) from the real lime sensor data (S(k)) and that the at least one combustion parameter estimate (X(k)) and/or the at least one combustion parameter measurement (Y(k)) is used in performing said engine control.
1. A method for control of a combustion engine comprising an engine, at least one sensor and at least one engine control unit ( 200 ) arranged to control the engine ( 100 ) by the use of real time sensor data S(k), comprising the steps of using a prior (P 1 (k,j;z), to derive at least one combustion parameter measurement (y(k,j)) from the real time sensor data (S(k)), and using the at least one combustion parameter measurement (y(k,j)) in performing said engine control, wherein within a single combustion cycle (k), the prior (P 1 (k,j;z)) is used in order to select a combustion parameter candidate (z(k,j,n)) from a set of candidates (z(k,j,1) . . . z(k,j,N)) as the measurement (y(k,j)) of the combustion parameter (v(k,j)).
2. The method according to claim 1 , wherein a candidate with the lowest risk is chosen as the combustion parameter measurement (y(k,j)).
3. The method according to claim 1 , wherein the combustion parameter estimates (X(k−1)) from former combustion cycles, real time sensor data (S(k)) and combustion parameter measurements (Y(k)=(y(k,1), . . . , y(k,J)) T ) and a prior (P 2 (k;X,Y)) are used to obtain a combustion parameter estimate (X(k)) of the current combustion cycle and the combustion parameter estimate (X(k)) is used in performing said engine control.
4. The method according to claim 3 , wherein the combustion parameter estimate (X(k)) of the current combustion cycle is derived using a risk function based on a prior (P 2 (k;X,Y)) and the combustion parameter estimates (X(k−1)) from former combustion cycles and the combustion parameter measurements (Y(k)) of the current cycle.
5. The method according to claim 1 , wherein the prior (P 1 (k,j;z), P 2 (k;X,Y) is to some extent derived from combustion theory.
6. The method according to claim 1 , wherein the prior (P 1 (k,j;z),P 2 (k;X,Y)) is to some extent derived from practical experiments in a model engine.
7. The method according to claim 1 , wherein the prior (P 1 (k,j;z),P 2 (k;X,Y)) is to some extent updated by analyzing the performance of the engine during the operation of the engine.