초록 ▼

In a method for controlling a thermodynamic process, in particular a combustion process, in which the system status (st) is measured, compared with optimization targets (rj), and in which suitable setting actions (ai) are performed in the system for controlling it, a process model (PM) is determined

In a method for controlling a thermodynamic process, in particular a combustion process, in which the system status (st) is measured, compared with optimization targets (rj), and in which suitable setting actions (ai) are performed in the system for controlling it, a process model (PM) is determined that is independent of the optimization targets (rj) and which describes the effects of actions (at) on the system status (st), and in which a situational evaluation (SB) that is independent of the process model (PM) evaluates the system status (st) by means of quality functions (ut) with regard to the optimization targets (rj).

대표청구항 ▼

That which is claimed: 1. A method for controlling a thermodynamic process of a system, the method comprising: providing optimization targets which relate to predetermined operation of the thermodynamic process of the system; determining a process model which is independent of the optimization targ

That which is claimed: 1. A method for controlling a thermodynamic process of a system, the method comprising: providing optimization targets which relate to predetermined operation of the thermodynamic process of the system; determining a process model which is independent of the optimization targets and which describes effects of actions on status of the thermodynamic process of the system; measuring the status of the system; determining actions to take in an effort to cause the thermodynamic process of the system to perform closer to the optimization targets, wherein the determining the actions is carried out using at least the process model, the measured status, and a situational evaluation, and wherein the situational evaluation is independent of the process model and evaluates status of the thermodynamic process of the system with regard to the optimization targets by using quality functions; and then carrying out the determined actions in the system, wherein at least partially by virtue of the process model being independent of the optimization targets, and the situational evaluation being independent of the process model, the determining the process model includes retaining past information about the system's status so that the process model can continue to use the past information also after major changes in the system's status. 2. A method according to claim 1, wherein the determining the process model and the determining the actions to take are carried out by a computer. 3. A method according to claim 1: wherein the situational evaluation using the quality functions includes determining, at time t, a quality function ut as follows description="In-line Formulae" end="lead"u t=Σu(rj,st);description="In-line Formulae" end="tail" wherein u(rj,st) represents the quality functions, st represents status variables, and rj represents the optimization targets; wherein the determining the actions to take using at least the process model includes determining, at time t+1, status variables st+1 as follows description="In-line Formulae" end="lead"s t+1=f(st,at); anddescription="In-line Formulae" end="tail" wherein at represents different actions that can change the thermodynamic process of the system. 4. A method according to claim 3, wherein: the determining the actions includes: numerically calculating, using the process model, effects of a plurality of different actions on the status of the thermodynamic process of the system, individually evaluating the effects using the situational evaluation, selecting, based upon the evaluating, actions of the plurality of actions which will more optimally cause the thermodynamic process of the system to perform closer to the optimization targets; and carrying out the determined actions in the system includes carrying out the selected actions in the system. 5. A method according to claim 4, wherein the effects are calculated several time steps in advance and an overall quality is determined. 6. A method according to claim 5, wherein the overall quality Q, for a number N of pre-calculated time steps, is determined as follows: 7. A method according to claim 3, wherein the process model is stored in a neural network which is constantly trained. 8. A method according to claim 3, wherein the system includes a boiler, the thermodynamic process of the system includes a combustion process in the boiler which is fired with solid fuel, and the actions include one or more actions selected from the group consisting of activation of a blower of the boiler and adjustment of particle size of the fuel. 9. A method according to claim 3, wherein: the system includes a boiler and the thermodynamic process of the system includes a combustion process in the boiler, the measuring the status of the system includes: monitoring the combustion process with at least one sensor while the combustion process is taking place in the boiler, evaluating the combustion process with a computer that is connected to the sensor, with the evaluating the combustion process occurring while the combustion process is taking place in the boiler, and taking a sample from the boiler using a sampling device, wherein the computer determines when the sample should be taken and the taking the sample is responsive to the computer determining that the sample should be taken. 10. A method according to claim 3, wherein the determining the process model and the determining the actions to take are carried out by a computer. 11. A method according to claim 10, wherein the process model is continually fine-tuned with information from ongoing measurements of the status of the thermodynamic process of the system. 12. A method according to claim 11, wherein the results of the ongoing measurements, prior to feeding them into the process model, are subjected to a noise filter. 13. A method according to claim 10, wherein prior to continuous use of the system, an exploration of a number of selected statuses of the system is performed to initialize the process model. 14. A method according to claim 13, wherein the exploration takes into account smoothed and weighted time behavior of the system in response to action upon the system. 15. A method according to claim 10, wherein the quality functions are set up according to fuzzy logic rules. 16. A method according to claim 4, wherein the thermodynamic process of the system is a combustion process. 17. A method according to claim 4, wherein the process model is continually fine-tuned with information from ongoing measurements of the status of the system. 18. A method according to claim 4, wherein the system includes a boiler, the thermodynamic process of the system includes a combustion process in the boiler which is fired with solid fuel, and the actions include one or more actions selected from the group consisting of activation of a blower of the boiler and adjustment of particle size of the fuel. 19. A method according to claim 4, wherein: the system includes a boiler and the thermodynamic process of the system includes a combustion process in the boiler, the measuring the status of the system includes: monitoring the combustion process with at least one sensor while the combustion process is taking place in the boiler, evaluating the combustion process with a computer that is connected to the sensor, with the evaluating the combustion process occurring while the combustion process is taking place in the boiler, and taking a sample from the boiler using a sampling device, wherein the computer determines when the sample should be taken and the taking the sample is responsive to the computer determining that the sample should be taken. 20. A method according to claim 1, wherein the measuring the status of the system comprises measuring the status of the thermodynamic process of the system. 21. An apparatus for controlling a thermodynamic process of a system, the apparatus comprising: sensors for measuring status of the system; a computer programmed for: determining a process model which is independent of optimization targets and which describes effects of actions on status of the system, wherein the optimization targets relate to operation of the system, determining actions to take in an effort to cause the system to perform closer to the optimization targets, wherein the determining the actions is carried out using at least the process model, the measured status, and a situational evaluation, wherein the situational evaluation is independent of the process model and evaluates status of the system with regard to the optimization targets by using quality functions, and providing feedback to the system for initiating a carrying out the determined actions in the system, wherein at least partially by virtue of the process model being independent of the optimization targets, and the situational evaluation being independent of the process model, the determining the process model includes retaining past information about the system's status so that the process model can continue to use the past information also after major changes in the system's status. 22. An apparatus according to claim 21, wherein: the determining the actions includes: numerically calculating, using the process model, effects of a plurality of different actions on the status of the system, individually evaluating the effects using the situational evaluation, selecting, based upon the evaluating, actions of the plurality of actions which will more optimally cause the system to perform closer to the optimization targets; and the providing feedback to the system includes providing feedback to the system for initiating a carrying out the selected actions in the system. 23. An apparatus according to claim 21: wherein the situational evaluation using the quality functions includes determining, at time t, a quality function ut as follows description="In-line Formulae" end="lead"u t=Σu(rj,st);description="In-line Formulae" end="tail" wherein u(rj,st) represents the quality functions, st represents status variables, and rj represents the optimization targets; wherein the determining the actions to take using at least the process model includes determining, at time t+1, status variables st+1 as follows description="In-line Formulae" end="lead"s t+1=f(st,at); anddescription="In-line Formulae" end="tail" wherein at represents different actions that can change the thermodynamic process of the system. 24. A method for controlling a thermodynamic process of a system, the method comprising: determining a process model which is independent of an optimization target and which describes effects of actions on status of the thermodynamic process of the system, wherein the optimization target relates to predetermined operation of the system, determining action to take in an effort to cause the system to perform closer to the optimization target, wherein the determining the action includes: using the process model to determine at least: a calculated first status of the system which should result from first action upon the system, and a calculated second status of the system which should result from second action upon the system, and using a situational evaluation to determine which one of the first status and the second status is more optimal with respect to causing the thermodynamic process of the system to perform closer to the optimization target, wherein the situational evaluation is independent of the process model and evaluates status of the system with regard to the optimization target by using at least one quality function; and then indicating either: that the first action should be taken if it was determined using the situational evaluation that the first status is more optimal than the second status with respect to causing the thermodynamic process of the system to perform closer to the optimization target, or that the second action should be taken if it was determined using the situational evaluation that the second status is more optimal than the first status with respect to causing the thermodynamic process of the system to perform closer to the optimization target, wherein at least partially by virtue of the process model being independent of the optimization targets, and the situational evaluation being independent of the process model, the determining the process model includes retaining Past information about the system's status so that the process model can continue to use the past information also after major changes in the system's status. 25. A method according to claim 24, wherein the thermodynamic process of the system is a combustion process. 26. A computer-readable medium comprising computer-executable instructions for performing the method of claim 24. 27. A method according to claim 24, wherein the determining the process model and the determining the actions to take are carried out by a computer. 28. A method according to claim 27: wherein the situational evaluation using the quality functions includes determining, at time t, a quality function ut as follows description="In-line Formulae" end="lead"u t=Σu(rj,st);description="In-line Formulae" end="tail" wherein u(rj,st) represents the quality functions, st represents status variables, and rj represents the optimization targets; wherein the determining the actions to take using at least the process model includes determining, at time t+1, status variables st+1 as follows description="In-line Formulae" end="lead"s t+1=f(st,at); anddescription="In-line Formulae" end="tail" wherein at represents different actions that can change the thermodynamic process of the system.