Control loop for regulating a combustion process

A control loop for regulating a combustion process in a plant, wherein the plant has a controlled system for converting material by way of the combustion process while at least one flame is formed and at least oxygen is supplied to the combustion, the control loop comprising:

• at least one observation device for imaging at least the flame;
• sensors for determining input data;
• wherein the sensors include the at least one observation device, and the input data comprises data from the imaging;
• at least one adjustment device for adjusting at least one supply selected from the group consisting of a supply of the material and the supply of at least oxygen, wherein the adjustment device is for being controlled by output data;
• and
• at least one computer for evaluating at least the input data with regard to target values and for determining the output data by using a current process model,
• wherein the computer has a feature extraction module, the feature extraction module is for extracting features from at least the input data by way of an information measure, the features are informative for the target values, and the features are for an alternative process model;
• wherein the computer generates the alternative process model using at least the feature extraction module;
• the feature extraction module has a feature selector;
• the feature selector generates combined information measures for the features and target values;
• and
• the feature selector searches for a maximum of the information measures with regard to the combination.

A method at least for preparing a process model for a control loop that is for regulating a combustion process in a plant, wherein the plant has a controlled system for converting material by way of the combustion process while at least oxygen is supplied and at least one flame is formed, the method comprising:

• using a current process model to provide output data for at least one adjustment device that is for adjusting at least one supply selected from the group consisting of a supply of the material and the supply of at least oxygen, wherein the using of the current process model includes
• (a) receiving input data at least from sensors, with the sensors including at least one observation device for imaging at least the flame, and
• (b) determining the output data, wherein the determining of the output data comprises evaluating the input data with regard to target values using at least one computer and the current process model, and the computer includes
• an action generator module that uses the current process model to determine from the input data suitable general actions for approaching the target values, and
• an adjustment variable selection module that defines the output data by selecting specific actions;
• and
• generating an alternative process model, wherein the generating of the alternative process model includes using the computer for extracting features from the input data by way of an information measure, wherein the computer includes a feature extraction module that extracts the features from the input data by way of the information measure, the features are informative for the target values, and the features are used in the generating of the alternative process model, and wherein each of the feature extraction module, the action generator module and the adjustment variable selection module is self-organized and operates automatically.