Abstract: A method of controlling a complex system and a gas turbine being controlled by the method are provided. The method comprises providing training data, which training data represents at least a portion of a state space of the system; setting a generic control objective for the system and a corresponding set point; and exploring the state space, using Reinforcement Learning, for a control policy for the system which maximizes an expected total reward. The expected total reward depends on a randomized deviation of the generic control objective from the corresponding set point.

Abstract: The embodiments relate to a method for the computer-aided control and/or regulation of a technical system, particularly a power generation installation. The actions to be performed in the course of regulation or control are ascertained using a numerical optimization method (e.g., a particle swarm optimization). In this case, the numerical optimization method uses a predetermined simulation model that is used to predict states of the technical system and, on the basis thereof, to ascertain a measure of quality that reflects an optimization criterion for the operation of the technical system.

Abstract: The invention concerns a method for the computerized control and/or regulation of a technical system (T). Within the context of the method according to the invention, there is implemented in a preset regulating process (CO1, CO2) an exploration rule (EP) by means of which new, as yet unknown, states (x) of the technical system (T) are started, a simulation model (SM) of the technical system (T) checking whether the actions (a2) of the exploration rule (EP) lead to sequential states (x?) lying within predetermined thresholds. Only in that case is the corresponding action (a2) performed according to the exploration rule (EP) on the technical system. The method according to the invention enables new states to be explored within the framework of the operation of a technical system, it being ensured through checking of appropriate thresholds that the exploration is carried out imperceptibly and does not lead to incorrect operation of the technical system.

Abstract: A computer-implemented method for controlling and/or regulating a technical system, in which actions to be carried out on the technical system are first of all determined using an action selection rule which was determined through the learning of a data-driven model and, in particular, a neural network. On the basis of these actions a numerical optimization searches for actions which are better than the original actions according to an optimization criterion. If such actions are found, the technical system is regulated or controlled on the basis of these new actions, such that the corresponding actions are applied to the technical system in succession. The method is suitable, in particular, for regulating or controlling a gas turbine, wherein the actions are preferably optimized with respect to the criterion of low pollutant emission or low combustion chamber humming.

Abstract: The invention concerns a method for the computerized control and/or regulation of a technical system. Within the context of the method according to the invention, an action-selection rule (PO?) is determined which has a low level of complexity and yet is well suited to the regulating and/or control of the technical system, there being used for determination of the action-selection rule (PO?) an evaluation measure (EM) which is determined on the basis of a distance measure and/or a reward measure and/or an action-selection rule evaluation method. The action-selection rule is then used to control and/or regulate the technical system. The method according to the invention has the advantage of the action-selection rule being comprehensible to a human expert. Preferably, the method according to the invention is used for regulating and/or controlling a gas turbine and/or a wind turbine.

Abstract: An interactive assistance system and method for computer-aided control optimization for a technical system is provided. For example, a gas or wind turbine, in particular for optimizing the action sequence or the control variables of the plant (e.g. gas supply, compression), wherein an input terminal is provided for reading at least one status parameter providing a first system status of the technical system, and at least one setting parameter for adapting a reward function. A simulation module having a pre-trained neuronal network, simulating the plant, serves to simulate an action sequence on the technical system, starting from the first system status and to the prediction of the resulting statuses of the technical system.

Abstract: A method for controlling a turbine is proposed, which is characterized at any point in the control by a hidden state. The dynamic behavior of the turbine is modeled with a recurrent neural network comprising a recurrent hidden layer. In this case, the recurrent hidden layer is formed from vectors of neurons, which describe the hidden state of the turbine at the time points of the regulation, wherein two vectors are chronologically linked for each time point with a first connection bridging a time and second connection bridging at least two points in time. Short-term effects can be controlled by means of the first connections and long-term effects can be adjusted by means of the second connections. Secondly, emissions and also occurring dynamics in the turbine can be minimized. Furthermore, a regulating device and a turbine with such a regulating device are proposed.

Abstract: For controlling a target system, e.g. a gas or wind turbine or another technical system, a pool of control policies is provided. The pool of control policies comprising a plurality of control policies and weights for weighting each of the plurality of control policies are received. The plurality of control policies is weighted by the weights to provide a weighted aggregated control policy. With that, the target system is controlled using the weighted aggregated control policy, and performance data relating to a performance of the controlled target system are received. Furthermore, the weights are adjusted on the basis of the received performance data to improve the performance of the controlled target system. With that, the plurality of control policies is reweighted by the adjusted weights to adjust the weighted aggregated control policy.

Abstract: For controlling a target system, e.g. a gas or wind turbine or another system, operational data of a plurality of source systems are used. The operational data of the source systems are received and are distinguished by source system specific identifiers. By a neural network a neural model is trained on the basis of the received operational data of the source systems taking into account the source system specific identifiers, where a first neural model component is trained on properties shared by the source systems and a second neural model component is trained on properties varying between the source systems. After receiving operational data of the target system, the trained neural model is further trained on the basis of the operational data of the target system, where a further training of the second neural model component is given preference over a further training of the first neural model component.

Abstract: A method for computer-assisted modeling of a technical system is disclosed. At multiple different operating points, the technical system is described by a first state vector with first state variable(s) and by a second state vector with second state variable(s). A neural network comprising a special form of a feed-forward network is used for the computer-assisted modeling of said system. The feed-forward network includes at least one bridging connector that connects a neural layer with an output layer, thereby bridging at least one hidden layer, which allows the training of networks with multiple hidden layers in a simple manner with known learning methods, e.g., the gradient descent method. The method may be used for modeling a gas turbine system, in which a neural network trained using the method may be used to estimate or predict nitrogen oxide or carbon monoxide emissions or parameters relating to combustion chamber vibrations.

Abstract: A method for the computer-supported generation of a data-driven model of a technical system, in particular of a gas turbine or wind turbine, based on training data is disclosed. The data-driven model is preferably learned in regions of training data having a low data density. According to the invention, it is thus ensured that the data-driven model is generated for information-relevant regions of the training data. The data-driven model generated is used in a particularly preferred embodiment for calculating a suitable control and/or regulation model or monitoring model for the technical system. By determining optimization criteria, such as low pollutant emissions or low combustion dynamics of a gas turbine, the service life of the technical system in operation can be extended. The data model generated by the method according to the invention can furthermore be determined quickly and using low computing resources, since not all training data is used for learning the data-driven model.

Abstract: The embodiments relate to a method for the computer-aided control and/or regulation of a technical system, particularly a power generation installation. The actions to be performed in the course of regulation or control are ascertained using a numerical optimization method (e.g., a particle swarm optimization). In this case, the numerical optimization method uses a predetermined simulation model that is used to predict states of the technical system and, on the basis thereof, to ascertain a measure of quality that reflects an optimization criterion for the operation of the technical system.

Abstract: The invention concerns a method for the computerized control and/or regulation of a technical system. Within the context of the method according to the invention, an action-selection rule (PO?) is determined which has a low level of complexity and yet is well suited to the regulating and/or control of the technical system, there being used for determination of the action-selection rule (PO?) an evaluation measure (EM) which is determined on the basis of a distance measure and/or a reward measure and/or an action-selection rule evaluation method. The action-selection rule is then used to control and/or regulate the technical system. The method according to the invention has the advantage of the action-selection rule being comprehensible to a human expert. Preferably, the method according to the invention is used for regulating and/or controlling a gas turbine and/or a wind turbine.

Abstract: The invention concerns a method for the computerized control and/or regulation of a technical system (T). Within the context of the method according to the invention, there is implemented in a preset regulating process (CO1, CO2) an exploration rule (EP) by means of which new, as yet unknown, states (x) of the technical system (T) are started, a simulation model (SM) of the technical system (T) checking whether the actions (a2) of the exploration rule (EP) lead to sequential states (x?) lying within predetermined thresholds. Only in that case is the corresponding action (a2) performed according to the exploration rule (EP) on the technical system. The method according to the invention enables new states to be explored within the framework of the operation of a technical system, it being ensured through checking of appropriate thresholds that the exploration is carried out imperceptibly and does not lead to incorrect operation of the technical system.

Abstract: For controlling a target system, such as a gas or wind turbine or another technical system, a pool of control policies is used. The pool of control policies including a plurality of control policies and weights for weighting each control policy of the plurality of control policies are received. The plurality of control policies is weighted by the weights to provide a weighted aggregated control policy. The target system is controlled using the weighted aggregated control policy, and performance data relating to a performance of the controlled target system is received. The weights are adjusted based on the received performance data to improve the performance of the controlled target system. The plurality of control policies is reweighted by the adjusted weights to adjust the weighted aggregated control policy.

Abstract: For controlling a target system, operational data of a plurality of source systems are used. The data of the source systems are received and are distinguished by source system specific identifiers. By a neural network, a neural model is trained on the basis of the received operational data of the source systems taking into account the source system specific identifiers, where a first neural model component is trained on properties shared by the source systems and a second neural model component is trained on properties varying between the source systems. After receiving operational data of the target system, the trained neural model is further trained on the basis of the operational data of the target system, where a further training of the second neural model component is given preference over a further training of the first neural model component. The target system is controlled by the further trained neural network.

Abstract: A computer-implemented method for controlling and/or regulating a technical system, in which actions to be carried out on the technical system are first of all determined using an action selection rule which was determined through the learning of a data-driven model and, in particular, a neural network. On the basis of these actions a numerical optimization searches for actions which are better than the original actions according to an optimization criterion. If such actions are found, the technical system is regulated or controlled on the basis of these new actions, such that the corresponding actions are applied to the technical system in succession. The method is suitable, in particular, for regulating or controlling a gas turbine, wherein the actions are preferably optimized with respect to the criterion of low pollutant emission or low combustion chamber humming.

Abstract: A method for computer-aided closed and/or open-loop control of a technical system is provided. A first value of an output quantity is predicted on a data-based model at a current point in time. A second value of the output quantity is determined from an analytical model. The state of the technical system at the current point is assigned a confidence score in the correctness of prediction of the data-based model. A third value of the output quantity is determined from the first and second value as a function of the confidence score for controlling the technical system. A suitable value for the output quantity can be derived from the analytical model even for regions of the technical system in which the quality of prediction of the data-based model is low because of a small set of training data. The technical systems can be turbines, such as gas turbines.

Abstract: The invention relates to a method for controlling a turbine which is characterized by a hidden state at each point in time of the control. The dynamic behavior of the turbine is modeled using a recurrent neural network comprising a recurrent hidden layer. The recurrent hidden layer is formed by vectors of neurons which describe the hidden state of the turbine at the points in time of the control. For each point in time, two vectors are connected chronologically to a first connection which bridges one point in time, and two vectors are additionally connected chronologically to a second connection which bridges at least two points in time. Short-term effects can be corrected by means of the first connections, and long-term effects can be corrected by means of the second connections. Emissions and occurring dynamics can be minimized in the turbine by means of the latter. The invention further relates to a control device and a turbine comprising such a control device.

Abstract: A method for the computer-supported generation of a data-driven model of a technical system, in particular of a gas turbine or wind turbine, based on training data is disclosed. The data-driven model is preferably learned in regions of training data having a low data density. According to the invention, it is thus ensured that the data-driven model is generated for information-relevant regions of the training data. The data-driven model generated is used in a particularly preferred embodiment for calculating a suitable control and/or regulation model or monitoring model for the technical system. By determining optimization criteria, such as low pollutant emissions or low combustion dynamics of a gas turbine, the service life of the technical system in operation can be extended. The data model generated by the method according to the invention can furthermore be determined quickly and using low computing resources, since not all training data is used for learning the data-driven model.