A method for computer-implemented identifying an unauthorized access to a wind farm is provided by obtaining environmental and operational data of the wind farm from a repository, which include technical data and organizational data of the wind farm indicating tasks to be dealt with in the wind farm in the future. Based on the environmental and operational data, for a predetermined time interval, a prediction of the operation and/or states of the wind farm is determined by a trained data driven mode. The trained data driven model provides a prediction of the operation and/or states of the wind farm as a digital output. The prediction is compared to operational conditions of the wind farm resulting from current or past user machine interactions of a user. An unauthorized access is identified in case of a predetermined deviation of the obtained operational conditions from the prediction of the operation and/or states of the wind farm.

A computer-implemented method for controlling and/or monitoring equipment (110) of a chemical plant is proposed. The method comprises the following steps: a) specifying at least one parameter of a production process to be optimized; b) receiving input data via at least one input channel (126), wherein the input data comprises operating conditions of the production process, physical properties of a plant equipment layout, and at least one predicted parameter determined from a physico-chemical white box model (128), wherein the physico-chemical white box model (128) comprises at least one thermodynamics model (130), at least one solid formation model (132) and a computational fluid dynamics (CFD) based numerical simulation (134) for predicting a precipitation process; c) optimizing the specified parameter via at least one processing device (138), wherein the specified parameter is optimized by adapting the operating conditions of the production process and/or the physical properties of the plant equipment layout based on the predicted parameter.

A non-transitory, computer-readable recording medium stores a program of reinforcement learning by a state-value function. The program causes a computer to execute a process including calculating a temporal difference (TD) error based on an estimated state-value function, the TD error being calculated by giving a perturbation to each component of a feedback coefficient matrix that provides a policy; calculating based on the TD error and the perturbation, an estimated gradient function matrix acquired by estimating a gradient function matrix of the state-value function with respect to the feedback coefficient matrix for a state of a controlled object, when state variation of the controlled object in the reinforcement learning is described by a linear difference equation and an immediate cost or an immediate reward of the controlled object is described in a quadratic form of the state and an input; and updating the feedback coefficient matrix using the estimated gradient function matrix.

A method of generating a controller for a continuous process. The method includes receiving from a storage memory, off-line stored values of one or more controlled variables and one or more manipulated variables of the continuous process over a plurality of time points. The off-line stored values are used to train a first neural network to operate as a predictor of the controlled variables. Then, the method includes training a second neural network to operate as a controller of the continuous process using the first neural network after it was trained to operate as the predictor for the continuous process and employing the second neural network as a controller of the continuous process.

An objective of the present invention is to provide a future state estimation method and future state estimation device with which, given that the space in which the estimation is carried out is finite, it is possible to rapidly estimate the states of a controlled object and the peripheral environment thereof in the form of probability density distribution for infinite future. Provided is a future state estimation device characterized by comprising: a model storage part for saving a model for simulating a subject of simulation and the peripheral environment of the subject of simulation; a future state forecast result storage part for storing information obtained by estimating future states of the subject of simulation and the peripheral environment of the subject of simulation within a finite space in the form of probability density distribution, for either an infinite time or a given time step in the future; and a future state forecast computation part for carrying out calculation, which is equivalent to a series, using the model for simulating the future states of the subject of simulation and the peripheral environment of the subject of simulation in the form of probability density distribution.

The present disclosure provides a method and an apparatus for adjusting a process control prediction model, and a process controller. In an embodiment, the method includes: determining, based on controlled variable data in process control data obtained through real-time monitoring, whether a prediction performance of the process control prediction model is lower than a reference performance; and when the prediction performance is determined to be lower than the reference performance, using manipulated variable data in the process control data monitored to adjust the process control prediction model. By way of the method, a re-test does not need to be executed to re-identify a model so as to eliminate a mismatch of the process control prediction model, thereby eliminating an influence of fluctuation introduced by addition of an excitation signal during the re-testing.

A computer implemented method for lateral control of a vehicle comprises the following steps carried out by computer hardware components: determining a location error of the vehicle; determining an orientation error of the vehicle; determining a cost function based on the location error and the orientation error using a circular transformation; and processing the cost function in a model predictive controller to control the vehicle laterally.

Systems and methods for training a reinforcement learning (RL) model for HVAC control are disclosed herein. A calibrated simulation model is used to train a surrogate model of the HVAC system operating within a building. The surrogate model is used to generate simulated experience data for the HVAC system. The simulated experience data can be used to train a reinforcement learning (RL) model of the HVAC system. The RL model is used to control the HVAC system based on the current state of the system and the best predicted action to perform in the current state. The HVAC system generates real experience data based on the actual operation of the HVAC system within the building. The real experience data is used to retrain the surrogate model, and additional simulated experience data is generated using the surrogate model. The RL model can be retrained using the additional simulated experience data.

A fluid balance management system includes: a detection device configured to output a detection signal corresponding to a load applied to a bed in which a subject is present; a prediction device configured to acquire load variation information indicating a variation over time of the load based on the detection signal, and predict an event that causes variation in fluid balance of the subject from the load variation information; and an output device configured to output prediction information corresponding to a prediction result of the event.

A fault prediction system for building equipment includes one or more memory devices configured to store instructions that, when executed on one or more processors, cause the one or more processors to receive device data for a plurality of devices of the building equipment, the device data indicating performance of the plurality of devices; generate, based on the received device data, a plurality of prediction models comprising at least one of single device prediction models generated for each of the plurality of devices or cluster prediction models generated for device clusters of the plurality of devices; label each of the plurality of prediction models as an accurately predicting model or an inaccurately predicting model based on a performance of each of the plurality of prediction models; and predict a device fault with each of the plurality of prediction models labeled as an accurately predicting model.