Industrial processes and equipment are prone to operational changes and faulty operation of such processes and equipment can adversely affect output of the overall setup. Existing systems for monitoring and fault detection consider individual instances of data for fault detection, which may not be suitable for industrial processes. Disclosed herein is a system and a method for anomaly detection in an industrial enterprise. The system collects data from a plurality of sensors as input. The system processes the collected data along temporal dimension, during which the data is split to multiple segments of fixed window size. Data in each segment is processed to identify anomalous data, and data in segments identified as containing the anomalous data is further processed to identify one or more sensors that are faulty and are contributing to the anomalous data.

A method for operating an industrial automation system may include receiving, via a first module of a plurality of modules in a control system, a plurality of datasets via at least a portion of the plurality of modules. The plurality datasets may include raw values without context regarding the plurality datasets. The method may then include identifying a subset of the plurality of datasets that influences a value of a target variable by analyzing the data without regard to the context, modeling a behavior of the target variable over time based on the subset of the plurality of datasets, and adjusting one or more operations of an automation device based on the model.

The present disclosure describes a sensor system for measuring a process value of a physical system, including: a plurality of sensors, wherein each sensor is configured to generate a sense signal as a function of the process value at a given time; a system state corrector configured to determine an actual system state of the physical system at a given state update cycle; a system state predictor configured to determine a predicted system state of the physical system at a given prediction cycle from a previous system state at a previous state update cycle; a sense signal predictor configured to determine predicted sense signals at the given prediction cycle from the predicted system state by applying a first operation to the predicted system state using a sense signal model of the physical system for predicting the sense signals.

A method for determining information (Info) about a state (Z) of a drive motor system (2) and/or of a drive battery pack (11) of a gardening, forestry and/or construction device (1), includes the steps of: acquiring at least one sensor temperature (TS) of at least one temperature component (3) of the drive motor system (2) and/or of the drive battery pack (11) by way of at least one component temperature sensor (4) at the same time as and/or at a time after operation of the drive motor system (2) and/or of the drive battery pack (11) and/or of a heater (300) and/or of a cooler (310) and/or of a fan (320) for heating and/or for cooling the drive battery pack (11), wherein the temperature component (3) heats up or cools down due to the operation; ascertaining operating data (BD) of the operation, wherein the operating data (BD) are of a different kind than the sensor temperature (TS); comparing the acquired sensor temperature (TS) or a variable based on the sensor temperature and the ascertained operating data (BD) or a variable (TM) based on the operating data (BD) by way of a temperature model (MOD), wherein the temperature model (MOD) is based on at least one model state (AZyes) of the drive motor system (2) and/or of the drive battery pack (11); and determining the information (Info) on the basis of a result of the comparison.

A system comprises one or more observation stations. Each observation station of the one or more observation stations comprises a corresponding set of one or more sensors. Additionally, the system comprises one or more physical machines that implement a computation engine configured to receive first observation data from the one or more observation stations. The computation engine may use the first observation data to train a machine learning system. The computation engine may subsequently use the trained machine learning system to provide feedback regarding an additional instance of the observation subject. The computation engine outputs the feedback.

Methods, mediums, and systems include use of a system manger application in a data processing system for fault detection a building automation system using deep learning, to receive point data for a hardware being analyzed, where the received point data is contaminated data, train a deep learning model for the hardware being analyzed, generate predicted data based on the deep learning model, analyze the predicted data and the received point data, identify a fault in the hardware being analyzed according to the received point data and the predicted data, and produce a fault report according to the identified fault.

A failure diagnostic system includes an instrument and a failure diagnostic device. The instrument makes measurement of a measured value regarding behavior of a diagnosis target. The failure diagnostic device has a model of the diagnostic target and performs simulation based on the model. The failure diagnostic device offers a user a proposal for execution of a special operation on the diagnostic target, on the condition that a difference between a result of the simulation and the measured value is greater than a predetermined error range but the difference provides an insufficient basis to determine whether or not the diagnosis target has a failure. The result of the simulation is calculated with the model supplied with a same input as an input to the diagnosis target at the time of the measurement by the instrument of the measured value regarding the behavior of the diagnosis target.

Methods, systems, and non-transitory computer readable medium are provided for correcting component failures in ion implant semiconductor manufacturing tool. A method includes receiving, from sensors associated with an ion implant tool, current sensor data corresponding to features; performing feature analysis to generate additional features for the current sensor data; providing the additional features as input to a trained machine learning model; obtaining one or more outputs from the trained machine learning model, where the one or more outputs are indicative of a level of confidence of a predicted window; predicting, based on the level of confidence of the predicted window, whether one or more components of the ion implant tool are within a pre-failure window; and responsive to predicting that the one or more components are within the pre-failure window, performing a corrective action associated with the ion implant tool.

Techniques for predictive water condition monitoring are described herein. An aspect includes a method that includes monitoring, by one or more processors, at least one water sensor to establish a baseline of a water condition model and monitoring one or more water conditions. A predicted water condition is determined based on the water condition model and the one or more water conditions. An alert is transmitted to one or more devices based on determining that the predicted water condition indicates a predicted contaminant level above a threshold.

A method includes converting time-series data from a plurality of prognostic and health monitoring (PHM) sensors into frequency domain data. One or more portions of the frequency domain data are labeled as indicative of one or more target modes to form labeled target data. A model including a deep neural network is applied to the labeled target data. A result of applying the model is classified as one or more discretized PHM training indicators associated with the one or more target modes. The one or more discretized PHM training indicators are output.