A system includes a production computing environment including a plurality of components, a centralized data repository that receives and stores data feeds relating to one or more components as a data log and at least one processor configured to obtain the data log of each component, generate a current state vector for the component based on the data log, compare the current state vector to a normal state vector of the component, determine that the current state vector deviates from the normal state vector, and in response, predict an anomaly associated with the component using an iterative machine learning method. The at least one processor may be configured to correct the predicted anomaly by taking at least one pre-configured action corresponding to the predicted anomaly.

Techniques for detecting and correcting for discrepancy events in a fluid pipeline are presented. The techniques can include obtaining a plurality of empirical temperature and pressure measurements at a plurality of locations within the pipeline; simulating, using a pipeline model, a plurality of simulated temperature and pressure measurements for the plurality of locations within the pipeline; detecting, by a discrepancy event detector, at least one discrepancy event representing a discrepancy between the empirical temperature and pressure measurements and the simulated temperature and pressure measurements; outputting to a user an indication that the at least one discrepancy event has been detected; accounting for the at least one discrepancy; determining, after the accounting and using an estimator applied to the pipeline model, a corrected branch flow rate for the pipeline; and outputting the corrected branch flow rate for the pipeline to the user.

Techniques for phase identification using feature-based clustering approaches are disclosed. Embodiments employ linear and nonlinear dimensionality reduction techniques to extract feature vectors from raw time series. In an embodiment, a constrained clustering algorithm separates smart meters into phase connectivity groups. Another embodiment clusters smart meter data, where voltage measurements are collected from smart meters and a SCADA system. Then, customer voltage time series are normalized and linear or nonlinear dimensionality reduction is applied to the normalized time series to extract key features. Next, constraints in the clustering process are defined by inspecting network connectivity data. Then, a constrained clustering method is applied to partition customers into clusters. Lastly, each clusters phase is identified by solving a minimization problem. In another embodiment, a machine learning algorithm generalizes a subset of phase connectivity measurements to a distribution network, the algorithm being an extension of a Mapper algorithm in topological data analysis.

Embedded convolutional data modeling is disclosed. A machine receives, as input, a plurality of data examples. The machine generates an embedded convolutional data model for the data examples. The machine outputs a source code for the model.

Disclosed are devices, systems and methods for the operational testing on autonomous vehicles. One exemplary method includes configuring a primary vehicular model with an algorithm, calculating one or more trajectories for each of one or more secondary vehicular models that exclude the algorithm, configuring the one or more secondary vehicular models with a corresponding trajectory of the one or more trajectories, generating an updated algorithm based on running a simulation of the primary vehicular model interacting with the one or more secondary vehicular models that conform to the corresponding trajectory in the simulation, and integrating the updated algorithm into an algorithmic unit of the autonomous vehicle.

According to an embodiment, a monitoring apparatus configured to generate time-series predicted data based on time-series measured data and a prediction model that generates predicted data including one or more predicted values predicted to be output from one or more sensors; and generate, for a first sensor among the one or more sensors, a displayed image including a measured value graph representing a temporal change in a measured value included in the time-series measured data in a second period after a first period, a predicted value graph representing a temporal change in a predicted value included in time-series predicted data in the second period, past distribution information representing a distribution of a measured value in the first period, and measurement distribution information representing a distribution of the measured value included in the time-series measured data in the second period.

The invent relates to an intelligent fault monitoring method based on high-order information enhanced recurrent neural network, for real-time fault monitoring of sewage treatment process. The invent includes two phases of offline modeling and online monitoring. In offline phase, the original data is extracted into high-dimensional high-order information features using OCIA, which can effectively deal with the non Gaussian feature of the data and solve the correlation between variables. Then the extracted features are trained by DRNN. In the online phase, the data are directly mapped to new high-order feature components, and to be discriminated in category by the DRNN network after trained offline. If there is no fault, then the results get into the monitoring model composed of simple OICA for unsupervised monitoring. If no fault is detected, it is determined that there is no fault in the process. On the contrary, the process fault is determined, and the fault information will be added to the training data of the network for training, so as to continuously improve the monitoring accuracy of DRNN.

System, Device and Method of detecting at least one abnormal datapoint in operation data (U) associated with an industrial environment (610) is disclosed. The method comprising iteratively applying one or more anomaly detection models (fi) to at least one subset (S) of the operation data (U), wherein the anomaly detection models (fi) are trained based on a training dataset (L) consisting of datapoints labeled as normal; classifying subset-datapoints in the subset (S) as one of normal datapoints (N) and abnormal datapoints (A) using the anomaly detection models (fi); updating the training dataset at least with the normal datapoints; retraining the anomaly detection models (fi) with the updated training dataset after expiration of a threshold time, wherein the threshold time is based on the number of updates to the training dataset; and detecting the at least one abnormal datapoint in the operation data (U) using the anomaly detection models (f′i).

The invention relates to a container treatment machine for treating containers, in particular in the beverage-processing industry, medical technology, or the cosmetics industry, the container treatment machine comprising a control unit for controlling the function of the container treatment machine and at least one treatment unit for treating the containers; the container treatment machine is designed to treat the containers in exactly one way; the container treatment machine comprises at least one component which can output data relating to its operating state and/or the operating state of the container treatment machine to the control unit; and the control unit comprises a neural network which is configured and trained to use the data to determine whether a deviation of the operating state of the container treatment machine from a normal state is present and/or imminent.

Targeted distributing of reports containing historical process control information to particular user devices via a communications network. A curating service permits assigning a score to each report based on an interest level value of the historical process control information to a user associated with each user device and/or an urgency value of the historical process control information. Routing reports to user devices based on the score raises visibility of the historical process control information without overburdening the communications network.