Siemens Aktiengesellschaft

Siemens Corporate Research, Inc.

Siemens Medical Solutions Usa, Inc.

Siemens Corporation

Massachusetts Institute of Technology

Siemens Energy, Inc.

Amit Chakraborty

Ioannis Akrotirianakis

Chao Yuan

Liang Hua Hsu

Holger Hackstein

Dorin Comaniciu

Peiya Liu

Lu-yong Wang

Zhijing Liu

Biswadip Dey

Motto Alexis Legbedji

Leif Wiebking

Jan Eggebrecht

Yu Sun

Rodrigo Carrasco

Yaofeng Zhong

Bogdan Georgescu

Anand A Kulkarni

Michael Rindler

Fabian Moerchen

Kai Zhang

Steven Zigler

Tolga Han Seyhan

Arindam Dasgupta

Swapna Somasundaran

Andrey Torzhkov

Eberhard Ritzhaupt-Kleissl

Anuradha Annaswamy

Paolo Favaro

Michael Lange

Georgia Olympia Brikis

Yangyang Xu

Xi He

Ruken Duzgun

Zhiwei Qin

Xiaocheng Tang

Dingcheng Li

Wotao Yin

Shiqian Ma

Hans Weber

Watao Yin

Nemanja Petrovic

Andre Werner

Venkatraman Gurumurthi

Ammaiappan Balasubramanian

Todd Putvinski

Robin Hoeher

Eric Greaser

Saikat Mukherjee

Sudarshan Sampath

Siong Thye Goh

Vikas Chandan

Alexis Motto

Matthew Evans

Yuan Chao

Yoav Sharon

Yingnan Cui

Xinmin Cai

A computer-implemented system and method for synthesizing a controller for an under actuated robotic manipulator includes a machine learning based model having a plurality of neural network modules. Each module is configured to approximate a function related to an underactuated controller for a robotic manipulator. Parameters of each function are learned during training of the model using a loss function that satisfies one or more conditions including structure preservation, integrability and equilibrium assignment.

Synthesizing machine learning-based controllers for underactuated robotic manipulators

Systems, methods, and computer-readable media are disclosed for generating and training a deep convolutional generative model for multivariate time series modeling and utilizing the model to assess time series data indicative of a machine or machine component's operational state over a period of time to detect and localize potential operational anomalies.

Deep convolutional factor analyzer

System and method for synthesizing a controller for a dynamical system includes a feeder neural network trained to estimate an ordinary differential equation (ODE) from time series training data (X) of a trajectory having embedded angular data and configured to learn dynamics of a physical system by encoding a generalization of a Hamiltonian representation of the dynamics using a constant external control term (u). A neural ODE solver receives the estimate of the ODE from the feeder neural network and synthesizes a controller to control the system to track a reference configuration.

Physics informed neural network for learning non-Euclidean dynamics in electro-mechanical systems for synthesizing energy-based controllers

A method for identifying underperforming agents in a multi-agent cooperative system includes receiving information relating to the performance of each agent in the multi-agent system, calculating an estimated extracted resource value of each agent based on the received information, comparing the estimated extracted resource value of each agent to a threshold value, calculating a performance index based on the comparison and identifying an agent as an under-performing agent based on the performance index. A system for identifying under-performing agents in a plurality of agents in a multi-agent cooperative system includes a performance analyzing processor, a communications port for receiving state information for each agent and control information for each agent, a classifier for identifying a subset of agents in the plurality of agents that are performance comparable and an optimizer configured to identify an under-performing agent of performance comparable agents and generate updated control information for the identified under-performing agent.

Methods and systems for performance loss estimation of single input systems

A computer-implemented method of scheduling jobs for an industrial process includes receiving jobs to be executed on machines within a manufacturing facility. A job schedule is generated based on an optimization function that minimizes total energy cost for all the machines during a time horizon based on a summation of energy cost at each time step between a start time and an end time. The energy cost at each time step is a summation of (a) a first energy cost associated with each machine in sleeping mode during the time step, (b) a second energy cost associated with each machine in stand-by mode during the time step, and (c) a third energy cost associated with each machine in processing mode during the time step. The jobs are executed on the machines based on the job schedule.

Energy efficient scheduling of industrial process by reducing idle time and tardiness of jobs

A computer-implemented method for performing machine condition monitoring for fault diagnosis includes collecting multivariate time series data from a plurality of sensors in a machine and partitioning the multivariate time series data into a plurality of segment clusters. Each segment cluster corresponds to one of a plurality of class labels related to machine condition monitoring. Next, the segment clusters are clustered into segment cluster prototypes. The segment clusters and the segment cluster prototypes are used to learn a discriminative model that predicts a class label. Then, as new multivariate time series data is collected from the sensors in the machine, the discriminative model may be used to predict a new class label corresponding to segments included in the new multivariate time series data. If the new class label indicates a potential fault in operation of the machine, a notification may be provided to one or more users.

Unifying semi-supervised approach for machine condition monitoring and fault diagnosis

A computer-implemented method for assessing material microstructure of a machine component involves obtaining a raw image of a section of the component captured via a microscope. The method further includes pre-processing the raw image to generate a ternary image defined by pixel data including three levels of intensities. The method further includes identifying, from the ternary image, phase boundaries delineating at a phase in a primary constituent material of the component. The method further includes determining a volume associated with the phase based on the identified phase boundaries. The proposed method may be utilized, for example, as an automated tool for assessing material degradation and for quality control of gas turbine engine components.

System and method for automated microstructure analysis

A method for training a deep learning network includes defining a loss function corresponding to the network. Training samples are received and current parameter values are set to initial parameter values. Then, a computing platform is used to perform an optimization method which iteratively minimizes the loss function. Each iteration comprises the following steps. An eigCG solver is applied to determine a descent direction by minimizing a local approximated quadratic model of the loss function with respect to current parameter values and the training dataset. An approximate leftmost eigenvector and eigenvalue is determined while solving the Newton system. The approximate leftmost eigenvector is used as negative curvature direction to prevent the optimization method from converging to saddle points. Curvilinear and adaptive line-searches are used to guide the optimization method to a local minimum. At the end of the iteration, the current parameter values are updated based on the descent direction.

Efficient calculations of negative curvature in a hessian free deep learning framework

A method for monitoring a condition of a system or process includes acquiring sensor data from a plurality of sensors disposed within the system (Sand S). The acquired sensor data is streamed in real-time to a computer system (Sand S). A discriminative framework is applied to the streaming sensor data using the computer system (Sand S). The discriminative framework provides a probability value representing a probability that the sensor data is indicative of an anomaly within the system. The discriminative framework is an integration of a Kalman filter with a logistical function (S).

Discriminative hidden kalman filters for classification of streaming sensor data in condition monitoring

A method and system for recognizing (and/or predicting) failures of sensors used in monitoring gas turbines applies a sparse coding process to collected sensor readings and defines the L-1 norm residuals from the sparse coding process as indicative of a potential sensor problem. Further evaluation of the group of residual sensor readings is perform to categorize the group and determine if there are significant outliers ('abnormal data'), which would be considered as more likely associated with a faulty sensor than noisy data. A time component is introduced into the evaluation that compares a current abnormal result with a set of prior results and making the faulty sensor determination if a significant number of prior readings also have an abnormal value. By taking the time component into consideration, the number of false positives is reduced.

Gas turbine sensor failure detection utilizing a sparse coding methodology

A method of predicting an amount of power that will be generated by a solar power plant at a future time includes: forecasting a value of a data variable at the future time that is likely to affect the ability of the solar power plant to produce electricity (S); computing a plurality of features from prior observed amounts of power generated by the power plant during different previous durations (S); determining a trending model from the computed features and the forecasted value (S); and predicting the amount of power that will be generated by the power plant at the future time from the determined model (S).

Combining multiple trending models for photovoltaics plant output forecasting

A method for solar forecasting includes receiving a plurality of solar energy data as a function of time of day at a first time, forecasting from the solar energy data a mode, where the mode is a sunny day, a cloudy day, or an overcast day, and the forecast predicts the mode for a next solar energy datum, receiving the next solar energy datum, updating a probability distribution function (pdf) of the next solar energy datum given the mode, updating a pdf of the mode for the next solar energy datum from the updated pdf of the new solar energy datum given the mode, forecasting a plurality of future unobserved solar energy data from the updated pdf of the mode, where the plurality of future unobserved solar energy data and the plurality of solar energy data have a Gaussian distribution for a given mode determined from training data.

Solar power forecasting using mixture of probabilistic principal component analyzers

The Support Vector Machine (SVM) has been used in a wide variety of classification problems. The original SVM uses the hinge loss function, which is nondifferentiable and makes the problem difficult to solve in particular for regularized SVMs, such as with l-norm. The Huberized SVM (HSVM) is considered, which uses a differentiable approximation of the hinge loss function. The Proximal Gradient (PG) method is used to solving binary-class HSVM (BHSVM) and then generalized to multi-class HSVM (MHSVM). Under strong convexity assumptions, the algorithm converges linearly. A finite convergence result about the support of the solution is given, based on which the algorithm is further accelerated by a two-stage method.

Proximal gradient method for huberized support vector machine

For generating a word space, manual thresholding of word scores is used. Rather than requiring the user to select the threshold arbitrarily or review each word, the user is iteratively requested to indicate the relevance of a given word. Words with greater or lesser scores are labeled in the same way depending upon the response. For determining the relationship between named entities, Latent Dirichlet Allocation (LDA) is performed on text associated with the name entities rather than on an entire document. LDA for relationship mining may include context information and/or supervised learning.

Information relation generation

A method for predicting failure modes in a machine includes learning () a multivariate Gaussian distribution for each of a source machine and a target machine from data samples from one or more independent sensors of the source machine and the target machine, learning () a multivariate Gaussian conditional distribution for each of the source machine and the target machine from data samples from one or more dependent sensors of the source machine and the target machine using the multivariate Gaussian distribution for the independent sensors, transforming () data samples for the independent sensors from the source machine to the target machine using the multivariate Gaussian distributions for the source machine and the target machine, and transforming () data samples for the dependent sensors from the source machine to the target machine using the transformed independent sensor data samples and the conditional Gaussian distributions for the source machine and the target machine.

Transferring failure samples using conditional models for machine condition monitoring

Growing community of inventors

East Windsor, NJ, United States of America

Amit Chakraborty