- Xu, Y., Xu, R., Yan, W. and Ardis, P., 2017, May. Concept drift learning with alternating learners. In Neural Networks (IJCNN), 2017 International Joint Conference on (pp. 2104-2111). IEEE.
- Harpale, A. and Srivastav, A., 2017. A textual transform of multivariate time-series for prognostics. arXiv preprint arXiv:1709.06669.
Predictive Maintenance technologies aim to detect, diagnose, and predict failures and degradation in machine components prior to criticality. The ultimate goal is to prevent downtime, identify root causes for follow-up action, and enable efficient evidence-based maintenance planning and optimization. GE Research has a rich history in the development of Predictive Maintenance technologies, with deployed tools managing over one hundred thousand assets across the GE business units as well as countless more for GE's customers in the aerospace, power generation, transportation, oil exploration, and healthcare domains.
The primary research topics being pursued by GE fall into two categories: Early Warning and Prognostics.
Early Warning pertains to detecting anomalous behaviors in a system's operation at the earliest possible time, regardless of the operational mode or context that the system is operating in. The intent here is to provide the maximum lead time for any potential action, with additional technologies brought to bear once sufficient evidence exists to suggest potential problems. This technology is applicable across industrial verticals and is traditionally the first item in combined Prognostics Health Management (PHM) deployments. GE Research software solutions for Early Warning are built around unsupervised, semi-supervised, and fully-supervised data exploration, enabling a broad span of solution complexities based upon the availability of ground truth feedback regarding normal and abnormal asset operations. As part of this, research has included an emphasis on fusion algorithms to integrate alerts and mixed-type information from multiple models to improve prediction accuracy and lead detection time as well as reduce alert fatigue. Similarly, research activities have yielded a robust multivariate time series search pipeline to speed up human interaction when searching for signatures, features, and patterns in massive time series data (for root cause and diagnostic reporting).
Prognostics go a step further to provide long term predictions of behavior and life. Taking over once Early Warning has occurred, Prognostic algorithms aim to forecast remaining useful life, time to reduced capability, and emergent fleet segmentation for planning inspection, maintenance, repair, and spare parts inventory. GE Research technologies in this space are built upon a suite of hybrid modeling techniques that use embodied domain physics along with condition monitoring data from fielded systems and simulations. This also allows learning systemic behaviors from entire fleets through techniques such as transfer learning. Furthermore, auto-inspection technologies allow us to inspect and label component condition (and quantify performance of associated prognostic models) without human bias, adapting the fielded models in a continuous learning mode.
GE remains committed to advancing the state of the art in the area of Predictive Maintenance and is always seeking new frontiers in its associated technologies.
GE Research's Predictive maintenance technologies have been integrated in a number of GE products/platforms such as Expert on AlertTM for GE Transportation, Tube WatchTM for GE Healthcare and Analytics Based MaintenanceTM (ABM) for GE Aviation.
Expert on AlertTM provides centralized monitoring of locomotive health status and performance in real time. It delivers real-time locomotive health checks and diagnostics of key locomotive components, enables proactive part and resource planning and provides recommendations validated by experts that ensure problems are fixed correctly the first time.
GE Healthcare Tube WatchTM delivers peace of mind by moving from potential unplanned downtime to planned events, helping to avoid patient and staff disruptions and associated revenue loss.
Tube WatchTM allows proactive part delivery and service scheduling to avoid tube failure with maximum lead time.
Analytics Based MaintenanceTM utilizes continuously-learning life prediction (Digital Twin) models of HPT blades, shrouds, and nozzles as well as combustion systems to provide usage-based inspection and engine removal recommendation to customers. These models help in reducing GE aviation business services costs as well as helping customers manage their aircraft fleet operations to maximize utilization.
Capabilities utilized for Predictive Maintenance project
Materials & Process Modeling
Combining the power of physics and data driven models to accelerate the discovery, development and servicing of material solutions and processesRead more
Knowledge Management & Big Data
Applying semantic modeling, text mining and Big Data to capture and digitize industrial domain knowledge for human and machine useRead more
Developing and scaling machine learning solutions for industrial applications to facilitate continuous learning, adaptation and improvement in dynamic operating environmentsRead more
Enhancing fundamental and applied research to mimic human visualization and interpretationRead more