July 15, 2017
COE-ASM at the World Remanufacturing Conference:
Dr. Mike Thurston on Product Monitoring & Prognostics
One of the most important ways remanufacturing can become more effective and efficient is to take out the guesswork. Currently, many, if not most remanufacturing processes require some type of assessment to determine product condition, detect system failures, and define the type and extent of required restoration processes. This takes significant development effort as well as process time, adding cost and variability to the remanufacturing process.
But what if remanufacturers could reduce this uncertainty? What if we knew how the product was used during its lifecycle, fom its operating environment to its performance and failure modes? The COE-ASM’s own Director, Dr. Mike Thurston, sees significant opportunities to increase the use of product monitoring as a way to improve the remanufacturing process. At the 2017 World Remanufacturing Conference, Dr. Thurston spoke to industry and business leaders on where some of the opportunities are and how they can be achieved.
Dr. Thurston’s discussion started with a simple premise: the work that goes into testing components for failures is necessary, but guaranteeing that there are no latent failures or components that may wear out in the next use cycle is challenging. To address this, Dr. Thurston highlighted emerging technologies in product monitoring and prognostics that can allow OEMs, service technicians, or remanufacturers to see what is going on during product use, and use that data to make better decisions on core and component disposition.
But these technologies go beyond onboard sensors that communicate real-time use data. Indeed, while physical sensors can detect anomalies and imminent failures, sophisticated models are required to understand how this data applies to the condition of the asset. Dr. Thurston suggests that companies explore model-based technologies such as General Electric (GE)'s Digital Twin. to bridge the gap. By using documented data on the product configuration (such as design and material composition data), engineers can build digital mathematical models that represent how real-life products are expected to perform. Then, by leveraging data about specific use contexts (such as environmental conditions and use intensities), these digital models can simulate actual performance in parallel to the physical operation of a deployed product. Beyond facilitating assessment, such digital models can project both the nature and likelihood of failures based on context-specific data. Combined with internet-connected analysis tools like GE’s Predix, these models can then be used to optimize intervention decisions, suggesting the best way and the best time to act. Applied across a network of deployed products, such advanced technologies reduce the uncertainties in determining when to remanufacture a product and which components can be reliably reused.
Today, with the communications and data analytics technologies that are available, data has never been more abundant nor more valuable to decision makers. With early versions of these technologies already on the market, and continuing to develop in the ways they can be applied, it is time for the remanufacturing industry to consider how to take advantage. In this light, Dr. Thurston left audience with some important homework to consider: How could you improve your operation if you knew more about your cores?