Now many of the industrial IoT adopters simply store raw sensor data in cloud. Some try to apply big data visualization and create some kind of reports using analytics software.
But raw sensor data does not provide useful information for the business decisions.
But the data can be processed in a specific way and become a digital twin of a real world entity. This phenomenon includes sets of data models that basing on sensor data feeds and real world objects simulations provide a real time digital image of the equipment or process.
As a result user can get a set of AI powered systems on top of diagnosis and prognosis solutions combining data from real world digital twins with third party inputs
- decision support system for the top management
- augmented guidance systems for service stuff based
For example in one of our projects the was the use case when model compared could recognize PV panels malfunctions like being covered with dust. Afterwards it took the the weather forecast feed together with the costs of cleaning PV panels from dust and potential losses that appear when you leave the panels uncleaned.
Such systems involve several components:
- connected equipment with sets of sensors and IoT modules
- cloud platform used for data storage and analytics (custom solutions for time series and unstructured data or third party platforms e.g. AWS IoT, Amazon Sumerian, Dell EMC, GE Predix, SAP IoT, Salesforce IoT explorer etc)
- set of mathematical, physical, statistical and econometrical models on top of deep learning algorithms used for digital simulation
- HMI solutions (web, touch-panels, AR/MR applications based on machine vision algorithms)
Digital twin might also involve additional functionality depending on the project needs
- Natural language processing solutions (voice-data)
- blockchain solutions for authentication, validation and verification of action
We can help to increase revenues from existing and future IoT customers by saving some of the customer’s money spent on third party maintenance services and business travel, while getting some fractions of the saved amounts for additional big data and cloud infrastructure used by diagnosis and prognosis digital twin predictive analytics models.