Artificial Intelligence for the Utility of the Future Hannover Messe – Smart Grid Forum 2016
Some basic facts € 47,000,000,000,000 * • An accumulative investment of € 6,037bn is expected to be invested in T&D (New Demand, Renewables and Refurbishment) till 2035. • On average € 350bn are invested each year in T&D. • Each day a utility creates the same amount of data they accumulated in a year only a couple of years ago and this will continue to increase. Exponentially. • 80% or more of that data is unstructured. * According to the IEA the accumulative amount that needs to be invested in energy supply and efficiency by 2035 with a strong emphasis on energy efficiency to meet the 2 ⁰C climate change target
For a small investment in AI, utilities can enhance the information embedded in all their data and significantly leverage their vast investment in IoT with: – Machine learning – Powerful visualization – Natural language querying – Capture experiential (tribal) knowledge from the workforce From data rich to knowledge rich
Traditional Analytics Analytics is the discovery, interpretation and communication of meaningful patterns in data User Visual Intelligence Traditional Analytics Extraction, Translation & Loading Real Time Big Data Graph Databases • Smart by not intelligent – no machine learning • Cannot acquire knowledge & model dynamically • Cannot interact with Users in a conversational fashion • Are not capable of being “app” driven
Next-Gen of analytics – multi dimensional Static System Adaptive System integrating multiple sources 9/18/2015 Confidential & Proprietary : 5
App Orchid’ Unique Approach. User Visual Intelligence Analytics ++ Knowledge Deep Learning Acquisition Knowledge Discovery Knowledge Curation Natural Language Interface Real Time Big Data Graph Databases Unstructured • Continuously learning and improving • Knowledge acquisition & Discovery is automated • Bringing structured data into play (80% of all data!!) • Capable of Human conversation.
Teaching machines to think like humans “has” Asset Performance • How are objects processed • How does one object influence “typeof” “typeof” another? • How do we visualize Objects? • How do we Analyze Objects? Solar • Risk How do we process Pane dependencies and study root l cause? • How do you make systems think like “Human Beings”? Inverter
Building a “ facebook ” for the Energy World. Asset Management Weather External Feeds Performance & Predictive
Analytical Apps Type a question When was the deviation from the DDP the highest ? Structured Data Apps Unstructured Data Results Capture Tribal Knowledge
Geospatial visuatization d d Temporal analysis of outages based on severity conditions. Regional Outage analysis d d based on the Severity index
Powerful suite of KPIs for Renewables.
A Learning System What is the expected forecast for a typical month of July
Simulate cloud cover action using a natural language based scenario builder Confidential & Proprietary : App Orchid Inc - 2014
Microgrid Analysis Ask me anything you want. For example, how many boilers are operating below capacity ? Clustering analysis for the “Rubber banded” area shows the Cost and Outage Implications around asset performance
Demo
Benefits to the customer • Greater forecasting accuracy • Faster analysis • Increased dispatching efficiency • Reduction of cost • A solution for retiring workforce
Contact Dan Goldenblatt Director, Business Development (EMEA) Email: dang@Apporchid.com Tel: +49 33203 269013 Mob: +49 173 3491154 www.apporchid.com 9/18/2015 Confidential & Proprietary : 17
Recommend
More recommend
