IBM Research ‘Big Bets’ in Sustainable Technologies: Smarter Water Management Sherif El-Rafei, Business Development Executive, IBM Research Middle East & Africa srafei@eg.ibm.com April 2013 © 2013 IBM Corporation Smarter Planet/Smarter City 2 © 2013 IBM Corporation Reimagining how science and technology can have impact • Managing human impact on rivers by streaming information • Reimagining the energy grid by synchronizing supply • Fighting infectious disease by spreading data • Reducing traffic jams by creating them • Reducing CO2 while boosting business efficiency • Improving communication by talking to the Web • Helping premature infants by sensing complications before they happen • Mapping beneath the seafloor to help reduce the risk of dry holes • Creating drinking water by filtering oceans © 2013 IBM Corporation Smarter Water Management Overview 4 © 2013 IBM Corporation Smarter Water Management means enabling higher levels collaboration and innovation across value chains and ecosystems Environment Regulation Supply Climate Change Demand Control Engagement Intelligence Infrastructure Recycled/Treated Natural Water Sources Raw Water Transport Clean Water Supply Consumers Sewage Treatment © 2013 IBM Corporation We Work at Three “Scales” Natural scale Water resource mapping and availability Water quality monitoring and management (surface and subsurface) Land use analysis Extraction monitoring (surface and subsurface) Flood control Utility scale Water quality and usage Discharge, combined sewer overflow Asset management “Smart levees” and levee monitoring systems Weather event assimilation Energy management Enterprise Scale Water usage tracking Water quality control (into and within plants, discharges) Supply chain optimization Energy management Business process improvements Metrics and management © 2013 IBM Corporation Smarter Water Source Management 7 © 2013 IBM Corporation Hydrogeosphere – an Integrated Computational Modeling Framework Water Cycle Hydrological model Water Cycle Watson hydrological model Deep Thunder Ocean Model Groundwater model basin model Weather / Climate / Atmospheric modeling New Insights come from integration of multiple disciplines Water Quality (Measurement Large River Basin Simulation Management Technology) © 2013 IBM Corporation Phase II - Large River Basin Simulation Cooperation between IBM Austin Research Laboratory & University of Texas. Full scale simulation of the Guadalupe River. – Demonstrating a predictive model with ~100X speedup. Width of each segment represents depth The color represents flow velocity Red: high velocity Blue: low velocity Availability of geographical and sensor data is crucial to success. Eventual goal: Mississippi River. – About 80X larger than the Guadalupe. Total number of reaches: ~3,900 Number of pour ports: ~1,800 Total length: ~15,000 km Modeled by 131K nodes, two unknowns at each node (depth and velocity). 262K unknowns solved at each time point © 2013 IBM Corporation Subsurface (Hydro-geological) Flow Model Variable-scale using unstructured (tetrahedral) meshes Time-dependent, model-based subsurface flow modeling Can be coupled with the surface flow model Model solved using: Locally conservative multiphase (water, air) Numerical model based on Control-Volume Finite Element discretization Can include geo-mechanical effects of elastic/plastic aquifers, and topography and density driven flows Transient temperature effects, fracture and faults can be specified Numerical kernel extensively used in basin modeling (scalable to from millions to billions of cells) © 2013 IBM Corporation 5-6 April 2010 Flooding Event Coastal storm with heavy rains (up to 284mm in 24 hours) starting at about 1700 BRT on 5 April 2010 – heaviest recorded compared to the previous 48 years One of the most significant global weather events of 2010 Local flooding leading to mudslides, killed over 200 people and left 15000 homeless Widespread disruption of transportation systems (e.g., road closures, airport and rail delays) Rio de Janeiro mayor Eduardo Paes admitted that the city's preparedness for heavy rainfall had been "less than zero," but added "there isn’t a city that wouldn’t have had problems with this level of rainfall." © 2013 IBM Corporation What is Weather Modelling? A mathematical model that describes the physics of the atmosphere –The sun adds energy, gases rise from the surface, convection causes winds Numerical weather prediction is done by solving the equations of these models on a 4dimensional grid (e.g., latitude, longitude, altitude, time) Complementary to observations (e.g., NWS weather stations) Solution yields predictions of surface and upper air –Temperature, humidity, moisture –Wind speed and direction –Cloud cover and visibility –Precipitation type and intensity © 2013 IBM Corporation Match the Scale of the Weather Model with the Client’s Needs 2km 2km “You don't get points for predicting rain. You get points for building arks.” (Lou Gerstner) Capture the geographic characteristics that affect weather (horizontally, vertically, temporally) Ensure that the weather forecasts address the features that matter to the business © 2013 IBM Corporation Short-Term Weather Event Prediction and Observation Nowcasting (Sensors) Forecasting (Modelling) NWS / Commercial Providers Deep Thunder Forecast for longer-term planning where decisions require days of lead time, but may not have direct coupling to business processes Forecast for asset-based decisions to manage weather event, pre-stage resources and labor proactively Continental to Global Scale 72-168 Remote Fine-tune approach based upon extrapolation from Doppler radar and satellite observations Local Scale 18-72 In Situ Near-real time revision Local Scale 3 0 Time Horizon for a Local Weather Event (Hours of Lead Time) © 2013 IBM Corporation Command Center for Rio de Janeiro © 2013 IBM Corporation The Importance of Real-Time Coastal Awareness Protecting coastal cities Protecting our environment Monitoring/managing coastal agriculture and industries Our vision: coastal awareness, weather prediction and flood prediction in concert to protect citizens, infrastructure, and the environment Managing maritime operations Tracking pollutant dispersion 16 © 2013 IBM Corporation Realtime Coastal Awareness • Collaboration with National University of Ireland, Galway • Objective: Real-time prediction of bay conditions (quality and circulation patterns) for environmental decision support • Challenges: – Noise and uncertainty in measurements – Model scale • Methodology: – Data assimilation for real-time modelling – HPC implementation • CODAR = high frequency radar for water surface speed © 2013 IBM Corporation CODAR project infrastructure CODAR CODAR `Assimilation of 10GB / hr. HF radar for water speed • CODAR adds to wealth of sensors in Galway Bay – Smart Bay tidal gauges and flow measurements – Sonars for water velocity at varying depth – Two weather stations • Ideal prototyping environment © 2013 IBM Corporation Smarter Water Distribution Management 19 © 2013 IBM Corporation A Measurement and Modeling Technology Platform Measurement Platform Management Environment Smart Sensor Bus Integrated Modeling Environment General Technology platform to deliver physical intelligence for smarter planet applications by leveraging state of-the-art metrology, a broad set of models and unique controls to different length & time scales of the physical world © 2013 IBM Corporation Leakage & Pipeline Failures… Water losses reduction – More than 32 billion cubic meters of treated – – water is lost annually through distribution network leaks [1] A conservative estimate of the total annual cost of water loss to utilities worldwide is US$14 billion [1] According to IWA, 15%~30% water is leaked [2] –15%~30% water leaking in the world[2] – 900 leakage/burst per year in big cities[4] Public image improvement – 250~300 pipe bursts per year in Trondheim City, – Norway [3] About 900 leakage per year in Hong Kong. [4] May 25, 2010, pipe burst at Beijing JingGuang Bridge causing a 5-hour water supply disruption and severe traffic jam in the business center Source: 1)From Bentley company 2) “Water Industry: Managing Leakage”. Engineering and Operations Committee, UK. 3)Jianhua Lei and Sveinung Segrov, Statical approach for describing failures and lifetimes of water mains. Wat. Sci. Tech. Vol. 38, No. 6, pp. 209-217. 4) Hong Kong Water Supplies Department Annual Report (2008) 5) A Lambert, (2001) What do we know about pressure-leakage relationships in distribution systems? IWA Conf. n Systems approach to leakage control and water distribution system management. Brno, Czechoslovakia. ISBN 80-7204-197-5 21 © 2013 IBM Corporation Addressing Non-Revenue Water using Analytics and Optimization Leakage or Theft Detection at the Residential Level Understand usage patterns and detect anomalies for low and high consumption to detect leakage, theft or faulty meters Leakage Detection at the Network Level using optimization Find “optimal” location of leak(s) to explain difference between actual measurements and model predicted measurements Leakage Reduction using Dynamic Pressure Control Optimal Valve Placement for Pressure Reduction Create optimization model to adjust the pressure dynamically so that only the required flow will be supplied yielding cost reduction in energy and water achieved. Use an optimization model to find the optimal number of valves, and their location, so as to enable the most effective pressure management 22 © 2013 IBM Corporation © 2013 IBM Corporation Business Innovation Architecture Demo Asset Lifecycle planning enables informed operational and strategic decision support annual cost Usage / Smart Meters Infrastructure Network Relationships Failure History Asset Condition Assessment Risk Estimation & Prediction Environmental Attributes replace repair Capital Budget failure rate {Labor, material, service interruptions, …} Replacement Cost Estimation Strategic Plan {Labor, routine disruptions, cost, material, ….} Maintenance Cost Estimation Spatial Coordinates Asset Attributes Periodic inspection Business Constraints Failure Impact Decision Support Operationa l Plan Operational Budget Backup Assets Strategic replacement in 2, 5, and 10 years Efficient use of crew and equipment © 2013 IBM Corporation Integrated Water Management 25 © 2013 IBM Corporation © 2013 IBM Corporation Strategic Water Information Management Platform 27 © 2013 IBM Corporation Water Resource Management 28 © 2013 IBM Corporation Strategic Water Information Management (SWIM) Platform Data types (as examples) (from multiple sources and systems) Run-off Visualization layer Quantity/Flo w Applications layer (Open) standards Business rules layer Security Usage and Discharge “An integrated set of technologies, data and tools” Quality Environment/ Ecology Models layer Data handling layer Climate Economic Data content layer Geology/ hydrology Network layer Energy data Sensing layer © 2013 IBM Corporation 30 © 2013 IBM Corporation Hindi Ευχαριστώ Thank You English Greek Dziekuje Polish شكرا Russian Thai Portuguese Arabic Gracias Merci Spanish Obrigado Teşekkür ederim Turkish French Mulţumesc Grazie Italian Tamil 31 Februar Romanian Simplified Chinese Traditional Chinese Danke Japanese German © 2013 IBM Corporation Environmental Analytics Platform Vineyard Factories, Bridges, Refineries, Airports etc. 32 © 2013 IBM Corporation Low-Power Mote Technology (LMT) LMT—a wireless data gathering technology A general IBM wireless sensor platform – – Highly robust and scalable sensing solution Forms Mesh Network World’s lowest power consumption – 5 to 7 year lifetime with two AA batteries Very flexible and modular design What are the benefits ? Means to maintain soil moisture while minimizing water usage for irrigation Prevent frost and/or fungal damage Alarm workers to take measures to save crops. Predicting local frost damage Determine optimum harvest point Optimize crop growing and food processing Sensors can be located with +/- 3 Improved asset and operational management feet Environmental sensing: 33 – – – – – – – – – Temperature and Humidity Soil Moisture and Temperature Sun light / irradiation Dew point Pressure, Air flow Carbon dioxide Presence and Occupancy Corrosion and Air quality Location © 2013 IBM Corporation LMT for Agriculture Applications What can we monitor ? • Soil temperature • Soil moisture • Air temperature • Humidity • Sunlight • ….. • pH ? • What would like to measure which you cannot do today ? What are the benefits ? • Means to maintain soil moisture while minimizing water usage for irrigation • Predicting local frost damage • Alarm workers to take measures to save crops. • Determine optimum harvest point • Prevent frost and/or fungal damage • Optimize crop growing and food processing • Improved asset and operational management 34 © 2013 IBM Corporation Soil Moisture Detection – Full field and largescale IR imaging Semi-spherical mirror IR camera Less moisture [1] Data from Iven Mareels’ IBM presentation in January 2011 35 © 2013 IBM Corporation Example – Crop Growing • Total of 35.3 acres over three fields in Eastern New York • 95 motes supporting 475 sensors • • • • • Soil temperature Air temperature Soil moisture Humidity Light • Data streamed back into a central gateway every 2 s • Software Solution allows remote monitoring and control • Deep Analytics • • • • • 36 Moisture Modeling Time Series Forecasting Optimization Statisical Correlation …. © 2013 IBM Corporation Example - Fungal Disease Detection • Phytophthora is a fungal disease in potatoes, which depends on temperature, humidity and whether the leaves are wet. • Extensive wireless sensing system in the Netherland measures air pressure, temperature, relative humidity and illumination • System alerts farmers of patches within his fields which are most susceptible and can be used to gauge the steps that need to be taken. 37 © 2013 IBM Corporation Research’s Strategic Disciplines Exploratory Business Analytics & Math. Sciences Industry Solutions Services Software Systems Technology © 2013 IBM Corporation