Data Driven Buildings Bharathan Balaji Amazon AI Labs Buildings - - PowerPoint PPT Presentation

Data Driven Buildings

Bharathan Balaji

Amazon AI Labs

Buildings are evolving..

8 November 2018 Bharathan Balaji, SBIoT 2018 2

Shelter Water Electricity Security Thermal Comfort Fire Safety Connectivity Sensors Software

Data: Interact. Optimize. Innovate.

Data Generation Bottleneck: Sensors

Reliable Expensive Not Flexible

8 November 2018 Bharathan Balaji, SBIoT 2018 3

Wired Wireless + Battery Battery Maintenance Inexpensive Flexible Wireless + Harvesting Unreliable Inexpensive Flexible

Pible: Perpetual Indoor BLE Sensor

• Sensors
• Light
• Temperature
• Humidity
• PIR
• Door Event
• Bluetooth beacon
• Bluetooth Low

Energy

• Solar harvesting
• Super capacitor

8 November 2018 Bharathan Balaji, SBIoT 2018 4

Limitation: Manual configuration

Pible: BuildSys 2018

Duty Cycle with Reinforcement Learning

• Automatically adapts to each lighting

condition

• 1 sample every 56 seconds on average

8 November 2018 Bharathan Balaji, SBIoT 2018 5

Scaling Energy Harvesting Configuration: EnsSys 2018

Reward Days

Best Demo Award: BuildSys 2018

Data Collation Bottleneck: Vertical Systems

November 6, 2018 Bharathan Balaji, SBIoT 2018 6

Lighting System Plug Loads Heating, Ventilation and Air Conditioning (HVAC) Enterprise Network Security System

Integration Platform for Applications

6 November 2018 Bharathan Balaji, SBIoT 2018 7

Building A Building B Building C Building D Data Connectors

Large amount of data generated in modern buildings

REST/ Native API

Apps

Visualize Maintain Analyze

Data management system for sensors and actuators Next generation building applications via standardized API

Ø Scalable, distributed data storage Ø Metadata and contextual tagging Ø Access control across users Ø REST API for app development

Control

Smart Building Applications

11/6/18 Bharathan Balaji, SBIoT 2018 8

Personalized Control BuildSys ’13, Ubicomp ‘16 Energy Disaggregation BuildSys ‘10, BuildSys ‘13 Occupancy Based Control IPSN ‘11, SenSys ‘13 Fault Detection and Diagnosis BuildSys ‘14

App Portability Bottleneck: Naming Semantics

8 November 2018 Bharathan Balaji, SBIoT 2018 9

Each Building is Different

• Equipment, Vendor, Institution
• Changes with time: Repairs, Retrofits

November 6, 2018 Bharathan Balaji, SBIoT 2018 10

v University, Hotels, Hospitals, Shopping Malls

Lecture Hall Bio Labs Mixed Use Library

Brick: Building Metadata Schema

11/8/18 Bharathan Balaji, SBIoT 2018 11

Brick

Infrastructure: Sensors Equipment Thermostat HVAC Luminaire Motion Sensor Smoke Detector Fire Safety Management Monitoring Access Control APIs Applications Occupant Interaction Fault Detection Demand Response

Brick: BuildSys 2016, Applied Energy 2018 https://brickschema.org

Brick Fundamentals

6 November 2018 Bharathan Balaji, SBIoT 2018 12

Zone Temp Sensor Sensor Temp_Sensor Zone_Temp_Sensor RM-3 ZNT-3

Tags TagSets

Location Room

Relationship

type type hasLocation

Brick Class Hierarchy

8 November 2018 Bharathan Balaji, SBIoT 2018 13

Equipment Fire Safety System HVAC AHU Terminal Unit VAV Point Command Sensor Temperature Sensor Room Temperature Sensor Location Floor Room Kitchen Lab

AHU

Power Meter Supply Fan

Lighting Controller

HVAC ZONE VAV

Lighting Zone

Damper Return Fan Thermostat

Temperature CO2 Sensor

Room 102 Room 101 Supply Air Return Air

An Example “Model” Building

Relationships in Example Building

11/6/18 Bharathan Balaji, SBIoT 2018 15

Temperature Sensor VAV Room 101 HVAC Zone AHU Damper

hasPoint hasPoint hasPart feeds feeds

Lighting Zone Lighting Controller Power Meter Room 102

isLocationOf

controls

hasPoint isLocationOf isLocationOf Location Equipment Point Relationship

Legend

11/6/18 Bharathan Balaji, SBIoT 2018 16

1.Complete Vocabs, Extensible Framework 2.Represent all necessary relationships -> Using RDF 3.Usable query mechanism -> SPARQL over RDF 4.Open Source -> BSD and RFC

https://brickschema.org

Need to Map Existing Building Metadata

8 November 2018 Bharathan Balaji, SBIoT 2018 17

Scrabble: Map existing buildings to Brick

• Learn from prior examples
• Ask expert when known examples are not enough

8 November 2018 Bharathan Balaji, SBIoT 2018 18

Source Building Target Building RM120.Temp Temperature Sensor in Room-120 Water Temperature Sensor in Room-3 R3.WtrTemp Known RM = Room Temp = Temperature Wtr = Water Buildings Expert

Scrabble: Multi-Stage Algorithm

8 November 2018 Bharathan Balaji, SBIoT 2018 19

Experts Give Examples at Two Stages

8 November 2018 Bharathan Balaji, SBIoT 2018 20

Dataset

• University of California, San Diego – 3 buildings, 3000 data points

{ ‘vendor_name’: ‘NAE 99 N2 0 VMA999 SA T’, ‘bacnet_desc’: ‘Supply Air Temp’, ‘bacnet_unit’: ‘64’, }

• Carnegie Mellon University – 1 building, 1000 datapoints

‘CMU/CMPS BLDG/First Floor/VAV Corridor 9900 Central/ Airflow Setpoint’

8 November 2018 Bharathan Balaji, SBIoT 2018 21

VAV Supply Air Temp Sensor Building Floor Corridor VAV Air Flow Setpoint

Results: Compared to State of the Art Baseline

• Faster learning speed than baseline
• Capable of accumulating examples given by expert

8 November 2018 Bharathan Balaji, SBIoT 2018 22

Data Driven Buildings

8 November 2018 Bharathan Balaji, SBIoT 2018 23

Equipment Sensor

Brick

Visualize Maintain Analyze Control Devices

“Alexa, join the meeting”

Alexa for Business

https://aws.amazon.com/alexaforbusiness

Alexa for Hospitality

https://www.amazon.com/alexahospitality

Thank you!

Amazon AI Labs – Reinforcement Learning Team LinkedIn: https://www.linkedin.com/in/bharathanbalaji/ Email: bhabalaj@amazon.com Acknowledgements Francesco Fraternali, Jason Koh, Anna Levitt, Gabe Fiero Mani Srivastava, Rajesh Gupta, Yuvraj Agarwal

8 November 2018 Bharathan Balaji, SBIoT 2018 25

Results: Buildings in the Same Campus

8 November 2018 Bharathan Balaji, SBIoT 2018 26