Senior Design Projects
ECS193 A/B Winter & Spring 2019
Help us make smart buildings smarter
| ********** | |
| Marco Pritoni | |
| LBNL |
Project's details
| Help us make smart buildings smarter | |
| Connected sensors and devices are now common in many new building and building managers are installing more of them in older buildings. A large wave of new data flows into databases, but their active use to improve energy efficiency is still limited. LBNL is leading several research projects aimed at optimizing building energy use using modern sensors and controls. For instance, buildings should reduce their energy use when they are unoccupied. Despite the importance of occupancy, energy information systems for buildings do not typically track occupants presence or location. The historical reason is that occupancy sensors ( e.g. Infrared, ultrasound or CO2 ) were expensive to install and to connect to a central system, especially in an existing building, and have often had questionable reliability. If readily available, occupancy data could be used in real-time applications, such as building controls, as well as in retrospective analysis, to understand past energy use and the factors driving it. In the former case the control system can dynamically adapt static schedules to respond to occupancy data. Examples of the latter are measurement and verification ( M&V ) or fault detection and diagnostic ( FDD ). | |
| The project involves estimating building occupancy using the Wi-Fi network and other data sources ( e.g. class schedules, calendars etc. ) and using this occupancy information to identify ways of improving building operations. The students will work closely with researchers at LBNL and assist them with: 1 ) development and use of “drivers” to communicate with databases or building energy management systems, 2 ) data processing and management 3 ) occupancy modeling. The buildings analyzed, nearly 100, are located in multiple campuses in California ( Davis, Berkeley and Ontario ). | |
| - A report containing important findings after detailed data analysis and possible avenues for improving building operations. - Easily shareable and replicable analytics code - An open source software package to obtains occupancy counts from the Wi-Fi network | |
| Necessary: Python ( jupyter notebooks, pandas ), databases, working with web APIs. Preferred: familiarity with machine learning tools ( sci-kit learn ), networking | |
| ********** | |
| 30-60 min weekly or more | |
| Open source project | |
| Attachment | Click here |
| Yes | |
| Team members | N/A |
| Ethan | |
| N/A |