Buildings Evolved attend the Mission Innovation Challenge 7 Priority Area 4 (Predictive Maintenance and Optimization) Planning Workshop in Montreal, Quebec, Canada. September 27-28, 2018

Table of Contents

• 1. Attendees
• 2. Topics for Research
  • 2.1. Predictive Control
  • 2.2. Control oriented emulator
  • 2.3. Data Clearing House (DCH)/ Open data platform
  • 2.4. Demand Response (DR) and Flexibility
  • 2.5. Fault Detection and Diagnosis (FDD)

1. Attendees

We were incredibly honoured to be invited by CSIRO once again to an international conference, or in this case a research planning workshop, following up on work done in Malmo in May 2018. This was an opportunity to do a deep dive into the R&D efforts of each nation with most of the attendees being scientists and researchers from institutions similar to CSIRO, including:

1. CanMet (Canada)
2. Lawrence Berkeley National Labs (USA)
3. Unina (Italy)
4. TNO (Netherlands)
5. JRAIA (Japan)
6. BEIS (UK)
7. Hydro Quebec (Canada)
8. Department of Energy (USA)
9. RISE (Sweden)
10. National Institute of Standards and Technology (NIST) (USA)
11. Pontifical Catholic University of Paraná (Brazil)
12. Pacific Northwest National Laboratory (USA)
13. University College London (UK)
14. Oak Ridge National Laboratory (USA)

2. Topics for Research

After running through presentations from each country involved in IC#7, we discussed in detail the following topic areas:

2.1. Predictive Control

The art of predictive control is taking in forecast data or historical data to estimate future requirements; this could range from model predictive control (MPC), weather forecasting, occupant preferences to a “climate box” that solves the control and modeling for small residential properties.

2.2. Control oriented emulator

Dr David Blum from Lawrence Berkeley National Labs (LBNL) gave us amazing insight into the work being done at LBNL and the development of open source building simulation and modelling tools such as BOPTEST, Energy Plus, Spawn of Energy Plus (SOEP), Modelica, FMI Standard, Open Building Control as well as moves in Control Description Language (CDL), Model Predictive Control, as well as work in the IEA annexes.

Other participants pointed to the inputs required to develop and build optimal control strategies that incorporate data from energy use, grid interaction, CO2 emissions and thermal comfort. Demand response, weather forecasts and electricity market spot prices were also discussed as inputs.

2.3. Data Clearing House (DCH)/ Open data platform

The DCH prompted the most wide ranging discussions as it is the “cross-cutting” activity that bridges together all other aspects of IC#7. Dr Stephen White lead the discussions with group sessions to flesh out the requirements, paths to market and challenges of such a venture.

2.4. Demand Response (DR) and Flexibility

With the integration of generation and storage technologies into buildings, the question of how best to manage these systems from a grid perspective is raised. The buildings should operate in concert with the grid, so that both sides are not working against each-other, rather they are working in concert to reduce costs, improve reliability and reduce emissions. Flexibility in consumption and demand is required to produce a responsive and adaptable grid.

2.5. Fault Detection and Diagnosis (FDD)

Currently FDD is done with heuristic rules but there is plenty of scope for AI/ML to take over and add significant value. The issue faced by researchers is the lack of available data from buildings from which to draw these conclusions. Buildings are not connected and this makes inroads into FDD difficult. It is estimated by the IEA that FDD can save 30% of energy costs. It was pointed out by TNO that FDD also improves occupant comfort.