Subtask A: Open Data and Data Platform (Stephen White, CSIRO and Chun-Ping Gao, BCA)

Data from a wide variety of sources (including Building Automation Systems (BAS), Internet of Things (IoT) sensors, and mobile devices) are now accessible in the cloud. The diversity and granularity of these data sources can be used to drive new data mining and machine learning ‘Applications’, that reduce energy consumption and enhance productivity in buildings.

Unfortunately, the HVAC industry has been a laggard in the adoption of these emerging digital technologies, and has not gained the full benefits of Information Technology/ Operations Technology (IT/OT) convergence. Companies struggle to consolidate data, from multiple disconnected systems and service providers. Lack of interoperability, and proprietary vertically-integrated building automation systems, often force innovators to invest significant extra effort getting access to sites, creating custom interfaces to equipment, decoding point labels etc.; all before they add any value to the data through the provision of new Application services.

Recent forums have recommended the use of ‘Open Data’ as way to address data-access barriers. This subtask aims to develop methodologies that support the implementation of real-time Open Data sharing in buildings, as a mechanism for stimulating energy productivity innovation in HVAC services.

Subtask A will provide the knowledge, standards, protocols and procedures for low-cost high-quality data capture, sharing and utilization in buildings.  Subtask Activities include

• A.1 Open Data Concepts and Data Governance (Stephen White, CSIRO)
• A.2 Open Data Platforms (Sofia Stensson and Rafael Gomez Garcia, RISE)
• A.3 Data Information Management (Gabe Fierro, Colorado School of Mines and Pieter Pauwels, TUE)
• A.4 Data Sets (Chun-Ping Gao, BCA)
• A.5 Open Data Platform Utilization (Stephen White, CSIRO)

 

Subtask B: Model-Based Predictive Control (José Candanedo, Canmet and Igor Sartori, SINTEF)

Model-based predictive control (MPC) is a technique whereby a mathematical model of a building is used to enable scheduling of the operation of the building as a function of predicted future conditions (such as weather and occupant behaviour). Whereas conventional building control systems react to changes in conditions without any foreknowledge, the proactive “look ahead” approach of MPC makes it possible to optimize the operation of HVAC equipment.  This results in significant improvements in energy efficiency, comfort conditions, load management, and building-grid interaction, among other Applications.

Despite its demonstrated potential, MPC has only rarely been implemented in the operation of real buildings. A practical framework is required for the rapid development and implementation of advanced control strategies in buildings, including control-oriented building modelling approaches, exchange of data and control sequences with physical buildings, and control performance assessment.

Subtask B will develop a roadmap for scaling the use of data-driven model-based predictive control in buildings, including tools for testing and implementing MPC, and methodologies for developing MPC algorithms

• B.1 Test Cases (Igor Sartori, SINTEF)
• B.2 Control Oriented Modelling Methods (José Candanedo, Canmet)
• B.3 MPC Simulation and Evaluation (David Blum, LBNL)
• B.4 MPC Implementation (Henrik Madsen, DTU)
• B.5 Roadmap for MPC (José Candanedo, Canmet)

 

Subtask C: Applications and Services (Jin Wen, Drexel and Ivo Martinac, KTH)

Having access to diverse high-quality data (subtask A) and control-oriented building models (subtask B), creates new possibilities for better understanding of building processes, and for holistic optimization of building performance using cloud based data mining and other software analysis tools.

A wide variety of building services 'Applications' are possible, limited only by the imagination of what could add value to the relevant commercial stakeholders (eg building owners, tenants, facilities managers, utilities etc) and their respective key performance indicators (energy, comfort/health, maintenance cost, etc).  

Subtask C focuses on the development of Applications.  It will support innovators to develop building energy efficiency (and related) Applications that can be used (and ideally commercialized) for reducing energy consumption in buildings and coordinating building energy demand to achieve additional electricity system benefits. 

Subtask Activities include

• C.1 Benchmarking Algorithms (Wim Zeiler, TUE)
• C.2 Automated Fault Detection, Diagnostics and Recommissioning Applications (Zheng O'Neill, TA&M)
• C.3 Building to Grid Applications (Anna Marszal-Pomianowska and Hicham Johra, Aalborg)

 

Subtask D: Case Studies and Business Models (Dimtrios Rovas, UCL and Bo Jorgensen, SDU)

Subtask D aims to compile evidence on the benefits of data-driven smart buildings applications, and disseminate it to stakeholders through the analysis of case studies and business models. Case studies will cover a wide range of representative building typologies, climates and occupant applications with detailed high-quality monitored data and relevant contextual information. Dissemination will include (funding permitting) organising prize based competitions for data driven innovation.

Subtask Activities include

• D.1 Case Studies (Dimitrios Rovas and Paul Ruyssevelt, UCL)
• D.2 Smart Data-Driven Innovation Strategies
• D.3 Dissemination (Clayton Miller, NUS)