3 ways building controls must evolve
With many exciting advances in data analytics, machine learning, grid integration, user interaction and renewable energy technologies entering the buildings industry, it is easy to get caught up in the vision and potential of high-performance buildings. The unfortunate reality is that many buildings designed to high-performance standards struggle to deliver on this vision.
In fact, studies in the U.K. have found that 75 percent of high-performance buildings are not meeting their energy goals (PDF) and are in fact using two to five times more energy than their models predicted (PDF). Instead of living the dream of a comfortable, super efficient and intuitive building, many owners are experiencing operational issues, excessive energy consumption and user complaints.
Many reasons for this are cited, from occupant engagement to lack of integrated design, but the fundamental reason often can be traced to the building controls. This system, typically the responsibility of a subcontractor, dictates the performance of buildings and yet is invisible to most owners and users.
High-performance buildings demand more functionality, system integration and dynamic computation, but market and technical barriers prevent control systems from keeping up with the rapid pace of building and energy system innovation. Unless these systems can evolve to support modern buildings, they will remain the missing link undermining the future of the high-performance buildings industry.
1. Controls must meet high-performing building needs
Modern, highly efficient buildings — and net-zero buildings — are not only fundamentally designed differently, but must operate in an infinitely more complex world of data, weather and occupancy prediction, grid interactivity and system integration. Yet the majority of today’s hardware and software controls still cling to approaches rooted in the pneumatic systems from the 1980s, dominated by centralized heating and cooling equipment.
This legacy approach prevents buildings from delivering all of the services required of modern, high-performing buildings: providing optimum occupant comfort, tracking and maintaining energy performance, interfacing with occupants and facilities personnel for efficient operation, providing grid services, integrating and managing all energy systems including electric vehicles (EVs) and energy storage, enabling flexibility to adapt to building retrofits and growth — and most of all doing this in a manageable, adaptable, easily implemented and operated fashion.
2. Support systems integration for strict building performance goals
As buildings become more passive to meet efficiency goals, thermal comfort is provided primarily through passive elements, such as solar control, natural ventilation and predictive thermal mass strategies. Instead of single, large HVAC equipment dominating the thermal control strategy, multiple, distributed passive systems interact within the space to control multiple variables simultaneously (such as thermal comfort, lighting, ventilation).
These complex interactions have many interdependent variables requiring integration of each subsystem. Similar to rapid expansion of smarter grid controls enabling the utility grid to move from centralized fossil fuel generation to distributed supply and demand assets, building controls must evolve to meet these fundamentally different paradigms. In fact, the Carbon Trust studied 28 buildings that were part of the Low Carbon Buildings Program and found that no building that integrated multiple low-carbon technologies met its design energy targets. The deviation was even greater when multiple technologies were integrated into one system, such as heating. Controls no longer can rely on straightforward single systems that can be oversized to compensate for poor integration.
To add to this complexity, buildings are now an integral part of a wider energy system encompassing EVs, battery storage and a next-generation grid that is developing interactive rate structures to manage the new demands of distributed renewables. A building’s energy consumption is a key variable to be balanced alongside battery storage, EV charging, or even bi-directional charging. These demand-response mechanisms are crucial to the developing smart grid. In the past few years, energy software companies have emerged, but each focuses on a component of this wider system. For instance, one can control building energy flows from solar photovoltaics, batteries and electric vehicles, but can give only simple commands to the existing building management system or lighting control system to control building load. To develop integrated demand-side management for net-zero districts, a fully integrated strategy encompassing all of these elements is needed.
3. Evolve beyond simple rules-based algorithms
In addition to these integration challenges, control strategy paradigms are shifting from the reactive, instant response of spaces dominated by large HVAC systems to lagging, highly variable passive responses requiring predictive control strategies.
Current rules-based control sequences predominantly react to real-time space conditions based on a limited number of variables. Newly emerging technology moves away from canned control sequences and instead models all possible future outcomes based on current conditions, possible inputs and real time — and uses this information to control the building.
Lawrence Berkeley National Laboratories (LBNL) is developing one application of this using Modelica, but the project is in early research. Small startups are also pursuing this approach, but they are encountering strong investment and market resistance due to lack of venture capital interest in hardware, entrenched incumbents and the risk-averse nature of the engineers and contractors specifying these systems.
There are also many other technical approaches for integrated, predictive building controls, such as the use of machine learning, which the industry has not fully explored due to historical barriers and lack of domain experience. Google (PDF) has even used neural networks to optimize its data centers, resulting in 40 percent savings in an already highly efficient system. However, these promising approaches require a large amount of good quality historical data, and although many new buildings have the extensive submetering and trending required to provide this data, the data quality and management have systemic issues. In addition, new buildings do not have a year’s worth of performance data to immediately control the building after handover.
A center of innovation
When Rocky Mountain Institute approached the design of its Innovation Center — which was occupied in 2016 — it wanted to push the boundaries of what a state-of-the-art beyond net-zero energy building could accomplish and understand the controls needed to meet the goals over time. To achieve net-zero energy, we needed a passive building that heated the occupants, not the air. We used a distributed approach with automated external shades, targeted under-floor radiant heating, automated natural ventilation, manual ceiling fans and dedicated outside air units. The high thermal mass of the Innovation Center uses weather predictions to precool the slab during a night flush and to warm the slab with large south-facing windows controlled by the external blinds.
Understanding the controls industry challenges mentioned above, we did everything an owner could do to ensure we would receive an integrated system that worked. We brought on the controls contractor and commissioning agent during early design using an integrated project delivery process. We worked with the design team to ensure all design specifications had the correct controls protocols and worked extensively with each subsystem manufacturer, from the blinds to battery controls, to ensure the integration was not only technically possible but would also be fully resourced and supported.
In spite of this level of planning, foresight and collaboration among all parties responsible for the system, we encountered many of the same issues heard about from industry colleagues time and time again, which — in our experience — was a clear illustration of today’s technology gaps and the state of the industry.
For example, the extensive energy submetering system has had consistent integration and reliability issues. While many of our systems were Modbus compatible, we found only limited points were uncovered and accessible by the central control system, resulting in reduced functionality. Subsystems with internal control algorithms have led to operational confusion around the source of system actions and difficulty in troubleshooting control sequences.
We have been able to draw on our internal technical resources to find solutions for many of these issues, but building owners normally would not have the resources or time to do this. Buildings should not require an entire engineering organization to understand the controls and ensure they run well. The year it took us to optimize the Innovation Center’s controls — a relatively small commercial building at 15,400 square feet — is not an aberration. We used the best technology available and the best partners we could find; in fact, the Innovation Center exceeds its designed energy use. Rather, our experience is an illustration of the systems-integration challenge facing the controls industry today.
Moving in the right direction
Significant strides have been made in the past 10 years to enable an open controls integration platform leveraging protocols such as Modbus and BACnet and open software such as Tridium. These approaches have taken big steps in proving the market desire for integrated systems that do not tie clients to specific providers. However, these approaches are falling short of seamless open integration, as many incumbents have no incentive to truly open their system beyond exposing a few token points. Most systems require extensive resources to integrate hardware using these protocols because they only provide simple inputs, such as on/off and hide the largely proprietary control algorithms.
Many startups and research institutions are banging on the door of the incumbents and pushing this revolution. In addition to work by LBNL, Pacifica Northwest National Labs (PNNL) has developed the open source Volttron platform, which "can independently manage a wide range of applications, such as HVAC systems, electric vehicles, distributed energy, or entire building loads" through an object-oriented software environment, allowing easy application development while being agnostic about programming language and controls protocol.
In addition, many new startups are introducing advanced data analytics, machine-learning-based controls and open integration of platforms. Many of these advances are not being driven by the value of enhanced building controls, but instead by capturing the value of integrating building loads, energy storage, solar power production and electric-vehicle charging in the rapidly evolving world of advanced utility rate structures. These approaches hold great promise, but face significant market barriers due to large incumbents and the presence of legacy systems in all portfolios.
The coming revolution
Given RMI’s experience with the Innovation Center, we passionately believe these controls integration and operational issues are one of the biggest barriers to the building industry evolving and joining the 21st century of "smart" buildings. Many startups and new software solutions deliver advanced analytical and optimization approaches along with grid integration, but all are built on the fundamental assumption that the base building controls and data-management systems enabling these approaches are working. Many of these startups are finding that their solutions cannot be fully applied or are failing because the buildings are not functioning at a high enough level to deliver the data and stability these systems need.
Controls have an opportunity to emerge as the single most important aspect determining a building’s operational performance, but only if the industry can deliver a fully integrated value proposition that encompasses predictive modeling, data analytics and energy system integration and that also fundamentally works for building owners and occupants. RMI believes that when these new approaches can provide a fully integrated value proposition, they will drive market demand. But until then, building controls will continue to be a missing link, limiting the forward progress of high-performing buildings.