The actual operation of any building is an undocumented art that is learned through years of working closely with the building systems. This begins before handover with designs that don’t account for system interactions or undocumented sequence tweaks by the controls contractor.
They are propagated through poor documentation and hand-over training procedures that expect the client’s facilities management teams to decipher complex strategies, unaddressed design issues and cryptic building management system (BMS) screens.
And when the facilities management team finally learns all the little tricks and interactions of the systems, it all changes with poorly documented retrofits, zones bisected by new office walls and user overrides.
Facilities teams must struggle to contain escalating operating expenses and tightening resources in the face of these challenges. Given the long-term use of buildings, the risks affiliated with maintenance increases over time due to several factors – isolated and captured knowledge within senior facility staff and loss of knowledge/expertise at each transition of people or systems.
Each has its own nuances that can make them greater than the others in contributing to the risk profile of the building.
Isolated and Captured Knowledge
Most facilities teams are severely understaffed. The natural consequence is that teams are forced into reactive mode with little ability to plan, strategize or look holistically at the building management environment. Reactive-mode work means that many issues are fixed in the most expedited manner. Some call that a Band-Aid. In tech companies, it is called accruing technical debt.
When teams are in reactive mode, priorities are often set by the loudest, most insistent complaint or most obvious negative consequence. A fault detection and diagnosis (FDD) service is a classic example of reactive mode. They are designed around what we know will go wrong – sort of like looking for a specific number of velociraptors in Jurassic Park.
As systems age and change occurs, FDD systems generate an overwhelming number of alarms. In portfolios, like a college campus, between FDD and BMS alarms, daily notifications can reach the hundreds. Teams often do not have the resources required to triage, prioritize and troubleshoot pages of alarms to translate them into actionable work.
Meet Bob, the Experienced Facilities Manager
Today’s answer to this is a facilities manager we’ll call Bob. Bob has worked the building system for 25 years. Bob keeps the lights on. Bob knows that zones One and Nine in Building L-2 have never, and will never, hold the right setpoint because their boxes are undersized.
He knows that opening the window in the hallway affects the pressure differential in the adjoining lab and that the open office always overheats because of the reflection off the steel-clad building across the street. Bob knows what you need to fix and when.
- How to we de-risk Bob when he gets sick or retires?
- How do we help Bob transition from fixing what is broken to maintaining what would break if ignored?
- And can we do that in a way that lets us document and share the knowledge?
Using a cloud-based asset modeling and diagnosis system is the answer. Bob is invaluable, and we will always need Bobs, but Bob cannot scale past one.
We need to scale Bob through harnessing the data, history and relationships that form this in-depth knowledge. There needs to be a central source of knowledge documenting overrides, performance issues, outcomes from troubleshooting, issues being tracked, approaching end-of-life issues for assets and system dynamics.
With data-driven machine learning tools, Bob’s knowledge that a certain window actuator affects pressure in another room could be captured in models that quantify and visualize the interactions between all points and equipment in the system.
Loss of Expertise During Transitions
Now that we have met Bob, what happens when Bob actually does retire?
Building owners live with the cost and complexity of re-attaining knowledge when Bob leaves. In a perfect world, every change and its justification has been accurately documented, sequences and tweaks are cataloged historically and nothing is ever deleted.
Without the cloud to de-risk, this just does not exist. The typical BMS has a two-week history of actions and data. Pushed, 30 days.
Cloud companies can hold data – all the data – for 10 years. Expand the definition of data from simply the sensors and energy usage from equipment to include the anomalies detected, troubleshooting and resolution, pre- and post-data trends for changes, all the messaging between building team members or a cloud service provider that led to a diagnosis, and more.
When Bob leaves, the cloud data is a substantial de-risking tool.
Building owner/operators are at financial risk of uncontrolled escalations in the cost of managing their buildings in the future.
The factors contributing to this risk are:
- Staff-based Knowledge – siloed knowledge within senior facility staff
- Knowledge Loss – loss of knowledge/expertise with each transition of people or systems
- Knowledge Depth and Breadth – staff need deep knowledge of complex technology and programming skills
- System Complexity – highly complex building system interactions and strategies
These issues drive buildings reporting operational energy up to 5-9 times design, and 75 percent of buildings not performing according to design.*
Info and diagram provided by Chris McClurg and Building IQ
*Allerd Stikker. Closing the Gap. Carbon Trust, 2002
Consider also new construction. Typically, a central repository of documentation on how the building actually works does not exist. Construction documents are often not updated with as-built changes or the reality of unaddressed design issues. The commissioning report may have glimpses, but this is often lost in a folder somewhere on someone’s hard drive soon after handover.
This evolving knowledge base, which now includes operational issues and user interactions, is then not captured in operational repositories like work tickets or space retrofit documentation. Again, it is up to the facilities management team and Bob to decipher this new living, breathing, evolving building and retain this knowledge in their heads.
However, if we embed cloud services in the commissioning process, we are documenting at every step.
For instance, cloud-based commissioning would automatically build a record of which equipment influences other equipment and to what degree. This is knowledge inherent to Bob over time.
The cloud-based systems de-risk both Bob’s departure and help reduce the cost of hiring and managing new people. Cloud systems, done right, mean that owner/operators do not have to find a new Bob. Instead, they can groom new Bobs that are all leveraging the knowledge held in the cloud.
Lack of Deep Knowledge in Ever-Expanding Areas
There is a subtle difference in knowing why things work as compared to how things work.
This is critical for getting to the root cause of system issues and interactions. Here is where the data science comes in. Building drawings and operational practices help us understand the how of things. Data science can help us understand the why.
Data science and deep analytics examine the hidden factors that only years of experience can tell us (e.g., Bob’s hunches). Cloud-based systems enable the discussion – in this case, looking at all the data from the BMS and applying algorithms that correlate how equipment relates to one another, or one to many others. Doing so over time builds the why of the building.
We can now not only see the chilled water secondary loop valve cycling, but also see that this point is being driven by upstream points, like boiler enabling, and also driving downstream system response, like AHU discharge air temperature cycling.
Adding to the challenge, technicians today have to be half mechanical experts and half programmers. It can be extremely challenging to find techs with deep knowledge in both areas. In fact, some universities are even partnering with local community colleges to create their own “supertech” apprentice programs just to create a pipeline of fully qualified techs.
Going back to the first de-risking scenario, deep system knowledge will help Bob justify that he is allocating his team’s resources most effectively. His team will spend less time troubleshooting and chasing down controls subs for answers. Ultimately, this deep knowledge resides with the building owner/operator, protecting them from expensive maintenance contracts and unnecessary retrofits.
Ever-Increasing Complexity in Building Systems, Interactions and Sequences
With the advent of high-performance buildings, passive design strategies and new technologies, facilities teams are being confronted with ever-increasing complexity and decreased system robustness.
Emerging research (including an examination of Loxford and Brine Leas schools in the UK) has shown that for every additional system interacting in the space, actual energy consumption over the design intent multiplies.
This compounds the problems around de-risking the operation and management of these buildings. Every new “smart” system is another opportunity to compartmentalize knowledge, increase the impact and cost of losing key personnel and further embed a reactive posture into building management.
One methodology that addresses complexity is to combine professional expertise and advanced machine-learning tools in such a way as to capture and document this complexity.
Instead of being constrained by the static, out-of-date sequence of operations (which may have been wrong from the beginning), solutions based on a human/cloud methodology could learn the actual interactions through both observations, training (by humans) and modeling a system’s response to all other variables in the system. This provides a critical tool for facilities teams to understand the complex interdependencies in their buildings and respond quickly and efficiently.
As the pace of technology and data integration with building continues to accelerate, paradoxically, so does the risk of operating and maintaining these highly complex systems.
Facilities teams’ knowledge must be simultaneously deeper, to understand complex system interactions, and wider, encompassing programming, complex controls strategies and new technology. It is extremely difficult to recruit and retain these resources, especially in the face of labor shortages as retirement age for much of the workforce approaches.
The Answer: Cloud-based Building Management System
Cloud-based analytics platforms go beyond traditional fault detection and energy optimization services to provide a central knowledge management platform to manage the significant risks of siloed knowledge in a workforce that’s difficult to train and retain.
By utilizing a cloud-based platform from initial handover, all issues, system dynamics and documentation can be held in one central source of truth.
As the team learns the building’s dynamics and quirks, and as the building inevitably accumulates overrides and issues, this central resource will document these changes and update the dynamic models to the new system relationships and interactions.
In the face of so many process and documentation challenges, building managers need a system that can grow and evolve with their portfolios and their facilities staff to de-risk the operational and cost issues tied to the siloed, undocumented knowledge base required to weather these challenges.
Two hand-picked articles to read next: