Fault Detection and Diagnostics (FDD) software has become a standard recommendation in commercial building energy programs. ASHRAE Guideline 36 references it. LEED EB credits reward it. Major BMS vendors offer it as an add-on. The premise is compelling: continuously analyze equipment sensor data, flag deviations from expected behavior, and surface actionable maintenance alerts before they become costly failures or extended energy waste.
We have no argument against FDD. It does what it claims. A properly tuned FDD system on a commercial HVAC installation will catch stuck dampers, failed economizer actuators, sensor offsets, and simultaneous heating and cooling (the infamous simultaneous H&C fault that can run undetected for months and waste 10–15% of HVAC energy). That is real value.
The question we want to address is different: once you have FDD deployed, is optimization still a separate, additive investment? Or does FDD coverage mean you have "solved" your energy management problem? In our experience, most facilities that have invested in FDD have not solved it — they have addressed one category of energy waste while leaving another entirely unaddressed.
What FDD Catches, and What It Misses
FDD is designed to detect equipment operating outside its normal performance envelope. The underlying model is: this equipment should behave like X under these conditions. When it deviates from X by more than a threshold, raise an alert.
This model is excellent at catching hardware and control faults: a cooling coil valve that is stuck open when it should be closed, a supply air temperature sensor that has drifted 4°F from calibration, a variable air volume (VAV) box terminal damper that fails to modulate. These are abnormal operating states — the equipment is behaving differently from design intent.
What FDD does not catch is normal equipment operating according to its original schedule and setpoints when those schedules and setpoints are no longer appropriate. A BMS programmed in 2016 with occupancy schedules based on a different tenant mix, cooling setpoints based on equipment that has since been replaced, and ventilation rates based on occupancy assumptions that have changed — this is not a fault. From the FDD system's perspective, everything is working as programmed. From an energy perspective, a significant fraction of that operation may be unnecessary.
The typical examples we encounter:
- AHUs starting 90 minutes before first occupancy because the original commissioning assumed a worst-case warm-up time that no longer reflects the building's current thermal characteristics
- Cooling setpoints unchanged since a major renovation that significantly improved the building envelope's insulation — the HVAC is working harder than necessary to cool a space that holds temperature much better than it did
- Ventilation airflow at design maximum rates on evenings and weekends when occupancy is a fraction of design values — code-minimum ventilation might call for 20% of design airflow, but the BMS was never reprogrammed
- Chiller staging logic unchanged since original commissioning, still starting a second chiller at load thresholds that made sense with the original equipment efficiency curve but are now incorrect after a compressor replacement
None of these conditions will appear in an FDD alert log. They are not faults — the equipment is doing exactly what it was told to do. The waste comes from instruction sets that no longer match operational reality.
The Two Sources of HVAC Energy Waste
It is useful to think about building energy waste as coming from two distinct sources that require different tools to address:
Equipment failure waste — energy consumed because equipment is malfunctioning or degraded. Stuck valves, failed sensors, degraded heat exchangers, refrigerant leaks. FDD is the right tool for this category. The economic impact is bounded by the scope of the malfunction — a stuck valve wastes energy proportional to the one valve's contribution to the system.
Scheduling and setpoint waste — energy consumed because the control strategy is not matched to current operational requirements. This waste is not bounded by a single component — it scales with the entire system. An oversized morning pre-conditioning window wastes energy across the entire cooling plant for the entire duration of the unnecessary pre-conditioning period.
In a well-maintained building with a recently tuned FDD system, equipment failure waste is often 5–15% of total HVAC energy. Scheduling and setpoint waste in the same building — with schedules and setpoints that were set at commissioning and not revisited since — is often 15–30%.
FDD addresses the smaller category. Optimization addresses the larger one. They are not substitutes for each other.
Why Scheduling Waste Persists Even in Well-Run Buildings
The reason scheduling waste persists even in facilities with competent engineering staff is not negligence — it is bandwidth. Reviewing and updating BMS schedules and setpoints is time-consuming work that requires expertise and careful testing. A facility engineer responsible for multiple buildings, managing contractors, and responding to maintenance requests does not have the time to re-analyze optimal pre-conditioning windows for 15 AHUs across 3 buildings every quarter.
The original commissioning schedules get copied forward indefinitely. They are conservative by design — commissioning engineers set schedules to guarantee comfort under worst-case conditions, knowing that no one will be blamed for starting HVAC too early, but complaints will come if it is started too late. That conservatism bakes in systematic over-conditioning that persists until someone explicitly revisits it.
Automation changes this. An optimization system that continuously models building thermal response and updates pre-conditioning start times based on current conditions effectively performs the schedule optimization work that would otherwise sit on an overworked engineer's backlog indefinitely. It is not replacing engineering judgment — it is applying engineering principles continuously at a rate that manual processes cannot match.
Can One System Do Both?
There is a genuine question about whether FDD and optimization should be the same software or different systems. The vendors in both spaces have opinions on this, and the answer is evolving.
Some FDD platforms have added optimization features, and some optimization platforms — including ours — surface fault-like alerts when observed equipment behavior deviates significantly from predicted behavior (which can indicate a hardware fault rather than a scheduling issue). There is real overlap in the sensing and telemetry layer: both categories need BMS telemetry, equipment state data, and historical performance data.
Our view: in the near term, they are functionally different tools with different user interfaces, different response workflows, and different value propositions. FDD outputs go to the maintenance workflow — someone needs to dispatch a technician, order a part, and close a work order. Optimization outputs go to the control layer — the system makes an automated schedule adjustment, or surfaces a recommended setpoint change for an engineer to approve. These are different operational workflows even if the underlying data is the same.
Longer term, the distinction will probably blur as platforms mature. For now, a building with FDD already deployed should evaluate optimization as an additive investment, not as a redundant one.
A Practical Decision Framework
If you are trying to decide where to invest first — FDD or optimization — here is how we think about it:
Prioritize FDD if: your building has older equipment with degraded components, you do not have active commissioning on your HVAC controls, or you have had significant comfort complaints or unexplained energy spikes that suggest equipment faults rather than scheduling issues. FDD gives you a foundation of equipment health before optimization has much to work with — trying to optimize an AHU with a stuck damper is fighting the underlying problem rather than addressing it.
Prioritize optimization if: your equipment is in good shape with recent commissioning, but your BMS schedules and setpoints have not been systematically reviewed since original commissioning or a major renovation. This is the common state for many buildings built or renovated 5–10 years ago — the equipment works fine, but the control strategy has drifted from optimal.
Deploy both if: you have the budget and an operational team that can act on alerts from both systems. The telemetry infrastructure is largely shared, so the marginal cost of the second system is lower than the first. The combined coverage — catching both equipment-level faults and system-level scheduling inefficiency — gives you a much more complete picture of your building's energy performance than either tool alone.
The honest summary: FDD prevents energy waste from things breaking. Optimization prevents energy waste from things running correctly but unnecessarily. Most buildings have both problems. Solving one without the other leaves real money on the table.