Case Study: Microsoft Redmond Campus
In or around 2010, Microsoft started a project focusing on energy management and improved building operations. The project involved the integration of seven building management systems used on the campus, and the deployment of fault detection and diagnostics across its campus. The campus has 15 million square feet of space in approximately 120 buildings.
The initial pilot of the FDD application consisted of 2.5 million square feet of space, where three potential contractors demonstrated their software, prior to selecting a contractor and deploying the application across the campus. Although Microsoft benefits from the third lowest utility rates in the country, they were able to save over $1 million in energy cost the first year from the use of FDD.
Microsoft invested 10% of the yearly energy costs in the deployment of the FDD application, and had a payback of less than 18 months. Microsoft previously had a process of recommissioning each of its buildings every five years. After one year with the FDD application they are not recommissioning each building every year; saving energy, and operational costs.
Definition of a fault and problems that can be detected with FDD
A fault is a deviation in the value of at least one characteristic variable from its normal expected behavior. Faults that can be detected through FDD include:
HVAC systems that improperly simultaneously heat and cool Excessive outdoor air intake and conditioning Under-utilized free cooling potential
Equipment malfunction (such as broken/leaking valves, broken/stuck dampers, sensors out of calibration)
Systems with the wrong setpoints and operating schedules Unintentional manual overrides
Lack of energy-saving control sequences (such as chilled water reset)
A bad bearing in a motor or compressor (the bearing then can be replaced
before the whole gear box fails and becomes an emergency repair)
Misaligned motor, rotor imbalance, or cracked rotor bar
Dirty filters or strainers
Incorrect refrigerant or oil levels
Pumps with throttled discharges
Short cycling of equipment
Excessive oscillation (hunting) of control points and/or control loop tuning needs
Incorrect fan and pump speeds, pressures, or low flow rates.
Improper building or space pressurizations (negative or positive)
Inefficient boiler combustion
Excessive building peak electrical demand
MICROSOFT Illustrative example of fault detection and diagnosis output (simplified). Microsoft is now collecting 500 Million data point values every day, and using that data to create energy and operational savings. The company expects to reduce energy consumption by 10 percent.
The use of analytic software applications have demonstrated positive measurable results. A study by Lawrence Berkley Labs titled "Automated Continuous Commissioning of Commercial Buildings in 2011, indicated 30% reduction in building total energy consumption and related costs over the baseline; and 30% reduction in building peak demand and CO2 emissions On top of that are operational savings related to increasing personnel efficiency and effectiveness; facility engineers and technicians being more quickly alerted to a fault in a building system, provided improved information on potential system remedies, and monetizing faults to indicate the wasted energy. Given their track record, analytic applications have been successfully used in lighting systems, electrical distribution, conveyance equipment, data centers, etc.