INTRODUCTION

Evidence of extensive performance problems in buildings shows that an efficient California building stock will not result solely from designing efficient buildings and installing efficient equipment in them. Comprehensive, top-down analyses by utility planning agencies of billing data for the population of new commercial buildings have shown consumption 10% higher than levels projected. This is determined by construction characteristics, even when based on metered end-use loads. These analyses support the conclusion that even newly constructed buildings are consuming more energy than they should.

These performance problems are not inherent in efficiency technologies themselves, but instead result from errors in installation and operation of complex building heating, ventilation, and air conditioning (HVAC) and control systems. These systems are becoming increasingly more sophisticated to obtain higher levels of energy efficiency, but this adds to the complexity and subtlety of problems. Problems are even more common in existing buildings, arising over time from operational changes and lack of maintenance. They often result in comfort control and indoor air quality problems that affect occupant health and productivity.

The traditional means of assuring efficient performance, commissioning of new buildings followed by regularly scheduled preventative maintenance, are clearly insufficient to address this issue. Manually commissioning buildings is valuable in terms of finding problems and developing techniques but does not yield persistent results.

Automated fault detection and diagnostics (FDD) methods are powerful tools that can extend the benefits of manual commissioning and regular maintenance. Element 2 had seven projects focused on finding ways to detect subtle, energy-wasting problems, or faults, within heating and cooling equipment and control systems. Finding a problem begs the question of what caused it, and much of the effort in these seven projects was devoted to isolating the causes of faults and estimating their impact on the energy and comfort performance of the heating, ventilating, and air-conditioning (HVAC) systems.

Project 2.1, Fault Detection and Diagnostics for Rooftop Air Conditioning, focused on FDD for packaged HVAC units, which use about 54% of heating, cooling, and ventilation energy use in California. Project 2.2, Equipment Scheduling and Cycling, used a novel electrical meter to look for scheduling and cycling faults in medium to large HVAC systems. Project 2.3, Air Handling Unit and VAV Box Diagnostics, continued development of automated rules to find faults in built-up air handlers and variable air volume boxes. Project 2.4, Demonstration of the Whole Building Diagnostician, gathered information about user interaction with an outdoor economizer diagnostic software module. Project 2.5, Pattern Recognition Based Fault Detection and Diagnostics, developed the software specifications to fully automate a set of diagnostic methods that had been successfully applied manually in the past. Project 2.6, Enhancement of the Whole Building Diagnostician, added new capability to the whole building energy diagnostic module making it more flexible and useful for most buildings with building automation systems. Project 2.7, Enabling Tools, integrated a new capability to accelerate developing and testing FDD software in a laboratory and field environment.

back to top

2-1. FAULT DETECTION & DIAGNOSTICS FOR ROOFTOP AIR CONDITIONING

Packaged air conditioners are the most poorly maintained type of HVAC system. In California, they use about 54% of the HVAC energy in the commercial sector. The Purdue research team developed thermo-fluids based fault detection and diagnostic (FDD) methods that can pinpoint five common maintenance problems.

• This project was highly successful, resulting in a cost-effective method to detect simultaneous faults using only temperature sensors and models of normal operations.

• The FDD methods are effective for packaged air-conditioning units with fixed orifice or thermal expansion valvles.
  
  
• Controllers that embed these diagnostics methods will save energy and maintenance costs by providing alerts only when maintenance is needed and giving the mechanic better information.
  
  
• The historical data from the diagnostic system will also serve as a database for manufacturers to improve the reliability of components.

Research Team: Jim Braun and Haorong Li with Purdue University conducted the research with support from Todd Rossi and Doug Dietrich with Field Diagnostic Services, Inc.

back to top

2-2. EQUIPMENT SCHEDULING AND CYCLING

• Motor loads such as reciprocating chillers, fans, and pumps can be identified and tracked using a single NILM.
  
  
• Loads turning on or off outside of the normal schedule can be detected, as well as abnormal cycling.
  
  
• The MIT research team achieved a major breakthrough by developing a method to detect certain loads driven by variable speed drives. In addition, they made considerable progress in automating the detection and tracking algorithms.

Research Team: Les Norford, Steve Leeb, K. Douglas Lee, Peter Armstrong, and Chris Laughman with MIT conducted the research. Lanny Ross with Newport Design Consultants provided field support.

back to top

2-3. AIR HANDLING UNIT AND VAV BOX DIAGNOSTICS

• Two rule sets (APAR and VPACC) were thoroughly tested in a NIST laboratory facility called the Virtual Cybernetic Building Testbed (VCBT) as well as using data from a half dozen field sites.
  
  
• NIST worked with three major building control manufacturers to embed these rules in their respective controller products using the native programming language of each. A fourth manufacturer recently expressed interest in the next phase of development, which will entail testing at dozens of facilities to prove the reliability of the algorithms in different HVAC systems and facility types.

Research Team: Steven Bushby, Natascha Castro, Jeffrey Schein, Cheol Park, and Michael Galler with NIST conducted the research, along with John House with the Iowa Energy Center.

Objective:

Develop and demonstrate diagnostic methods for air-handling units and variable-air-volume boxes in real buildings, including buildings owned and operated by control system manufacturers.

Baseline Conditions:

Current automated fault detection and diagnostics of AHUs and VAV boxes typically consists of single point alarms on critical temperatures and pressures. These alarms will protect equipment from certain types of catastrophic failures, but in general will not detect more subtle problems that lead to energy waste, equipment wear and occupant discomfort. The commercial penetration of automated FDD technology today is near zero.

back to top

2-4. DEMO OF THE WHOLE BUILDING DIAGNOSTICIAN

Air handlers in commercial buildings often do not function properly due to sensor faults, control problems or scheduling errors. The objective of this project was to evaluate the usability of the Outdoor Air Economizer (OAE) module of the Whole Building Diagnostician (WBD) software, and to field test the software under three types maintenance management arrangements. Battelle, which developed the WBD under contract with US DOE, trained and supported use of the OAE by the operations staff of a large commercial office building, the energy manager of a government building campus, and the controls manager for a mechanical contractor providing services to a large commercial office building owner.

• The OAE was successful in detecting sensor and hardware problems as well as control setting problems at all of the demonstration sites.
  
  
• Fully automated data collection from some building automation systems was a challenge for the WBD. This can be overcome by working with BAS vendors and will, over time, likely disappear as better communications standards come into use.
  
  
• Once problems were identified by the OAE, too often no action was taken to make repairs. This suggests that a mechanism is needed for delivering the results to users in a way that better encourages them to correct the problems found, which may require changing incentives and rewards to inspire action by building staff.

Research Team: Michael Brambley, Srinivas Katipamula, Rob Pratt, and Nathan Bauman with Battelle conducted the research. David Jump with Nexant and Lanny Ross of Newport Design Consultants provided field support to the participants.

back to top

2-5. PATTERN BASED FAULT DETECTION AND DIAGNOSTICS

Pattern Recognition Diagnostics, led by Battelle with participation by Architectural Energy Corporation, pursued the automation of proven diagnostic methods that were manually exercised by an expert engineer using short-term data collection. The manual methods were embodied in Architectural Energy Corporation's ENFORMA® diagnostic software.

• The research team selected several diagnostic problems associated with chillers, boilers, cooling towers, and pumps, since other projects within the program were focused on air handlers, economizers, and VAV boxes,
  
  
• Review of the visual diagnostic process for these components indicated that the best automation method would be to use rule-based methods.
  
  
• A complete specification for the automation process was produced and a limited demonstration in prototype software was completed illustrating the efficacy of this approach. Field testing was not possible because of a lapse in co-funding in the second year of the project.

Research Team: Rob Briggs and Michael Brambley with Battelle, and Stuart Waterbury with Architectural Energy Corporation, conducted the research. S. Gaines and R. Lucas with Battelle provided project support.

back to top

2-6. ENHANCEMENT OF THE WHOLE BUILDING DIAGNOSTICIAN

Enhancement of the Whole Building Diagnostician, provided a significant improvement to the WBD. Increases or decreases in whole building, or building systems, electrical or gas energy use may be caused by system faults or changes in occupant activity. For example, energy use will increase with increased sales in a restaurant, but it will decrease if an HVAC unit went off line due to compressor failure. The Whole Building Energy (WBE) module of the Whole Building Diagnostician, which is designed to flag anomalies in energy use patterns, was improved with changes to allow it to be used on a wide spectrum of building types.

• The added feature allows the user to specify any BAS variable or other accessible variable (such as sales volume) as one of up to five independent variables in the WBE module. The previous versions allowed only outside air temperature, outside humidity, and building schedule as independent variables. This will allow operators of similar buildings, such as chain retail stores and restaurants, to compare energy performance and spot positive and negative trends over time. The value to California chain managers could be a significant motivating factor to use or install energy management control systems.
  
  
• Methods and user interfaces for a second planned improvement, a module for energy use comparisons among peer buildings that could be used in near real time, were developed and partially implemented, but full implementation was not possible because of a second-year lapse of co-funding.

Research Team: Michael Brambley, Krishnan Gowri, and David Chassin with Battelle Memorial Institute, Northwest Division led the research team.

back to top

2-7. ENABLING TOOLS

As more FDD methods and products become available, selecting the most effective ones will be difficult because of the complexity of products and the lack of repeatability in faults. NIST previously addressed this problem by developing a laboratory, called the Virtual Cybernetic Building Testbed (VCBT), in which controllers and faults could be tested using building simulations. The FDD Test Shell was developed in this project to test FDD methods. These tools were then used to test the Whole Building Diagnostician.

• The project proved that VCBT/FDD Test Shell can independently and objectively assess the capability of new FDD tools quickly in a controlled environment. This tool will allow manufacturers and their prospective customers to test variations of FDD tools required for specific installations.
  
  
• The tool will accelerate the acceptance of FDD methods in the marketplace by giving building owners confidence that spending limited budgets on FDD features for their building automation systems will be a good investment.
  
  
• Testing of the Whole Building Diagnostician showed that it was successful in detecting eleven of fifteen faults without false alarms under normal sensitivity settings. Two of the undetected faults were due to "climate screening" (for example, a drift in the calibration of a return-air temperature sensor when the outdoor-air temperature and return-air temperature are nearly equal).

Research Team: Steve Bushby, Natascha Castro, Michael Galler, and Cheol Park with NIST created the enhancements to the VCBT and the FDD Test Shell, along with John House of the Iowa Energy Center. Srinivas Katipamula and Michael Brambley of Battelle participated in the joint research task of blind testing the WBD, along with Jeffrey Schein of NIST.

back to top