Get the latest updates on the Coronavirus impact on engineers.Click Here
Mechanical

Leveraging data to make mechanical systems smarter

Analyzing building data can help to uncover and predict mechanical problems, as well as equip building operations and maintenance staff with and information to be more proactive and efficient while optimizing resources.
By Jason Gladney March 12, 2020
Courtesy: Envise

Learning objectives

  • Understand how automated fault detection and diagnostics and building automation systems can enhance a building’s performance.
  • Leverage the available data in buildings to uncover mechanical problems that may be masked by operating conditions.
  • Analyze data in buildings to predict failures so repairs can be planned.

ASHRAE describes smart building systems and technologies as those that interpret data related to the historical and ongoing building operation and autonomously draw conclusions from that information. One of the more mature aspects of these technologies is related to automated fault detection and diagnostics, which is typically a software solution that digests historical and ongoing building automation system data and automatically processes it through rules and algorithms to detect anomalies such as faulty operation and degraded performance.

Mechanical designers and heating, ventilation and air conditioning contractors don’t like to get call backs for design issues, especially if they become cost-prohibitive fixes like rightsizing undersized equipment that has already been installed. To minimize this likelihood, they often skew diversity factors during the design process as a form of risk mitigation.

There are a variety of reasons that buildings can end up with oversized systems based on the above circumstances. Oftentimes, various trades are working in parallel on the building design and some factors that can affect the required HVAC design load are not fully vetted at the various design iterations, including the building envelope insulation ratings, the number of people occupying spaces that could change significantly over time once occupied and the amount of heat-generating equipment that will ultimately end up in the space (e.g., computers, dual monitors, copiers, plug load variations).

The designer must often make educated guesses at these variables and often this preliminary design is what ends up in the final building due to coordination issues or the fast-track nature of many projects. When a building does have more capacity than required, that additional capacity can potentially mask issues that would otherwise become apparent, such as energy drift due to dirty coils, leaking valves, leaking dampers, overridden BAS or equipment parameters.

Automated detection

A frequently implemented smart building technology is AFFD to determine and present an overview of potential faults or anomalies detected on the equipment and systems operation. Usually a description of each fault is displayed along with the data used to determine the fault and a potential root cause with possible solutions to fix them.

Figure 1: Service technicians are checking a chiller after automated fault detection and diagnostics detects a condenser fouling fault. Courtesy: Envise

Figure 1: Service technicians are checking a chiller after automated fault detection and diagnostics detects a condenser fouling fault. Courtesy: Envise

To help with prioritizing the issues and to help justify the allocation of resources to further investigate and repair the systems, the AFFD typically will rank the discoveries based on avoidable cost resulting from the fault for a specified period, the potential energy and cost savings over time, an analysis of the maintenance impact and a comfort index rating. The information from AFFD can often even be used to develop energy conservation projects from the generated diagnostic findings. If implemented, these projects can then be tracked throughout completion to validate the work done and show cost savings associated with each project or energy conservation measure to help validate the return on investment.

A few of the issues AFFD can usually uncover include:

  • Rogue variable air volume zones falsely driving supply air resets or after-hours air handling unit schedules.
  • Simultaneous heating and cooling in AHUs or in overlapping VAV zones.
  • Energy drift from degradation of heating or cooling performance based on weather normalized meter data.
  • Overridden setpoints, equipment and schedules based on established operational guidelines.
  • Equipment short cycling or fan belt slipping.
  • Leaking or stuck outside air damper, economizer inefficiencies and indoor air quality concerns.
  • Nonfunctioning sensors or controllers which could also result in missing demand response events.

As building systems are becoming more complex and as smart building technologies are evolving and becoming more mainstream, it is necessary for all of the appropriate stakeholders to have the right information, expectations and accurate data to properly understand and maintain their buildings to protect their investment while providing the best environmental conditions for the occupants.

Many building operations and maintenance engineers still rely on their intuition and override building operating setpoints based on their fear of losing temperature control in a building instead of relying on the automation system to make the right decisions. Their fear, often based on misinformation or a negative past experience, is holding them back from reaping the rewards of a more data-driven smart building technology approach. The occupants may remain comfortable, but often at significant unnecessary energy, maintenance and/or equipment costs.

Finding the problem

Several service call examples help illuminate the importance of conducting a robust initial and ongoing commissioning effort on a building project to help unmask the above potential design issues, in addition to the ability of using smart building technology to predict failures and ultimately generate a return on investment for the first and ongoing cost of implementing the technologies.

In the first scenario, on a warm spring afternoon, the chief building O&M engineer of a good manufacturing practice pharmaceutical site in Silicon Valley scrambled to determine a path to increase cooling in his process environment that would not compromise the validated conditions. While everything had been working fine until now, he had reached the edge of his operating parameters and knew the temperature would rise as the week progresses. He needed to find a solution or his company would have to discard hundreds of thousands of dollars in defective products.

Figure 2: When the temperature drop of water flowing through a cooling coil decreases, the required waterflow would need to increase to maintain the same energy exchange. For example, assuming a constant entering water temperature of 41°F, the system would need to flow 171 gallons per minute to deliver 100 tons of cooling at 14°F delta T. If the coil lost 4°F of delta T, due to dirty coils effecting heat transfer for example, the system would need to pump 240 gpm to deliver the same 100 tons of cooling (40% more gallons per minute). Courtesy: Envise

Figure 2: When the temperature drop of water flowing through a cooling coil decreases, the required waterflow would need to increase to maintain the same energy exchange. For example, assuming a constant entering water temperature of 41°F, the system would need to flow 171 gallons per minute to deliver 100 tons of cooling at 14°F delta T. If the coil lost 4°F of delta T, due to dirty coils effecting heat transfer for example, the system would need to pump 240 gpm to deliver the same 100 tons of cooling (40% more gallons per minute). Courtesy: Envise

Meanwhile, a nearby tech startup was ducting temporary cooling into a small server room to keep the systems up and running after a compressor burnout contaminated the cooling system. If the server overheated, the company would lose its local network and the office would shut down until repairs could be made.

Later that week, building service technicians reported to a local government building that had become stuffy and stale. The situation had been escalating for more than year without resolution from the building operator.

All three of the above service calls could have been avoided if their buildings had been a little smarter, as described by the ASHRAE Handbook Chapter 63 description of smart building systems, by automatically analyzing the overlooked data and presenting potential diagnosis. These sites all had a BAS, but none of the owners had leveraged the capabilities to uncover underlying issues. Most buildings that have a BAS are designed and built in a competitive environment to meet minimum code or design requirements. Often due to budget constraints and pricing pressure, contractors make use of their discretion to leverage the greyness of what is specified and sold.

For example, there are subtle nuances such as a system being capable of performing a function and actually implementing that function. These systems typically have enough sensors and control capability to perform at a higher level than what is in effect delivered, but often fall short of expectations due to competitive bid affairs and sometimes contractor competency issues.

Programming the controls

During the pharmaceutical site service call, it was determined that the site was running out of cooling capacity because of low delta-T at the coil and across the central plant serving the entire building. The building O&M engineer was convinced there were system sizing issues or a mechanical failure, but the diagnosis came down to a perfect storm of sensor calibration issues, programmed reset errors and hunting valves that combined to stress the system capacity to its limit under these nondesigned operating conditions.

Figure 3: Building automation system point-to-point checks of wiring and sensor inputs to controllers are key to ensuring control points are reading and controlling correctly and as assigned in programs before testing sequences of operation and commissioning systems. Courtesy: Envise

Figure 3: Building automation system point-to-point checks of wiring and sensor inputs to controllers are key to ensuring control points are reading and controlling correctly and as assigned in programs before testing sequences of operation and commissioning systems. Courtesy: Envise

Fixing this problem during a GMP-monitored production run was very demanding, stressful and costly because changing parameters on the system to get back to the design conditions required a strict and solid documented plan. If a smart building technology such as AFFD had been implemented, the degraded operating conditions could have triggered notifications that the delta-T at the central plant and at several of the AHUs were deviating from acceptable values and could have predicted, alerted and even prevented this loop tuning and energy drift issue before escalating to an emergency service situation.

The situation was most likely the result of a poorly programmed and commissioned system that somehow made it past the proper design and commissioning gates and got signed off without being flagged. Most likely due to the rushed nature of construction projects, the limited resources available in the market and the lack of experience across the board, poor sequences of operation are often implemented — which, after dialog with the building O&M engineer, appeared to have been the case here.

Some consulting engineers will leave the written sequences of operations to the controls contractor to determine. Oftentimes even if the consultant provides written sequences, the controls programmer will still rely on a previously programmed sequence code from their library of projects that may match the request only slightly at best. Most programmers rely on “copy-paste” from a previous project to expedite the process, which can have mixed results, depending on their source codes related sequence, validation of operation, quality control checks and their internal company policy on revision history and preservation.

Figure 4: This shows the cooling valve at 100% open but only about 2°F air temperature drop across the cooling coil. The space temperature is satisfied so no one would take a closer look at this until the temperature isn’t maintained and people start complaining. This graphic is just a snapshot in time; the problem cannot be diagnosed from this data alone. Courtesy: Envise

Figure 4: This shows the cooling valve at 100% open but only about 2°F air temperature drop across the cooling coil. The space temperature is satisfied so no one would take a closer look at this until the temperature isn’t maintained and people start complaining. This graphic is just a snapshot in time; the problem cannot be diagnosed from this data alone. Courtesy: Envise

If the mechanical design is more complex and requires specific criteria be followed, the programmer may end up creating a patchwork of pasted programs and workarounds from their library, which could cause significant future problems and deviations from design performance when the building is fully occupied.

The keys to ensuring a positive outcome involve providing clear expectations for a better engineered, documented and coordinated design along with careful criteria for the selection of the controls contractor and commissioning professional.

Controls expertise

A controls technician must understand the dynamics of the system to be controlled and the nuances of what can cause failures or unexpected deviations. Some of this expertise can come from a well-rounded background, but most comes from experience. On some occasions, basic systems that get poorly implemented can cost the end user significant money in high operating costs or premature failures.

Inexperienced programmers easily can miss a bevy of best practices that typically are learned over time. Something as simple as not implementing a regular cycling of floating actuators can cause significant problems when they eventually get lost by not having an accurate fully open and fully closed reference point. In the case of VAV boxes, this can be magnified by multiple zones fighting each other on both the airside and water side. A simple after-hours cycle from full-open to full-closed every night can ensure that any floating actuator properly knows its position and controls appropriately the following day, but this simple piece of code can cause large problems if missed.

Other lessons learned can be a little more difficult to spot. As an example, an often-overlooked rogue zone issue can cause equipment to work harder than necessary as it drives resets depending on the sequence of operations. This waste propagates all the way up the energy chain overdriving the AHUs and potentially the entire central plant.

For example, a rogue zone trying to cool a space that it will never be able to satisfy can cause an AHU supply air temperature reset to drive down to its lowest cooling setpoint, potentially forcing other zones on that same AHU into a reheat situation, wasting both cooling and heating energy. That same rogue zone also could cause the entire system to operate after hours if the setback temperatures are exceeded, driving unnecessary operating time and burden on the entire system.

Commissioning

In the second service call scenario, the server room compressor burnout was likely caused by “misses” in the final startup and commissioning of the installation. The unit was designed to be oversized to leave capacity for expected future server rack additions. The diagnosis was that the compressor was short-cycling due to improper time delays and safeties. Each time a compressor starts there is a buildup of heat that requires system operating runtime to dissipate.

When compressors short-cycle, there is not enough runtime to allow for the heat to dissipate that was generated during each restart. The wear and tear of each start and the buildup of heat eventually resulted in a motor winding failure. This unit was a small piece of a much larger project and did not get the commissioning attention required.

The nature of traditional construction projects and the difficulty of closing out the last 10% of projects can cause the commissioning and close-out process to fall short. A lot of the HVAC controls work can’t be completed until other work is installed and operational. As project schedules shift to the right, the controls time gets compressed (because the occupancy date typically remains the same) and there just isn’t enough time to properly startup and commission many projects and a lot of this work can’t be stacked simultaneously. Even if the initial commissioning is perfect and sets a baseline of operation, the building evolves over time and often changes drastically in the first months of occupancy. Commissioning is not only about quality assurance, but also about establishing the optimum baseline of a building’s operation. In that sense, commissioning is becoming more prevalent and also trending more toward a long-term, systematic, data-driven approach to optimize the balance of energy consumption and occupant environmental conditions to identify potential future energy saving opportunities. Leveraging the data and smart building technologies through ongoing commissioning after the building has evolved past the initial startup phase can help to predict potential equipment failures based on actual operating conditions.

Modeling the data

Benchmarking the building operation and comparison to energy models is becoming more integral to the commissioning process and leverages smart building infrastructure and technology. Ongoing commissioning instead of spot checking with retro-commissioning is also favorably contributing to the life cycle cost of buildings by not only reducing energy drift but potentially finding more opportunities for further energy reductions and extending the operating life of equipment, while ensuring the occupants remain comfortable and productive. This deeper and ongoing commissioning also tends to look at how systems operate as a whole, which again broadens the overall optimization and ensures the savings in one area don’t cause a larger increase in another.

The third service call example involving a service call at the government office was determined to be the result of an inadequate delivery of the project. Ultimately the building began to suffer from indoor air quality issues and the building O&M staff overrode systems in attempts to compensate. The system design was complex for the low-bid winning contractor and the implementation was poor, leaving the building O&M engineers with issues they were not trained or knowledgeable enough to understand.

One of the issues involved the heat wheels associated with the two large AHUs that served the building. Perhaps due to poor installation or lack of maintenance, the heat wheels began to fail in a couple different ways, including seal damage, which negatively affected the separation between the opposing airstreams causing airflow short cycling and abrasion to the wheel media, which restricted airflow through the media and reduced AHU performance.

Many of the control system points throughout the building were manually overridden as the building O&M engineers decided they had a better feel for how the building responded to weather and occupancy conditions and didn’t trust the BAS. Compounding the problems was an outdoor air intake that was at street level, situated close to a group of homeless people occupying a nook in the building, which introduced foul-smelling human odors from the area into the building. The building O&M staff locked those fresh air dampers closed, basically eliminating about 50% of the building’s designed outside air intake and flush resulting in the occupants ultimately complaining about poor indoor air quality in several areas of the building.

By the time of the service call, several tenants had complained of stale air and odors causing headaches and other maladies. If the designers had required smart building technologies to be implemented such as AFFD, especially due to the complexity of the mechanical systems involved, many of the root cause problems could have been automatically discovered and presented to the building O&M engineers for resolution before reaching its ultimate fate.

Codes and standards

The competitive bid nature of this government building also could have contributed to the above result. Many of these issues can be addressed by documenting the appropriate baseline of expectations and verification requirements and working with trusted companies in a best value instead of low-bid situation.

Establishing a baseline of expectations currently is being addressed with some newer code requirements and guidelines. Examples are California Title 24, ASHRAE Guideline 36: High-Performance Sequences of Operation for HVAC Systems and ASHRAE Standard 202: Commissioning Process for Buildings and Systems.

The competency issue is another concern often driven by engineers and technicians unfamiliar with best practices. This causes the shortcutting of new technicians getting to the front lines without proper knowledge, experience or oversight. Smart building technologies in the forms of integrated and third-party BAS analytics, AFFD, artificial intelligence and machine learning in various formats for design and commissioning — including the use of digital twins — are also in stages of development and deployment and will help to cover and uncover some of these current pitfalls.

Figure 5: Some mechanical systems contain a lot of data that can fill an entire screen to view at one time. There are many related parameters driving this system, which can be a lot to digest and interpret visually as snapshots in time. In some cases, analytics can help users zero in on potential issues, deficiencies and opportunities more effectively than graphical user interfaces implemented with a standard building automation system. Courtesy: Envise

Figure 5: Some mechanical systems contain a lot of data that can fill an entire screen to view at one time. There are many related parameters driving this system, which can be a lot to digest and interpret visually as snapshots in time. In some cases, analytics can help users zero in on potential issues, deficiencies and opportunities more effectively than graphical user interfaces implemented with a standard building automation system. Courtesy: Envise

One advancement that will help deliver better results involves the work of ASHRAE Guideline 36. The efforts behind ASHRAE Guideline 36 are to provide uniform and vetted best-in-class sequences of operation designed with specific intention for the most common systems.

Although this will not solve all of the associated competitive and competency issues and all buildings require some nuanced programming modifications, the guidelines will provide common terminology and consistencies in implementation that are intended to maximize HVAC systems’ energy efficiency and performance while also providing control stability. Among other reforms, the guidelines aim to balance all of the above with good indoor air quality, which can also help make occupants more comfortable and productive.

Although code may require certain sensors and sequences, such as demand-control ventilation requirements in California Title 24, the sensors typically are used only for those isolated control sequences along with basic trending and simple alarming.

Using data wisely

There is an opportunity to better use the existing sensor data and a case could be made to engineer systems with additional sensors beyond the minimum requirement and include additional smart building technologies such as AFFD; control contractors often will only include the minimum sensors required by the sequence of operations. Most BAS are designed, engineered and programmed for basic local equipment control rather than leveraging sensors and data from across the entire energy chain.

However, installing more sensors and gathering more data doesn’t always directly translate to better results and the ROI for going above and beyond should be taken into consideration. Most systems already have enough data to provide good results. A good base design will allow for economically scalable smart building technologies that have the potential to mitigate a lot of the abovementioned issues.

Unfortunately, these implementations are often engineered out of projects due to a poor value proposition along with a tight budget. Commissioning doesn’t catch everything, buildings are always evolving, the weather changes from season to season and no building O&M engineer has the time to continually evaluate his systems or dig in to find the patterns and connections in the data that a computer can do on a continual 24/7 basis.

Smart building technologies can help find wasted energy and reduce equipment life cycle cost while keeping the occupants comfortable. Because smart building technologies can help predict failures before they become problems and can help building O&M engineers prioritize alarms and issues based on comfort or cost impacts while issues are not resolved, the engineers can be more proactive and efficient and the occupants can remain comfortable.

The pharmaceutical site, server room and government office example failures cost the building owners more than $55,000 in unplanned emergency service cost due to the failures. Leading up to those events, the equipment consumed untold excess energy and the equipment ran harder than necessary also increasing the operating cost of the building and life cycle cost of the equipment. On top of that, their own employees incurred increased stress, downtime and productivity was affected and they accrued various incidental direct and indirect cost.

These impacts could have been avoided with proper system design and implementation efforts at an added cost less than the above impacts just by leveraging what they had and implementing smart building technologies.


Jason Gladney
Author Bio: Jason Gladney is vice president and northern California branch manager of Envise. He has 25 years of experience in HVAC controls, system integration and performance contracting.