Unlocking the Value of BAS Data: Why Smarter Trending Matters

Most BASs allow for logged data to be downloaded and/or displayed in table or graphical format. Graphical displays typically show information against a timestamp. Single-point trends show one variable plotted against timestamps. Multi-point trends allow related variables to be plotted together. For example, supply fan speed command, supply fan status, discharge air pressure, and discharge air pressure setpoint can all be plotted together. This allows for a quick review of the control loop that is maintaining discharge air pressure at setpoint. Figure 1 below shows how various issues can be deciphered by using a multi-point trend.

How Can Trending Be Used?

Verifying System Performance: Trending can be used to verify the performance of a system. Key performance indicators can be identified and reviewed on a regular basis. For example, an AHU’s discharge air temperature, discharge air pressure, and outdoor airflow rate trends can be reviewed and compared against their respective setpoints to provide an overall snapshot of system performance. This review can be performed by building operators working with existing systems or by commissioning providers confirming the systems are truly ready prior to arriving onsite for functional performance testing.

Troubleshooting Issues After the Fact: Oftentimes, issues in our HVAC systems occur without someone watching the situation unfold in real-time. It might not be obvious what the root cause of the issue was without the presence of trends. Having a solid understanding of the sequence of operation, along with a substantial amount of properly configured trends, is often all one needs to properly diagnose the issues or at least zero in on possible culprits.

Capacity Verification for Future Rightsizing: Far too often, equipment is grossly oversized. This may lead to premature failure of said equipment, and often that failed equipment is then just replaced in kind. That is, the same costly and oversized equipment goes back into the system, only for history to repeat itself. Strategic trending can allow the building to “speak” to us—to let us know the maximum capacity it is demanding from various components of the HVAC system. I say strategic trending because it can often be a little more involved. For example, if a chilled water plant has a flow meter, as well as plant entering and leaving temperature sensors, the plant cooling output can be computed by the following equation:

• Cooling Output [tons] = 0.042 × flow [GPM] × (Entering Temp [°F] – Leaving Temp [°F])

The BAS can compute cooling output in real time and trend that software variable. Additionally, it would be best to see the maximum value registered in that trend over many years of cooling operation. BAS trends are often configured to retain only a set amount of data (e.g., a year’s worth) and erase the oldest data to make room for new data. This approach could result in critical information about maximum building cooling demand not being available when it comes time to replace the chiller. As such, a separate trend, which only logs the maximum value of cooling demand in a given month, could be configured. Then the BAS could be set up to retain that much more limited data set for 10 years or more.

Streamlining Functional Performance Testing: Actively putting a system through its paces is what functional performance testing is all about. Budget constraints, project schedule constraints, uncooperative ambient conditions, and a host of other obstacles may prevent us from thoroughly testing the system through every scenario. Trending can be used in conjunction with active manipulations to the system to verify that the system continues to perform in scenarios that are not possible to test during initial onsite testing.

Additionally, there are some portions of the automated sequence of operation that are inherently very cumbersome to actively test. For example, imagine the specified sequence of operation calls for the rotation of the lead cooling tower to occur at 9 a.m. on a Tuesday, only when the current lead tower exceeds the current lag cooling tower’s operating hours by a set threshold of 720 hours. The difficulty of actively testing this lead cooling tower designation rotation really depends on what, if any, of the following levers were provided by the controls contractor:

• Is the rotation schedule (i.e., 9 a.m. on a Tuesday) adjustable?
• Is the rotation threshold of 720 hours adjustable?
• Can the cooling towers’ accumulated operating hours be temporarily overridden?

If insufficient levers were provided, it may be most efficient to verify this portion of the sequence through trending alone.

Where Are the Pitfalls?

In my experience, trending is rarely completed correctly, completely, or at all at the time of functional performance testing during new construction commissioning. We find the same situation when working in existing buildings during re- or retro-commissioning endeavors.

Here are some of the reasons why I think this situation is so common:

• No specification for trending. Designers may assume configuring the trends of key points of information would come standard. Unfortunately, it is time-consuming and deemed a low priority for a temperature controls programmer who is under pressure to finish the project on schedule. And if the project isn’t going well for them, they may resist doing anything that isn’t explicitly specified. Additionally, many controls programmers have expressed to me that they believe it is best for the owner to set up the trends they want after the project is turned over to them, since they “know what they want.” For the record, I don’t agree with such an approach and feel it is a cop-out.

• Trending data is intentionally limited so as “not to bog down the servers” or “reduce network traffic.” I cannot speak to the validity of these claims, but it seems every time a temperature controls contractor is pressed on the issue, they figure out some solution to provide adequate trending without these concerns becoming a reality.

• Some BASs can have multi-point trends easily customized by selecting any of the individually trended points. Other BASs require the grouping of points for a multi-point trend to be predetermined. The latter situation requires some forethought into what groupings of variables would be appropriate, and I have yet to ever see such groupings specified in design documents.

• There is widespread misunderstanding of when interval trends and COV trends are appropriate. Interval trends are appropriate for analog variables, and COV trends are appropriate for binary values. COV trends are also important for analog values on critical systems, such as temperature and humidity in an operating room, where these key variables could change drastically in a short timeframe due to the high amount of air changes occurring in the space. I have seen many projects where binary points are incorrectly configured for interval trending. It makes troubleshooting with trends much more difficult. If you are trying to figure out why an AHU was off when scheduled to be on, it is easiest to narrow in on the root cause if you can tell the moment the fan was disabled through the binary output (BO) and the moment the fan lost status through its binary input (BI). If the BI changed state before the BO, then it seems like a fan failure scenario or a hardwired safety may have triggered, and it is best to send someone to the AHU to investigate. If the BO changed before the BI, then the automated program is shutting it down, and efforts are best put toward sifting through the programming logic to determine what could have caused the BAS to send a fan disable command.

Tips for Cx Providers

Here are a few tips to ensure trends are properly configured, as well as how to better leverage trends in the commissioning process:

• Push for trending requirements to be clearly documented in the design documents.
• Require two weeks of trend data to be made available prior to functional testing. This will help verify the system is truly ready for testing and will force the contractor not to overlook this requirement.
• Request the level of BAS access that allows you to create trends yourself during the functional testing phase. This will allow you to more easily configure some trends that may not have been specified but will greatly assist in your testing.
• Consider how trends can be used to augment or replace active manipulations to the system to verify various aspects of the sequence of operation.
• Keep in mind that temporary trends at tighter intervals may need to be configured for key variables when troubleshooting. This is to avoid the concept of aliasing,[1] where the data set you collect does not match what is really going on, since the low frequency of data entries masks issues occurring between the data entries.

Conclusion

Ensuring appropriate trending of BAS points is part of setting clients up for success. Trends can also be used by commissioning providers for a variety of scenarios. If you know the sequence of operation for a given system, and appropriate trends have been configured, then the value you can bring to the project is only limited by your creativity in how you analyze the data.

References:

1. Sellers, D. (2003). Installation of Data Loggers: “Aliasing” and other pre-installation considerations. HPAC Engineering