Learn how to do a trend analysis in HVAC systems

HVAC trend insights

• The ability to analyze HVAC trend chart data in building management systems for commissioning professional, facility managers and building operators can be a potent tool, but despite its potential, some find the task daunting.
• The purpose of this article series is to help readers conquer the fear of data analysis so that they can act on opportunities for building retuning and troubleshooting. This is part one.

Any avid sports fan knows that analytics are a great tool to evaluate athletic performance, but only in support of the eye test applied when watching games. The best building engineers, facility managers and commissioning professionals use a similar combination of a snapshot in time and historic data analysis to evaluate building and system performance. For example, they may supplement the review of alarm pages, hours of field troubleshooting and graphic observations with the analysis of trend data.

Comfort with data analysis is more urgent than ever: There’s a silver tsunami of legendary heating, ventilation and air conditioning (HVAC) personnel retiring and taking along with them a wealth of systems knowledge that some as-built documents just can’t communicate as effectively. This exodus, juxtaposed with the rapid growth of technology in the mechanical, electrical, plumbing and building automation industry, creates a need for understanding the deeper meaning sensor and equipment data can provide.

Figure 1: Multiple zone variable air volume air handling unit with relief fan and differential pressure outside air measurement. Courtesy: Lionheart EMC

Line graphs and scatterplots can be intimidating, but the capability to understand them will steady the same hands that will have to manage the components of an upgraded building automation system (BAS) or software analytics package to drive initiatives in energy efficiency and building decarbonization.

Let’s walk through the process together, using the real-life example of an older 1990s building. This building is at the center of pressures from local law mandates such as New York’s Local Law 97 and Boston’s BERDO. The building management system is older and the sequences have not yet been conformed to more modern ones such as ASHRAE Guideline 36: High Performance Sequences of Operation for HVAC Systems.

This situation requires us to dig deeper into the data provided by the building management system to find opportunities for energy saving. A common place to start is with a review of a multiple zone variable air volume (VAV) air handling units (AHU) with an integrated economizer, equipped with a relief fan, mixed air damper, outside air damper, exhaust air damper, heating and a chilled water control valve. Through this analysis, we can better see what the system alarm page doesn’t show — e.g., key parameters such as occupancy, cooling coil control and economizer operation — and brings us a step closer to determining how effective the AHU been at saving energy (see Figure 1).

A mental shift in HVAC systems

Using a BAS to do more than change schedules and setpoints is a legitimate mental shift, but also a valuable one for keeping up with the speed of technological changes. Before making any attempt to set up trends on a BAS, ensure that the proper training, system access and privileges have been assigned. Omitting this step can result in serious impacts to the current system configuration because not all automation systems have the same process in place for setting up trends.

An important question to ask is whether trending of points would require a download of the new parameters at the supervisory level. If the answer is yes, this may mean that during downloading the new trend configuration, visibility to the current automation network can be lost.

Figure 2: Air handling unit occupancy trend. Courtesy: Lionheart EMC

Check the sequence of operations

In verifying the reliability of the AHU data, the sequence of operations should be checked to confirm enabling and disabling conditions for temperature control and operation. This will help guide the process of selecting the right data points for trending. Some additional information on the sequence of operations can shed some light on the intended occupancy schedule the equipment supports.

To add more context to the AHU’s sequence of operations in this case, we outline the parameters of the economizer and chilled water control. Differential dry bulb economizer control was implemented here, however, a word of caution to those using older sequences: Ensure your climate zone aligns with the economizer strategy.

In economizer control, there are various strategies deployed by designers depending on the local climate. ASHRAE Standard 90.1: Energy Standard for Buildings Except Low-Rise Residential Buildings sheds light on the different control types: fixed dry bulb, differential dry bulb, fixed enthalpy, electronic enthalpy, differential enthalpy, dewpoint and dry bulb.

Figure 3: Chilled water control valve signal. Courtesy: Lionheart EMC

For this case, the dry bulb economizer control sequence is enabled at an outside air temperature of 55°F, disabled at 70°F and only at points in time in which the outside air temperature is below the return air temperature. The economizer damper is dedicated to maintaining a supply air temperature setpoint of 60°F and because this is an integrated economizer, there are periods of time in which it works in tandem with the chilled water control valve to support the supply air temperature control toward the setpoint.

HVAC trend types: COV and interval

Two important terms to define are change of value (COV) and trend intervals. Trending data at a change of value means that data from a specific point is pulled only when that value is changed at or beyond a preset threshold. Trending via intervals involves data snapshots at specific timeframes.

For software values such as setpoints or on/off points (start/stop and status or occupancy), logging in these values each time they change can provide meaningful insight into issues related to erratic control (for example, hot and cold complaints due to multiple setpoint changes). Pulling a trend on a setpoint over a period of a day or a week can allow for visibility into the specific dates and times in which these setpoints were changed.

Figure 4: Supply air temperature versus supply air temperature setpoint. Courtesy: Lionheart EMC

More specific applications of trends using a COV feature would be for parameters that fluctuate frequently but need to be logged, such as fan airflow in cubic feet per minute (cfm). For a fan capable of supplying 20,000 cfm, we may only want to log values every 1,000 cfm or greater.

For trending using a time interval, a common starting point in the industry is to trend data every 15 minutes. We may see this with dynamic values such as temperature, humidity or modulating signals such as fan speed, compressor output, etc.

Trend analysis of building occupancy

Buildings, occupant patterns and room uses change over time. In the case of the building containing the AHUs reviewed here, a simple trend was set up to show the unit occupancy status over the course of one week, from Monday to Sunday (see Figure 2).

The trend shown in Figure 2 has two values representing the different states of the AHU: 1 for on and 0 for off. For some building types such as hospitals, an AHU running 24/7 is not uncommon at all and doesn’t initially raise any flags. What if it were discovered however that all of the AHUs within this building behaved this way, even when a number of them were in service to administrative offices with fixed occupancy periods that ended around 6 p.m.? A significant opportunity would exist to reduce energy consumption by adjusting the nontypical AHU.

If the AHU is serving VAV boxes or dedicated to a particular area in the building, it would be important to verify the unoccupied setpoints before making any modifications. Shutting down a unit without consideration for setback temperatures or for spaces with a higher propensity for freezing conditions can have serious consequences. In such instances, it is prudent to work with the facility operations team and BAS contractor to ensure that critical processes aren’t interrupted by making schedule adjustments.

Control loop tuning in HVAC systems

Even though many interval-based HVAC trends will be initially set at 15 minutes, certain issues aren’t easily detectable at these points in time and smaller intervals may be required for insight into tuning issues that won’t show up on an alarm view. In the case of the AHUs, we consider a situation where the economizer has hit the high limit and the chilled water control valve is now responsible for maintaining the supply air temperature setpoint. A trend review of the chilled water control valve modulating signal (Figure 3) is shown alongside a trend of the supply air temperature and the supply air temperature setpoint (Figure 4) over the course of a working day at 15-minute intervals.

At first glance, the trends may not pique any interest. The chilled water control signal appears to be normal, aside from some early afternoon spikes that may represent an influx of people returning from lunch or some other variation in load. This must be considered carefully because the 15-minute interval is only a snapshot and the capture of this spike in control valve signal and temperature may indicate a deeper issue, because at no point during the day (disregarding the supply fan motor heat), is the supply air temperature at or below setpoint.

We then adjust the trend intervals from 15 minutes over the course of a day, to one minute over the course of 15 minutes (see Figure 5).

Figure 5: Chilled water control valve signal at one-minute intervals. Courtesy: Lionheart EMC

The spikes every minute show the control valve constantly changing position, which is a symptom of control valve hunting. A system’s inability to provide stable supply air temperatures can be a symptom of an improperly tuned control loop and, if not investigated, could lead to premature failure of control actuators.

Using trends to verify sensor accuracy

On a day forecast to have weather conditions favorable for economizer control, an opportunity may exist to verify the accuracy of the outside air temperature sensor. For such a day, a quick check of the weather says that the highest temps we will experience will be 62°F, which means that for a significant portion of the day we should be economizing. By selecting just the outside air temperature, we create a single point trend bound between 7 a.m. and 7 p.m. (see Figure 6).

Figure 6: Building management system outside air sensor temperature trend. Courtesy: Lionheart EMC

Let’s read between the lines here: The valley or lowest point of data is at about 50°F, which is a few degrees behind the economizer enable point. The peak, however, appears to be at above 80°F, which is very different from the weather forecast for the day. Take another step by pulling a trend on the economizer damper position on the same working day period — from 7 a.m. to 7 p.m. (see Figure 7).

Figure 7: Economizer command. Courtesy: Lionheart EMC

A constant signal of 20% open is likely for the code-required minimum ventilation per ASHRAE Standard 62.1: Ventilation for Acceptable Indoor Air Quality. As expected, the economizer damper is opening within the limits prescribed in the sequence of operations with respect to the installed outdoor air temperature sensor between 7:45 and 9:30 a.m. for free cooling. This provides data to show that the control loop for the damper is tracking the outside air sensor properly. This doesn’t, however, negate the fact that the high of the day was 62°F and the outdoor air sensor is reading above 80°F.

Many resources for gathering data for outdoor weather conditions are publicly available, such as from Weather.gov. For older buildings, it isn’t uncommon to find a single outside air temperature sensor being broadcast across a variety of air handlers, boilers and chillers for control and reset sequences. The engineer can gather some additional data and take a step forward by creating a two-point trend series of the public weather data for this region and compare it to the outside air temperature provided by the building management system (BMS) sensor (see Figure 8).

Figure 8: Web outside air temperature data versus building management system outside air temperature data. Courtesy: Lionheart EMC

It is completely reasonable to have just a few points in time during a day when the outside air temperature sensor is not at the same value as public web data, especially when considering where the weather station is.

However, in this case the locations are extremely close. This two-point trend shows that the BMS sensor is more than 10°F above the actual outside air temperatures at times. Let’s assess how this has affected the economizer damper with a three-point trend, placing the web data for outside air temp, the BMS sensor data and economizer command (see Figure 9).

Figure 9: Web outside air temperature data versus building management system outside air temperature data and economizer command. Courtesy: Lionheart EMC

From the web data provided on outside air temperature, the dry bulb economizer control loop should have been enabled closer to 10:30 a.m. until the end of the trend period at 7 p.m. The damper remained at minimum position, but the use of free cooling was low, which in turn causes the chilled water control valve to open. The upstream impact for this is felt at the central plant as chilled water pumps increase in speed to maintain system differential pressure.

At the minimum, this information allows for a rather informed service call to the BAS contractor. The temperature trends from the BMS appear to track closely to that of the web data, so the question is posed as to what can cause inconsistent readings from a dry bulb temperature sensor.

Mounting of the outside air temperature sensor is a critical detail that can have incredible impacts to HVAC equipment performance. In the Northern Hemisphere, the ideal location for an outdoor air temperature sensor would be on the north side of a building in a shaded region. Mounting it on the east or west side of a building could lead to high levels of sun exposure, which can cause up to a 30% increase in sensor readings. Locations close to exhaust or other sources of ambient heat can also lead to false readings.

It is important to understand that the peaks and valleys in data analyzed for various systems at the surface reflect inconsistencies with their design intent. The deeper impact that these deviations may speak to, however, can go beyond comfort control and reveal the story behind hours or late nights of troubleshooting, as well as why such large capital investments in HVAC equipment may not live up to expectations.

With so much information made available, the ability to read trend charts and see their deeper meaning can help engineers and building professional part the seas of data and walk the path that helps us best assist the people we have been given the opportunity to serve.

Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our WTWH Media editorial team and getting the recognition you and your company deserve. Click here to start this process.