5. Component-Level Integration

For most projects, the first step down the road towards integration will occur at the component level. Systems that lend themselves to component and subassembly level testing are shown in Table 1. Issues associated with this process have been touched upon in the planning example associated with the Student Center AHU1 preheat coil (refer to Section 3.1).

Table 1: Components and subassemblies that can be tested at the component-level

 

Air handling Systems

Pumping Systems

Cooling Systems

Heating Systems

Components
and
Subassemblies

Economizers

Pumps

Compressors

Boilers

Coils

Active Filtration

Chillers

Heat Exchangers

Humidifiers

Valves

Cooling Towers

Furnaces

Drive Systems

Drive Systems

Condensers

Drive Systems

Fans

Coils

Drive Systems

Sensors

Dampers

Tube Cleaners

Sensors

 

Valves

Sensors

 

 

Terminal Units

 

 

 

Sensors

 

 

 

 

Integration and functional testing at the component level have the following general characteristics and goals:

1   All necessary parts need to be in place and properly installed.

For example, an effort to test a preheat coil and its associated control loop would be futile if the necessary control valve was not installed or if the connection from the DDC controller to the valve had not been verified.

2   The testing will target adjusting and tuning small assemblies of highly related components.

Continuing with the Student Center AHU1 preheat coil example, the component-level testing targets the integration of the components and processes directly related to the preheat function. For AHU1, this would involve:

·       Verifying the operation and control of the preheat coil’s circulation pump.

·       Verifying the initial tuning of the preheat coil’s discharge-temperature control loop.

3   Minimizing the potential for interaction with other components can be desirable.

Ultimately, the goal of integrated testing and operation is the exact opposite of this. However, at the beginning of the integration process, taking steps to minimize the potential for ripple effects of a downstream or upstream process from impacting the process under test allows the component-level effort to identify and focus on component related issues.

There are many issues that can be encountered during component-level integration and testing. The best way to instruct providers on them is through several detailed examples. The Student Center project offers the following examples.

5.1. Preheat Coil Test

The first is a good example of the need to minimize the interaction of components under test with other parts of the system or other systems, which was discussed above. When the commissioning provider’s team initiated a preheat coil test at the Student Center project, AHU1 was being used to provide temporary heating and cooling to facilitate the installation of finishes on the project. Under the circumstances, the provider decided that he would take advantage of the fact the unit would be operating during the cooler spring weather, and planned to perform the component-level testing of the preheat coil while on site in March to observe construction activities (see March data in Figure 8). When the provider’s team arrived at the site on the day of the test, they spent time touching base with the contractors to verify that everything was ready and found that:

·       The preheat coil installation was substantially complete, including the pre-start checklist.

·       The design control sequence for the preheat coil’s discharge temperature control loop was in place. In addition to serving the preheat valve, the output had been temporarily piped to the economizer dampers for temporary operation.

·       The preheat coil’s circulation pump was running and under the control of the DDC system.

·       Inlet connections were not made to the terminal units, so any air reaching them was simply dumped to the area in the vicinity of the open duct taps.

·       AHU1 was adjusted to run at a fixed speed of 30% to circulate air through the project area, but minimize the potential for duct or fan casing damage from excessive positive or negative pressures should a fire damper slam shut or the fan start without the outdoor air dampers being open.

·       To facilitate the drying of joint compound, the control contactor indicated they had forced the unit to 100% outdoor air by removing the feedback spring from the pilot positioning relay on the outdoor air-damper actuator, causing it to run full stroke regardless of the control input.[13]

Temperatures were in the mid 40s, which are ideal conditions to test the functionality of the preheat coil control loop with out endangering the coil. The details of the preheat coil control loop are outlined in Table A-1 and illustrated in Figure A-1. In general terms, the pump would run continuously while the control valve modulated to control the coil’s discharge temperature. The commissioning team’s test procedure was straightforward and included the following basic steps.

1   Verify system readiness.

2   Load the preheat coil by allowing it to see 100% outdoor air.

3   Trend the performance of the coil.

4   Tune the control loop using the closed-loop method[14] to provide stable performance.

5   Change the load on the preheat coil by allowing the system to operate on 50% outdoor air.

6   Verify stable performance of the control loop, re-tuning if necessary.

Anticipating the unavailability of the control system operator’s workstation at this point in the project, the provider’s team had brought along data loggers, which captured operating data unattended while the provider team observed construction activities. Upon retrieving the data, they discovered that the preheat coil pump appeared to be performing well, but the control loop was very unstable. Concluding that tuning was in order, they proceeded to work with the control contractor to adjust gains in an effort to achieve stability. After several attempts to stabilize the control loop, they decided to take a closer look at what the control valve was actually doing in response to the control system. When the lead provider climbed a ladder to observe the preheat valve’s stem motion, it was noted the feedback spring on the outdoor air damper’s pilot positioner was actually hooked up and not disconnected as the contractor had indicated. The outdoor air dampers were not in a fixed position at all, but were following the same signal that was being sent to the preheat valve. At the low flow rate associated with the low fan speed, the dampers had very non-linear characteristics. The instability that the team was fighting was being driven by the economizer non-linearity, not the actions of the preheat valve. The team soon discovered that when they isolated the economizer dampers from the preheat-coil control loop and locked the system in the 100% outdoor air position, the process came under control, but with significant proportional offset. Some additional tuning effort to tighten the proportional band and add some integral gain eliminated the offset and provided a stable responsive preheat-coil control loop. The tuned loop would now serve as a good starting point when it was time to move from component-level integration to system level integration.

5.2. Economizer startup

The experience with the preheat coil coupled with the seasonal change towards the need for temporary cooling rather than temporary heating caused the control contractor to focus their efforts on installing the remaining hardware and software necessary to control the economizer dampers and chilled water valve.

The details of the economizer and chilled water valve control via discharge temperature are outlined in Table A-1 and Figure A-1. In general terms, the economizer dampers were to be controlled in sequence with the cooling-coil valve as required to maintain discharge temperature. The discharge temperature was to be reset based on the position of the terminal-unit reheat valves, but for the purposes of temporary operation, the control contractor programmed a fixed set point of 55°F into the system. The economizer was to be disabled based on enthalpy comparison, but because the commissioning provider had requested to defer the installation of the humidity sensors until later in the project, when there was not so much construction dust, the control technician programmed the change-over to occur any time the outdoor air temperature was above 75°F. A mixed-air low-limit control loop was arranged to override the signal to the economizer dampers if the mixed air temperature dropped below 45°F. Interlocks with the fan operation and fire and smoke control system were not in place. The fan was simply shut down at the end of the day by the electrician when he left the site and restarted in the morning when he arrived.

As part of the team effort, the control technician made sure things were properly working and that there were no temporary “fixes” while he performed point-to-point checks and economizer function start-up. Satisfied with his effort, he called the lead commissioning provider to say that the subassembly was ready for testing.

The cool mornings and warm afternoons that are typical of the area’s weather pattern in the spring (see Figure 6 and Figure 8) provide an ideal opportunity to verify the sequencing of the economizer dampers with the chilled water valve, and verify the functionality of the mixed-air low-limit cycle. Granted, the sequencing would probably need to be spot-checked later because the damper interlocks were not in place, but the hardwired interlocks would not likely impact the software controlling the process or the loop-tuning parameters. Additionally, the integrated test of the air handling system and subsequent trend analysis would provide ample opportunity to pick up any problems that were introduced.

The provider had also anticipated some retesting. By working with the team to test the subassembly on the next site visit, he would have a good chance to verify sequencing, limit functions, and change-over under real-world conditions instead of through manipulation of set points due to the daily temperature swings that were typical for that time of year (see April in Figure 8).

To definitively capture the sequencing that should occur as the outdoor air temperature rose and fell above and below the discharge-temperature set point, the provider preprogrammed data loggers and sent them to the control technician so they could be installed and running for several days before his arrival. The commissioning specifications had included assistance of this type in the control contractor’s scope of work.

The focus of the control tech paid off for everyone. The control loop was stable under all operating modes observed, including shutdown and restart. The temperatures had varied enough over the days preceding the visit to allow the sequencing issues to be verified by the trend data. A relatively quick manual test verified the change-over and mixed-air low-limit process.

The only item that was of any concern to the provider was the lack of a freezestat to protect the system during testing. Earlier, this was not an issue because the chilled water coil had not been filled. Also, spring was normally considered to be well under-way in the region by April. However, a quick trip to the National Weather Service web site confirmed his suspicions regarding the potential for freezing weather (see Figure 11) and he elected to be conservative. He shared his concern and the weather data with the owner, project manager, mechanical contractor, and control contractor, they all agreed it would be prudent to install and test the freezestat as soon as possible and pay close attention to the weather forecast in the meantime.

Figure 11: St. Louis weather statistics and record temperatures for April

Click figure to display it as a PDF.

5.3. Variable Frequency Drive Startup

The decision to install the freezestat caused the contractor to re-think their installation plan. As a result, they decided that construction was to the point where they could install the permanent variable frequency drive (VFD) interlock wiring, including connections and interlocks with the smoke isolation dampers, instead of making a temporary connection to just the freezestat. This allowed the commissioning provider to verify the VFD interlocks during the May construction visit. The factory start-up of the drive had been coordinated by the contractor to occur when it was first energized for temporary service back in January to ensure that the warranty was not violated. At that point, he made sure the manufacturers representative reviewed the connection points for interlock and control wiring with the electrical and control contractor. Thus the providers test only needed to verify that the factory start-up had occurred and then verify the interlocks. Things proceeded fairly smoothly and in a manner similar to the other component level tests described in the preceding paragraphs. A few key issues are worthy of mention.

1   VFD interlocks were verified in all possible combinations of selector switch positions.

The VFD’s were equipped with bypass contactors and hand-off-auto switches. As was require by good practice and the sequence of operation (see Table A-1), the safeties needed to work regardless of the position of any of these selector switches. The commissioning provider included positive verification of this contingency in the test procedure.

2   Installation and functional testing of the pressure relief doors was coordinated to occur before allowing VFD testing to occur.

One of the dangers inherent with the installation of smoke isolation dampers directly on the discharge of the fan was air hammer. To prevent damage due to this effect during VFD testing, pressure relief doors had been installed at critical points in the system. These doors had been functionally tested before testing the VFD controls and placing the system into temporary operation by a special fitting that allowed the contractors duct leakage test machine to subject the doors to the trip pressure. But, because that had occurred early in the construction process, the commissioning provider took a few minutes to re-inspect the doors and make sure that nothing had been installed in a manner that would obstruct there operation during VFD testing.

3   Installation of motor/VFD protection equipment installed prior to VFD start-up and testing.

Occasionally, stray currents generated by a VFD can actually travel through the motor shaft and the voltage discharge from the current passing through the bearings can cause them to be damaged or fail completely over time. This phenomenon can occur in both new and existing motors if the motor is not rated for VFD operation. Various methods of eliminating shaft currents are available, such as insulated bearings or shaft grounding devices. To illustrate the concept, the facility operators for the Student Center had finally determined that the rash of bearing failures they had experienced in the mid 90’s correlated very well with the point in time when the University had started wide spread use of VFDs. While skeptical at first, their maintenance records convinced them: 99% of the failures were on existing motors where a drive had been retrofitted or on new motors where nobody, including the facilities group, had paid much attention to matching the drive and the motor. Actual measurement of the voltage and current between the shaft and ground by a recent project’s electrical commissioning provider and recent experience with shaft grounding systems had convinced him that there was an issue and a solution. All motors and VFD’s must be adequately protected prior to start-up, testing, and sustained operation.

5.4. Trim Humidifier Startup

The piping for the trim humidifier had also been completed by the time of the May construction visit and the fan in the terminal unit associated with it was capable of operating. Ambient conditions and the construction schedule would prevent a functional test of the units performance until the following fall and winter. However, verification of the interlocks could be accomplished if steam and airflow were available. These interlocks were particularly critical if problems with water being expelled from the duct were to be avoided. Additional information on this topic can be found in Chapter 7 – Humidification.

The three interlocks to be tested were: 1. the warm up interlock, which prevented steam injection until the manifold jacket was warmed up; 2. the airflow switch, which prevented the humidifier from operating without air flow; and 3. the high limit switch which shut down the unit if the discharge humidity got out of hand. A description of the airflow-interlock test procedure is outlined below.

Because the 95% filters located downstream of the terminal units serving the museum had not been installed at the time of the test, the provider had to temporarily obstruct a portion of the filter housing to simulate the pressure drop that would be created by a filter. This pressure drop was the signal used to prove flow for the airflow interlock.

Despite a successful test of the interlock circuitry, the provider discovered an important oversight. The control sequence required a two-position isolation valve be installed to shut down steam flow to the humidifier when it was not required. This was an independent device from the modulating valve contained in the unit itself and had the advantage of eliminating any heat gain into the air stream from the jacket manifold during the summer. The unit would operate satisfactorily with out it, but energy would be wasted and the added load represented by the jacket heat may not have been included in the cooling capacity provided to serve the space. The problem existed because even though the control sequence required the interlock, the drawings didn’t show the second control valve. The mechanical contractor, working from the piping plans had not realized it was missing and the valve supplied by the control contractor was sitting back in the shop, awaiting return as excess material. The good news was that the test effort had revealed the problem in a timely manner. There was plenty of time to modify the piping and control connections before the ceiling was scheduled to go in, which would have made access more difficult (but not impossible). From the provider’s perspective, the test was a success though a bit of retesting would be required to verify that the repair had not invalidated the existing interlocks.