Introduction

This case study involves a subsystem with several zones that are heated indirectly through a flat-plate heat exchanger. These zones are part of a greenhouse complex. The greenhouse subsystem had been in operation in this configuration for about a year, but there were numerous control and performance problems. This case study will describe the original configuration, performance problems, the solution, and the results. There are two appendices which cover some technical aspects in more detail.

The greenhouse is not heated year-round, so the zones are filled with Glycol. A heat exchanger separates the greenhouse zones from the rest of the heating system, which is filled with water.

The greenhouse is part of a much larger system that includes multiple heat sources and heat loads that are not part of this case study. It is important to note that the greenhouse and other heat loads are heated from a heat storage subsystem which consists of about 1000 gallons of hot water.

Greenhouse Subsystem Details

The greenhouse subsystem has the four circulators, two zone valves, a heat exchanger, and four zones with differing properties:

1. Office Radiant: Small heat load (<5000 BTU/hr), low temperature (<120℉). Controlled by standard thermostat.
2. Mechanical Room Radiant: Small heat load (<5000 BTU/hr), low temperature (<120℉). Controlled by standard thermostat.
3. Greenhouse Slab: Large heat load (>110,000 BTU/hr). Desired Glycol temperature dependent on outdoor temperature, wind speed, and sunlight. Controlled by a ‘dry contact’ in greenhouse management system.
4. Benches: Moderate heat load (10,000-70,000 BTU/hr). Ideally operates at higher temperatures (120℉ - 150℉). Controlled by a ‘dry contact’ in greenhouse management system.

Originally this subsystem was controlled with a collection of commercial zone controllers. Demand from any zone would turn on the appropriate circulator. If any zone was calling for heat, circulator P15 would also run to provide hot water to the heat exchanger. P18 is a Grundfos Alpha 15-55, configured to provide constant pressure. The other circulators are fixed-speed units. The mixing valve (V1) was controlled based on outdoor temperature in order to provide warmer water to the radiant zones when outdoor temperatures were cooler.

System Performance

The system should meet a basic list of performance goals:

1. Correct Control. In response to zone demand, the necessary pumps and valves should turn on, and be off otherwise.
2. Heating Performance. The system should deliver Glycol at the appropriate temperature to each of the zones.
3. Maintainability. The system should provide a mechanism to determine that it is operating properly, with tools to determine both the desired and actual state of pumps, valves, and system temperatures.
4. Heating System Compatibility. This subsystem is part of a larger system. It should operate so as to maximize performance of the larger system.

Taking these goals in sequence, the original implementation did not meet expectations.

1. Correct Control - The commercial zone controllers are a very effective solution for straightforward installations, but the installers struggled for months to get all of the pumps and valves working together. For instance, pump P15 would not always come on, and P18 would sometimes run when there was no demand. Despite repeated efforts, the controls never worked properly.
2. Heating Performance - The ‘outdoor reset’ function governing the mixing valve was buried in a set of very complex configurations on a propane backup boiler. There was no way to see what temperature it was trying to achieve, and in this application outdoor temperature alone is not a good predictor of desired water temperature as sunlight has much more influence. As a result the mixing valve very often did not provide reasonable temperatures. The mixing valve was often manually over-ridden in order to get enough heat.
3. Maintainability - When there were problems, it was very difficult to see and understand what was happening. There’s no practical way to draw an overall control schematic for a system of this type since the manufacturer does not provide internal schematics for their controls. Additionally, the number of iterations of control wiring attempted in the course of this project made documentation impractical. As a result, there were no documents describing the wiring or control of this subsystem.
4. Heating System Compatibility - Effective use of heat storage depends on maintaining maximum thermal stratification in the storage tanks. This in turn requires the lowest possible circulation speeds and the highest possible ‘Delta T’ - difference between supply and return temperatures. Since P15 is a fixed-speed circulator, it had to be set to a relatively high speed in order to transfer enough heat at times of peak demand. That means that most of the time it was running too fast and returning water at very high temperatures. This would very quickly destratify the heat storage tanks.

In this chart, P15 came on at about 9:00, injecting very hot water into the bottom of the storage tank. This destroys stratification by dropping the temperature at the top of storage and raising the temperature at the bottom of storage. This creates two problems:

1. The drop in temperature at the top means that the heat source (an automatic pellet boiler in this case) must come on-line sooner than it would have.
2. The increase in temperature at the bottom means that the heat source is heating warm rather than cool water. That results in a shorter and less efficient cycle.

Shorter and more frequent cycles are a much less efficient way to utilize the heat source. Feeding the boiler with relatively hot water also forces it to reduce its output which reduces efficiency still further.

Solution

The solution to this problem was to give the Vesta controller information about the demand from the greenhouse zones and let it control P15 to achieve the desired Glycol temperature.

The commercial zone controllers were removed along with the motor drive for the mixing valve. The mixing valve was set to straight-through (no mixing).

The following Vesta hardware was used:

• Vesta controller (already on-site and in use for other functions)
• RI-024A (formerly RM-1210) relay input box to sense demand from each zone
• VS-1108 controller to vary the speed of P15
• Two temperature sensors to measure Glycol and return water temperatures

The RI-024A relay box has four relays with 24VAC coils. Each relay has a contact that's sensed by the Vesta and an extra contact that can be used to control an external device. In this application, the spare contacts are used to control circulators P16, P17, and P18. Here’s a schematic showing the connections in the RI-024A relay box:

In each case, the thermostat (or dry contact) completes a low-voltage circuit that activates a relay. The Vesta can sense the relay closure. The relay contacts also provide power to the appropriate circulator. In the case of the office and mechanical room, the thermostat also provides power to a motorized zone valve. This wiring replaces the functions provided by the commercial zone controllers, while adding the ability for the Vesta to monitor and respond to zone demands.

The wiring in the relay box takes care of all devices except P15, which will be managed by the Vesta controller.

Vesta Setup

The RM-1210 relay box, the VS-1108 variable speed control, and the two sensors were connected to the Vesta. All of the signals relating to these devices were given user-friendly names:

• GH Office Tstat - Vesta discrete input, TRUE if thermostat is calling for heat
• Mech Room Tstat - Vesta discrete input, TRUE if thermostat is calling for heat
• Slab Heat Demand - Vesta discrete input, TRUE if greenhouse controller is calling for heat
• Bench Heat Demand - Vesta discrete input, TRUE if greenhouse controller is calling for heat
• GH Glycol Hot Supply - Temperature of Glycol being supplied to the greenhouse
• GH Return To Tank - Temperature of the water being returned to storage
• P15 GH HX Pump - Vesta variable speed output driving P15

In addition to these inputs and outputs, three internal variables were created to support management of the greenhouse:

• GH Low Demand - TRUE if any low-temperature greenhouse zone is calling for heat
• GH Demand - TRUE if any greenhouse zone is calling for heat
• Greenhouse Target - desired Glycol temperature

Vesta Rules

There are a few rules necessary to obtain the desired system behavior. First, the Vesta has to determine if there is greenhouse (GH) demand, and whether that demand includes a low-temperature zone. The two ‘state variables’ (GH Low Demand and GH Demand) are set:

Set ~GH Low Demand to TRUE if GH Office Tstat is true

Set ~GH Low Demand to TRUE if Slab Heat Demand is true

Set ~GH Low Demand to TRUE if Mech Room Tstat is true

Set ~GH Demand to TRUE if ~GH Low Demand is true

Set ~GH Demand to TRUE if Bench Heat Demand is true

The greenhouse target temperature is set to 120 unless the bench is the only demand - then, it’s set to 150.

Set Greenhouse Target to 120.0 if ~GH Demand is true

Set Greenhouse Target to 150.0 if Bench Heat Demand is true and ~GH Low Demand is not true

Finally, a PID rule is used to modulate P15 to hold the Glycol supply temperature at the desired target temperature any time there’s greenhouse demand:

Control P15 GH HX Pump to make GH Glycol Hot Supply close to Greenhouse Target when ~GH Demand is true

Vesta User Interface

In addition to the built-in Vesta user interface, this site had a custom GUI that provided monitoring for the whole system. This was updated to show detail of the greenhouse subsystem.

The circulator pumps are not numbered, but P15 is at the top right. The bar graph next to it shows that it’s running at just over 50% of full speed. The greenhouse benches and radiant slab are active as is the office radiant zone.

Results

The system now provides consistent Glycol temperature. The chart below shows the controls responding to changes in zone demands in the greenhouse.

At first, the temperature of hot water available from storage (magenta line) is a consistent 180℉, and the greenhouse office, bench, and slab are all calling for heat (it’s below zero outside). Circulator P15 (green line) is running at about 30%. Glycol supply to the greenhouse (yellow line) is stable at 120, which is the target temperature. Return water going back to storage is about 110℉.

At about 6:15, the available temperature in storage starts to drop. P15 speeds up to compensate, maintaining a steady 120℉ Glycol supply temperature.

At just after 8:00, the office is satisfied. A few minutes later, the slab is satisfied. At this point, the bench is the only zone calling for heat, and the target temperature is raised to 150℉. P15 ramps up to about 90%, and the Glycol temperature overshoots briefly. By 8:30, the Glycol temperature has stabilized and P15 is running at about 20%.

Finally, the bench zone is satisfied at about 9:40, and P15 shuts off.

As a separate note, it was determined that the heat exchanger was mis-plumbed with parallel flow rather than counterflow. When that’s fixed, return water temperatures will likely be even lower.

Appendix A

P15 was originally a Grundfos Alpha 15-55. The plumbing between this circulator and storage is almost all 2” copper, presenting a very low head loss. At full fixed speed, the circulator delivered a flow rate of about 10gpm.

P15 was replaced with a Grundfos 15-58FRC. This is a fixed-speed pump,with three switch-selectable speeds. It was installed with the highest speed selected, but connected to a variable-speed control unit that can vary the speed continuously from 0% to 100%. At 100%, the 15-58 has a flow rate nearly the same as the 15-55.

Appendix B

Heat transfer in heat exchangers is a very complex topic. In this case, we did not have any specifications for the heat exchanger, so we can only model heat transfer one each side of the heat exchanger independently, and make reasonable assumptions about the performance of the heat exchanger itself.

Heat Transfer Equations

Water weighs 8.34 lbs/gal and 1 BTU will raise 1 lb of water 1℉. The equation for heat transfer is therefore:

BTU/hr = Gallons/Minute * Minutes/Hour * Pounds/Gallon * (T1-T2)
BTU/hr = Gallons/Minute * 60 * 8.34 * (T1-T2)

Solving for DeltaT (T₁-T₂):

(T1-T2) = BTU/hr / (Gallons/Minute * 60 * 8.34

Solving for flow rate:

Gallons/Minute = BTU/hr / ((T1-T2) * 60 * 8.34)

Greenhouse Zones

The greenhouse has four zones. BTU loads are from the design documents. Office and bench flow rates are actual measured values. Note that the bench zone can be supplied with water at two different temperatures and presents a different BTU load depending on supply temperature.

Using the equation for Delta T above, we can calculate the drop in temperature that we would expect from each zone, and the resulting return temperature:

Under peak load conditions, the greenhouse is supplied with hot Glycol at 120℉, and the return temperature will be about 100℉, with a total heat load of about 150,000 BTU/hr.

Supply (Hot Water) Side

The supply side of the heat exchanger is fed with hot water at about 170℉. In the original configuration, P15 was a fixed-speed circulator delivering about 10gpm. This was reasonable at peak load, but with smaller heat demands it resulted in 10gpm of very hot water being returned to the bottom of storage:

Predicted Values

The ideal flow rate under these conditions would be slow enough to minimize the return temperature to storage. If the heat exchanger has adequate surface area, we might be able to achieve a return temperature on the hot water side that's close (perhaps 10℉ above) to the return temperature on the Glycol side:

Observed Values

Logged data for office only and combined slab and bench zone operation shows good correlation with predicted values. Note that circulator speed is only available as a percentage of electrical drive signal, which does not have a linear relationship with actual flow rate. At this time, peak load conditions have not been observed.