08 Hearing Exhibit 108_Attachment JTS-5 - Public_Salas O_Brien_Ambient_Loop_Study.pdf

Denver District Steam and Chilled Water Systems Ambient Temperature

Ground Loop Feasibility Study

December 13, 2023

Prepared By

Brian Urlaub, AVP | Director of Geothermal Operations

Megan Oelze, Mechanical Engineer | Project Manager

Josh Plummer, Mechanical Engineer, EIT

Josh Hays, Energy Engineer, PE

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5 Proceeding No. 22A-0382ST

Denver Ambient Loop salasobrien.com 2

Table of Contents

Denver District Steam and Chilled Water Systems Ambient Temperature Ground Loop Feasibility Study

Ambient Loop

Goals

Criteria

Methodology

Thermal Load Profiles

Steam Data

CHW Data

New CHW Data for Steam Buildings

Analysis

Steam Loop Network

TRNSYS Model

Summary of GHX Sources

Hydraulic Model

CHW Loop Network

TRNSYS Model

Summary of GHX Sources

Hydraulic Model

Capital Ambient Loop Network

TRNSYS Model

Summary of GHX Sources

Hydraulic Model

Aggregate energy source requirements

Sewage Heat Recovery

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 3

4-pipe HHW & CHW Plant

Conclusion

Other Considerations

Energy Sources

CAPEX

Phasing

Appendix

TRYNSYS Schematic - Denver steam Pipe Ambient Loop………………………………………...35

Denver Site - Steam_Existing………………………………………………………………….……...36

Denver Site - Chilled Water_Existing…………………………………………………………………37

Ambient System - Downtown Network Piping (1-Pipe) …………………………………………….38

Ambient System - Downtown Network Building Groups (1-Pipe) ………………………………...39

Ambient System - Capital Loop Piping (1-Pipe) …………………………….………………………40

Ambient System - Capital Loop Building Groups (1-Pipe) …………………………….……..……41

Ambient System - Downtown Network Piping (2-Pipe) …………………………….………………42

Ambient System - Downtown Network Building Groups (2-Pipe) …………………………….…..43

Denver Hydraulic Model - Steam Lines to Ambient …………………………….………………….44

Denver Hydraulic Model - CHW Lines …………………………….…………………………………47

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 4

EXECUTIVE SUMMARY

Salas O’Brien was contracted to perform a study to determine the technical feasibility of converting both the chilled water (CHW) systems and the steam piping systems to an ambient temperature loop (ATL) served by a vertical ground loop heat exchanger system (GLHE) or geothermal system. RMH was contracted by Xcel Energy, and tasked with developing the building heating and cooling loads as well as analyzing different technologies at the building level to determine what systems would be technically feasible. Salas O’Brien received all the building thermal load profiles from RMH for this study, and then provided RMH the temperature profile of the ATL for them to use to determine the efficiencies of the geothermal heat pump systems at each building.

The main objective was to determine if the existing piping infrastructure, both CHW and steam, are able to be converted to an ATL system. It had been determined that laying new pipe for an ATL network in the streets of Denver would be cost prohibitive. To determine this, Salas O’Brien analyzed both the hydraulics required of the system and the amount of energy required of the system in order to heat and cooling each building that would be connected to the ATL. The study would also make some conclusions as to whether or not the existing piping network size and routing would be sufficient for the transition to an ATL.

After the analysis was complete, it has been determined that the existing steam and CHW piping could be converted and utilized as an ATL system to serve the buildings noted in the study. These buildings would need to be converted to a geothermal heat pump (GHP) system to take advantage of the ATL.

As you will find in the report, there are some caveats to this determination such as a few areas in the steam piping network that would need new pipes added, or in the CHW piping network a few buildings would need to be served by a different piping network or system. Lastly, based on the existing pipe sizing, expansion of new building connections would be limited based on the maximum flow due to the size of the current piping.

AMBIENT LOOP

An ATL system utilizes an ambient ground temperature between 40°F-90°F. The piping network is filled with water and potentially a food grade antifreeze (Propylene Glycol) (maximum of 30% mix in northern climates, or where the GLHE/ATL minimum temperatures are less than 36°F) that interconnects buildings and can share thermal energy between them. The balance of the energy would then come from some other source, such as a vertical GLHE, wastewater heat recovery, surface water (river or lake), or other types of renewable energy that may be available on or near the site. These sources could be either heat absorption or heat rejection opportunities. The benefit of a networked system is the sources do not have to be adjacent to the building it serves but can be located down the street or in another part of the community that is all interconnected. Another benefit of ambient temperature loops is that they can be uninsulated and with minimal heat loss to the ground vs. steam or hot water. In this regard, ATL piping networks are very similar to city water supply systems. Each building would utilize an electric geothermal heat pump system for heating, cooling, and DHW. The Figure 1 below is representative of an ambient loop geothermal network utilizing a single pipe to interconnect buildings with vertical ground loop heat exchangers as the source energy. The single pipe offers lower cost of installation and less space than a traditional two pipe system.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 5

Figure 1

Goals

Salas O’Brien was contracted to perform a feasibility study to evaluate converting the downtown Denver district steam and chilled water piping systems into an ambient temperature system. Salas was also contracted to investigate the magnitude of the ground heat exchanger needed to handle the entirety of all building loads.

Criteria

Xcel Energy would like to determine a solution that could reuse as much of the existing piping network as possible in order to reduce the amount of construction and capital expenditure that would be needed to decarbonize downtown Denver.

The geothermal system in this study is being designed with a minimum ATL temperature in heating of 42°F so antifreeze solution is not needed in the system. This saves on cost due to the large volume of fluid in the total ATL system, and also increases the capacity of the GHP’s at the building level which can reduce the overall size required for each building.

METHODOLOGY

To answer the pressing questions on the feasibility of utilizing an ambient loop in the downtown Denver area, our team quickly recognized that a detailed model of the buildings and equipment on the proposed ambient loops was needed. Unlike most HVAC systems and central hot water and chilled

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 6 water plants, ambient loop systems are not easily modeled using simple spreadsheet techniques or even many of the more commonly used modeling packages on the market today for several reasons:

1. The borehole ground heat exchangers common to ambient loop systems act as large thermal batteries and it is critical to know how much energy has been stored and/or retrieved from these batteries to accurately predict their performance in both the short- and long-term time frames. In addition, the ground heat exchangers themselves are relatively difficult to model as they can exhibit both vertical and radial/horizontal stratification, multi-mode heat transfer including conduction, convection, radiation and advection, and localized and global thermal storage.

2. The performance of the heat pumps (typically in each building) depends heavily on the temperature of the water/glycol entering the heat pump from the ambient loop. But the temperature at any point in the ambient loop is a function of all the buildings and equipment upstream of the current location. And with these systems truly being a closed loop, the performance of a single piece of equipment is also impacted by the performance of this same piece of equipment a few minutes before as the water circulates through the loop and returns to the equipment. In effect, the performance of any piece of equipment on the ambient loop is then a function of the performance of every other piece of equipment on the ambient loop; including the large thermal storage ground heat exchangers which dampen, but do not eliminate any large temperature fluctuations. This tightly coupled dependency requires a solution technique designed to solve these types of problems quickly and efficiently.

3. The controls on ambient loop projects are somewhat unique; often requiring local control decisions on the operation of the individual ground heat exchangers and ancillary heating absorption/rejection devices as well as an over-arching master control for operation of the ambient loop pumps. These controls can become even more involved if seasonal loop controls are utilized – for example using a cooling tower in the cooler spring months to pre-cool the ground heat exchangers for the hot summer months ahead - without cooling the fields too much and negatively impacting their ability to heat the buildings in the short term.

4. The distribution and number of the heat rejection/heat absorption devices is absolutely critical.

In an ambient loop system, the customer directly in line after a ground heat exchanger sees the most advantageous entering fluid temperatures to the heat pumps; with heavy dampening of the peak highs and lows and therefore allowing for the most efficient operation. The customer last in line sees the impact of all of the customers upstream of them and sees the lowest entering fluid temperatures in heating mode and the highest entering fluid temperatures in cooling mode – leading to a decrease in heat pump efficiency. To combat this large temperature change for the last customer in line, higher ambient loop flows can be utilized, or the number of customers between heat injection/extraction points can be reduced. Each of these options results in a significant change in either the first cost or operating cost and needs to be carefully studied.

With these requirements in mind, the team chose to use the TRNSYS simulation software package.

TRNSYS is a powerful and incredibly flexible modeling tool that is ideally designed for simulating ambient loop projects. The vast libraries of heating, cooling, and power generating equipment let experienced users create custom, detailed, dynamic, system models of complicated real-world projects and estimate their performance over a wide range of operating conditions. The program has been used for more than 40 years in projects all over the world and has a strong reputation for its proven performance in solving difficult modeling tasks.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 7

The TRNSYS simulation software package was chosen as the modeling platform for this project for several reasons:

• The program has an extensive library of excellent component models for the study of this project including ground heat exchangers, pumps, buried pipes, heat pumps, chillers, cooling towers, heat pumps, boilers, controllers, etc.

• These models can be tuned/calibrated to measured or manufacturer-provided data to improve their predictions.

• The program is incredibly flexible and allowed the team to build custom ambient loop distribution systems to study the different layouts for downtown Denver.

• New component models can easily be added to the package, and we used that capability to write custom control models for the ambient loop and ground heat exchanger pumps.

• The program uses a dynamic solution technique to solve the resulting system of interconnected components such that the impact of small changes in one model get reflected in the operation of the entire system. For example, a change in the load of one of the buildings impacts the flow that gets to that building, the thermal losses from all the pipes that deliver water to that building, the pumps that circulate water around the loop, and the heat exchange between the ambient loop fluid and the soil in the ground heat exchanger.

• The team has used TRNSYS successfully in several other ambient loop projects; including the

Berczy Glen development in Canada which is currently under construction and is nearing full operation. In that project, TRNSYS was extensively used to study different design options including ground heat exchanger placement and design, various pump control strategies, supplemental heat production including solar thermal collectors and natural gas boilers, and supplemental heat rejection using fluid coolers and cooling towers.

To study the performance of the proposed ambient loop systems, detailed system simulations were first created in the TRNSYS simulation software. Component models representing the pipes, pumps, heat exchangers, heat pumps etc. were connected together in the simulation program much like the real system is connected together via pipes and wires. TRNSYS then simulated these designs at 3-minute time steps for a period of 10-years to evaluate the transient performance of the systems; carefully watching for excessive heat build-up in the ground heat exchangers or periods where low-temperature operation might cause freezing issues in the ambient loop.

For this project, we originally created six different design scenarios that we studied in TRNSYS before ultimately settling on three designs for further analysis. The original six designs included:

1. A new large-diameter ambient loop through downtown Denver to serve all the buildings currently being served by steam.

2. An ambient loop utilizing as much of the existing steam network of pipes as possible and serving all the customers currently on the steam distribution network.

Denver Ambient Loop salasobrien.com 8

3. A central plant option where chilled water and hot water is made at a central downtown location using heat recovery chillers and distributed to the current CHW customers using a combination of existing and new piping in the downtown area. The heat rejection and heat absorption requirements of the heat recovery chillers are met by a large, centralized, ground heat exchanger field. This option allows the simultaneous heating and cooling loads to be met concurrently by the chillers and was analyzed to see if it could provide a reduction in the number of ground heat exchanger bores required.

4. An ambient loop system through downtown Denver that utilizes the existing steam lines as much as possible but eliminates several buildings that are farther off the ambient loop than would be practical to efficiently serve them.

5. A smaller ambient loop in the Capitol region that meets the loads of just a dozen buildings in the surrounding area.

6. An ambient loop using as much of the existing CHW lines in the downtown area as possible that serves the customers on the existing CHW lines.

Option 1 was intriguing because it allowed for a more compact ambient loop through the downtown area. And with large ambient loop pipes (24”), the pumping power was greatly reduced, and the number of required heat injection points was also reduced (large flows means more customers between heat injection/extraction devices). But ultimately the cost of an all-new piping network was deemed to be prohibitively expensive, and this option was shelved despite its performance potential.

Option 2, which is very similar to Option 4, was ultimately rejected for practical reasons. It just didn’t make a lot of sense to serve several of the buildings on the existing steam network with an ambient loop when they were located far from the main ambient loop lines and required modifications to the existing piping network in order to make them work.

Option 3 was ultimately shelved after an extensive analysis revealed that the simultaneous heating and cooling benefit of the centralized heat recovery chillers did not significantly reduce the number of bores needed in the ground heat exchangers for the central plant. When you add in the need to install all new hot water piping to the buildings being served by the CHW network to make this option work, a cost intensive exercise, the option was deemed not likely to be beneficial.

This left three options that made the final cut; two typical single-pipe ambient loop designs that serve most of the buildings on the existing steam network, and a two-pipe ambient loop that serves the majority of the buildings on the existing CHW network - plus a few building from the steam network that made more sense to be lumped into the CHW design than to be met by the steam line system. It should also be noted that the two systems being served on the steam lines could be combined into a single loop – we simply severed them into two projects to demonstrate how a small single loop in the Capitol region could operate independently.

Steam Line Design Methodology

When it became clear that the focus was to re-use as much of the existing steam line network as possible, and not gravitate towards a solution with new larger piping, we spent considerable time

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 9 reviewing the existing steam network and laying out potential routes for the ambient loop. The keys for designing the flow network were:

1. Re-use as much of the existing steam network as possible and limit the number of new pipes that need to be laid down. Avoid pipes crossing over each other and unnecessarily long runs between buildings.

2. Try to avoid the use of parallel flow loops that require additional pumps/vaults/control valves.

3. Figure out which of the outlying buildings on the steam network should not be served by the ambient loop due to cost/installation concerns.

4. Determine the minimum pipe size that we’ll allow in the loop. In a closed loop system, the flow must be able to pass through each section of the loop without excessive pressure drop and out-of-range flow velocities. Larger pipes require less heat injection/extraction points and have a lower pressure drop but replacing large sections of the existing network is going to be cost prohibitive.

5. Set the maximum allowable temperature drop along the loop between heat injection/extraction points to minimize the impact of a building’s location along the loop.

6. Attempt to keep the ground heat exchangers at the same size along the loop. This standardizes the pump and valve sizes and reduces the required inventory of replacement parts to be kept on hand in case of a failure.

7. Attempt to keep several of the ground heat exchanger locations close to each other so that a single ideal location in the downtown area could potentially serve multiple heat injection/extraction points.

After careful consideration of the above design criteria and a careful review of the existing network of steam pipes, we chose a 12” pipe as the minimum size that we’ll allow for full flow through the ambient loop. Using typical sizing criteria for pipes, this equates to a peak flow rate of approximately 3600 gpm.

To balance the need to keep the number of ground heat exchangers low against the need to keep the temperature drop/rise along the loop between injection/extraction points reasonable, we decided upon a design temperature drop/rise of 10°F. At this temperature difference, we can meet approximately 1500 tons of load between injection/extraction points. 1500 tons is also large enough to cover the largest single building peak load such that we don’t have to split flows into a building and straddle a ground heat exchanger location.

With this sizing, the next task was to lay out a downtown loop that created similarly sized loads between injection/extraction points, kept several fields close together, and minimized the amount of new piping needed. On paper this exercise seems easy, but it’s harder to achieve than it seems due to each building having a different peak load – with the larger buildings providing the biggest challenge. If all of the buildings had identical 100-ton peak loads, you could simply put an injection/extraction loop every 15 buildings and be done. But a 1000-ton peak building located next to an 800-ton peak building means there has to be a ground heat exchanger located between those buildings on the loop and you have to work upstream and downstream from there to find the next ground heat exchanger locations.

Luckily, we were able to find a solution that met many of our design goals; resulting in ten 300-bore

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 10 ground heat exchangers being required around the main ambient loop and two 300-bore ground heat exchangers being required around the Capitol ambient loop. With the ambient loops built within TRNSYS, the models were then driven with hourly net loop heating and cooling loads that were provided to us by RMH and are specific to the needs of each building. These two ambient loop designs, along with results from the analysis of these designs, are presented later in this report.

Chilled Water Line Design Methodology

The existing chilled water lines serving the downtown Denver area provided an interesting design opportunity for an ambient loop. With their two-pipe layout (one supply pipe and one return pipe for each building) serving more than thirty downtown buildings, we were able to build an ambient loop system where each of the buildings are connected in parallel off the ambient loop instead of in series off the ambient loop like in most typical ambient loop systems. This design means that the upstream buildings on the ambient loop do not impact the performance of the current building as directly as they would in a typical ambient loop with series connections. Instead, each building’s return water mixes together and is returned to a central location where it passes through the central ground heat exchangers and is then returned to the buildings as conditioned supply water. In this design, all the ground heat exchangers can be centrally located and do not need to be distributed along the ambient loop. The other benefit of this design is that not every pipe in the ambient loop must be sized for the full ambient loop flow rate. The supply pipes are sized larger near the central ground heat exchangers and get smaller as more and more buildings get served off the main line. Conversely, the return lines grow in size as more and more building returns get mixed back together. With the existing network already designed to serve these 30+ buildings with chilled water, it is a simple conversion to change the network from a chilled water distribution network to an ambient loop network – provided the heating load in these buildings is not significantly larger than the current cooling load (a much larger heating load would mean that bigger supply/return lines are needed to serve that building than they currently are sized for).

After careful review of the existing CHW piping network, the following design criteria and methodology was used to build the ambient loop system on the existing CHW network:

1. Replace the CHW distribution systems in each building with typical water-to-air heat pumps.

2. Serve the building heat pumps from an ambient loop and keep the inlet temperatures ideally in a range between 40°F and 90°F.

3. Utilize a centralized ground heat exchanger field to dampen the loop temperatures and act as a source and sink for the heat pumps.

4. Check to make sure that all of the pipes in the existing piping network are sized sufficiently large to avoid high pressure drop / high velocity conditions at peak heating and cooling loads.

5. Serve all the buildings on the current CHW distribution network; plus, several of the buildings on the steam network that are located close to the existing CHW network and far from the existing steam network lines.

With these design goals in mind, we built a detailed ambient loop system in TRNSYS using the existing CHW pipe sizes and layout, typical water-source heat pump performance, and models of the ground

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 11 heat exchanger system and pumps, and then drove the simulations with the estimated hourly heating and cooling loads for each building. We then carefully analyzed the results from this system over a 10-year period and looked for problems in the initial design. We found that we needed to split the convention center and two other buildings off one of the current CHW distribution loops and put them on their own distribution network to reduce the peak flows through the supply and return lines near the central ground heat exchangers. As these buildings are located close to the central location, and already piped in a favorable way, this change should not be overly difficult/expensive. The model predicts that we’ll need approximately three thousand, 600’ bores to meet the cooling and heating needs of these buildings. This improved design, along with results from the analysis of this design, is presented later in this report.

THERMAL LOAD PROFILES

Three different subsets of thermal load profiles were used in the development of the TRNSYS model for the ambient loop study. Steam data from metered steam customers, metered CHW data, as well as added CHW load to steam buildings that currently aren’t CHW customers were all used in the modeling. These chilled water profiles were provided by RMH Engineering.

Understanding the thermal profiles and how they influence the direction of the ambient loop design is important. An ambient loop system is ideal in a location where you have diverse loads on the system.

When you look at the steam and the chilled water profiles below, you’ll notice a large year-round heating and cooling load throughout the downtown area which shows great potential for networking these buildings together. Networking buildings together allows for sharing of energy between the cooling and heating loads happening at the same time which is referred to as simultaneous load. This simultaneous load indicates periods of the year where the ambient loop network can transfer energy from one building to another without need a heat source or heat rejection source. The aggregate load profile in Figure 3 shows this simultaneous band, along with the remaining unbalanced heating and cooling loads which need a heat source or heat rejection source paired with the ambient loop network.

Steam Data

Xcel Energy serves roughly 120 customers in downtown Denver with low pressure steam. The steam data was evaluated at the building level. The data was filtered to remove any negative values, as well as normalizing any peaks that seemed unreasonably high. When filtering the data, we noticed a few meters had data points showing usages of over 10 times higher than any other data point during snapshots of time. After verifying these as outliers with Xcel, these points were normalized to match the data in the meter.

It was important to filter these load profiles due to the peak usages seen in the ambient loop network to determine the feasibility of using the existing pipe network from calculated flow rates. These outliers drove the peak use in the TRNSYS model, which in turn drove all aspects of the simulation from flow capacities, pump power, assets for heat rejection/injection, etc. These steam meters were then aggregated to form the load profile network used in the TRNSYS model. The sum of all steam meters can be seen in Figure 2.

Denver Ambient Loop salasobrien.com 12

Figure 2. Total District Steam Load

The steam customer data received was filtered to remove any negative values, as well as removing any peaks that seemed unreasonably high. These unreasonably high data points showed usages of over 10 times higher than any other data point. It was important to filter these load profiles due to the peak usages seen in the ambient loop network to determine the feasibility of using the existing pipe network from calculated flow rates.

CHW Data

The Denver district CHW plant serves roughly 45 customers. Annual load profiles of these CHW meters were aggregated for individual buildings for modeling purposes. The sum of all meters can be seen in the figure below.

One major advantage of connecting all of these buildings together is the simultaneous heating and cooling loads happening at each building. These loads will be picked up by the heat pumps and moved throughout the ambient loop network without needing any energy sources. The simultaneous load is shown in purple in Figure 4.

Denver Ambient Loop salasobrien.com 13

Figure 3. Total district chilled water loads.

Figure 4. Aggregate load profile with simultaneous load

Denver Ambient Loop salasobrien.com 14

New CHW Data for Steam Buildings

In conjunction with the ambient loop study, RMH was contracted to work on the customer-sited equipment needs. Along with this, RMH developed load profiles for buildings connected to this ambient network which includes steam customers that are not current CHW customers. These load profiles were included in the TRNSYS model on a building-by-building basis. A summary of the added loads to the ambient network can be seen below.

Figure 5. Annual Load Profile for Added Loads to the Ambient Loop

Denver Ambient Loop salasobrien.com 15

ANALYSIS

Our process for analyzing how to determine the feasibility of an ambient loop in downtown Denver was determined in part by the parameters set by Xcel Energy. Those parameters were evaluating the possibility of reusing the existing chilled water and steam piping, determining what pumping power would be needed, and the size of the assets needed to support the ambient loop (geo-exchanger, cooling towers, etc.)

To Model an ambient loop system and determine its feasibility, a complex software is needed to accurately reflect the flow of energy. For this project, a model was developed using transient system simulations tool (TRNSYS). TRNSYS can model real time results of large complex networks, providing results that can help determine power consumed by the system, loop temperatures at various locations, and size of assets needed to reject/inject heat into the system. This information then gets used to build a Pipe Flow model that can be used to identify the total pumping needs in the system, and ultimately determine if the current pipe sizes are adequate to supply the flows required for the system.

Three different ambient loop networks were identified and modeled:

(1) Steam loop network (single pipe)

(2) CHW loop network (two pipe)

(3) Capital loop network (single pipe)

Steam Loop Network

TRNSYS Model

The buildings included in the steam loop network can be seen identified in the figure below. For detailed layout of the TRNSYS model, see Appendix. Numerous piping scenarios were considered when trying to identify a solution that, implemented the least amount of new piping infrastructure underground. The final piping scenario that optimized the existing piping while meeting all of the building loads can be seen in the map below.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 16

Figure 6. Steam loop network buildings with ambient loop pipe.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 17

The piping segments that are reused are represented in green, while the pipes that will need to be either replaced or added are shown in red. All new piping is estimated at 12” which shows sufficient size for capacity needed from the current building loads. One thing not considered in this analysis is the ability for the system to meet the demand considering future growth or added customers.

Existing segment pipe sizes were used in order to determine the feasibility of the ambient loop meeting the customer loads. A detailed map showing the segment pipe sizes can be found in the Appendix.

Summary of GHX Sources

For an ambient loop concept to work, the temperature bounds of the loop must stay between 46oF (assumes no glycol) and 95oF. This requires multiple re-injection sources at various locations. The hourly GHX temperature profile throughout the year for the various GHX’s can be seen in Figure 7.

Figure 7. GHX Annual Temperatures – Steam pipe ambient network.

Denver Ambient Loop salasobrien.com 18

The worst loop temperatures modeled for the simulation year for heating and cooling can be seen in the figure below. This figure shows how the loop temperature progresses from building to building and was used to determine the number of injection points to make sure the ambient loop stayed within the bounds as required.

Figure 8. Temperature gradient throughout steam network ambient loop pipe.

Denver Ambient Loop salasobrien.com 19

The re-injection points are summarized in the table below with estimated peak and annual loads for each re-injection point. Ground heat exchangers (GHX) were sized to show the scale of a possible solution for each injection point. The summary of the number of closed loop vertical bores and the depth of bore can be seen in the table below.

Table 1. Summary of Re-injection sources

GHX

# of bores Bore depth

Annual heat rejection (MMBtu)

Peak heat rejection

(tons)

Annual heat injection (MMBtu)

Peak heat injection

(MBH)

1 300 600 20,756 1,206 8,221 5,040

2 300 600 20,548 1,210 7,520 5,176

3 300 600 20,333 1,201 5,376 5,504

4 300 600 20,593 1,229 6,146 5,193

5 300 600 20,733 1,233 6,202 4,896

6 300 600 21,137 1,284 5,816 4,971

7 300 600 22,403 1,315 5,549 5,118

8 300 600 21,461 1,268 5,061 4,282

9 300 600 19,892 1,193 5,716 4,238

10 300 600 18,809 1,141 4,759 3,886

Total 3,000 600 206,665 12,280 60,366 48,304

The injection points could also be served by other sources. For instance, if the Cherokee plant waste heat was utilized and piped from the plant to these locations, it could serve as the heat source. The cooling source could then be done locally with cooling towers since those are electric devices, and do not emit any onsite carbon, this would still achieve the main goal of decarbonization. The peak heating load is what would be required from that other source to provide the necessary energy to do the same job as the GLHE system. Another option could be utilizing wastewater heat, which again doesn’t have to be at that direct location and could be piped from wherever that resource would be within reason.

Hydraulic Model

Hydraulic modeling was critical in assessing the feasibility of re-using existing steam piping infrastructure. In order to re-use existing pipe infrastructure, we have to understand how much load/flow of energy is going through the system at the worst-case conditions either in the winter or summer. For

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 20 this modeling, the peak loads/flows through the ambient loop network were mapped out to start and assess the flow through pipes, water velocity, and overall pressure in the system.

The Steam Loop Network consists of 2 hydraulically connected parallel loops; a main loop on the right side (southeast) and a sub-loop on the left side (northwest) of this piping network. Flows and pumping requirements are shown in Table 2 below with all new piping being 12” schedule 40 steel. A couple of challenges to note about this Steam Loop Network are flow control of parallel loops and high pumping head requirement of the main loop.

Table 2. Summary of flow & pump sizing

Ambient System

Ambient Loop

Pumping Requirements

GPM HL

Total

HP

Main Pumping kWh

Geo Pumping kWh

Steam Network (1-Pipe)

Main Loop (Chiller Plant Loop) 3,600 585 626 502,046 938,014

Sub-Loop (Steam Plant Loop) 2,200 85 56

To ensure ambient fluid flow can be properly controlled in both parallel loops, multiple pumping locations or modulating control valves will be required. For the multiple pumping location option there would be at least one pumping location on the main loop and another pumping location on the sub-loop. For the control valve option there would be line-sized modulating control valves on the sub-loop and a 2nd on the parallel loop’s common pipe.

To address the high pumping head requirement of the main loop, multiple pumping locations should be spaced around this loop to keep system pressure below standard HVAC equipment pressure rating (150 psig), Figure 9 shows these pumping locations highlighted and shows system pressure as a color gradient.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 21

Figure 9. Hydraulic model of steam loop network.

Denver Ambient Loop salasobrien.com 22

CHW Loop Network

TRNSYS Model

The CHW loop network was analyzed using the existing CHW supply/return pipe and turn it into a source loop with minimal new piping. The piping segments that are reused are represented in green, while the pipes that will need to be either replaced or added are shown in red. Source loop would be a supply and return line with the same temperature bounds as the ambient loop network, but instead of needing multiple re-injection points, there only needs to be one located at the CHW central plant.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 23

Figure 10. Buildings along with piping included in CHW loop network.

On the CHW loop network, the US Mint building was removed, and instead the Denver Justice Center was added to the load profile due to far proximity from the steam piping. The connection to the Denver Justice Center piping would have to be new, but all other piping is existing.

Summary of GHX Sources

The CHW loop already has supply and return piping, so instead of needing multiple re-injection points throughout the piping system like the steam ambient loop network, all re-injection points need to be located at the chilled water plant. The hourly GHX temperature profile can be seen in Figure 11.

Through the TRNSYS modeling, the GHX was broken up into (4) different loops just in order to feasibly try and locate them in downtown Denver due to the large overall sizing.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 24

Figure 11. Hourly GHX temperature for CHW network.

Denver Ambient Loop salasobrien.com 25

A summary of loads as well as the size of GHX estimated for this system are summarized in Table 3.

Table 3. Summary of Re-injection sources – CHW Loop Network

GHX

# of bores Bore depth

Annual heat rejection (MMBtu)

Peak heat rejection

(tons)

Annual heat injection (MMBtu)

Peak heat injection

(MBH)

1 750 600 67,361 3,171 37,163 27,505

2 750 600 67,361 3,171 37,163 27,505

3 750 600 67,361 3,171 37,163 27,505

4 750 600 67,361 3,171 37,163 27,505

Total 3,000 600 269,444 12,684 148,652 110,020

Hydraulic Model

The CHW Loop Network consists of 4 separate (2-Pipe) loops all being pumped from a common location (Denver Chilled Water Center). Pumping requirements are listed below in Table 4 below with all new piping being schedule 40 steel, see hydraulic model PDFs in appendices for suggested sizes of new piping. The high pumping power requirement for Pod 1 is apparent when looking at Figure 12 below that shows hydraulic head loss as a color gradient with red being sections of piping that have a high head loss. This high head loss is not enough to warrant multiple pumping locations – unlike the Steam Loop Network – but does raise some concern about electrical cost to power Pod 1’s pumping requirements.

Table 4. Summary of flow & pump sizing

Ambient System Ambient Loop

Pumping Requirements

GPM HL

Total

HP

Main Pumping kWh

Geo Pumping kWh

CHW (2-Pipe)

Pod 1 26,739 221 1,756

8,657,298 3,469,424 Pod 2 18,208 140 757

Pod 3 10,427 25 62

Pod 4 10,236 25 15

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 26

Figure 12. CHW loop network hydraulic model.

Denver Ambient Loop salasobrien.com 27

Capital Ambient Loop Network

TRNSYS Model

A single pipe ambient loop network was analyzed for the capital loop network. Piping as well as re-injection sources were located for this district. The network piping as well as GHX locations can be seen in Figure 13.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 28

Figure 13. Buildings along with piping included in capital ambient loop network.

Since no steam piping exists for these buildings, all piping shown in red would have to be buried. This piping is all 12” schedule 40 carbon steel.

Summary of GHX Sources

Similar to the steam ambient loop network, the capital ambient loop network requires multiple locations of re-injection sources. This loop will require two GHX or re-injection sources to maintain ground temperature profiles within the given bounds. The GHX temperature profile on an annual basis can be seen in the figure below.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 29

Figure 14. GHX Annual Temperatures – Capital ambient loop network.

Denver Ambient Loop salasobrien.com 30

Figure 15. Temperature gradient throughout Capital network ambient loop pipe.

Denver Ambient Loop salasobrien.com 31

The summary of loads met by these re-injection sources, as well as quantity and depth of a GHX solution can be seen summarized in Table 5.

Table 5. Summary of Re-injection sources – Capital Ambient Loop Network

GHX

# of bores Bore depth

Annual heat rejection (MMBtu)

Peak heat rejection

(tons)

Annual heat injection (MMBtu)

Peak heat injection

(MBH)

1 200 600 12,893 763 5,344 3,254

2 200 600 13,561 790 5,758 3,398

Total 400 600 26,454 1,553 11,102 6,652

Hydraulic Model

Like the Steam Loop Network, this Capital Ambient Loop will circulate up to 3,600 GPM but one difference is that total pumping power for Capital Ambient Loop is significantly lower primarily due to shorter loop length. The Capital Loop requires 200 ft (87 psi) of head pressure at design flow which is much more manageable to achieve with a single pumping location than what the Steam Loop Network requires. The hydraulic results of Capital ambient loop are summarized in Table 6 below with all new piping being 12” schedule 40 steel.

Table 6. Summary of flow & pump sizing

Ambient System Ambient Loop

Pumping Requirements

GPM HL

Total

HP

Main Pumping kWh

Geo Pumping kWh

Capital (1-Pipe) Capital 3,600 200 214 396,920 495,698

Denver Ambient Loop salasobrien.com 32

AGGREGATE ENERGY SOURCE REQUIREMENTS

For all three loops, each one needs to be accompanied by an energy source to either reject heat or inject heat into the loop to maintain the minimum and maximum loop temperatures. For all three loops, an initial sizing of a ground heat exchange system was performed. A summary of the peak and annual loads needed for each loop are summarized in Table 7.

Table 7. Summary of Re-injection sources – Capital Ambient Loop Network

Loop # of bores Bore depth

Land area required

Annual heat rejection (MMBtu)

Peak heat rejection

(tons)

Annual heat injection (MMBtu)

Peak heat injection

(MBH)

Steam Loop

3,000 600 1.2M ft2

(27.5 acres) 206,665 12,280 60,366 48,304

CHW Loop 3,000 600 1.2M ft2

(27.5 acres) 269,444 12,684 148,652 110,020

Capital Loop

400 600 160k ft2

(4 acres) 26,454 1,553 11,102 6,652

Combined 6,400 600 2.6M ft2

(60 acres) 502,563 26,517 220,120 164,976

Sewage Heat Recovery

Platte River interceptor

- 19.9 MGD

- 18.2 MW

Cherry Creek interceptor

- 7.3 MGD

- 6.7 MW

Given the two interceptors, there would be a total heating capacity of 25 MW (85,000 MBH). This heating capacity would be able to cover 51% of the heating peak for the three combined loops. From an annual heating perspective, this could most likely handle more than 51% of the annual load from the heat injection standpoint, however, there would have to be additional heat sources provided at all GLHE locations in the three various loops to make up the remaining peak capacity. There would also have to be a way for the sewage/sewage source to make it all the way to each injection point in each loop as well. This could be accomplished with a HX and circulating water to each source, however that would be a significant amount of piping due to the widespread location of all of the heat injection sources.

Hearing Exhibit 108, Second Supplmental Direct Testimony of Joseph T. Schwark, Attachment JTS-5

Denver Ambient Loop salasobrien.com 33

4-PIPE HHW & CHW PLANT

An initial consideration for the project was to convert the district CHW plant into a HHW & CHW plant serving both HHW & CHW to every building connected to either the CHW network or the steam network. This system was not preferred or attainable due to the piping network that would have to be installed. The HHW piping alone would have to run down almost every street in downtown Denver.

From a utility coordination perspective, as well as an implementation perspective knowing every street would have to be dug up to locate piping was determined to not a viable solution.

CONCLUSION

After evaluating both the steam and CHW piping networks, Salas O’Brien has concluded that there is a viable technical solution so that the steam and CHW system can be reused as ATL systems serving the buildings connected to them. The pipe sizes are sufficient to provide enough energy to the buildings within that network, however the pipe sizes dictate the capacity available, and this will limit future expansion of the system, primarily the steam piping network. The steam piping network does require some new interconnecting piping to make a “loop,” and a few buildings would need to be removed off the SE CHW network to bring the hydraulic performance within good design practice. There will need to be additional infrastructure i.e., pump stations, which would need to be located on both the steam and capital loop network.

The last conclusion is that the large number of vertical GLHE’s required to do this solely with geothermal energy is not feasible as it would require acres of accessible land to install those systems.

Our assessment is that other potential sources such as the Cherokee Energy plant waste heat and sanitary wastewater heat would be the most cost effective and closest heat sources to the ATL’s. Our recommendation is to look at those resources and determine if they have the ability to provide what is needed to meet Xcel Energy’s goals of decarbonizing downtown Denver.

There are many other considerations to determine the feasibility of implementing an ATL in downtown Denver. Even though Salas O’Brien has determined that there is a technical solution to reusing existing pipe networks, there are many other factors which may prohibit this solution from being practical.

Other Considerations

Energy Sources

GLHE

The use vertical GLHE’s to handle the entirety of the heating loads for the loop are going to require 6,400 vertical bores which is equivalent to roughly 60 acres of open and accessible space. 60 acres of open space is rather unrealistic given the dense building landscape, and therefore not a realistic solution for an energy source. Our recommendation is that other energy sources would have to be analyzed since using only geothermal as a resource would require too much land area.

Denver Ambient Loop salasobrien.com 34

Sewage Heat Recovery

Sewage heat recovery was another potential heat source analyzed for the project. Given the capacity from the Denver Metro Wastewater main interceptor lines, it was determined that this could handle 51% of the overall heating capacity for the three different loops. This would also require the sources to be scattered to the 16 location points of re-injections sources on the maps. In order to make the sewage heat exchange source a viable solution, there would have to be a significant amount of piping installed, as well as the need for supplemental heat sources at each injection point. From an implementation perspective, it seems difficult to make such a solution work knowing how many different locations ultimately this heat source would have to make it to. Also from a regulation perspective, it’s possible that the wastewater would not be a viable source for heat rejection but only heat injection. This would require heat rejection sources, like cooling towers, to be located at every single injection point as well.

Our recommendation is that such a system would be unlikely to serve an ATL loop at such a large scale.

CAPEX

Salas O’Brien was not tasked with costing the piping implementation or other costs associated with converting buildings to be heat pump compatible. A rough order of magnitude cost of the GLHE’s for the entire loop however would fall somewhere around $200M range based off a drilling cost of $50/ft drilling depth. For this project to considered feasible from an implementation perspective, a full cost of the overall system would have to be analyzed.

Phasing

One important consideration on the ultimate feasibility for converting the steam and CHW loops into an ATL would be how to phase such a system. From a phasing standpoint, each building would have to be heat pump compatible, the new piping sections for the steam loop would have to be installed, and then the system would have to be converted from circulating steam to circulating water immediately.