Add Coil:Cooling:ITE:ColdPlate for direct-to-chip cooling

design/FY2026/NFP-Datacenter_LiquidCooledCoil.md

@@ -0,0 +1,268 @@
+Liquid-Cooled Plant Equipment
+===============================
+
+**Jeremy Lerond, Pacific Northwest National Laboratory**
+
+ - Original Date: June 2026
+ - Revision Date: 07/15/2026
+
+## Justification for New Feature ##
+
+The rapid expansion of datacenter projects across the US has driven industry stakeholders to request native modeling capabilities for liquid-cooled IT equipment (ITE) in EnergyPlus, as documented in [[0]][Issue #11408]. Currently, EnergyPlus lacks both an IT equipment object for liquid-cooled ITE and plant equipment capable of modeling the resulting loads. The `ElectricEquipment:ITE:LiquidCooled` object is being developed concurrently. This NFP proposes a new plant equipment object designed to capture those loads and interact with supply-side equipment to satisfy them.
+
+## E-mail and Conference Call Conclusions ##
+
+## Overview ##
+
+[[0]][Issue #11408] provides an overview of current liquid-cooled datacenter technologies. Expanded information is available in [[1]][Datacenter Liquid Cooling Market Characterization, Final Report, 12/12/2025, CalNEXT]. Based on these resources, rear-door heat exchanger (RDHx) and direct-to-chip (D2C) technologies have emerged as the dominant solutions. Immersion cooling remains limited to experimental applications and is not yet widely deployed. Consequently, this NFP focuses on developing a new plant equipment object capable of simulating D2C cooling. RDHx will be considered in a subsequent NFP.
+
+## Approach ##
+
+This NFP introduces a plant object for modeling D2C cooling of `ElectricEquipment:ITE:LiquidCooled` objects, which are being developed concurrently. While research [[1]][Datacenter Liquid Cooling Market Characterization, Final Report, 12/12/2025, CalNEXT] and [[2]][Fernandes et al.] demonstrate that D2C and RDHx systems can work together to cool a single server rack, the proposed object supports only one cooling system per rack. Hybrid cooling configurations could be modeled in a future NFP once RDHx support is available.
+
+### D2C
+The new object models D2C devices as cold plate heat exchangers. The heat transfer through the cold plate is characterized by the case-to-fluid thermal resistance, $R_{th}$, expressed as follows (see [[5]][Wang et al.], [[6]][Cheng et al.], [[10]][Shaeri et al.], and [[7]][Martinez et al.]):
+
+$R_{th} = \frac{T_{case} - T_f}{\dot{q}}$
+
+or $T_{case} = T_f + \dot{q} \cdot R_{th}$
+
+where:
+- $T_{case}$ is the chip case temperature (labeled $T_c$ in Figure 1, with TIM denoting the thermal interface material)
+- $T_{f}$ is the liquid coolant inlet temperature
+- $\dot{q}$ is the heat transfer rate
+- $R_{th}$ is the cold plate thermal resistance
+- $R_{th,nom}$ is the nominal thermal resistance
+
+<img src="./cold_plate.jpg" alt="Cold plate cross section" width="400">
+
+_Figure 1 - Cross Section of a Cold Plate_ (source: [[7]][Martinez et al.])
+
+The maximum heat transfer rate at the nominal flow rate is therefore:
+
+$\dot{q}_{max} = \frac{T_{case,max} - T_f}{R_{th,nom}}$
+
+where $T_{case,max}$ is the maximum allowable chip case temperature.
+
+As shown in [[8]][Eaton], cold plate selection is driven by $T_{case,max}$, which together with $R_{th,nom}$ determines $\dot{q}_{max}$. As mentioned in [[0]][Issue #11408], D2C systems are not always able to address all IT-generated loads, and air systems are used to meet the remainder. When $T_{case,max}$ is reached, the remaining heat generated by the IT equipment will be added to the zone heat balance.
+
+[[5]][Wang et al.], [[6]][Cheng et al.], and [[10]][Shaeri et al.] show that the thermal resistance varies as a function of coolant flow rate (and, for two-phase flow, heat flux as well), so the thermal resistance will be adjusted from its nominal value using a dimensionless multiplier from a user-defined empirical curve. For the single-phase flow supported in this NFP, only the flow rate dependence is relevant.
+
+A target case operating temperature is used to size the nominal liquid coolant flow rate through the cold plate. During operation, the flow rate is then adjusted to maintain this target temperature. If the maximum flow rate is reached, the case temperature may rise up to $T_{case,max}$.
+
+The proposed object, together with a `HeatExchanger:FluidToFluid` object and a pump (on the supply side of the same `PlantLoop`), will represent a coolant distribution unit (CDU). Users will add it to the demand side of a `PlantLoop`. As requested by [[0]][Issue #11408], this approach allows D2C to "_be connected to existing EnergyPlus PlantLoop architecture_". Additionally, this approach supports the requirement that liquids with other heat transfer properties "_be handled at the loop level with the fluid_type and user_defined_fluid_type fields_". Coolant properties can be specified using the `FluidProperties` family of objects. Note that PG25 (a mixture of 25% propylene glycol and water) is a widely used single-phase coolant (see [[5]][Wang et al.]) natively supported by EnergyPlus when using the `FluidProperties:GlycolConcentration` object. Two-phase heat transfer is not currently handled in EnergyPlus; this NFP therefore focuses on single-phase D2C. Two-phase D2C can be modeled using the `Coil:Cooling:ITE:UserDefined` object being developed concurrently, or addressed in a future NFP.
+
+As an alternative to the $R_{th}$ approach, [[7]][Martinez et al.] introduce an LMTD (Log Mean Temperature Difference)-based thermal resistance ($R_{LMTD}$) definition for single-phase liquid-cooled cold plates. By default, the $R_{th}$ approach is used.
+
+If the $R_{LMTD}$ approach is used, then:
+
+$R_{LMTD} = \frac{1}{\dot{m} \cdot c_p \cdot \text{NTU}_{LMTD}}$
+
+and $\varepsilon = 1 - e^{-\text{NTU}_{LMTD}}$
+
+so $T_{case} = T_{in} + \frac{T_{out} - T_{in}}{\varepsilon}$
+
+where:
+- $T_{case}$ is the chip case temperature
+- $\varepsilon$ is the heat exchanger effectiveness
+- $\varepsilon_{nom}$ is the heat exchanger effectiveness at the nominal flow rate
+- $\text{NTU}_{LMTD}$ is the number of transfer units as defined in [[7]][Martinez et al.]
+- $R_{LMTD}$ is the LMTD-based thermal resistance
+- $\dot{m}$ is the liquid coolant mass flow rate
+- $\dot{m}_{nom}$ is the nominal liquid coolant mass flow rate
+- $c_p$ is the liquid coolant specific heat capacity
+- $T_{in}$ is the liquid coolant inlet temperature
+- $T_{out}$ is the liquid coolant outlet temperature
+
+The maximum heat transfer rate at the nominal flow rate is therefore:
+
+$\dot{q}_{max} = \dot{m}_{nom} \cdot c_p \cdot \varepsilon_{nom} \cdot (T_{case,max} - T_{in})$
+
+where $T_{case,max}$ is the maximum allowable chip case temperature.
+
+For both approaches, the outlet fluid enthalpy is:
+
+$h_{out} = h_{in} + \frac{\dot{q}}{\dot{m}}$
+
+where:
+- $h_{out}$ is the specific enthalpy of the fluid leaving the cold plate
+- $h_{in}$ is the specific enthalpy of the fluid entering the cold plate
+- $\dot{m}$ is the liquid coolant mass flow rate
+
+Temperature and other properties will subsequently be calculated from the fluid property data.
+
+## Testing/Validation/Data Sources ##
+
+Performance curves for the thermal resistance modifier will be developed based on data from [[5]][Wang et al.], [[7]][Martinez et al.], and other peer-reviewed publications.
+
+## Input Output Reference Documentation ##
+
+TBD
+
+## Input Description ##
+
+```
+Coil:Cooling:ITE:ColdPlate
+    A1, \field Name
+        \required-field
+        \type alpha
+    A2, \field Availability Schedule Name
+        \note Availability schedule name for this cold plate. Schedule value > 0 means the cold plate is available:
+        \note - If this field is blank, the cold plate is always available
+        \note - If the cold plate is not available, all heat generated by the ElectricEquipment:ITE:LiquidCooled object
+        \note   will be added to the zone specified in this object
+        \type object-list
+        \object-list ScheduleNames
+    A3, \field Inlet Node
+        \note Liquid coolant inlet node
+        \required-field
+        \type node
+    A4, \field Outlet Node
+        \note Liquid coolant outlet node
+        \required-field
+        \type node
+    N1, \field Nominal Thermal Resistance
+        \note Cold plate thermal resistance at the nominal liquid coolant flow rate
+        \default 0.025
+        \minimum> 0.0
+        \units K/W
+    N2, \field Maximum Case Temperature
+        \note Maximum allowed chip case temperature
+        \required-field
+        \minimum> 0.0
+        \units C
+    N3, \field Target Case Operating Temperature
+        \note Targeted operating case temperature, used to size the nominal liquid coolant flow rate; during operation, the flow rate is adjusted to maintain this target temperature
+        \note If not specified, defaults to the Maximum Case Temperature; the Target Case Operating Temperature cannot exceed the Maximum Case Temperature
+        \minimum> 0.0
+        \units C
+    A5, \field Thermal Resistance Modifier Curve Name
+        \type object-list
+        \object-list UnivariateFunctions
+        \object-list BivariateFunctions
+        \note For univariate curves, flow rate ratio (Flow rate / Nominal Flow Rate) is used as the independent variable
+        \note For bivariate curves, flow rate ratio is used as the first independent variable
+        \note and heat transfer rate ratio (Heat transfer rate / Nominal Heat Transfer Rate) as the second
+    A6, \field Thermal Resistance Method
+        \note If Standard, the thermal resistance is defined as follows:
+        \note R = (Tc - Tf) / q_dot
+        \note where R is the cold plate thermal resistance (equal to Nominal Thermal Resistance at nominal conditions)
+        \note Tc is the chip case temperature
+        \note Tf is the liquid inlet temperature
+        \note q_dot is the heat transfer rate
+        \note If LMTD, a LMTD-based thermal resistance is used instead
+        \note R = LMTD / q_dot
+        \type choice
+        \key Standard
+        \key LMTD
+        \default Standard
+    N4, \field Nominal Liquid Flow Rate
+        \note Nominal liquid flow rate through the cold plate
+        \autosizable
+        \minimum> 0.0
+        \units m3/s
+    N5, \field Maximum Liquid Flow Rate
+        \note Maximum liquid flow rate allowed through the cold plate at all times
+        \note If not specified, the flow rate is unconstrained
+        \autosizable
+        \minimum> 0.0
+        \units m3/s
+    N6, \field Auxiliary Electric Power
+        \type real
+        \units W
+        \minimum 0.0
+        \ip-units W
+        \default 0.0
+        \note Auxiliary power consumed when the cold plate is available
+    A7, \field Zone Name
+        \note Zone where heat not captured by the cold plate will be added
+        \type object-list
+        \object-list ZoneNames
+    A8; \field End-Use Subcategory
+        \note Any text may be used here to categorize the end-uses in the ABUPS End Uses by Subcategory table
+        \type alpha
+        \retaincase
+        \default General
+```
+
+## Outputs Description ##
+
+The following outputs will be added:
+- Chip case temperature
+- Heat transfer rate
+- Heat transfer energy
+- Inlet liquid temperature
+- Outlet liquid temperature
+- Liquid mass flow rate
+- Auxiliary electric power
+- Auxiliary electric energy
+
+## Engineering Reference ##
+
+TBD
+
+## Example File and Transition Changes ##
+
+This object, along with the other liquid-cooled data center objects, will be integrated into a single example file. This effort will be coordinated with other entities involved in their development.
+
+No transition will be required.
+
+## References ##
+
+### General
+
+- [0] Issue #11408
+
+[Issue #11408]: https://github.com/NatLabRockies/EnergyPlus/issues/11408
+
+- [1] Datacenter Liquid Cooling Market Characterization, Final Report, 12/12/2025, CalNEXT
+
+[Datacenter Liquid Cooling Market Characterization, Final Report, 12/12/2025, CalNEXT]: https://calnext.com/wp-content/uploads/2025/12/ET24SWE0065_Datacenter-Liquid-Cooling-Market-Characterization_Final-Report.pdf
+
+- [2] ACS Door Heat Exchanger Requirements for Open Rack, Fernandes et al.
+
+[Fernandes et al.]: https://www.opencompute.org/documents/acs-door-hx-open-compute-requirements-for-open-rack-rev1-0-1-pdf
+
+- [3] Rear-Door Heat Exchangers Smart Cooling at the Rack Level
+
+[Legrand]: https://www.legrand.com/datacenter/at-en/news/rear-door-heat-exchangers-rdhx-smart-cooling-at-the-rack-level
+
+- [4] Liquid to Liquid CDU Test Methodology and Performance Rating, Wondium et al.
+
+[Wondium et al.]: https://www.opencompute.org/documents/ocp-wp-l-lcdu-test-methodology-performance-rating-r1-pdf
+
+### Single-Phase D2C
+
+- [5] Universal Direct-to-Chip Cold Plates for Single- and Two-Phase Cooling, Wang et al.
+
+[Wang et al.]: https://accelsius.com/wp-content/uploads/Universal-Direct-to-Chip-Cold-Plate-2.pdf
+
+- [6] OCP OAI System Liquid Cooling Guidelines, Cheng et al.
+
+[Cheng et al.]: https://www.opencompute.org/documents/oai-system-liquid-cooling-guidelines-in-ocp-template-mar-3-2023-update-pdf
+
+- [7] Experimental validation of the effectiveness-NTU approach for single-phase liquid cold plates and a consistent definition of thermal resistance, Martinez et al.
+
+[Martinez et al.]: https://www.sciencedirect.com/science/article/pii/S0017931025004673
+
+- [8] Selecting a Liquid Cold Plate Technology
+
+[Eaton]: https://www.eaton.com/us/en-us/products/thermal-management-solutions/cold-plate-heat-exchanger/selecting-a-cold-plate-technology-and-performance-comparison.html
+
+- [9] Experimental evaluation of direct-to-chip cold plate liquid cooling for high-heat-density data centers, Heydari et al.
+
+[Heydari et al.]: https://www.sciencedirect.com/science/article/pii/S1359431123021518?casa_token=yMz4c59NaAoAAAAA:njacRyk_Tos7PSyigDx1yWpRixkV3-Vbiqv1IR7vkgFqLkSTbfue6P6pvFMKMt6bWbwFj4f9aQ
+
+### Two-Phase D2C
+
+- [10] Demonstration of CTE-Matched Two-Phase Minichannel Heat Sink, Shaeri et al.
+
+[Shaeri et al.]: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10177605&utm_source=sciencedirect_contenthosting&getft_integrator=sciencedirect_contenthosting
+
+- [11] Performance Comparison of R1233zd(E) and R515B for Two-Phase Direct-to-Chip Cooling, Wang et al.
+
+[Wang et al. 2]: https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=11235787
+
+- [12] A Practical Metric for Cold Plate Thermal Performance in Two-Phase Direct-to-Chip Cooling, Wang et al.
+
+[Wang et al. 3]: https://accelsius.com/wp-content/uploads/A-Practical-Metric-for-Cold-Plate-Thermal-Performance-in-Two-Phase-Direct-to-Chip-Cooling.pdf