Apparatus and method for coolant distribution

Abstract

Relates to a system, apparatus, and method for providing a passive coolant distribution unit (pCDU) that promotes refrigerant flow circulation and manages and monitors cold storage in a closed-loop system.

Description

Cross - reference to related applications

[0001] This application is a continuation - in - part of U.S. patent application Ser. No. 18 / 198,522, filed May 17, 2023, which claims the benefit of priority to U.S. Provisional Application No. 63 / 344,291, filed May 20, 2022, and this application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 353,224, filed Jun. 17, 2022. The entire disclosure of each of the applications listed in this paragraph is incorporated herein by specific reference.

Technical Field

[0002] The present invention relates to a passive coolant distribution unit (pCDU), and more particularly to an apparatus, device, system, and method for providing a passive coolant distribution unit (pCDU) for promoting refrigerant flow circulation and managing and monitoring refrigerant inventory within a closed - loop system.

Background Art

[0003] Anthropogenic gas emissions are shifting the Earth's energy balance, thereby causing climate change. By 2050, global warming is estimated to drive one - third of animal and plant species to extinction, and the world economy's production is inevitably expected to decline by 11 - 14%, which is equivalent to approximately $23 trillion. See Non - Patent Document 1.

[0004] To address these severe predictions and avoid irreversible economic and environmental consequences, countries have introduced carbon credits to encourage companies to reduce or capture carbon emissions.

[0005] Despite strong economic incentives to reduce carbon emissions, most data centers and communication facilities are designed with and still operate on air-cooling technology, and their energy consumption for cooling hardware equipment is on average 40% of the total energy consumed within the facility. Further, the waste heat is in the form of hot air with low energy density, which requires expensive upgrades to air treatment systems and infrastructure to utilize that heat for other purposes.

[0006] As the awareness of social responsibility to address climate change grows, there is a global consensus on actions to improve energy efficiency and reduce carbon emissions across all industries. Therefore, it is extremely important to develop new high-efficiency cooling technologies and appropriate heat recovery / reuse technologies that meet sustainability requirements and market introduction.

[0007] Moving away from conventional air-based cooling technologies, liquid / refrigerant-based cooling technologies with superior cooling performance compared to conventional air-based technologies have attracted attention. This specification discloses methods, systems, and apparatuses for optimizing a new cooling system to achieve the next level of thermal performance while reaching sustainability goals through smart energy management solutions.

[0008] Heat exchanger within a coolant distribution unit (PDU) for liquid cooling The heat exchanger is part of a liquid-based electronic equipment cooling system. In most implementation forms, the heat exchanger is enclosed together with a pump unit fluidly coupled to the heat exchanger. The pump unit drives the coolant throughout the cooling system to remove heat from the electronic equipment. Depending on the implementation form of the system, the heat exchanger can discharge heat to air or liquid.

[0009] For example, Patent Document 1 of Steinke et al., incorporated by reference, discloses a coolant distribution unit for a multi-node chassis. In this implementation, the air-coolant heat exchanger is in fluid communication between the inlet conduit and the outlet conduit, and the pump circulates the coolant throughout the system.

[0010] Similarly, Patent Document 2 of Novotny et al., incorporated by reference, discloses a coolant distribution unit comprising a liquid-liquid heat exchanger and a pump for driving the coolant within the system.

[0011] Heat exchangers for liquid cooling systems always require at least one pump to drive the coolant. However, features enabling a pump-less coolant distribution unit have not been disclosed.

[0012] IoT systems / subsystems implemented to manage infrastructure and improve overall efficiency IoT devices are emerging as a technology of choice for companies to support carbon emissions reduction by adding intelligence to optimize operations. Of increasing importance for sustainable industrial activities, the global IoT market is expected to reach $390 billion to $1.11 trillion by 2025. See Non-Patent Document 2.

[0013] In essence, an IoT device is network-enabled hardware capable of sending / receiving data via the Internet and can be incorporated into a wide range of industrial equipment, sensors, actuators, or devices to optimize industrial processes. Recent advances in cloud-based technologies and machine learning (AI / ML) are seamlessly integrated with IoT devices to perform autonomous tasks and provide intelligent analysis. Along with the advancement of 5G technology, IoT devices are leading to a paradigm shift towards a sustainable Industry 4.0 ecosystem, which encompasses overall smart manufacturing and supply chain management based on big data to achieve sustainability goals.

[0014] Liu et al. (see Non-Patent Document 3) disclosed a method for constructing a green data center using a remote sensor network (Fig. 3). The system architecture consists of a temperature control system and a cloud management platform. The temperature control system provides cooling, ventilation, environmental signal acquisition and transmission, and the cloud management platform is responsible for data storage and query, data monitoring, accessory device management, and system evaluation of the big database.

[0015] Similarly, Evans et al. disclosed a method, system, and apparatus for improving the operating efficiency within a data center by modeling data center performance and predicting power usage efficiency (see Patent Document 3 of Evans et al. incorporated by reference) (Fig. 4). The state data (140) from Fig. 4 includes temperature, power, pump speed, and set values, all of which may be accessible via IoT devices.

[0016] A group from Chengdu Qinchuan Technology Development filed a patent application (see Patent Document 4 of Zehua Shao titled Intelligent Object Network Calorimeter and Management System).

[0017] This patent application discloses a method, system, and apparatus that have a simple structure, are functionally realized based on the technology of the Internet of Things, are suitable for implementation and use in various fields, and have a wide range of applications. Essentially, this is an IoT-compatible calorimeter comprising a CPU controller, temperature, and a flow measurement device.

[0018] Patent Document 5 by Reichmuth et al., incorporated by reference, discloses a method, system, and apparatus comprising one or more energy load meters coupled to an energy supply system of a facility. This invention enables accurate modeling of past and future usage, measurement of current usage and savings, diagnosis of current problems, and implementation of adjustments when appropriate, as well as monetization of energy savings.

[0019] Patent Document 6 by Victor et al., incorporated by reference, discloses a method, system, and apparatus including a heat exchanger and one or more sensors associated with many other components in a chemical / petrochemical plant and the heat exchanger.

[0020] A plant or refinery may include equipment such as reactors, heaters, heat exchangers, regenerators, separators, etc. Types of heat exchangers include shell and tube, plate and shell, plate fin, air-cooled, wet surface air-cooled, etc. The operating method may affect the deterioration of equipment conditions, extend equipment life, extend product operating time, or provide other benefits. Mechanical or digital sensors may be used to monitor the equipment, and sensor data can be analyzed programmatically to identify problems occurring. For example, the sensors may be used in conjunction with one or more system components to detect and correct conditions such as uneven distribution, cross leakage, strain, pre-leakage, thermal stress, fouling, vibration, problems of liquid lift, conditions that can affect air-cooled heat exchangers, conditions that can affect wet surface air-cooled heat exchangers, etc. The operating conditions or modes may be adjusted to extend equipment life or avoid equipment failures.

[0021] The use of IoT devices, cloud technology, AI, and machine learning for system-level monitoring and optimization is widely seen in the published literature. For example, Patent Document 7 by Qin, incorporated by reference, discloses a simulation model based on conditions related to building infrastructure for optimizing overall energy efficiency.

[0022] None of the above prior arts disclose features that enable the construction of a coolant distribution system that supports active and passive two-phase and liquid-phase cooling systems. Also, none of the above prior arts disclose a system that provides information regarding the amount of waste heat energy captured from an ICT system by a coolant distribution system. Further, none of the prior arts disclose a method, system, and apparatus that collectively acquire data from multiple cooling systems and enable an overall understanding of the entire system. These realized functions encourage data centers and telecommunications carriers to adopt the inventions disclosed to achieve sustainability goals.

[0023] The prior art does not disclose the hardware and software design features that are important for constructing a passive coolant distribution unit (pCDU), which is extremely difficult without a pump to drive the flow circulation in a thermosiphon loop. The implementation of a passive coolant distribution unit requires a number of novel features.

[0024] Therefore, a solution to the above problems of the prior art is needed.

Prior Art Documents

Patent Documents

[0025]

Patent Document 1

Patent Document 2

Patent Document 3

Patent Document 4

Patent Document 5

Patent Document 6

Patent Document 7

Non-Patent Literature

[0026]

Non-Patent Literature 1

Non-Patent Literature 2

Non-Patent Literature 3

Summary of the Invention

Problems to be Solved by the Invention

[0027] The main object of the present invention is to provide a passive coolant distribution unit (pCDU), and in particular, to provide an apparatus, device, system, and method for promoting refrigerant flow circulation in a closed-loop system and for managing and monitoring refrigerant inventory for a passive coolant distribution unit (pCDU).

[0028] A second objective of the present invention is to provide a passive coolant distribution unit (pCDU) designed to promote refrigerant flow circulation, manage and monitor the refrigerant inventory within a closed-loop system, release heat to facility service lines such as chilled water or any other air / liquid loop, and collect data using IoT-based sensor technology and analytics.

Means for Solving the Problems

[0029] The present invention provides a promising thermal energy management approach of extending passive two-phase heat transfer to the rack scale. A thermosyphon loop (TL) combined with high-performance evaporators and condensers enables an expandable rack cooling system that operates with high thermal performance. This is achieved without the need for active pumping / control and operates within a hybrid cooling paradigm that complements the significant investments in current air-cooling technologies. This cooling technology offers quantifiable advantages including higher revenues for data center / communications carriers' operators, significant reduction in cooling energy, lower carbon footprint, higher reliability, ease of maintenance, environmentally friendly, zero-maintenance using low-cost cooling fluids, extended product life cycle, and ease of implementation in existing and new data center / communication sites, providing an important differentiator.

[0030] Furthermore, passive two-phase cooling technology using a refrigerant with zero ozone depletion potential and low global warming potential as the working fluid provides excellent key metrics (e.g., heat density, efficiency, reliability, sustainability, etc.) and represents a viable long-term solution regarding the hardware high-density and thermal performance required by next-generation telecommunications and computing environments. Also, it is extremely important that the cooling technology is very efficient, as it will address positive environmental and sustainability (ESG) requirements for wider adoption of the technology.

[0031] The two-phase cooling system can be implemented in a passive mode (e.g., a thermosiphon where gravity / buoyancy is the driving force) or in an active mode with a mechanical driver (e.g., a pump that enables fluid flow circulation). Here, the inventors disclose the features of a (semi)-passive coolant distribution unit that can be adapted to both active and passive two-phase cooling systems.

[0032] This patent application discloses a passive coolant distribution unit (pCDU) that is designed to facilitate refrigerant flow circulation, manage and monitor the refrigerant inventory within a closed-loop system, release heat to a facility service line such as chilled water or any other air / liquid loop, and collect data using IoT-based sensor technology and analytics. The data stream from the pCDU can assist an end-user in monitoring, managing, optimizing, and detecting anomalies during the operation of a thermal management system. In one implementation, the user can quantify the energy flow from the cooled hardware to the secondary heat removal side to address sustainability requirements by tracking capture efficiency, heat removal temperature, and CO2 footprint. The F-gas accounting system collects and stores data regarding the system refrigerant charge. The system interlock requires a service technician to enter the charge mass during initial filling and any subsequent refilling.

[0033] The prior art does not disclose the hardware and software design features that are important for constructing a passive coolant distribution unit (pCDU), which is extremely difficult without a pump to drive the flow circulation within the thermosiphon loop. The implementation of a passive coolant distribution unit requires a number of novel features.

[0034] Embodiments of a passive coolant distribution unit (pCDU) system include: a heat exchanger, a coolant distribution manifold, a fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit, an auxiliary positive displacement pump downstream of the fluid inventory storage unit, and a sensor selected from at least one temperature sensor and at least one pressure sensor that provides a feedback loop for maintaining the liquid level in the fluid inventory storage unit.

[0035] The system can further include a chassis for housing the pCDU system, the chassis having a removable bracket for installing the chassis adjacent to a computer server rack.

[0036] The system can further include a second coolant distribution manifold and a quick connect for adding the second coolant distribution manifold to the passive coolant distribution unit (pCDU) system.

[0037] The system can further include a pump unit for providing a thermosiphon loop within the system.

[0038] The pump unit can be fluidly installed in parallel with the main liquid coolant flow path within the system.

[0039] The system can further include a backflow prevention device that stops the pumped liquid from returning to the fluid inventory storage unit.

[0040] The fluid inventory management system can include a level sensor to ensure that the coolant level is maintained at a selected level to prevent fluid cavitation within the pump unit.

[0041] The fluid inventory management system can include a modular and stackable liquid accumulator.

[0042] The system can further include a reserve tank that is fluidly connected to the fluid inventory management system, and a valve for transporting the coolant fluid to a main accumulator within the fluid inventory management system.

[0043] The system can further include a leak detection monitor for detecting a sudden drop in temperature from a temperature sensor within the system that constitutes a major leak, and a safety isolation valve triggered by the leak detection monitor to contain the coolant leak.

[0044] The system can further include a leak detection monitor for detecting data from a chemical sensor for detecting coolant leakage from the system, the chemical sensor being selected from at least one of metal oxide, infrared, and MEMS-based sensors, and the chemical sensor being useful for detecting minor leaks, including small leaks (usually due to diffusion of coolant molecules through polymer seals) that do not immediately affect the overall performance of the system.

[0045] The system can further include a control module for actuating an isolation valve to contain the coolant leak if a major leak, including a sudden drop in temperature, is detected within the system, and an emergency alert generated from the control module to inform the system operator of the major leak detection and provide for a shutdown of the system.

[0046] The system can further include air ingress management control for separating air ingress into the system and accumulating the separated air ingress in a second fluid inventory storage unit.

[0047] The system can further include a prediction controller for estimating the CPU (Central Processing Unit) heat load based on the PDU (Power Distribution Unit), and the prediction controller is used when data from the IT equipment is not accessible.

[0048] The system can further include a pump control system that optimizes pump speed based on a predicted coolant mass flow rate derived from a machine learning model that uses an input function from at least one of the IT equipment and the PDU (power distribution unit).

[0049] The system can further include data collected from the PDU (power distribution unit) and the heat exchanger to calculate rack-level heat recovery efficiency, and data storage in a database for saving energy outside the system for the end user to monitor and account for (sustainability accounting) carbon offset credits.

[0050] The system can further include server startup detection and a heater block starter function.

[0051] The system can further include a flow stabilization method during startup (using a PCDU).

[0052] The system can further include load distribution in a heat reuse application, where the demand load varies over time to ensure sufficient heat removal from the equipment to be cooled, such as IT equipment, servers.

[0053] The system can further include a water leak detection monitor for detecting accidental water leakage inside or around the heat exchanger.

[0054] Further objects and advantages of the present invention will become apparent from the following detailed description of the presently preferred embodiments schematically shown in the accompanying drawings.

[0055] The drawings illustrate one or more implementations in accordance with this concept by way of example and not limitation. In the drawings, like reference numerals refer to the same or similar elements.

Brief Description of the Drawings

[0056]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Mode for Carrying Out the Invention

[0057] Before detailing the disclosed embodiments of the present invention, it should be understood that since the present invention allows for other embodiments, the present invention is not limited to the details of the specific configurations shown in its application. Also, the terms used in this specification are for explanatory purposes and not for limiting purposes.

[0058] In the above summary and detailed description of the preferred embodiments and the accompanying drawings, specific features of the present invention (including method steps) are referred to. It should be understood that the disclosure of the present invention in this specification does not include all possible combinations of such specific features. For example, when a specific feature is disclosed in the context of a particular aspect or embodiment of the present invention, that feature can also be used, to the extent possible, in combination with, and / or in the context of, other particular aspects and embodiments of the present invention, and throughout the present invention.

[0059] In this section, some embodiments of the present invention are described in more detail with reference to the accompanying drawings in which the preferred embodiments of the present invention are shown. However, the present invention may be embodied in many different forms and should not be construed as limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete and will convey the scope of the present invention to those skilled in the art. Like numbers refer to like elements throughout, and prime notation is used to indicate like elements in alternative embodiments.

[0060] Other technical advantages will become readily apparent to those skilled in the art after consideration of the following drawings and description.

[0061] Exemplary embodiments are shown in the drawings and described below, but it should first be understood that the principles of the present disclosure may be implemented using any number of techniques, whether currently known or not. The present disclosure should in no way be limited to the exemplary implementations and techniques shown in the drawings and described below.

[0062] Unless otherwise specified, the articles shown in the drawings are not necessarily drawn to scale.

[0063] In a typical implementation, a passive coolant distribution unit consists of multiple sub-components such as a heat exchanger with sensors for energy measurement, two or more pairs of quick connectors for fluid connection, a fluid inventory management system, an optional pump to supplement the gravity for coolant circulation for coolant storage, sensors, a microcontroller, an IoT system with firmware and network modules for data input / output via data processing and networks, and backend software to provide additional software functions such as monitoring and optimization of data centers.

[0064] A list of components will be described.

[0065] Here, a list of abbreviations in this application is defined. AI Artificial Intelligence API Application Programming Interface BMC Baseboard Management Controller BTU British Thermal Unit pCDU Passive Coolant Distribution Unit

[0066] A list of component numbers in the figures will be described below. 800 pCDU 801 Server 802 Coolant Distribution Manifold Pair of quick disconnects for coolant 803 Fluid inventory management module 804 Heat exchanger 805 Pair of quick disconnects for facility-side coolant 806 Facility-side heat removal site 807 Sensor 808 MCU / Firmware 809 Network module 810 Data center infrastructure management software 811 Optional pump 812 Server rack 900 Computer server 901 Coolant distribution unit 902 Rack 903 Rack installation 904 pCDU unit 907 1100 pCDU Multiple servers 1101 Coolant distribution unit 1102 Second coolant unit 1103 Pair of quick disconnects 1104 Pair of quick connectors for fluidly connecting the pCDU to the second coolant distribution unit 1105 1200 pCDU Coolant distribution manifold 1201 Server 1202 Pair of quick connectors 1203 Redundant pair of quick connectors 1205 Second pCDU 1205 Heat exchanger 1301 "Down" pipe 1302 Evaporator 1303 "Up" pipe 1304 Three-way valve 1305 Parallel fluid path 1306 Positive displacement pump unit 1307 Backflow preventer 1308 Down pipe 1400 Temperature sensor 1401 1402 Controller 1403 Positive Displacement Pump 1404 Branch Point 1405 Backflow Preventer 1500 Downcomer 1501 Peripheral Pump 1502 Pressure / Temperature Sensor Unit 1503 Controller Unit 1601 Heat Exchanger / Condenser 1602 Evaporator 1603 Downcomer 1604 Uptake Pipe 1605 Liquid Accumulator - Low Level 1606 High Level 1608 Optional Pump Unit 1700 Metal Plate 1701 Predetermined Hole 1702 Second Portion 1703 Internal Geometry 1704 Metal Port Defining Port Position (e.g., Tap Hole or Through Hole) 1705 Modular Liquid Accumulator 1706, 1707, 1708, 1709 Ports 1800 IT Equipment 1801 Heat Exchanger 1802 Liquid Accumulator 1803 Sensor 1804 Control Unit 1805 Power Distribution Unit 1900 Liquid Level Sensor 1901 IT Equipment 1902 Power Distribution Unit 1903 Control Unit 1904 Valve 1905 Liquid Reservoir for Storing Coolant 1906 Heat Exchanger 1907 Port 2000 Open System 2001 Temperature Sensor 2002 Chemical Sensor 2100 Large-scale Leakage Response System 2101 Control Module 2102 Isolation Valve 2103 pCDU 2104 Emergency Alert 2105 IT Equipment 2200 Liquid Coolant Reservoir 2201 Second Reservoir 2202 Downcomer 2203 Heat Exchanger 2204 Upper Access Valve 2205 Lower Valve 2300 Data Center Environment 2301 Training Dataset 2302 Collocation Data Center 2500 Heat Recovery Monitoring System 2501 Heat Exchanger 2502 Heat Metering Module 2503 PDU (Power Distribution Unit) 2504 IT Equipment 2505 Power Cable 2506 Chilled Water Inlet 2508, 2509 Temperature Sensors 2510 Flow Meter 2600 pCDU 2601 Heat Exchanger 2602 Control Module 2603 PDU 2604 Server Rack 2605 Multiple Servers 2606 Downcomer 2607 Upcomer 2608 Heating Element 2609 PDU Power Usage Monitoring 2700 Server Rack 2701 PDU (Power Distribution Unit) 2702, 2703, 2704 Three Power Banks 2705 Lowest Server Group 2706 Middle Server Group 2707 Topmost Server Group 2801 IT Equipment Rack 2802 Primary Cooling Loop 2803 Facility-Level Secondary Loop Manifold 2804 Automatic Valve 2805 Heat Reuse Application 2806 Liquid / Liquid Heat Exchanger 2807 Variable Speed Pump Driven Chiller Loop.

[0067] Figure 1 shows the overall system architecture of a passive coolant distribution unit (pCDU) 800 for the present invention. Referring to Figure 1, the pCDU 800 is fluidly connected to a coolant distribution manifold and building facilities via quick connectors.

[0068] Heat Exchanger (Condenser) 805 The main function of the heat exchanger 805 is to transfer heat from IT equipment to the facility-level cooling infrastructure. Heat exchanger technology can include plate heat exchangers, tube-in-tube heat exchangers, shell-and-tube heat exchangers, or any other type of heat exchanger.

[0069] Particular attention is required when sizing the heat exchanger 805 to ensure that the pressure drop does not become prohibitively high for passive coolant circulation.

[0070] When forming a fluid connection between the heat exchanger 805 and the flow distribution unit (i.e., the manifold) 802, it is necessary to carefully design the pressure drop across the connecting pipes / tubes in a passive thermosiphon system to ensure that excessive pressure drop does not adversely affect the flow circulation.

[0071] To ensure proper flow circulation in the desired direction, different sizes of tube / hose diameters can be used to form a fluid connection between the heat exchanger 805 and the rest of the system. The heat exchanger 805 can be equipped with additional sensors 808 such as temperature measurement probes and flow meters to measure the energy flow rate through the heat exchanger.

[0072] Referring to FIG. 1, a pair of quick connectors 803 fluidly connects the coolant distribution manifold 802 and the fluid inventory management module 804.

[0073] Filling Port and Quick Connect Ports (e.g., Schrader valves) for filling or discharging coolant can be introduced anywhere along the inner or outer tubes or hoses of the enclosure. Quick connectors for easily forming or releasing fluid connections can be installed on or outside the enclosure for easy installation. In a typical implementation, there is a pair of quick connectors for the coolant side and a second pair of quick connectors for the facility coolant side (e.g., chilled water line). Optionally, additional pairs of quick connectors can be implemented to modularly expand the cooling system.

[0074] Fluid Inventory Management System 804 A cube, rectangular, or irregularly shaped container with mechanical specifications that allow for the storage of coolant is installed inside the pCDU enclosure. In a typical implementation, the fluid inventory management system 804 is installed between the heat exchanger 805 and the coolant distribution manifold 802, and the condensed liquid accumulates in the fluid inventory storage unit rather than stagnating inside the heat exchanger 805.

[0075] Optional Pump System 812 A pump unit, such as a positive displacement pump, can optionally be added to the system inside the enclosure of the pCDU system. The pump unit is installed downstream of the liquid accumulator to ensure that the inlet side of the pump unit is always in the liquid phase. In a typical implementation, the pump unit is connected in parallel with the rest of the system, and the mechanical displacement of the coolant supplements the gravity driving force that drives the coolant circulation. Additional features may include passive / active hydraulic components to prevent reverse flow within the system. During operation, the pump unit can remain in a standby state to conserve energy, and the pump is only activated when an auxiliary driving force is promoted to improve the overall performance of the cooling system. A control loop mechanism can be implemented to further optimize the system operation.

[0076] Microcontroller (MCU) 809 and sensor 808 The microcontroller is the central component that orchestrates the data streaming process. In a typical implementation, multiple modules, such as a power module, a network module, an energy measurement module, and analog / digital circuits enabling temperature sensors and mass flow meters, are mounted on the same PCB together with the microcontroller chip.

[0077] The PCB can have multiple I / O ports for sensor heads / probes that can provide accurate measurement values, which are then processed within the microcontroller and can be exposed to the end user via a standard data transfer protocol such as HTTP API or SNMP.

[0078] In a typical implementation, the pCDU system can include a network module that enables a wired (e.g., Ethernet (R)) and / or wireless connection (e.g., WIFI or BLE) to a local / remote network. The microcontroller can be flashed with firmware that can be updated manually programmed or remotely over-the-air (OTA). One of the main functions of the firmware is to host a web server that enables an end user to provision the pCDU.

[0079] Backend software service The backend software for the pCDU can provide a system monitoring platform, which can include a GUI-based dashboard for visualizing real-time data streams. The backend database can communicate with a machine learning module to provide further analytical insights such as predictions, estimations, or forecasts, all of which can be derived based on sensor measurements from single or multiple pCDU units.

[0080] It is important to monitor the hydraulic and thermal performance of a liquid cooling system for electronic device cooling applications. In mission-critical infrastructure (i.e., data centers), real-time monitoring and anomaly detection are important to minimize system downtime and perform early maintenance. In a passive coolant distribution system (i.e., a gravity-driven thermosiphon loop), it is even more important to perform real-time remote measurements to ensure robust coolant circulation within the system without using a pump unit.

[0081] In a representative implementation, the pCDU hardware has a unique identifier such that each pCDU hardware can be a node within a local area network.

[0082] Through this local network, sensor readings, as well as other industry-standard data stream sources such as Redfish, IPMI, or SNMP, can be captured for back-end software functionality. Optionally, these data sets can be pre-processed and packaged into a JSON-formatted payload, which is wirelessly broadcast via standard IoT protocols (e.g., MQTT, AMQP, WebSocket, etc.).

[0083] The payload ultimately reaches a local or cloud server for further data consumption and analysis. Lightweight analysis can be performed on the microcontroller layer (e.g., embedded AI / ML artificial intelligence and machine learning), while more complex analysis is typically performed back-end.

[0084] The back-end software captures data streams from the pCDU as well as other IT equipment and analyzes the operating state of the data center, such as temperature or energy usage.

[0085] Software functionality based on real-time telemetry data can be integrated with other software platforms such as data center infrastructure management (DCIM) software. Additionally, the software functionality can analyze the data and update the operating parameters within the data center accordingly to maximize overall energy efficiency or any other relevant figure of merit (FoM) of the system.

[0086] Removable universal installation hardware To utilize gravity to drive the coolant circulation, a passive coolant distribution unit (pCDU) is installed on top of the IT equipment.

[0087] A chassis containing various sub-components such as heat exchangers, sensors, microcontrollers, metering modules, and quick connectors can be mounted at various locations on top of the IT equipment.

[0088] For example, the chassis can be installed inside a server rack (rack mount), on top of a server rack (top of rack), on a structural ceiling (above the rack), inside a plenum space (remote installation), or even on a building roof.

[0089] Figure 2 shows the external position of the pCDU unit 907 relative to the computer server 901. The pCDU unit 907 can be deployed in various locations. The server rack 900 has a plurality of servers 901 equipped with a coolant distribution unit 902.

[0090] In the case of a rack mount deployment, the pCDU 907 can be installed at the highest point within the rack 903. By placing the chassis outside the rack 903, an IT operator can add more computing units to maximize the computing density within the server rack 903. Also, by moving the coolant distribution unit outside the IT envelope, facility-side chilled water is removed remotely. Above the rack equipment 904, the server rack 900 is attached.

[0091] The remote pCDU 905 can be installed on a suspension ceiling structure such as a Unistrut 906 to utilize the space above the server rack 900. The rooftop pCDU 907 can be placed outside the data center building 908.

[0092] To ensure that the pCDU chassis is pre-designed to adapt to flexible deployment scenarios, screw holes and mounting accessories can be mounted at multiple positions throughout the chassis.

[0093] Figure 3 shows the removable hardware for attaching the pCDU chassis 1000.

[0094] The present invention can use removable universal hardware for the flexible deployment and installation of the pCDU (1000). The pCDU chassis 1000 can have a plurality of tapped screw holes 1001 for removable mounting hardware 1002, 1003, and 1004.

[0095] In the case of a rack-mount installation, the mounting bracket 1002 can be installed on the side of the pCDU chassis. In the case of an installation using a suspended ceiling structure, the pCDU can be suspended from a rigid structure via the bracket 1004. In the case of an installation above a rack or on the roof, the bracket 1003 can be installed.

[0096] Modular deployment IT equipment within a data center continuously undergoes refresh cycles to upgrade its computing power. A liquid-based cooling system must be flexible to accommodate these continuous changes while minimizing server downtime resulting from refrigerant recovery, filling, and reconfiguration of fluid connections.

[0097] One strategy for minimizing such downtime is to avoid the need to recover refrigerant when only sub-components such as a coolant distribution manifold need to be replaced due to an increase in power output requirements. This can be achieved by providing a second set of quick-connect couplings on the pCDU.

[0098] Figure 4 shows a layout for connecting the pCDU to a coolant distribution unit and a server.

[0099] The pCDU1100 can include a plurality of servers 1101, a coolant distribution unit 1102, and a pair of quick disconnects 1104, all of which are fluidly connected. For ease of maintenance, a second pair of quick connectors. A redundant manifold can be installed during server operation. Once all servers are connected to the new coolant distribution manifold, the old manifold can be removed. This procedure minimizes the server downtime during this manifold replacement operation.

[0100] Similarly, if the system requires a heat exchanger with a larger capacity, a second pCDU can be added via a second pair of quick connects to increase the capacity.

[0101] Figure 5 shows a layout for connecting the pCDU1200 to a coolant distribution manifold 1201 and servers 1202.

[0102] The pCDU1200 can be fluidly connected to a coolant distribution manifold 1201 and a plurality of servers 1202. A first pair of quick connectors 1203 is used to make the connection between the pCDU120 and the coolant distribution manifold 1201. A redundant pair of quick connectors 1204 can be used to add a second pCDU1205 to increase the cooling capacity in a thermosiphon loop system or to increase the internal volume for an increase in coolant storage capacity. Multiple pCDU units can be daisy chained.

[0103] Pump-assisted gravity-driven coolant flow Figure 6 shows a representative implementation of a "hybrid" active / passive pCDU system.

[0104] Referring to Figure 6, an auxiliary pump can be added to the coolant distribution unit to further enhance the coolant circulation within the system.

[0105] In a typical implementation, a liquid accumulator is installed on the outlet side of the heat exchanger, where excess liquid is stored and maintained to provide a hydrostatic pressure. This hydrostatic pressure mainly serves to drive the coolant circulation inside a thermosyphon loop-based system. By providing a liquid accumulator, entry of vapor-phase coolant, which has a negative impact, into the pump unit is also prevented.

[0106] A positive displacement pump can be installed in parallel with the outlet of the liquid accumulator, thereby ensuring that the pump always receives liquid at a substantial inlet pressure and preventing cavitation.

[0107] When the positive displacement pump is not operating, the hydraulic path through the pump is closed, and all the liquid flows through the bypass. When the pump is not operating, the coolant circulation can be driven solely by gravity in a thermosyphon-based system ("gravity-only mode"). When the pump is turned on and becomes operational, the liquid flow is driven by both gravity and mechanical force. In a typical implementation, the liquid bypass requires a backflow prevention mechanism to prevent the pumped liquid from returning to the liquid accumulator and the pump suction side. Passive components such as check valves, or active components such as three-way valves or solenoid valves, can be deployed to achieve the required functionality.

[0108] Referring to FIG. 6, the heat exchanger 1301 provides the condensed coolant in the liquid phase through the "downward" pipe 1302. The coolant flows through the evaporator 1303, and the liquid-vapor mixture enters the heat exchanger through the "upward" pipe 1304.

[0109] A branch point 1305, which can be a simple T-valve or a three-way valve, can be added to the downward pipe to form a parallel path 1306. A pump unit 1307, such as a positive displacement pump unit, can be installed in series with the parallel fluid path 1306.

[0110] When the pump unit 1307 is turned off, the parallel fluid path 1307 along 1306 is blocked, and the coolant is driven only by gravity. When the pump is activated, the coolant flows through the pump via the pump-assisted gravity-driven flow.

[0111] To prevent unwanted backflow, a backflow prevention device 1308 can be added to the system. The backflow prevention device can be an active on / off valve or a passive device such as a check valve with a low operating pressure differential. Alternatively, the branch point 1305 can be a three-way valve that can be actuated by an external trigger.

[0112] Additional control system Figure 7 shows an additional control system used for pump-assisted gravity.

[0113] To maximize energy efficiency by turning off the pump unit when not needed, an additional control system can be implemented. Also, the pump speed can be optimized based on input data from sensors. In a typical implementation, temperature and / or pressure sensors can be placed to obtain the necessary information used to build a feedback loop. For example, liquid subcooling, i.e., the temperature difference between the coolant inlet temperature and the coolant outlet temperature in the heat exchanger, can be used in the feedback to optimize the system's volumetric flow rate by controlling the pump speed. When the subcooling is small, the pump operates at a low speed.

[0114] Referring to Figure 7, a control system for pump-assisted gravity-driven coolant flow can be used. The downcomer 1400 with a parallel fluid path is equipped with a pressure and / or temperature sensor 1401.

[0115] Pressure and / or temperature data can be sent to a controller 1402 that determines the ideal pump speed of the positive displacement pump 1403 and sends a control signal. When the branch point 1404 and the backflow prevention device 1405 are implemented with operable components, the controller unit 1402 also sends control signals to these components.

[0116] The pump control system optimizes the pump speed based on a predicted refrigerant mass flow rate derived from a machine learning model using input functions from at least one of the IT equipment and the PDU (power distribution unit).

[0117] The coolant mass flow rate can be approximated using CPU power consumption as an input. The CPU power consumption can be measured or predicted.

[0118] In a typical implementation, the absolute pressure or the subcooling temperature or both can be maintained through the thermosiphon operation by adjusting the pump speed to ensure that the onset of coolant boiling is optimized. Also, a control system is utilized to ensure that the pump unit is rested as much as possible to extend the life of the pump.

[0119] FIG. 8 shows an alternative control system having a peripheral pump.

[0120] In an alternative implementation, a pump with an impeller rotating within a concentric channel can be installed in series with the downcomer. For example, a peripheral pump, also known as a regenerative turbine pump, can be implemented.

[0121] Since the regenerative turbine pump does not block the refrigerant flow when off, it does not require a parallel flow path for installing the pump unit and can simplify the system implementation. When the system operates in the "gravity mode", the pump is stopped and the coolant flows freely through the pump.

[0122] When the pump is operating, the control system evaluates whether it can turn off the pump based on system performance to maximize energy efficiency. If the thermosiphon loop can maintain strong coolant circulation without a pump, the pump is turned off to save energy and extend the pump's lifespan. Also, the pump speed can be optimized based on input data from sensors. In a typical implementation, temperature and / or pressure sensors can be placed to obtain the necessary information used to build a feedback loop. For example, the subcooling temperature can be used as a proxy to optimize the system's volumetric flow rate by controlling the pump speed.

[0123] Figure 8 shows a typical implementation of a peripheral pump for a hybrid (active / passive) CDU. The downcomer 1500 can include a peripheral pump 1501 and a pressure / temperature sensor unit 502. Pressure and / or temperature data is processed by the controller unit 1503, and a control signal is sent to the pump unit 1501.

[0124] When the system operates in "gravity mode", the pump unit is stopped, and the coolant freely flows through the pump unit via gravity. If additional static pressure is required to improve thermosiphon system performance, the pump unit can be operated to apply additional static pressure and increase the overall coolant mass flow rate within the system.

[0125] Coolant Inventory Management To ensure proper operation of the system, an optimal amount of coolant must be maintained in a passive thermosiphon system. Coolant inventory management consists of a reservoir ("liquid accumulator") placed above the evaporator to ensure that a sufficient gravity driving force proportional to the height difference is generated.

[0126] Also, the reservoir is disposed below the outlet of the heat exchanger so that the condensed coolant from the heat exchanger is properly discharged into the reservoir. When a pump unit is implemented, the liquid accumulator is disposed above the pump to ensure that the liquid always flows through the inlet port of the pump unit to prevent fluid cavitation.

[0127] FIG. 9 is a side view of the liquid inventory management system.

[0128] In a typical implementation, heat exchangers / condensers 1601 and one or more evaporators 1602 are fluidly connected to downcomer 1603 and riser 1604. When heat is applied to evaporator 1602, the liquid coolant is displaced and reaches the heat exchanger, which can lead to non-ideal operating conditions with a lower heat transfer rate.

[0129] To prevent the condenser from "flooding", a component with a significant internal volume, namely a "liquid accumulator" 1605, can be installed in series with downcomer 1603. To ensure proper discharge of the condensed liquid into the liquid accumulator, the liquid accumulator can be positioned lower than the outlet port of the heat exchanger, but maintaining a sufficient distance between evaporator 1602 and liquid accumulator 1605 and having sufficient static pressure to drive the coolant out of the thermosyphon system.

[0130] Throughout the thermosyphon loop operation, the liquid level can vary from a low level 1605 under low heat load to a high level 1606 under high heat load. The liquid accumulator is sized appropriately so that the highest liquid level does not reach the heat exchanger. An optional pump unit 1608 can be installed below the liquid accumulator to ensure that the coolant entering the pump unit is always in the liquid phase.

[0131] As the height increases (rack unit), valuable rack space that could otherwise be used to add more computing resources is occupied, so rack-mounted devices must always be optimized to minimize the overall height. Therefore, it is desirable to have a liquid accumulator with an overall shape that resembles a "pizza box" with minimal height while ensuring that sufficient internal volume is provided for storing the coolant. Depending on the implementation form of the entire system, the required internal volume can vary, and thus, a modular liquid accumulator can be envisioned to minimize manufacturing costs. For example, a modular liquid accumulator can be constructed by brazing or welding components pre-defined with connector ports.

[0132] Figure 10 shows a modular fluid inventory system. As the height increases (rack unit), valuable rack space that could otherwise be used to add more computing resources is occupied, so rack-mounted devices must always be optimized to minimize the overall height. Therefore, it is desirable to have a liquid accumulator with an overall shape that resembles a "pizza box" with minimal height while ensuring that sufficient internal volume is provided for storing the coolant. Depending on the implementation form of the entire system, the required internal volume can vary, and thus, a modular liquid accumulator can be envisioned to minimize manufacturing costs. For example, a modular liquid accumulator can be constructed by brazing or welding components pre-defined with connector ports.

[0133] Referring to Figure 10, a metal plate 1700 having pre-defined holes 1701 can be brazed with a second component 1702 having recesses or cutouts that define an internal geometry 1703. When these two components are assembled with a second metal plate (1704) with varying numbers of repetitions, a modular liquid accumulator 1705 with four ports 1706, 1707, 1708, and 1709 can be constructed.

[0134] In a typical implementation, one of the upper ports 1706 can be fluidly connected to the outlet port of the heat exchanger, and the second upper port 1707 can be fluidly connected to the inlet port of the heat exchanger via a capillary tube.

[0135] The rising pressure from the liquid-vapor mixture can act as a pressure source to ensure that the liquid coolant is properly flushed during thermosiphon operation. If a single fluid distribution manifold is provided, the bottom port 1708 can be fluidly connected to a downcomer pipe, and the second bottom port 1709 can be capped.

[0136] If two fluid distribution manifolds are implemented, the second bottom port 1709 can be fluidly connected to the second manifold. Alternatively, a single component with ports and recessed surfaces can be used to further simplify the manufacturing process.

[0137] In a typical thermosiphon-based electronic device cooling application, a non-hermetic seal is introduced into the system, which can potentially lead to minor coolant losses over a long period of time. Therefore, it would be advantageous to be able to monitor the coolant inventory within the system.

[0138] Minor coolant leakage (minor coolant loss) refers to minor coolant leakage due to molecular diffusion of the coolant across non-metallic seals and materials over a long period of time.

[0139] To monitor the amount of coolant in the reservoir, a liquid level sensor such as a capacitance sensor can be implemented within the liquid accumulator. Since the level of coolant in the reservoir tank varies dynamically depending on the heat load, the ideal level can be determined for different heat loads to the system. The amount of heat load can be determined directly or indirectly from the IT equipment via a baseboard management controller (BMC) or a power distribution unit (PDU). If the liquid level falls below the ideal level, coolant needs to be added to the system.

[0140] Figure 11 shows a fluid inventory monitoring system for the present invention. In a typical implementation, heat from IT device 1800 generates liquid vapor, and the liquid vapor is condensed by heat exchanger 1801. This displaces the coolant within liquid accumulator 1802. Throughout system operation, the liquid level within the liquid accumulator can be monitored by sensor 1803.

[0141] To determine whether the liquid level is properly maintained, additional information such as the amount of heat load from the IT device should also be monitored. Controller unit 1804 collects data from IT device 1800 or power distribution unit 1805 and compares the actual measured values to the expected ideal levels. If the liquid level falls below a predetermined safe threshold level, appropriate maintenance protocols must be executed.

[0142] Figure 12 shows a fluid inventory management system having a reserve tank for easy and simple maintenance. To fully rationalize the coolant filling protocol as part of the system maintenance protocol, an additional reserve tank can be added as a backup. This second tank can be fluidly connected to the main liquid accumulator using a solenoid valve and / or a pump that can convey coolant from the reserve tank to the main liquid accumulator. If a pump is installed between the two tanks, the pump can be used to remove coolant from the system when system requirements change and a smaller amount of coolant within the system is required.

[0143] Referring to Figure 12, when data from liquid level sensor 1900, IT device 1901, and / or power distribution unit 1902 indicates a low liquid level, controller unit 1903 can activate valve 1904, such as a normally closed solenoid valve, that fluidly connects liquid reservoir 1905 to the rest of the system.

[0144] The reservoir is placed at the highest position above the heat exchanger 1906 to ensure that when the valve 1904 is opened, the liquid is properly discharged into the system. The port 1907 is provided in the reservoir tank for filling or discharging. Instead of a fully automatic maintenance approach, a manual isolation valve can be implemented instead.

[0145] The coolant can leak into the atmosphere from the non-sealed thermosiphon loop over time. This results in a lower fluid filling ratio and thus a degradation of thermal performance. Therefore, it is important to monitor any leaks within the system. As part of the leak monitoring system, multiple sensors can be installed throughout the system.

[0146] In a typical implementation, live data streams from temperature or chemical sensors (e.g., metal oxide, infrared, or MEMS-based sensors) can be monitored during operation.

[0147] The system can use a refrigerant leak detection system such as the system shown and described in Rinehart's U.S. Patent No. 6,772,598, which is incorporated by reference.

[0148] Chemical sensors can be used to detect anomalies within the system. Although the probability is extremely low, in the case of a catastrophic failure (i.e., pipe rupture), the liquid-phase coolant will rapidly leak out and begin to evaporate. This will result in a rapid drop in temperature. By monitoring the temperature data collected throughout the system, this behavior can be captured and accidental coolant leaks can be detected.

[0149] The refrigerant-based cooling system needs to be designed to always contain the coolant during operation. However, due to mechanical defects such as cracks or polymer seal failures, the coolant may be lost from the system.

[0150] Anomaly detection algorithms such as isolation forest, local outlier factor, robust covariance, support vector, or other machine learning methods can be applied to the data to identify abnormal behavior in the system. If any anomaly is detected, such as a sudden drop in temperature related to the rapid evaporation of the coolant to the surroundings, a maintenance alert is triggered and the system operator is notified.

[0151] Figure 13 shows a leakage detection system based on a live data stream from sensors. The non - sealed system 2000 has sections within the system that are "potential leakage points" where coolant leakage is more likely. Temperature sensors 2001 and / or chemical sensors 2002 can be installed at or around these locations where coolant leakage is more likely. The monitoring module 2003 flags system anomalies when leakage is detected, along with an anomaly detection analysis method.

[0152] If any of the temperature sensors detect a sudden drop in temperature due to the rapid evaporation of the coolant to the surroundings, this indicates a major leak, and the safety isolation valve activates to prevent the coolant from leaking to the surroundings. At the same time, the IT equipment undergoes a gentle shutdown sequence.

[0153] A major leak can include a large rupture or polymer seal failure that forces the system to shut down. A major coolant leak leads to a sudden temperature drop and can be used as a proxy for detecting such an accident.

[0154] A minor leak includes refrigerant diffusion through the polymer seal and can potentially lead to a slow but steady loss of refrigerant. Despite a minor leak, the cooling system functions as expected until the amount of coolant in the system drops below the minimum required amount.

[0155] The use of certain flammable coolants requires a chemical sensor (leak detector) by law for safety reasons.

[0156] Figure 14 shows the critical leakage response system of the pCDU. When a significant coolant leakage is detected from system 2100, the control module 2101 activates the isolation valve (2102) that isolates the pCDU 2103 where most of the fluid is stored. While an emergency alert 2104 is notified to the system operator, the IT equipment 2105 undergoes a gentle shutdown including the redistribution of the workload to prevent data loss and service interruption.

[0157] The system can use a water leakage detection monitor to detect accidental water leakage inside or around the heat exchanger. See the US Patent No. 4,835,522 of Andrejasich et al. incorporated by reference.

[0158] Research has shown that the diffusion of ambient air into the system can have an adverse effect on the overall cooling performance. For example, due to the intrusion of air, an increase in thermal resistance or a decrease in mass flow rate can occur within a thermosyphon loop (TSL) for a given heat flux. Due to the non-hermetic nature of the thermosyphon loop, air can diffuse into the system through potential leakage paths such as elastomeric seals within the pipe joints.

[0159] Based on the real-time data stream, the software can analyze, monitor, and detect the intrusion of air into the coolant loop. Anomaly detection algorithms such as isolation forest, local outlier factor, robust covariance, support vector, or other machine learning methods can be applied to the data.

[0160] To remove the air diffused from the system, a small component with a defined internal volume can be placed above or near the top of the TSL that can function as an air trap. This component can selectively accumulate air on the refrigerant vapor. Non-intrusive sensors such as ultrasonic Time of Flight sensors or temperature sensors can be equipped around the air trap to evaluate and monitor the amount of trapped air. As part of the maintenance procedure, the collected air can be removed by a discharge mechanism.

[0161] Figure 15 is a schematic diagram of an air ingress management system for a pCDU. When air enters the system, due to its low density and non-condensable nature, the air can be separated from the condensed coolant from the coolant reservoir 2200, connected to the downcomer 2202, and accumulated in a second reservoir 2201 located at an appropriate height, for example, above the heat exchanger 2203.

[0162] During regular maintenance, the air reservoir can be exhausted via the upper access valve 2204 with the lower valve 2205 closed to prevent loss of the working fluid.

[0163] Features related to CPU / GPU heat load prediction and analysis Recent IT equipment supports detailed real-time telemetry. For example, the latest server motherboards are equipped with a BMC (Baseboard Management Controller) that publishes various data related to computing hardware such as CPU temperature and fan speed for consumption through industry-standard data transfer protocols such as IPMI and Redfish.

[0164] Real-time telemetry data provides essential information for data center operators to plan facility upgrades or optimizations. However, in colocation facilities where access to data from IT equipment is not feasible, the insights derived from such data are lost.

[0165] When access to IT equipment is restricted, the pCDU also has problems and cannot obtain real-time data, perform predictive analysis, and provide the end user with operational insights related to the performance of the cooling system. For example, using the heat load and temperature measurements from the CPU / GPU, the coolant mass flow rate can be predicted without using expensive flow measurement devices.

[0166] The coolant mass flow rate can be approximated using the CPU power consumption as an input. The CPU power consumption can be measured or predicted.

[0167] In the absence of heat load information, the CPU / GPU heat load can be predicted using a machine learning-based model that uses some input features such as CPU / GPU specifications and data related to detailed power consumption, which can be obtained from a "smart" PDU. The machine learning model can be deployed on the pCDU installed in the co-location data center to provide operational insights despite limited access to data from IT equipment.

[0168] The system can use a prediction model such as the model shown and described in U.S. Patent No. 10,581,974 to Sustaeta et al., which is incorporated by reference.

[0169] A prediction controller can be used to estimate the CPU (Central Processing Unit) heat load based on the PDU (Power Distribution Unit), and the prediction controller is utilized when data from IT equipment is not accessible. A machine learning model can be used to predict the CPU power consumption using the time-series data from the PDU as an input.

[0170] The machine learning model correlates the real-time power consumption from individual servers to the CPU power consumption.

[0171] Figure 16 shows a schematic diagram of the prediction model development for the pCDU when access to data from IT equipment is not achievable.

[0172] In a data center environment 2300 where access to real-time telemetry data from IT devices is possible, a training data set 2301 can be easily obtained. In an environment where data access is restricted, such as a colocation data center 2302, a prediction model can be deployed to predict data related to IT hardware.

[0173] A colocation data center is a facility where IT equipment is owned by individual tenants and the data center provides only cooling / power as a service. Since the IT equipment is owned by the tenant, the data center operator has no operational visibility into what is going on inside the IT equipment.

[0174] As a result, the predicted CPU / GPU heat load or the actual CPU / GPU heat load based on telemetry data can be further utilized to derive the coolant mass flow rate or the quality of the vapor within the thermosiphon loop. Using these parameters, it is possible to determine whether to turn on an auxiliary pump to switch from a gravity-driven coolant flow to a pump-assisted gravity-driven coolant flow to ensure that an appropriate amount of coolant is circulating to cool the computer chip.

[0175] FIG. 17 is a schematic diagram of the conditional pump operation process flow of the pCDU. Based on the CPU heat load from a prediction model or direct measurement, the coolant mass flow rate and the quality of the vapor can be derived. Using these operating parameters, the pump within the pCDU can be activated and the pCDU can be switched to operate from a gravity-driven coolant flow (passive) to a pump-assisted gravity-driven flow (active).

[0176] Function related to monitoring of sustainability indicators Unlike an air-cooled based system, a liquid / refrigerant cooling system enables an accurate measurement of the amount of waste heat captured in warm water.

[0177] The pCDU is equipped with an industry-standard energy metering unit (BTU meter), which can not only provide the accurate amount of heat captured in the facility-side coolant loop, but also provide waste heat recovery efficiency based on the heat load to IT equipment via the BMC or PDU.

[0178] For example, the amount of heat captured can be calculated based on a simple energy balance calculation, and the input energy can be extracted from the CPU / GPU heat load or derived from the PDU. The heat recovery efficiency can be monitored over time to improve the operating energy efficiency. Additionally, a significant decrease in the heat recovery efficiency inside the heat exchanger can be used as a proxy to detect heat exchanger fouling.

[0179] Figure 18 shows a heat recovery monitoring system 2500 for the pCDU. The pCDU includes a heat exchanger 2501 and a heat metering module 2502. In a typical system implementation, the PDU 2503 provides power to the IT equipment 2504 through a power transmission line, i.e., a power cable 2505.

[0180] The heat generated from the IT equipment is transferred to the facility-side refrigerant loop inside the heat exchanger 2501. For example, cold water enters at 2506 and warm water exits the pCDU (2505). The temperature sensors 2508, 2509 and the flow meter 2510 in the BTU meter can calculate the amount of heat transferred to the facility-side refrigerant loop using the energy balance equation.

[0181] The input energy to the heat exchanger can be directly extracted from the heat load from the CPU / GPU or derived from the PDU. Monitoring the heat recovery efficiency can assist data center operators in accessing quantitative information to improve the overall operating energy efficiency. Additionally, the heat recovery efficiency can be used as a proxy to detect heat exchanger fouling.

[0182] Sustainability-related data streams such as heat recovery efficiency can be linked to an auditable or immutable database, which is an append-only database that cannot be deleted or modified. In combination with a standard-based metering system from pCDU hardware, the database can be managed, authenticated, and verified by a third party.

[0183] End-users of the system can develop carbon offset projects based on this end-to-end platform to generate permanent additional verifiable actual carbon offsets. Carbon offsets can be monetized in the form of carbon tax credits. A SaaS (Software-as-a-Service) platform or a license-based subscription model can orchestrate the end-to-end process from data generation to the submission of rationalized carbon credit claims. Quantification of heat collection and energy recovery from computer racks / communication equipment cabinets for carbon tax credit claims.

[0184] Using data collected from PDU (Power Distribution Unit) and heat exchangers, the rack-level heat recovery efficiency can be calculated.

[0185] Also, for end-users to monitor and account for carbon offset credits, data is stored in a database for storing energy outside the system (sustainability accounting). Archiving energy efficiency and carbon offset information can use the techniques employed in U.S. Patent No. 8,000,938 to McConnell et al., which is incorporated by reference.

[0186] Function for improving the thermosiphon loop startup operation To overcome issues related to flow instability during cold start of a two-phase thermosiphon loop (e.g., new deployment or rack-level maintenance), the startup process may include an active pump to establish a robust flow. However, in the case of a gravity-driven passive thermosiphon loop without a pump, it can be very difficult to establish a robust flow under low heat.

[0187] In a typical thermosiphon loop, the mass flow rate of the low-boiling refrigerant tends to increase as the amount of heat introduced into the system increases. When a server rack equipped with a two-phase thermosiphon loop-based cooling system is first turned on, the heat from the IT equipment during cold start may not be substantially sufficient to reach the ideal conditions for two-phase flow. Additionally, there is a possibility of temperature overshoot before a robust coolant circulation is established. This issue is partly due to the lack of control of IT equipment power, which is typically defined independently at the server level.

[0188] To mitigate issues related to the cold start-up scenario, it would be advantageous to establish circulation within the thermosiphon loop in a manner that is controlled independently of the heat output supplied by the server. Resistance heating elements (localized or distributed) located on the riser pipe and controlled by the pCDU can provide an accurate amount of heat to the thermosiphon loop during system operation, but are primarily envisioned to be utilized during initial startup. When the IT equipment turns on and begins to provide substantial power exceeding a threshold heat amount, the additional heating elements can be automatically turned off. It is worth noting that idle IT equipment still consumes a significant amount of power far exceeding the heat amount required to establish two-phase flow, and thus, the resistance elements consume power only during initial startup or maintenance.

[0189] The power supply to the heater is controlled by the pCDU, for example, using a solid-state relay packaged in the pCDU and powered by AC or DC power supplied from a rack PDU. During startup, heat input to the riser tube initiates stable recirculation of the two-phase refrigerant mixture in the thermosiphon loop, while avoiding uncontrolled heat sources, such as unstable flow patterns that are not controlled in the presence of servers / IT hardware.

[0190] A control system, such as the pCDU, detects a robust two-phase flow, and when the server is operating, the heater automatically turns off, improving the energy efficiency of the system. Specific examples of implementations are shown in the following figures. Here, the starter heater is located at the bottom of the riser tube where the static pressure is maximum.

[0191] Figure 19 is a schematic diagram of features for improving the thermosiphon loop startup operation of the pCDU. The pCDU 2600 has a heat exchanger 2601 and a control module 2602. The PDU 2603 supplies power to a server rack 2604 having a plurality of servers 2605.

[0192] Each server can be fluidly connected to the heat exchanger via a downcomer 2606 and a riser tube 2607 that form a thermosiphon loop. A heating element 2608 can be installed at the bottom of the riser tube. When the server is first turned on with little heat load on the thermosiphon loop, flow instability can occur.

[0193] Monitoring 2609 the PDU power consumption can detect the moment the server is first turned on. The control module 2602 can activate the heating element 2608 through a control signal 2610 to provide sufficient coolant driving force by generating a liquid-vapor mixture in the riser tube column. Once the system has established a robust coolant circulation, the heater can be turned off.

[0194] Furthermore, by precisely controlling the startup sequence of the servers within the rack, it would be advantageous to establish circulation within the thermosiphon loop in a controlled manner. In a thermosiphon loop-based cooling system, the bottommost server has the greatest static pressure head to assist in maintaining a stable flow within the thermosiphon loop and is thus relatively more advantageous than any of the upper servers for initiating two-phase flow.

[0195] As the mass flow rate slowly increases during initial startup, the microcontroller within the pCDU communicates with the power distribution unit to adjust the startup process, powering on the servers starting from the bottom and proceeding in order to the upper levels, while taking in the threshold mass flow rate as an input for accurately triggering the sequence of events.

[0196] Once a stable coolant flow is established, the overall system becomes resilient to external perturbations such as fluctuations in heat flux caused by changes in computational demands. This process can be carried out with or without using data center infrastructure management (DCIM) software or data center management (DCM) software.

[0197] The resistive heating element (2608) provides thermal energy to circulate the coolant within the thermosiphon loop.

[0198] The system can use server startup detection and heater block starter functionality. When the server rack is powered on, the computer servers begin consuming power from the PDU (power distribution unit). This initial "startup" can be detected simply by monitoring PDU usage. The control module can detect the "startup" and activate and turn on the heater block. The heater block supplies thermal energy to the thermosiphon loop to generate coolant circulation. When the control module detects substantial coolant flow circulation, the heater block is turned off since the thermal energy from the computer chips provides sufficient heat to maintain the flow.

[0199] Figure 20 is a schematic diagram of the server rack startup sequence of the pCDU. The startup sequence can be adjusted by starting the servers sequentially from bottom to top. For example, when the server from server rack 2700 is turned on, the PDU 2701 including three power banks 2702, 2703, and 2704 can provide power to the bottom server group 2705 by starting the power bank 2702.

[0200] The bottom server group 2705 with the maximum static pressure head can help initiate a stable annular two-phase flow in the thermosiphon loop.

[0201] As the mass flow rate increases and the passive two-phase thermosiphon loop becomes stronger, the other power bank 2703 can be turned on to supply power to the middle server group 2706. Finally, the last server group 2707 can be turned on.

[0202] The system can use a flow stabilization method (using the PDU) during startup.

[0203] Furthermore, it would be advantageous to establish circulation within the thermosiphon loop in a controlled manner by precisely controlling the startup sequence of the servers within the rack. In a thermosiphon loop-based cooling system, the bottommost server has the greatest static pressure head to assist in establishing a stable flow within the thermosiphon loop and is thus relatively more advantageous than any of the upper servers for initiating two-phase flow. As the mass flow rate slowly increases during initial startup, the microcontroller within the pCDU communicates with the power distribution unit to adjust the startup process, powering on the servers starting from the bottom and proceeding sequentially to the upper levels while taking in the threshold mass flow rate as an input for precisely triggering the sequence of events. Once a stable coolant flow is established, the overall system becomes resilient to external perturbations such as fluctuations in heat flux caused by changes in computational demands. This process can be carried out with or without using data center infrastructure management (DCIM) software or data center management (DCM) software.

[0204] System optimization by secondary side heat load balancing In heat reuse applications, the heat load from IT equipment and the heat reuse requirements do not always match well. For example, discharging all of the heat from IT equipment to space heating applications may only be feasible under certain environmental conditions, such as in winter rather than summer.

[0205] However, provisioning IT equipment based on baseline heat requirements can be overly restrictive for provisioning. Furthermore, heat reuse applications may require a well-controlled supply temperature.

[0206] To mitigate this problem, it is advantageous to control the secondary loop working fluid supply temperature and the secondary side fluid flow rate through the condenser. The supply temperature from the heat reuse application to the IT equipment rack is set to a maximum value at the design stage.

[0207] However, when the heat reuse load is low, the supply temperature may exceed the specified maximum value. By providing a centralized liquid / liquid heat exchanger, heat can be diverted to the facility loop (e.g., dry cooler or chiller) to maintain the supply temperature within specifications. At the same time, server utilization, and thus the heat load, in the rack unit can vary over time.

[0208] To ensure a designated return temperature designed for the heat reuse application, a fail-open automatic valve controlled by the pCDU is provided at the inlet to the secondary condenser to reduce the flow rate through the condenser when server utilization, and thus the heat load, is low.

[0209] Figure 21 is a schematic diagram of system optimization by heat load balancing on the secondary side of the pCDU.

[0210] The IT equipment rack 2801 incorporates a primary cooling loop 2802 that interfaces with the facility-level secondary loop manifold 2803. An automatic valve 2804 controlled by the pCDU adjusts the flow of the secondary flow through the condenser to ensure that the temperature exiting the condenser on the secondary side matches the requirements of the heat reuse application 2805.

[0211] The liquid / liquid heat exchanger 2806 is provided to provide supplementary heat rejection to a variable speed pump dry cooler loop 2807 that has sufficient capacity to reject the designed IT heat load, which is controlled to achieve a specific secondary side supply temperature to the IT equipment rack, for example.

[0212] Certain advantages are listed above, but various embodiments may include some, none, or all of the listed advantages.

[0213] Without departing from the scope of the present disclosure, modifications, additions, or omissions may be made to the systems, apparatuses, and methods described herein. For example, the components of the systems and apparatuses may be integrated or separated. Further, the operations of the systems and apparatuses disclosed herein may be performed by more, fewer, or other components, and the methods described may include more, fewer, or other steps. Further, the steps may be performed in any suitable order. As used herein, "each" refers to each member of a set or each member of a subset of a set.

[0214] To assist any reader of the patent office and any patent issued on this application in interpreting the claims appended hereto, Applicant notes that unless the terms "means for" or "step for" are expressly used in a particular claim, Applicant does not intend for any of the appended claims or elements of those claims to be construed under 35 U.S.C. 112(f).

[0215] The term "about" may be + / −10% of the referenced amount. Further, preferred amounts and ranges can include amounts and ranges referenced without the prefix about.

[0216] Although the invention has been described, disclosed, illustrated, and shown in various terms of certain embodiments or modifications which are actually contemplated, the scope of the invention is not intended to be thereby limited nor should it be construed as being limited, and other modifications or embodiments as may be suggested by the teachings herein are reserved especially if they are within the breadth and scope of the claims appended hereto.

Claims

1. A passive coolant distribution unit (pCDU) system, heat exchanger; Coolant distribution manifold; A fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit; Auxiliary positive displacement pump downstream of the fluid storage unit; A sensor selected from at least one temperature sensor and at least one pressure sensor, which provides a feedback loop for maintaining the liquid level in the fluid storage unit; and Leak detection monitor for detecting leaks within the aforementioned system A passive coolant distribution unit (pCDU) system equipped with the following features.

2. The system further comprises a chassis for housing the pCDU system, the chassis having removable brackets for mounting the chassis adjacent to a computer server rack. A passive coolant distribution unit (pCDU) system according to claim 1, characterized in that...

3. Second coolant distribution manifold; and A quick connect for adding the second coolant distribution manifold to the passive coolant distribution unit (pCDU) system. The passive coolant distribution unit (pCDU) system according to claim 1, further comprising the above.

4. The passive coolant distribution unit (pCDU) system according to claim 1, further comprising a pump unit for providing a thermosiphon loop within the system.

5. The passive coolant distribution unit (pCDU) system according to claim 4, characterized in that the pump unit is fluidly installed in parallel with the main coolant flow path in the system.

6. The passive coolant distribution unit (pCDU) system according to claim 4, further comprising a backflow prevention device that prevents the pumped liquid from returning to the fluid inventory storage unit.

7. The aforementioned fluid inventory management system is A level sensor ensures that the coolant level is maintained at a selected level in order to prevent fluid cavitation within the pump unit. The passive coolant distribution unit (pCDU) system according to claim 4, characterized by comprising the above.

8. The passive coolant distribution unit (pCDU) system according to claim 1, characterized in that the fluid inventory management system comprises a modular and stackable liquid accumulator.

9. A reserve tank fluidly connected to the fluid inventory management system; and A valve for transporting the coolant fluid to the main accumulator in the fluid inventory management system. The passive coolant distribution unit (pCDU) system according to claim 1, further comprising the above.

10. The leak detection monitor is for detecting a rapid drop in temperature from a temperature sensor in the system that constitutes a serious leak, A safety isolation valve, triggered by the leak detection monitor, contains the coolant leak. The passive coolant distribution unit (pCDU) system according to claim 1, further comprising the above.

11. The leak detection monitor is for detecting data from a chemical sensor for detecting coolant leakage from the system, the chemical sensor being selected from at least one of metal oxide, infrared, and MEMS-based sensors, and the chemical sensor is useful for detecting minor leaks, including small leaks (mostly due to the diffusion of coolant molecules through a polymer seal) that do not immediately affect the performance of the entire system. A passive coolant distribution unit (pCDU) system according to claim 1, characterized in that...

12. The leak detection monitor, A control module for activating an isolation valve to contain coolant leakage if a significant leak, including a rapid drop in temperature, is detected within the system; and An emergency alert generated from the control module to notify the system operator of the aforementioned critical leak detection and provide a system shutdown. The passive coolant distribution unit (pCDU) system according to claim 1, characterized by comprising the above.

13. A passive coolant distribution unit (pCDU) system, heat exchanger; Coolant distribution manifold; A fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit; Auxiliary positive displacement pump downstream of the fluid storage unit; A sensor selected from at least one temperature sensor and at least one pressure sensor, which provides a feedback loop for maintaining the liquid level in the fluid storage unit; and Air ingress management control for separating air ingress into the aforementioned system and accumulating the separated air ingress in a second fluid inventory storage unit. A passive coolant distribution unit (pCDU) system characterized by comprising the following:

14. A passive coolant distribution unit (pCDU) system, heat exchanger; Coolant distribution manifold; A fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit; Auxiliary positive displacement pump downstream of the fluid storage unit; A sensor selected from at least one temperature sensor and at least one pressure sensor, which provides a feedback loop for maintaining the liquid level in the fluid storage unit; and A predictive controller that estimates the CPU (Central Processing Unit) thermal load based on a PDU (Power Distribution Unit), wherein the predictive controller is used when data from IT equipment is inaccessible. A passive coolant distribution unit (pCDU) system characterized by comprising the following:

15. A passive coolant distribution unit (pCDU) system, heat exchanger; Coolant distribution manifold; A fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit; Auxiliary positive displacement pump downstream of the fluid storage unit; A sensor selected from at least one temperature sensor and at least one pressure sensor, which provides a feedback loop for maintaining the liquid level in the fluid storage unit; and A pump control system that optimizes pump speed based on a predicted coolant mass flow rate derived from a machine learning model that uses input from at least one of the following: IT equipment and / or a PDU (power distribution unit). A passive coolant distribution unit (pCDU) system characterized by comprising the following:

16. A passive coolant distribution unit (pCDU) system, heat exchanger; Coolant distribution manifold; A fluid inventory management system between the heat exchanger and the coolant distribution manifold for accumulating condensed liquid in a fluid inventory storage unit; Auxiliary positive displacement pump downstream of the fluid storage unit; A sensor selected from at least one temperature sensor and at least one pressure sensor, which provides a feedback loop for maintaining the liquid level in the fluid storage unit; Data collected from the PDU (Power Distribution Unit) and the heat exchanger to calculate rack-level heat recovery efficiency; and Storage of data in a database for storing energy outside the system so that end users can monitor and account for carbon offset credits (sustainability accounting) A passive coolant distribution unit (pCDU) system characterized by comprising the following:

17. The passive coolant distribution unit (pCDU) system according to claim 1, further characterized by comprising server startup detection and heater block starter functions.

18. The passive coolant distribution unit (pCDU) system according to claim 1, further comprising a method for stabilizing the flow during startup (using a PCDU).

19. The passive coolant distribution unit (pCDU) system according to claim 1, further characterized by load balancing in heat reuse applications, wherein the demand load fluctuates over time and sufficient heat removal is ensured from the equipment to be cooled, such as IT equipment and servers.

20. The leak detection monitor comprises at least one liquid detection sensor installed in the system, wherein a single sensor generates an output signal proportional to the severity of the leak, or multiple sensors jointly output signals and estimate the severity of the leak by analysis thereof, and use the signals to warn of an abnormal situation and to provide an estimated maximum response time required to deal with the leak. A passive coolant distribution unit (pCDU) system according to claim 1, characterized in that...

Images