With the proliferation of Internet of Things (IoT) devices and real-time applications, edge computing has emerged as a prominent trend. As a result, smaller and distributed data centers have been developed at the network edge.
Over time, data centers have experienced significant changes in size, design, efficiency, and management. Initially, data centers were confined to small rooms or dedicated spaces within office buildings. However, as data storage and processing demand grew, data centers expanded into massive facilities spanning thousands or even millions of square feet. Some of today's data centers are so large that they are commonly referred to as "hyperscale" data centers.
Data centers consume a significant amount of energy to power and cool the IT equipment. To address this, there has been a focus on improving energy efficiency with
more efficient cooling sources like free cooling chillers, and more energy-efficient server hardware capable of operating at higher temperatures.
Overview of a data center cooling system
The schematics above describe a typical HVAC plant for a data center, and introduce the main elements of Daikin’s solution, including several components:
- Air-to-water chiller
- Computer Room Air Handling (CRAH) unit
- Control systems
- After-sales services
In this article, we will discuss the main challenges in choosing control solutions for data centers and how Daikin can support overcoming those challenges.
Challenge #1: How to ensure cooling system proper and continuous operation
In a data center HVAC system, effective control of multiple units, including CRAHs and chillers is critical to ensure the proper running of the cooling system. Monitoring and controlling multiple units allow for the quick detection and resolution of malfunctions. This helps to reduce the risk of operational disruption due to cooling system issues.
Reason #1: To overcome this challenge, Daikin offers an intelligent Data Center Manager (iDCM), an external panel that controls up to 20 units, including Primary pump management and free cooling mode. iDCM is designed to ensure maximum reliability and leverages an advanced machine learning algorithm, working based on manufacturer data and live data recorded on site.
iDCM identifies the status of the units, and defines actions according to the control logic, while also calculating the capacity of the chiller and estimating the flow delivered.
The chiller group controller switches on or off chillers based on specific conditions depending on the supply and return water temperatures in both sections of the primary water loop. We have two stages:
Stage-Up: occurs when the controller decides to switch on more chillers. The conditions for activation include the primary supply water temperature being above the supply water temperature setpoint, plus a certain margin called the "Stage up Offset." This margin can be user-configured and ensures that the chillers only activate when necessary, avoiding system overload.
Stage-Down: is when the controller decides to switch off the chillers. The conditions for deactivation include the primary supply water temperature being below the supply water temperature setpoint, minus a certain margin called the "Stage Down Offset". The user configures this margin and ensures that the chillers are switched off when they are no longer needed, saving energy.
Challenge #2: How to ensure compatibility of control with chosen plant design
A control system should be compatible with different projects designs to be flexible and tailored to the customer’s needs. This approach allows systems to be designed and implemented specifically for each project, ensuring an appropriate response to the unique requirements of each situation.
Reason #2: Daikin offers various equipment solutions specifically designed for data centers, and carefully tailored to each customer's needs. IDCM can integrate many different plant devices such as sensors, water pumps, hydronic valves, depending on circuit type and overall capacity of the plant. Furthermore, building on this customised approach, Daikin provides a comprehensive customer support solution, starting from the solution design and extending throughout the sales process. These involve adapting and managing the system to meet the customer's evolving requirements. In conclusion, Daikin provides a control system that offers unparalleled flexibility and versatility.
Daikin control system can be integrated with the Building Management System (BMS), which enables high-level monitoring and centralised control of the entire data center infrastructure, ensuring efficient management and operation of the facility. With the BMS protocol, the system can be fully interoperable with the customer's BMS IT management logic.
Among the range of features that this Intelligent Data Center Manager (iDCM) offers that are specifically tailored to meet specific needs, we can list:
Sequencing and Staging: The system enables sequencing and staging of up to 20 units, which is critical for efficiently managing cooling loads in large data center. It also supports free-cooling mode, which significantly reduces energy consumption.
Complete control and monitoring: iDCM solution provides a comprehensive control of the primary loop and detailed monitoring of the units. This includes monitoring power consumption, capacity, and flow rate of the chillers to ensure optimal performance and efficiency.
AI Optimisation: By integrating Artificial Intelligence (AI), advanced optimisation algorithms can be applied in real-time to further improve energy efficiency and overall system performance.
Tier III Loop Compatibility: Fully compatible with Tier III data center loops (N+1 System), as special execution also for TIER IV (2N System) on demand.
Our solution meets the stringent requirements for redundancy and fault tolerance in critical infrastructure.
Integrated sensors: With up to 8 integrated sensors, including temperature sensors, differential pressure sensors and flow meters, our system provides detailed insight into environmental conditions and system performance, facilitating proactive maintenance and optimisation.
Together, these features make Daikin's control solution the ideal choice for data center applications, providing robust control, monitoring, and optimisation capabilities to ensure reliable and energy-efficient operation of critical infrastructure.
Challenge #3: Control compatibility with cooling system units
When it comes to implementing a data center solution, in most cases the control solution is provided by a system integrator, which is an external company from the units' supplier. In that way, the system integrator develops a control solution without having the information about the internal logic of the unit and optimal point of operation.
Reason #3 In contrast, Daikin can offer a 'complete' data center solution by integrating its units into an in-house control system. Thanks to this approach, the control system can integrate the logic already designed for the single unit control software, to maximize the efficiency of the entire system according to real-time load and environmental conditions.
The Complete integration of unit software on the iDCM unlocks some additional features in terms of unit management, increasing reliability and in the control of free cooling mode.
The control of supply temperature is more stable than logic of standard BMS Limiting operating mode changes (mechanical/mixed and free cooling only).
Challenge #4: How to optimize the efficiency of a large HVAC system
An efficiency-focused approach can significantly enhance the overall performance of the data center, proven by reductions in energy consumption and associated operating costs.
Load balancing for optimal units configuration
Adaptability to changes in load and external ambient temperature based on the number of active devices is important. By controlling many units, it is needed to quickly adapt to such changes while maintaining optimal cooling conditions at all times.
Reason #4: Daikin iDCM incorporates a sophisticated logic of energy optimisation called Smart Sequencing. This is an advanced control function that relies on Machine Learning models of the chillers to select the optimal combination of units ensuring the lowest energy consumption.
The machine learning algorithm builds the performance profile of the chillers based on historical data from the manufacturer’s know-how and data recorded by the iDCM on site. The controller finds a different optimal number of units which have maintained their performance, prioritizing running them at higher load to maximize the efficiency.
Keeping the units in optimal condition with predictive AI
In a data center cooling system, it is extremely important to keep the units in optimal condition at all times to ensure reliable operation, and thereby reduce the risk of downtime. Regular maintenance also extends the life of the equipment, reducing the need for costly replacements and protecting the company's investment. Finally, it helps to ensure the safety of the equipment, preventing damage and ensuring continuity of operations.
Reason#5: Daikin iDCM’s Artificial Intelligence stores and monitors units' parameters related to its life cycle. These parameters are then used to sequence machines, prioritising those with a better efficiency curve and favouring those with less wear and tear.
One of the main advantages is that it continuously updates its Machine Learning model of the plant's chillers based on the data recorded on-site, allowing to learn performance over time and detect performance degradation.
In a standard system, usually the controller runs the chillers in parallel, spreading the load across all the machines, without knowing which unit is the most efficient during their life cycle. With AI, however, load curves are dynamically updated based on real-time data, optimising the system more effectively. AI-driven load curve updates enable dynamic optimisation based on constantly updated real-application data. iDCM intelligently shuts down unnecessary units to optimise free cooling benefits and takes machine degradation into account. On average, a machine loses 1% of its performance each year. This device sequences units prioritizing those with the best efficiency curves for energy savings.
Optimising for all operating conditions, including free-cooling mode
Another requirement is to achieve maximum efficiency and performance in every operating mode. Reason #6 With the iDCM, specifically designed for data centers, it is possible to ensure both high efficiency and cooling performance in both standard and free cooling modes. The iDCM selects the operating mode based on the following criteria:
Free cooling activation: When the outdoor air temperature is lower than the incoming water temperature by a preset free cooling delta T (adjustable), the system initiates free cooling. Once this is confirmed as feasible by the chiller group controller, a simultaneous free cooling command is given to all chillers.
Free-cooling mode switching: If necessary, a free cooling mixed mode command is sent from the controller to the chiller unit, allowing maximum free cooling utilisation while meeting system capacity requirements.
Full compressor mode activation: The chiller operates in full compressor mode when the delta T between the incoming water temperature and the ambient air temperature is less than 2°C (adjustable).
Challenge #5: Consider control systems’ costs
Reducing operating costs can bring a few benefits. Firstly, it improves operational efficiency by optimizing the use of resources. It also ensures greater long-term economic sustainability.
Reason #7: Daikin control system solution is integrated of primary loop control and chiller control. By having the control function within the unit, such as performance monitoring, the number of controlled field devices has been effectively reduced while simultaneously enhancing the value proposition.
Performance monitoring is an integrated algorithm into the software of every unit that makes up the Daikin data center chiller solution. This enables each unit to calculate efficiency and energy data, assess performance and identify potential areas for improvement and anomalies without the need to install additional devices such as energy or flow meters.
In terms of cost benefits, iDCM offers a double advantage. Firstly, significant benefits in terms of installation costs. iDCM requires less sensors than traditional HVAC plant control systems, maximizing the use of sensors installed on the units. As a result, installation-related costs are lowered. Furthermore, the time required for installation is also streamlined, as there are fewer cables to manage and fewer connections to make.
On top of that, maintenance costs are also reduced. With fewer sensors installed, there are fewer parts wearing out and fewer components to replace over time. This translates into lower long-term maintenance costs and a further reduction in overall costs.
Reason #8: Not all data center installations have the same number of units. To meet this challenge, Daikin iDCM is available in different sizes depending on the number of units within the facility design, these are:
-
IDCM004 up to 4 sequenced units;
-
IDCM008: up to 8 sequenced units;
-
IDCM012: up to 12 sequenced units;
-
IDCM016: up to 16 sequenced units;
-
IDCM020: up to 20 sequenced units.
This allows to exploit a scalable price solution that fits and is specific to a precise number of units. This modular approach ensures a competitive alternative in the market by providing a cost-effective solution that is adapted to each type of plant, making it easier to configure the control system.
Do you want to know more about Daikin’s data center solutions? Then, get in touch with the form below.