A Model-Driven Approach for Design & Operation of Data Centers

With the advent of microgrid technology in mission critical markets, electrical power system modeling, analysis, and simulation have become even more vital in design and operation of data centers. Hence, making the proper design decisions early on during the system planning stages are now paramount to support the energy efficiency strategies to significantly reduce the total cost of ownership (TCO) and improve the reliability of engineering and facility operations.

Today, the owner-operator of modern data centers is expecting a Smart Data Center Solution where the facility is designed, analyzed, validated, and then taken on-line to provide a system with increased efficiency and life cycle management; from design to deployment to real-time operation and back to capacity planning.

Model-Driven Design and Operation

A model-driven approach codifies the design team’s intent to ensure that facilities operate precisely as they were intended to, or better:

1. The power system model is validated for optimal performance at the outset of the project (power flow, short circuit, protective device coordination, arc flash, reliability assessment, and more).
2. It continues to function in online mode once the facility is operational, comparing 'as-is' versus 'as-designed' data.
3. It serves as the basis for real-time simulations and 'what-if' studies, e.g. capacity, upgrades, faults, reliability, etc.
4. As the facility evolves over time, the model can be easily updated to serve as a digital asset repository.

Design and Analysis

Utilizing the model-driven design approach helps determine the electrical behavior through creating various scenarios which allows the engineer to predict future behavior, analyze expansion projects, or identify improvement opportunities of the power system.

Creating an electrical model of the data center offers a centralized location for a system knowledge base and electrical network assets. The model includes parameters of the electrical components and how these components are inter-connected to form the complete power system layout. 

Using pre-built data center one-line diagram templates based on the commonly used ‘Tiered’ concept for data center design allows for streamlining the modeling and design process and save time during the process. These pre-built templates contain typical electrical data information which can be easily and quickly adjusted to customize the design to meet the specifications of the facility.

The “perfect system” starts during the design phase.

Intelligent Monitoring and Visualization

Utilizing the system model along with real-time data integration with metering devices, data acquisition, and archiving systems, the data center operator can then accurately monitor, visualize, predict, and operate the facility.

Monitoring energy usage provides up-to-date data on power consumption, carbon emissions, and temperature and humidity settings. This information is essential to the operator to identify opportunities for quick reduction in power consumption, energy-saving initiatives, and ultimately reduce OPEX costs.

In addition, the indication of abnormal conditions is an important function of a monitoring system. Alarm and warning schemes provide immediate signals for abnormal parameters, including areas in the mission critical facilities that are not directly metered. Dedicated data center interfaces, thin-client dashboards and web-enabled reports provide operational efficiency and effectiveness of new energy saving projects.

System Prediction, Forecasting and Risk Assessment

The ability to use real-time and archived data to predict the system behavior in response to operator actions and/or system disturbances and events is a powerful capability of a model-driven Smart Data Center Solution. Analytical applications allow for pre-configured scenarios to be assessed and analyzed to deliver real-time predictive simulations and decision solutions as well as to provide a platform for Operator Training Simulators (OTS) to accelerate the traditional training method and to make the training an ongoing process.

The ability to recover from a system disturbance depends on the time it takes to establish the cause of problem and take remedial action. This requires a fast and complete review and analysis of the sequence of events prior to the disturbance. 

Forecasting system response will eliminate errors and system downtime. The system should be configured to allow the operator to re-play previously recorded message logs while controlling the playback of archived data to re-run at original or accelerated speeds.

Combining distributed energy technologies with an intelligent power management simulation application will assist operation and engineering staff to quickly identify the cause of operating problems.  These applications should assist in increase facility reliability and control risk and costs.

The ability to playback historical events is especially useful for root cause and effect investigations, improvement of system operations, identification of potential security vulnerabilities related to the electrical network, and exploration of alternative actions via replay of operating scenarios.

Optimization and Automation

Intelligent monitoring is the base of a data center system.   Besides improving data gathering capability and real-time prediction of system response, the model-driven data center solution is used for reducing losses and determining where energy supply and demand can be optimized.  The monitored data and collected information are used to provide supplemental automation via user-defined actions that can be added or superimposed on the existing system.

Features of supplemental automation must contain the following controls:

• Automatically determine the loading requirements and manage the loading demand for each Power Distribution Unit (PDU).
• Switching sequence and work order management to verify whether the sequence is compliant with safety switching procedures and requests confirmation during execution simulation of each step before proceeding to the next step in order to avoid inadvertent actions.
• Based on the generation cost, availability, efficiency curves, determine the amount and mix of available distribution energy resources (DER) needed to run the server racks without an unnecessary amount of power abundance.
• Load preservation via load curtailment application that monitors system’s conditions and make decisions about when and how load shedding will be required. This intelligent load shedding application then executes the desired optimal strategies and performs the controls necessary to shed the minimum required load.

Adopting a smart data center solution takes the guesswork out of design and vastly improve operations.

A Smart Data Center Solution can be used from the design stage to optimize power usage, increase energy savings and reduce the TCO while maintaining high availability, reliability and security during the facility's operation.

Extending the power monitoring system by equipping it with an appropriate electrical system model (context and knowledge), simulation, predictive analysis, playback applications, and load preservation system will provide the data center designer, engineer, owner-operator with a powerful set of intelligent tools and enterprise solutions.

Finally, all of these capabilities and applications should be included in one integrated platform with the flexibility and compatibility that allows you to expand and upgrade your data center design and power management system as your needs grow.