Model Predictive Control Enabling Flexible Operation of Data Centers

Table of Contents

1. Introduction and motivation

2. Outline of the thesis

3. Literature Review

4. Implementation background

5. Data Center Control Strategy Methodology

5.1. Coordination by Model Predictive Control

5.2. Real-time MPC

5.3. Prediction based on ARIMA Model

5.4. Optimal Control

5.4.1. Objective Function

5.4.2. Equality constraints

5.4.2.1. DC power consumption

5.4.2.2. Real-time regulation violation

5.4.2.3. Predicted regulation violation

5.4.2.4. Power consumption of Server Rack System

5.4.2.5. Power consumption of UPS system

5.4.2.6. Storage level deviation

5.4.2.7. Power consumption of CRAC

5.4.2.8. Targeted power level deviation of CRAC system

5.4.3. Inequality constraints

5.4.3.1. Server Rack System

5.4.3.2. UPS System

5.4.3.1. CRAC system

5.4.4. Key Performance Indicators of Simulation Result

6. Data Center Model

6.1. Workload Profile

6.2. UPS Capacity

6.3. CRAC System information

7. Case Study

7.1. Parameters

7.2. Simulation result

7.2.1. Scenario 1

7.2.1.1. Regulation violation

7.2.1.2. Workload queue

7.2.1.3. State of charge

7.2.1.4. Data center temperature

7.2.2. Scenario 2

7.2.2.1. Regulation violation

7.2.2.2. Workload queue

7.2.2.3. State of charge

7.2.2.4. Data center temperature

8. Conclusion and Outlook

Objectives & Key Themes

This thesis explores the integration of Data Centers (DCs) into the smart grid to provide Ancillary Services (AS) through Demand Response (DR). The primary research objective is to develop an online, energy-aware scheduling algorithm based on Model Predictive Control (MPC) that adjusts DC power demand to track reference signals while maintaining operational stability.

• MPC-based energy-aware scheduling for data centers.
• Co-scheduling of IT equipment, cooling (CRAC) systems, and uninterruptible power supplies (UPS).
• Use of ARIMA models for predicting regulation signals.
• Ensuring adherence to operational constraints like server room temperature and workload requirements.

Excerpt from the Book

Summary of Chapters

1. Introduction and motivation: Outlines the growing energy demand of data centers and their potential as flexible resources for grid stability in the smart grid.

2. Outline of the thesis: Provides a roadmap of the research structure and the focus of subsequent sections.

3. Literature Review: Surveys existing research on data center demand response, model predictive control in industrial applications, and current energy-aware cooling strategies.

4. Implementation background: Details the software tools, including MATLAB and YALMIP, used for modeling and optimization.

5. Data Center Control Strategy Methodology: Presents the mathematical framework for the MPC algorithm, including load modeling, thermal control, and constraint formulation.

6. Data Center Model: Describes the physical parameters of the data center testbed and the components (servers, UPS, CRAC) included in the simulation model.

7. Case Study: Validates the proposed MPC strategy through simulation results under different scenarios and parameter settings.

8. Conclusion and Outlook: Summarizes the key contributions of the thesis and suggests future directions, such as refined thermal modeling and battery cell optimization.

Keywords

Data Center, Model Predictive Control, Ancillary Service, Demand Response, Thermal Modeling, Deferrable Workload, UPS, ARIMA Prediction, Regulation Command, Load Following, Grid Stability, Energy Optimization.

Frequently Asked Questions

What is the core problem addressed in this research?

The research addresses the high energy consumption of data centers and the challenge of integrating renewable energy sources into the power grid, proposing that data centers act as flexible demand-response resources to stabilize the grid.

What are the primary components studied within the data center architecture?

The study focuses on the combined operation and scheduling of IT server equipment, cooling (CRAC) systems, and the uninterruptible power supply (UPS) system.

What is the main goal of the proposed scheduling algorithm?

The primary goal is to use an online MPC-based algorithm to adjust total DC power consumption to track grid regulation signals without compromising the operational safety of the servers.

Which scientific method is utilized for the controller design?

The research utilizes Model Predictive Control (MPC), supported by an ARIMA model to predict future regulation signals for the optimization horizon.

How is the server rack system modeled for flexibility?

The model treats the server rack system as a flexible load by implementing a workload queueing approach, allowing delay-tolerant tasks to be shifted to off-peak times.

Which key performance indicators are used to evaluate the results?

The performance is assessed using Mean Absolute Error (MAE), Mean Squared Error (MSE), Maximal Absolute Deviation (MAD), and Maximal Relative Deviation (MRD).

How does the research handle the thermal constraints of the server room?

The study implements a lumped parameter thermal model that accounts for the interaction between server heat production, air circuit temperatures, and the CRAC system's cooling efficiency (EER).

What is the significance of the UPS system in this framework?

The UPS serves as a secondary, passive reinforcement tool for operational flexibility, helping to manage regulation deviations when primary components reach their limits.