Mengxuan Song

Thermal-Aware Energy Management of an HPC Data Center via Two-Time-Scale Control

This paper presents a two-time-scale control method to optimize the energy consumption of high-performance-computing data centers through dynamic frequency scaling of processors, tasks assignment, and cooling supplement. First, the steady and dynamical models of the data center are built, which reflect the computational interactions and thermal relationship among the components of the data center. Next, the energy minimization problem for processing a parallel task is divided into two parts that correspond to the steady thermal model and the dynamic thermal one. Then, the problem is solved in a two-time-scale manner, i.e., the optimization of task assignment and processing frequency is considered in steady thermal environment, and the optimization of cooling supplement is achieved in dynamic thermal environment. Finally, simulations of a real task trace are carried out, which demonstrate that the proposed method can significantly improve energy efficiency while guaranteeing the thermal constraints of the data center.

Flow-field guided steering control for rigid autonomous ground vehicles in low-speed manoeuvring

ABSTRACT This paper studies the low-speed manoeuvring problem for autono-mous ground vehicles operating in complex static environments. Making use of the intrinsic property of a fluid to naturally find its way to an outflow destination, a novel guidance method is proposed. In this approach, a reference flow field is calculated numerically through Computational Fluid Dynamics, based on which both the reference path topology and the steering reference to achieve the path are derived in a single process. Steering control considers three constraints: obstacle and boundary avoidance, rigidity of the vehicle, plus the non-holonomic velocity constraints due to the steering system. The influences of the parameters used during the flow field simulation and the control algorithm are discussed through numerical cases. A divergency field is defined to evaluate the quality of the flow field in guiding the vehicle. This is used to identify any problematic branching features of the flow, and control is adapted in the neighbourhood of such branching features to resolve possible ambiguities in the control reference. Results demonstrate the effectiveness of the method in finding smooth and feasible motion paths, even in complex environments.

Anemometer Positioning Optimization for Flow Field Calculation in Wind Farm

AbstractIn a wind farm, the wind data of the anemometers are used to evaluate the local wind resource. The position of the anemometer in a wind farm influences the wind resource evaluation signific...

Thermal-aware load balancing in a server rack

Cooling availability of data centers has drawn considerable attention from academia. This paper considers temperature control of a cluster of servers with different inlet air temperatures. A rule-based control approach and a model predictive control approach are proposed to find the optimal load distribution, which can lower the outlet air temperature of servers. Web request traces of WorldCup 98 are used as workload in the simulation. The simulation results indicate that the proposed control strategies achieved significant drops in temperature than that of other competitors.

A combined AR-kNN model for short-term wind speed forecasting

Wind speed forecasting is critical to and challenging for wind energy industry. We present a combined AR-kNN regression model for short-term wind speed forecasting. Historical samples are selected to train the coefficients of a k-nearest-neighbor (kNN) regression model in order to capture the current variation pattern of wind speed. The training samples of the kNN model are combined with the recent samples of an autoregressive (AR) model tracking recent correlation of wind speed. To verify the performance of the proposed model, we apply it to the data of 10-min wind speed, and compare it with the persistence model, the single AR model and the single kNN regression model. The simulation results demonstrate that the combined AR-kNN model is effective and generates the most accurate forecasting of wind speed in several cases.

Wind farm micro-siting based on auto-regressive wind prediction

The variety of the wind resource brings much uncertainty to the long-term wind condition, on which wind farm micro-siting is based. In this paper, a novel method based on the long-term prediction of wind distributions is proposed to optimize the micro-siting of wind farms. Long-term wind resource is described by the annual probability of wind speeds and directions. The auto-regressive model is applied to predict the wind characteristics and the genetic algorithm is used to optimize micro-siting. The data of a wind farm in Netherland is used to test the validity of the proposed method. The simulation results show that the method maximizes the generated power and improves the efficiency of utilization of wind energy.