Computational Fluid Dynamics Assisted Control System Design With Applications to Central Processing Unit Chip Cooling

Author(s):  
R. Zhang ◽  
C. Zhang ◽  
J. Jiang

In this paper, a computational fluid dynamics (CFD) assisted control system design methodology has been described in detail. The entire design and evaluation procedure has been illustrated through a feedback control system synthesis for a central processing unit (CPU) chip cooling system. The design methodology starts with a full-scale CFD simulation of the nonlinear dynamic process to generate the input and output databases of the process. Using this data set, linear dynamic models around specified operating points are obtained using system identification techniques. Based on these models, one can design appropriate control systems to meet the required closed-loop control system specifications. To illustrate the effectiveness of this technique, it has been used to design a controller for a PC chip cooling system. In particular, the coupling issues between ‘real-time’ dynamic controllers with non real-time CFD simulation have been resolved. A physical experimental test bench based on a cooling system of a Pentium III CPU has been constructed. The feedback linear control systems designed by the proposed CFD approach have been evaluated experimentally for six CPU load conditions.

Author(s):  
R. Zhang ◽  
C. Zhang ◽  
J. Jiang

In this paper, a computational fluid dynamics (CFD) assisted control system design methodology has been described in detail. The entire design and evaluation procedure has been illustrated through a feedback control system synthesis for a CPU chip cooling system. The design methodology starts with a full-scale CFD simulation of the nonlinear dynamic process to generate the input and output databases of the process. Using this data set, linear dynamic models around specified operating points are obtained using system identification techniques. Based on these models, one can design appropriate control systems to meet the required closed-loop control system specifications. To illustrate the effectiveness of this technique, it has been used to design a controller for a PC chip cooling system. In particular, the coupling issues between ‘real-time’ dynamic controllers with non real-time CFD simulation have been resolved. A physical experimental test bench based on a cooling system of a Pentium III CPU has been constructed. The feedback linear control systems designed by the proposed CFD approach have been evaluated experimentally for six CPU load conditions.


2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Lin Meng ◽  
Yang Gao ◽  
Yangyang Liu ◽  
Shengfang Lu

Purpose As a short take-off and landing aircraft, FanWing has the capability of being driven under power a short distance from a parking space to the take-off area. The purpose of this paper is to design the take-off control system of FanWing and study the factors that influence the short take-off performance under control. Design/methodology/approach The force analysis of FanWing is studied in the take-off phase. Two take-off control methods are researched, and several factors that influence the short take-off performance are studied under control. Findings The elevator and fan wing control systems are designed. Although the vehicle load increases under the fan wing control, the fan wing control is not a recommended practice in the take-off phase for its sensitivity to the pitch angle command. The additional pitch-down moment has a significant influence on the control system and the short take-off performance that the barycenter variation of FanWing should be considered carefully. Practical implications The presented efforts provide a reference for the location of the center of gravity in designing FanWing. The traditional elevator control is more recommended than the fan wing control in the take-off phase. Originality/value This paper offers a valuable reference on the control system design of FanWing. It also proves that there is an additional pith-down moment that needs to be paid close attention to. Four factors that influence the short take-off performance are compared under control.


2006 ◽  
Vol 129 (2) ◽  
pp. 221-231 ◽  
Author(s):  
André Burdet ◽  
Reza S. Abhari ◽  
Martin G. Rose

Computational fluid dynamics (CFD) has recently been used for the simulation of the aerothermodynamics of film cooling. The direct calculation of a single cooling hole requires substantial computational resources. A parametric study, for the optimization of the cooling system in real engines, is much too time consuming due to the large number of grid nodes required to cover all injection holes and plenum chambers. For these reasons, a hybrid approach is proposed, based on the modeling of the near film-cooling hole flow, tuned using experimental data, while computing directly the flow field in the blade-to-blade passage. A new injection film-cooling model is established, which can be embedded in a CFD code, to lower the central processing unit (CPU) cost and to reduce the simulation turnover time. The goal is to be able to simulate film-cooled turbine blades without having to explicitly mesh inside the holes and the plenum chamber. The stability, low CPU overhead level (1%) and accuracy of the proposed CFD-embedded film-cooling model are demonstrated in the ETHZ steady film-cooled flat-plate experiment presented in Part I (Bernsdorf, Rose, and Abhari, 2006, ASME J. Turbomach., 128, pp. 141–149) of this two-part paper. The prediction of film-cooling effectiveness using the CFD-embedded model is evaluated.


1988 ◽  
Vol 31 (1) ◽  
pp. 56-61
Author(s):  
Atsushi Takahashi ◽  
Takao Okada

This study discusses various control systems that can keep the room pressure and supply/exhaust airflow rate at constant levels in "other rooms" of a highly airtight containment facility when the supply/exhaust airflow is shut off in one of the rooms for decontamination purposes. This study has shown that the constant air volume control system (CAV) allows hysteresis to occur at small differentials on the performance curve of the static pressure differentials and that this hysteresis can cause wide fluctuations in room pressure. In contrast, the variable air volume, central processing unit (VAV-CPU) control system can maintain both airflow rates and room pressures. Each room pressure was controllable to the set level, with an error of less than ±0.5 mmH2O even during transient distur bances. This control system limited fluctuations in the airflow to and from each room to 5 percent during the transient responses. This control system also allows power savings in the operation of supply/exhaust fans, because of the reduced airflow rate and the static pressure of the fans, and is considered to be an excellent control system.


Sign in / Sign up

Export Citation Format

Share Document