Technical Program


< Go back to the papers' list


Title

As Good As It Can Be – Ventilation System Design by a Combined Scaling and Discrete Optimization Method


Topic

F3 - Fan Design Methods


Authors

SCHÄNZLE Christian
Technische Universität Darmstadt, Chair of Fluid Systems

Darmstadt - Germany
christian.schaenzle@fst.tu-darmstadt.de
ALTHERR Lena
Technische Universität Darmstadt, Chair of Fluid Systems

Darmstadt - Germany
lena.altherr@fst.tu-darmstadt.de
EDERER Thorsten
Technische Universität Darmstadt, Chair of Fluid Systems

Darmstadt - Germany
thorsten.ederer@fst.tu-darmstadt.de
LORENZ Ulf
Technische Universität Darmstadt, Chair of Fluid Systems

Darmstadt - Germany
ulf.lorenz@fst.tu-darmstadt.de
PELZ Peter F.
Technische Universität Darmstadt, Chair of Fluid Systems

Darmstadt - Germany
peter.pelz@fst.tu-darmstadt.de

Abstract

Designing complex ventilation systems is a challenging task. A required operation can be fulfilled with numerous different combinations of fans, and the application engineer has to choose an adequate set-up. The focus often remains on the functional quality of the system whereas its energy consumption plays a subordinate role.
Fans are well-engineered products i.e., their individual energy efficiency is already highly optimized. It is common practice to choose one fan that fulfills the peak operation load and that is driven in its best operating point in this maximum loading condition. When it comes to a temporal distribution of different loading conditions multiple smaller fans may be more appropriate considering energy consumption. However, due to the multitude of different fan combinations and operating strategies even for an experienced engineer it is impossible to keep track of all possibilities.
In this paper we use the methods of the new research direction at TU Darmstadt called Technical Operations Research (TOR) for the design of energy optimal technical systems. It is based on mixed integer programming, a discrete optimization technique known e.g., from the field of Operations Research. This approach follows the understanding that optimized components do not imply energy-efficient systems. Instead of optimizing the energy efficiency of single components we set the attention on the entire technical system. With TOR we design the most energy-efficient ventilation system out of a given kit of fans and derive the optimal control settings for different probability distributed loading conditions.
We illustrate our optimization approach by designing an energy optimal ventilation system for an office floor. The required operation contains four different volume flow rates that result from the time dependent occupation density of the office rooms. The ventilation ducts are fixed i.e., the system characteristic is given as an input to the optimization model. The task of our optimization algorithm is to find the energy optimal combination of fans and their control settings for each loading condition. The fans are represented by dimensionless head curves: pressure coefficient and efficiency versus flow coefficient. By applying affinity and scaling laws the variation of fan size and rotating speed allows one to find the optimal ventilation system out of a huge number of fans.