Signed in as:
filler@godaddy.com
Signed in as:
filler@godaddy.com
Cooling tower transfer characteristics is one of the most, if not the most important parameters in cooling tower thermal design. Unlike heat exchangers, where transfer coefficients for various configurations are widely published, cooling tower transfer characteristics are not commonly found in the open literature. One of the reasons for this is that fill manufacturer’s generally regard the transfer characteristics of their fills as confidential. The following two sources do give transfer coefficients of various fills:
The fill transfer characteristics of the two sources above are available from a database in the software. The fill database of the software is expanded as fill manufacturers make the characteristic curves available. The fills characteristic of Brentwood is included in the software for counterflow fills.
The only way to be certain of what the transfer characteristics are for a fill, is to do an independent fill test. This can be a costly undertaking, but if you take the penalties involved if a cooling tower does not meet the guaranteed performance, this is a small price to pay. The only problem is that there are only a few facilities globally where tests like these can be conducted.
The software package does not use multidimensional CFD, instead it is a one dimensional analytical model. The multi-dimentional effects are addressed in the semi-empirical equations of the transfer and pressure drop coefficients. These semi-empirical equations are derived from numerical and experimental studies. The software is therefore very cost effective when compared to CFD.
Accuracy
The big question is how do the accuracy of the software compare to the accuracy of CFD. Experience has shown that the analytical based models (like the current software) give results of the same order of magnitude than full-blown CFD simulations. The results of the software are, however, obtained at a fraction of the time and cost of the CFD simulations.
Why?
Why are CFD models of cooling towers not necessarily more accurate than analytical models? The geometry of a cooling tower is generally very complex. The water distribution system with the support structure, pipes and nozzles are difficult to accurately model in CFD. Add the water droplets, fill, fill support structure and fan to the CFD model and you can appreciate the increase in complexity of the model. CFD models are therefore simplified by employing many simplifying assumptions. Some transfer and loss coefficients are obtained from experimental research. These are the same semi-empirical models that are employed in the analytical models. Therefore, the same models are used in both CFD and the analytical models and the end results are therefore not too different.
CFD has its place
Having said the above, CFD is still a very powerful and useful tool to analyse cooling tower performance. It can run many scenarios that will be impossible to do with analytical tools. When you consider the cost and time to do CFD analyses, analytical models like the current software, will always be in demand.
The software is based on a semi-empirical analytical approach. The semi empirical equations are used to calculate the following:
These semi-empirical equations are only valid for certain parameter ranges and operational conditions. The software may therefore not cater for all possible cooling tower sizes and operational conditions. The software will give an answer, but the equations will be extrapolated where the accuracy is unknown. The software will, however, warn the user if some equations are used outside the specified ranges for the design parameters. The software was specifically designed for relatively large industrial cooling towers and should work without any problem for these towers.
Many different numerical schemes are used in the software to obtain convergence of the equations. There may be some cases where numerical instabilities may occur, and the solution process will fail. There are, however, corrective measures implemented in the code that automatically prevents these instabilities. It must be stressed that these cases are relatively rare. A user familiar with the software can solve the numerical instabilities when they occur by changing the solution control parameters. This is similar to steps taken in CFD software packages to ensure stable convergence of solutions
The airflow through mechanical draft cooling towers can determined/specified in two different ways:
Firstly, the fan curve can be entered into the software by specifying up to a sixth order polynomial. This case is used when the fan curve is known. The polynomial (equation) that gives the fan static pressure vs volumetric flow can be entered into the software. The fan curve should be given by the fan manufacturer. The fan curve is only valid for one specific fan configuration, i.e. if the number of fan blades or the blade angle changes then a new curve should be entered into the software for the new configuration. Some fan manufacturers provide software with their fans which generate fan curves for all possible fan configurations. The polynomial that represents the fan curve should be determined by the user by using a curve fitting procedure. Microsoft Excel is commonly used for curve fitting. By using the fan curve the software will determine the operating point of the cooling tower by iteratively solving the draft and energy equations. The fan speed or other cooling tower parameters, like the fill height, can then be varied to achieve the required cooling load.
The second approach is to specify the air flow through the cooling tower that will achieve the required cooling load. The pressure drop through the cooling tower is calculated by the software. The pressure drop and flow rate can then be used to specify a fan that will meet all the geometrical and operational requirements.
Over the years, the software was enhanced to cater for numerical instability. However, in some rare cases (especially those at extreme operating and atmospheric conditions) numerical instability may still occur during program execution. There are however many measures that can be taken to fix the problem. Here is a short summary of steps to follow to obtain solutions (numerical convergence):
The following sources were consulted in the development of the software. They are presented in chronological order of the publication date.
Copyright © 2024 Anthelia (Pty) Ltd - All Rights Reserved.
Powered by GoDaddy
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.