VARC(1.2): A SOFTWARE PACKAGE ON VALVE DISSIPATIVE COEFFICIENTS

José Carlos Lauria

Escola de Engenharia Mauá

Praça Mauá, 1

09580-900 São Caetano do Sul – SP

Brazil

Phone +55 11 4239-3066    Fax +55 11 4239-4041

E-mail: jclauria@maua.br

Abstract: “VARC – Valve Representative Coefficients” is a PC-based package developed to assist users in analysing and comparing valves as elements alone. It is a supporting tool for managing an extensive collection of files containing data on the dissipative characteristics of valves gathered from the current literature. Values are recorded to text files in the form of the distinction coefficient and can be displayed both according to this parameter and in terms of the loss coefficient. 

Keywords: valves; loss coefficient; dissipative characteristics

1    INTRODUCTION

Computer simulations are more and more employed to predict head and flowrate variations in fluid distribution installations. Although highly representative models have been developed, the results they provide are no more precise than the experimental data input.

Some quantities as diameter and length of pipes can be accurately identified; others, however, are surrounded by great uncertainty, mainly because of lack of comprehensive databases. Valve performance coefficients are a case in point.

Reference to the technical literature shows that an extensive amount of valuable data on valve performance is available, but they are spread out in many sources and presented in a wide variety of forms. Ordered and detailed records of valve experimental information, along with proper means of managing them, are essential for reaching the maximum effectiveness in flow-control studies.

The interest of the author in describing the operating behaviour of valves has lead the need for compiling data on valve dissipative characteristics and reducing them to a form convenient to computer simulations. It became clear at a fairly early stage of the information search that a computer tool was required to support its organization and recording.

This paper reports on a software package developed to implement and facilitate the usage of the material obtained and to assist the analyst in comparing and discussing limits of applicability of valves as elements alone. It comprises a computer program, called VARC (VAlve Representative Coefficients), a data bank composed by individual files registering dissipative data of valves to magnetic media, a utility program to display short descriptions of valve features and the source the data were collected from, and a utility program to print this information out.

To date (November 8, 2000) the data bank is formed by 404 files, spanning 44 different valve styles.

As the purpose of the work is to discuss concepts related to valve influence in the flow control, the package is distributed free of charge to those interested on the subject.

2    WORKING FIELD

Pertinent and essential pieces of information are available on the recorded data files. They are valuable for selecting the adequate valve type and size for each kind of installation, for describing the flow characteristics of valves as elements integrated with the piping system or for simulating representative valve-generated transient flows.

The collection of data makes it possible to crosscheck several different information sources, besides supporting the validation of new experimental results on the theme.

However, it is not the purpose of the work a critical analysis of valve performance. The data are registered for electronic processing without paying attention to theoretical aspects, practical details or conceptual refinements.

As any mechanical component, valves have their merits and limitations due to their own design and must be evaluated according to the task to be performed.

3    DATA REPRESENTATION

Expressing the loss characteristics of the flow across valves can be made in several ways. For calculating the flow under specified conditions (which is not the same as valve selection) the C_v-term is widely employed. It is a dimensional coefficient expressed in customary U.S.-units and is defined as the number of gallons of water (q) that would pass through a valve with a one pound per square inch pressure drop (DP)

                             (1)

On the other hand, in fluid mechanics developments it is widespread the representation of the head loss (DH) in nondimensional form through a loss coefficient. In SI-units and being Q the flowrate, A the cross-sectional area and g the acceleration of gravity, one has

                         (2)

Both the loss coefficient and the C_v-term present the inconvenience of not having fixed limits. For purposes of standardization, a distinction coefficient can be defined

                      (3)

related to the loss coefficient as

                                   (4)

Tullis [1] calls eqn (2) “discharge coefficient”, while Cohn [2] prefers “flow coefficient”. Both terms are valid in broad sense because any ratio involving flow conditions referred to an area is either a flow or a discharge coefficient, depending on the area. However, these names for eqn (2) can give rise to misinterpretations with the conventional definition for the parameters. Miller [3], on his side, adopts the expression “flow factor”, which seems to be somewhat generic.

The term “distinction” is chosen because such a coefficient provides a direct comparison of valves as isolated devices, allowing to identify whether the dissipative effects (~DH) or the inertia effects (~Q) is the dominant part in the flow across the valve–

                    (5)

            (6)

In this fashion, the distinction coefficient ranges from zero to one for all valve styles, a feature very convenient also for computer-simulation purposes. Fig.3 illustrates the distinction-coefficient behaviour for some valves.

There are several ways to describe the open area of a valve (angle, distance, area). A common ground is obtained dividing the generic aperture by the value corresponding to the fully open aperture. The ratio varies always from zero to the unit and is here referred as reduced aperture A_r.

4    DATA ACCURACY

As geometric similarity is rarely maintained in industrial production, fluctuations in data are identifiable even for valves of the same style but of distinct size in a given source. Many times guiding values for all kinds of nominal diameters are given for a specific valve type with no mentions of the range of data variations.

Rare are the references that completely describe the valve configuration.

The available data come from tests performed under markedly different conditions and are originally reported on a wide variety of forms. Only a few sources give the upper and lower limits of validity for the coefficient under consideration. No reference was found which specifies the uncertainty related to the data presented. As a consequence, the registered values must be viewed as indicative elements and should be adopted with great care.

Driskell [4] comments that, as manufacturing tolerances are undocumented, the uncertainty is at least of ±10 percent. However, according to Wing Jr [5], in some cases the data may differ by as much as 50 percent. Ruus [6] suggests a variation of ±15 percent of the reported values for sensitivity analysis.

If accurate information is required, specific measurements must be made. Palmer et alii [7] measure the loss coefficient for a butterfly valve and verify that the values quoted from the manufacturer’s catalogue are significantly higher at large valve openings and it may be that they include pipe loss between the valve and the downstream pressure tapping.

5    THE DATA BANK

During the last 13 years more than eight hundred references on valves and fluid flow were examined by the author to form the present data bank. A total of 404 sets of data for 44 different valve styles are registered to date (Nov. 8, 2000), as summarized in Table 1. Each data file occupies about 0,5 kbyte.

For convenience of use, the data for each valve are recorded to individual text files. They are external to the supporting program so as to provide immediate access to any application software and to allow inclusions of new files.

Values are registered in the standardized form of the distinction coefficient against reduced aperture. Each file consists of a discrete set of points with uniform spacing at five percent increments in the reduced aperture. As the shut-off aperture for some valves is at values different from zero, there are at most 21 co-ordinates for describing the valve loss curve.

Table 1    Overview of the data-bank components

6    THE SOFTWARE

VARC is a specific utility-program for managing a collection of files containing data on the dissipative characteristics of valves via the distinction coefficient.

VARC is intended to assist users in analysing and comparing valves as components alone in graphical- or numerical form.

6.1    Specifications

VARC is written in Turbo C++ language for personal computer environments and has been tested under MS-DOS 6.2, Windows 3.1 and Windows 98.

VARC itself is composed by a 142-kbyte executable file and is accompanied by a reference manual recorded to a 32-kbyte text file.

Via menu selection, VARC presents two kinds of graphical output related to the dissipative characteristics of valves (distinction or loss coefficient), provides tabular display of data sets and accepts input of valve performance data. Numerical outputs can be stored on floppy or hard disks and printed out.

The documentation is complemented by four text files (total of 40 kbytes) given information on the data source, valve features, codification of both valve and source.

Two utility programs were written to display (40 kbytes) and to print out (40 kbytes) the content of all the text files.

The total storage area is of approximately 400 kbytes.

Besides the available experimental data, the user may supply additional information or create new files.

By linear interpolation VARC creates 101 pairs of plotting values. It was verified that this arrangement leads to an excellent representation of the loss curve for a minimal stored-space required. Data may be displayed both according to the distinction coefficient or the loss coefficient. Up to five curves for recorded-file data plus a curve for keyed-in values can be displayed at a time.

6.2    Program execution

VARC operates interactively and the user is prompted to all required inputs. The program begins by checking the video card adapter and the graphics driver. Then, the directory containing the executable file is identified and kept as a reference from which other files will be read.

In the sequence, the validity of the user’s identification file (VARC.USR) is checked out. If the file is not found or is invalid, the program is terminated. In the case of a positive result, a presentation screen is displayed, followed by a screen identifying the user, to finally the command screen to be shown.

VARC terminates any time on both text or graphics mode when the combina­tion ALT_X is typed.

6.3    Program organization

The command screen (Fig. 1) consists of three parts: the main line that describes the root leading to each one of the tasks possible to be performed; a communication window with secondary menu tables, whose options are activated by pull­down menus; and a help line.

Main menu line

VARC is organized around seven chief options, as seen in Fig. 1:

Fig. 1 Command screen

(1) VALVE. Segment that deals with the various forms of entering the experimental data required to the performance curves to be drawn.

When the option for reading file is activated, the program goes by default to de location where the available files are stored; any other address can also be specified. All the files found are listed in alphabetical order in presenting pages of up to 147 elements (Fig. 2).

(2) GRAPHICS. Makes it possible to plot up to six curves for the distinction coefficient (Fig 3) or for the loss coefficient (Fig. 4). Organized to present the curves one at a time or in sequence without pause and to allow changing the extreme value of the ordinate axis.

(3) COEFFICIENTS. Sets, in global terms, the type of the coefficient to be considered when the plotting of the curves is invoked.

Fig. 2    Selection of data file

Fig. 3    Screen for the distinction coefficient

Fig. 4    Screen for the loss coefficient. (same valves as Fig. 3)

(4) DISPLAY. Designed for a tabular presentation of the co-ordinates of the distinction coefficient or loss coefficient for each valve selected.

(5) WRITE. Records all the pertinent data to disk files. Allows setting a reference path through which the data will be recorded and a reference extension for the files.

(6) PRINT. Prints out matrices containing numerical information on distinction or loss coefficient.

(7) SCREEN. Specifies colour or black-and-white text- and/or graphics output.

7    CONCLUSIONs

Of primary importance in simulating valve-controlled flows is the availability of reliable data on valve performance as elements alone. They are the starting point for numerical and conceptual analysis in the design, operation and maintenance of fluid installations.

VARC and the associated data bank provide background information for the analyst gain knowledge on performance characteristics of a broad range of valve styles and perform sensitivity analysis on the limits of controllability for flow control problems.

From using VARC it becomes clearer that a “universal” valve does not exist. Each one has its strengths and limitations and should be seen as a mechanical component that comes in discrete sizes and behaves according to the installation it is placed in.

The data gathered through the work give a fairly good coverage of valve styles. Notwithstanding, it should be kept in mind that this is not an exhaustive compilation or traces all the contributions made on the subject.

References

[1]    Tullis, J. P. Hydraulics of Pipeline, John Wiley: New York, pp. 87-92, 1989.

[2]    Cohn, S. D. Performance analysis of butterfly valves. Instruments, 6(8), pp. 880-884, 1951

[3]    Miller, E. Some recent developments in flow-regulating valves. Water Power, 21(9), pp. 355-359, 1969.

[4]    Driskell, L. Predicting flow through control valves. Chemical Engineering, 90(18), pp. 94-100, 1983.

[5]    Wing Jr., P. Practical determination of control valve Cv. ISA Journal, 7(9), pp. 90-94, 1960.

[6]    Ruus, E. Head losses (Chapter 2). Closed-conduit flow, ed. M. H. Chaudhry & V. Yevjvich, Water Resources Publications: Littletown, pp. 13-37, 1981.

[7]    Palmer, M. H.; Butterworth, B. & Sprague, M. Auto self-closing valves: site measure­ments and analysis. Proc. of 6^th BHRA International Conference on Pressure Surges, Cambridge, United Kingdom, pp. 187-208, 1990.