After we substitute Eq. (3) into Eq. (i) and assume that the turbulent diffusion coefficient of the drops is constant and equal to the turbulent viscosity from data in [9] UtD r
~t=-7~-udr
U~U t
u/8 , we obtain j = --p~.(u'v'/u).(C 0 - C l) + p~'C 1"[{(2a2p~'u2)/(9DR)}
+ Vd],
where 6 is the thickness of the liquid film. Calculations show that if the outer wall is permeable, the local droplet flow is 1.45 times higher than for a solid wall. Thus, gas extraction increases separation efficiency. LITERATURE CITED 1 2 3 4 5 6 7 8 9.
V. G. Sister, "The hydrodynamics of centrifugal separators, ~' Dissertation for Candidate of Engineering [in Russian], Moscow (1974). Yu. D. Ovchinnikov, "Hydrodynamics of centrifugal separators in ammonia productions" Dissertation for Candidate of Engineering [in Russian], Moscow (1987). B. J. Azzopardi, "Measurement of drop sizes," Int. J. Heat Mass Transfer, 22, No~ 9, 1245-1279 (1979)." P. A. Kouzov, Fundamentals of Analyzing the Dispersion of Industria! Powders and Pulverized Materials [in Russian], Khimiya, Leningrad (1971). A. M. Kutepov, L. S. Sterman, and N. G. Styushin, Hydrodynamics and Heat Transfer in Steam Formation [in Russian], Vysshaya Shkola, Moscow (1983). ~. I. Levdanskii and I. M. Plekhov, "Improving elementary centrifugal separators," Khim~ Promst'., No. 2, 40-42 (1981). A. S. Nazarov, "Hydrodynamics of a radial reactor with a permeable catalyst and layer," Dissertation for Candidate of Engineering [in Russian], Moscow (1982). L~ D. Landau and E. M. Lifshits, Hydrodynamics [in Russian], Nauka, Moscow (1986)o K. V. Grishanin, Dynamics of Waterways [in Russian], Gidrometeoizdat, Leningrad (1979)o
AUTOMATED DATA BASE FOR THE FLOW-THROUGH PART OF CENTRIFUGAL COMPRESSOR STAGES A. G. Nikiforov
UDC 621.515:681o3.016
The complexity of the physical phenomena in the operation of centrifugal compressors and the absence of a complete theory leads to the necessity of long and expensive tests of the flow-through part in searching for an optimum design. As a result, various organizations are constantly updating large amounts of technical documentation and test data which are obtained in designing and testing centrifugal compressors. The following features are inherent to designing centrifugal compressors: wide application of developed and verified design solutions, research on finding and justifying new solutions~andno unique optimization criteria~ Study of existing designs and analysis of previous test data is required, even at the first planning stage. However, even analyzing published research takes a lot of time, and unpublished data are not available to most specialists. Moreover, organizations that achieve test results irretrievably lose them over time. Therefore, plans for centrifugal compressor stages are traditionally based on comparing only a Small number of variants. It should also be noted that, even with current information retrieval systems, simultaneous analysis of multiple data sets with respect to several interacting parameters is almost impossible without using computers. Thus the problems of systematization, centralized acquisition, and storage of experimental data are urgent and cannot be solved to meet modern requirements without using computer techniques. In other words, there is a requirement to automate the processes of assembling, accumulating, and processing data and to going to paperless information storage on
Translated from Khimicheskoe i Neftyanoe Mashinostroenie, No. 6, pp. 4-7, June, 1993.
0009-2355/93/0506-0255512.50
9 1993 Plenum Publishing Corporation
255
iCompressor I
Level I Level
I Section o . ~ _ ~ ] Heat exchanIger No. 1
N
I1
Heat axe-banI ~" N I I Section
I S e ~
ger
I
Level
9 --
I
~'
No. 2
.....
I
]e
9
,"t
"] I
I
1
iFinalheat .exchanger I
!
'.!. I
Level /F
I Inlet l
Level .F"
!
I IInlet6~ '
[tr~
I Fig. i.
!
I J
Rotor
[ II
I
[Diffuser
I
,,' I
I I Reverser] I Outlet I
, I
= I {', Wne
.
.
.
.
.
.
I
Level of detail for a centrifugal compressor.
computer-readable media. Moreover, automated data bases are a necessary first stage to implement any required automated design system that is to be applied to centrifugal compressors. In the process of developing an automated data base for the flow-through part of the stage, an information base and a programming package were created to allow that data base to function: an information logic structure was developed, a data base operating system was chosen, data base and external formats were chosen, and programs were developed to service the information base. The information logic structure of the automated data base provide for multipurpose utilization of the data base as an information system for I) the compressor as a whole, 2) for sections, 3) the cooling between sections, and 4) each stage and its elements. The automated data base was programmed for various levels of computers. For the [Russian] Unified System (ES) computers we used the original RELBAZ [I, 2] system which has been used effectively for many years to record and manipulate data. This is a relational program; it communicates with the data base via a simple language that is easy to teach nonprogrammers. Other versions of the automated data base are programmed in FoxBASE and its analogs for IBMcompatible personal computers. After the programs were chosen, the engineering part of the problem was attacked: detailing the flow-through part of the stage, describing the geometry and gas dynamic characteristics of the stage elements, determining the data base rel~tionships [interfaces] and the structural connections between them, developing the required codifiers, and determining the content of commands to search for information. The actual object - the compressor - is represented in the data base in five levels of detail (Fig. i): the compressor, a section (a group of uncooled stages), the stages, elements [inlet, inlet control equipment, rotor, diffuser, reverser, and outlet), and details [rotor disk, rotor vanes, diffuser, and reverser]. The first four levels characterize a set of relationships that reflect the structural assembly, geometry, optimum parameters, thermodynamics and gas dynamics, and test data. The relationships, which make up the five levels of detail, contain information only on the geometry of the design elements. The data base for one stage consists of 38 constant relations, which include over 600 aspects. To date, research on over 200 centrifugal compressor stages, designed at the St. Petersburg State Technical University (including both overall characteristics and characteristics of separate elements of the flow-through part) and 40 stages of Dresser Clark compressors (USA) (only integral gas dynamic characteristics) have been processed and entered into the automated data base. The following point on the organization of automated data bases must be emphasized. On the one hand the variety of geometric designs of the flow-through part sets a trend to the most complete and detailed description of the centrifugal compressor design. On the other hand, the volume of information must be held to a reasonable level to reduce computer memory requirements, data preparation time, and search time for required data. Because optimizing
256
[
L_ Inlet
Meridional
ro~ I section
of
of diffuser
]
ch~acter-
I
pt bypass valve]
/
[of reverser
of outlet
Bypass valve + reverser powe~ character-
Radial cross section of
[ inlet
__i___
I
characteristics
!
I Rotor power i charac her[ istics
Fig. 2.
Vanes
~ h c a~a' cter-[1 ......
istics
Information logic structure of the data base.
this problem is difficult, it is imperative to have a simple method to reorganize the data base. One of the distinguishing features of this automated data base is the simplicity in changing its contents: once the format of the data base or one of its parts has been corrected, it is easy to include (exclude) new relationships. This possibility of improving and developing the automated data base in response to user requirements results from using a data base operating system that allows existing data in the base to be reorganized. The contents of the automated data base for the flow-through part of centrifugal compressors and the structural connections between relationships is shown in Fig. 2. 0nly constant relationships, into which prepared information is written, are shown. For example, the relationship "Experimental research," which contains test data, is connected to programs to process results, several working formulas, and programs to write calculated results into other relationships. Now we examine the basic information contained in the the automated data bases The relationship "Stage" includes basic handbook geometric parameters of the stage and codes for the elements comprising the stage. However, the relationship contains inlet and outlet angles and the numbers of rotor, diffuser, and reverser vanes; the diameters Do, DBT , DI, and other geometric dimensions, fn total, the relationship contains 48 aspects. The celationship '~Power characteristizs" contains tables of the efficiency, loss, and p~essure-drop coefficients as functions of flow rate [or various stage operating regimes (the number of regimes can vary, but cannot exceed ten). There is provision for entering various efficiencies (polytropic, adiabatic, overall~ etc.), pressure drops, and flow rates. Rotor information is contained in four relations~ The relationship "Rotor" (41 aspects) includes basic geometric parameters of the flow-through part, codes for the disks and vanes, nomenclature and notations for the rotors, information sources, data entry dates, and the surnames of the people who entered information into the data base. The relationship "Rotor disk" (25 aspects) contains data on the geometry of the meridiona! cross section of the rotor in the form of coordinates describing the flow-through part. The relationship "Optimum rotor parameters" (16 aspects) contains the basic gas dynamic parameters Mu~ , ~opt, ~2opt, Nopt, ~pump, ~max, Reu2, etc.). The contents of the relationship "Rotor characteristics"
(45
aspects) are similar to "Stage characteristics."
The relationship "Vanes" (38 aspects) contains data on the profile of the cylindrical vanes. Into it are written the vane coordinates of the rotor, diffuser~ and reverser vanes. The relationships describing the other stage elements have a structure and contents similar to the rotor relationship.
257
Automated data base Geometry of the flow-through section
1
Process test data
Test results
I
i
Power charaeteristics
Calculate gas J thermodynamics
Gas thermodynamic parameters
Calculate flow ] I Plot velocity~ around rotor ~ envelope along rotor vanes
Calculate meridional flow Select and calculate layer thickness
1
i
Calculate flow ]I Plot velocity around diffuser ~ envelope along vanes I ~iffuser vanes
|
Calculate flow I [ Plot velocity ] ~.around reverser H envelope along vanes I [ reverservanes] Create work file
t:ttoen
Modelefficiency: ~iffuser diffuser
~ser
Random descent
Combined identification
j-
~
~Analyze loss] components
Modelpressure
J
drop
Fig, 3. Consolidated block diagram of the programming package for identifying mathematical models.
For ES computecs, the automated data base for the flow-through part is written in the form of two independent data bases controlled by one data base operating system, due to the fact that one data base cannot contain more than 256 aspects. The first data base includes 16 relationships containing information on the geometric dimensions of the flow-through part of the stage and its elements; the second has 12 relationships containing gas dynamic test results, optimization parameters, characteristics, and some geometric dimensions of the flowthrough part of the stage and its elements. In order to operate the automated data base, we created and used a library for storing the data base formats, the source code of the formats and headers for the relationships, the source code of user programs, applications programs, and operating instructions for the programs. A FoxBASE-type data base operating system does not have the above [aspect] limitation, so the automated data base for personal computers is organized as a single program package. A special program was developed for various data cequests to the base, and service programs were written to organize the data base as a system of menus. All the programs developed to operate the automated data base on-line can be divided into three groups: data input, information extraction, and special requests. The programs to search, process, and select information for output according to user requests are designed for designing multistage compressors on the basis of available stage data or for designing stages on the basis of data in the automated data base for specific elements. For example, after calling up one of the most general programs (in terms of requests) starts a dialog with the user for simultaneously satisfying 17 limitations on the range of the Mach number, the optimum flow rate, and the width of the rotor outlet, the minimum bushing diameter, the maximum possible relative diameter of the stage, the range of the geometric diffusivity, etc. After the initial limitations are input, the program produces a choice of stages or their elements that satisfy the input conditions. The choices are written into the working relationship and can be sent either to the printer or the display according to the user's request. This information can be used like handbook material to design a stage that meets specified conditions. Data on how one parameter or another affects the efficiency of the stage
258
or one of its elements can be obtained easily and rapidly by using the system to do a special sampling and then a statistical processing of the resultant information. It should be noted that using this data base only partially lightens the design effort. It does not exclude using practical experience and intuition to make decisions on selecting one flow-through part or another. Complete automation of this process is difficult due to specific limitations: the impossibility of developing unique criteria for estimating the adequacy of the results obtained on different test stands, the use of various test methods, the possibility of entering erroreous data, etc. Creation of the automated data base made it possible to automate the process of generalizing a large amount of test data based on single methods to model the power characteristics of a centrifugal compressor and its elements [3]. Here the basic stages of constructing mathematical models includes the following: examining the stage element by element and summing [pressurel losses along the moving gas; calculating and analyzing the three-dimensional gas flow in the flow-through part; representing the power characteristics of an element as a function of similitude criteria, velocity distribution parameters, and a small number of characteristic geometric dimensions: using already known representations of the physics of the gas dynamics; and doing statistical processing of test results in order to obtain mathematical functions. This program package, a consolidated block diagram of which is shown in Fig. 3, was developed at the Smolensk Branch of the Moscow Power Institute for IBM-compatible personal computers. This package includes: calculating the meridional flow in the flow-through part by a quasi-orthogonal method; calculating flow around the vane equipment of the stage by the method of characteristics; various methods for graphically representing the velocity distribution; identifying the mathematical model by random descent and by a combination of the methods of Hooke and Jeeves, of combined gradients, and of random and deterministic descent; and generalized mathematical modeling of the efficiency and pressure drop. LITERATURE CITED .
2.
E. A. Barsukov, P. I. Komarov, V. M. Ryvkin, et al., "Realizing the RELBAZ data base operating system," Uprav. Sist. Mashiny, No. I, 98-102 (1984). V. N. Denisov, L. N. Ershova, W. I. Kislitsin, and A. G. Nikiforov, "Creating an automated data base for predicting the technical status and for planning repairs on the main thermomechanical equipment of a thermal power station," Energetik, No. 7, 22-23
(Z985). 3.
K. P. Seleznev and Yu, B. Galerkin, Centrifugal Compressors [in Russian], Mashinostroenie, Moscow (1982).
259