Full report

Original page 1

( " NASA TECHNICAL MEMORANDUM Nfl-51f T/{--7697Y NASA TM-76979 ~ . NASA-TM-7697919830025636 - "'. DEVELOPMENT OF A PROCESS CONTROL COMPUTER DEVICE FOR THE ADAPTATION OF FLEXIBLE WIND TUNNEL WALLS Jurgen Barg Translation of " Entwicklung einer ProzeBrechnergesteuerten Regeleinrichtung fur die Adaptation von Flexiblen Windkanalwanden " ~ Institut fur Luft-und Raumfahrt Technische Universitat. Berlin. ILR Mitt. 70. July 1980. 51 pp. , . t' " ' C( ,-; LANCLCY FE::';::,;:'C I C::::lE:1 Ll2;i :,i''', ::~:.; HM,iPTON, VIRGIr~!A NATIONAL AERONAUTICS AND SPACE ADMINISTRATION ~'IASHINGTON, D. C. 20546 SEPTEMBER 1982 .. /111111111111 1111111/111111111111111111111111 NF00286

Original page 2

\ \ " " STAH()A$30 TITLe PAGI 1. R." H.. "'2. G.'P9_1 Acc...' He. ,. Au','...", C.,.I•• H., . NASA TM -76979 •• 1111..... s.."'I. S. R.,."0.,. CONTROL COMP- September 1982 DEVELOPMEnT OF A PROCESS .. UTER DEVICE FOR THE ADAPTATION OF FLEX- 6. P.,'.""I", 0"."1,.,,... c.. I IBLE WIND TUNNEL WALLS 7. .l...h.r{.1 " Jurgen Barg II t. p.,ro",,,", O,....I ..'j." N_....4 1.44,••• SCITRAN !ox 5456 Sant"~ R::u·h",,.,,, r& 0'11 nA 12. Sf/;,.,j", Al."A' H_....4 AtU,... I. p.rr.,,..,,, O"."i,."." R.,N' fl •• 10. W••1r U"U H•• '. It. C.,,"n'., (i,.", H•. NAS"" iS42 . ft.,.,. ... 11 T"••f 4,.,1.4 c....,•• Tranalat100 tiona eronaut1c8 and SpAce AdQio!atratioo Wasn1ngtoo. DC. O546 IS. s....1_...'., H.... ... ,....Iff... 1..-, c.H Translation of "(Entwickluhg einer ProzeBrechnergesteuerten Regleinrichtung fur die Adaptation von Flexiblen Windkanalwanden)". Institut fur Luft-und Raumfahrt Technische Universitat . Berlin. ILR Mitt. 70. July 1980. 51 pp. •6. Ah't9Ct' tests the problems arise '. In wind tunnel of determining the wall pressure distrlbution, calculating the wall contour~ and controlling adjustment of the walls. This .report shows . how these problems have been solved for the high speed wind tunnel or he Technical University of.Berlin. 17. It., W..4. (Set..I.. ~ A.INr( " -;-~ '. - '1. 0. •..,,,,"_ s........ - --. Unclassified - Unlimited I ! I It. ri" Cl...... (.f ..... ..,... -. 1.-,,, CI...... tef .... ~ . Uuc!ua1f1d Uuc:l...1f1_ I I"'.....'.... In,-.. ~~ 51 t • AJ/J53/'5S .~~~ Ng3-33907#-

Original page 3

" I' , '. ",' '. ILR-Communication ILR 70 (1980) " '. Development of a Process-Computer-Controlled Control Device for the Adaptation of Flexible Wind Tunnel Walls by Jurgen Barg" ,~ '. Berlin, July 1980 ,. '. ..t -,Or,.

Original page 4

\1,1 (' Table of Contents Symbols 1. INTRODUCTION " LL Page 2 3 DEVELOPMENT OF A CONTROL 4 2. FUNDAMENTALS FOR THE DEVICE '. 2.1 Principle of the adaptive flexible walls 4 2.2 Control algorithm 6 2.3 Theory for calculating the external field 8 2.4 Numerical method 3· CONTROL DEVICE 3.1 Total buildup 3.2 Servo-system 9 10 10 11 3.3 Microprocessor-process computer-coupling 13 3.4 Test value acquisition device 14 I", 4. SOFT'!ARE FOR THE CONTROL DEVICE 15 ." 4.1 Control sequence 4.2 Microprocessor-software 15 16 4.2.1 Expansion of the microprocessor- 19 software 4.3 Process computer-software 20 4.3.1 Coordination program INTR 20 4.3.2 Test program MESS 22 4.3.3 Computer program 5· SUMMARY ,J 6. REFERENCES FIGURES, TABLES APPENDIX EXINT 23 24 26 27 47 1

Original page 5

h' " Symbols a. C pressure coefficient p h wall contour i run variable . I length of the flexible wall u tangential component of the disturbance velocity " v normal component of the disturbance velocity x coordinates y DAC digital information of the D/A-W F total calibration factor for displacement pickup F I calibration factor of the D/A-W H wall contour, wall position AH difference of the should- and actual- wall position 1M motor current M Mach number K control factor U velocity (tangential), electrical voltage ,. V adjustment speed of the adjusting member a angle of incidence ',. integration intersection values angle of the transmission factor of the displacement IP pickup g integration variables velocity potential Indices a initial value c computer e final value L1 m measured n control step .'- o offset 00 free incident flow '- 2

Original page 6

I', '\ ,; !, 1. INTRODUCTION /4* Because of the finite dimensions of a wind tunnel-test section the test results are falsified in model investigations. Adaptive flexible wind tunnel walls offer the potential to avo,id to a ,large extent such false results. Beyond that a blockage of the test section can be avoided during transsonic flow. The principle '. of the adaptive walls has already been successfully tested in different research institutes [1,2,),4J. These first tests made it clear that the time expenditures for control adjustments of the wind tunnel walls must be reduced if this principle is to be applied in commercially-used wind tunnels. For this a control device is necessary which can adjust the flexible walls in the shortest time possible. Here we find three problem areas: determination of the wall pressure distribution calculation of the wall contour controlled adjustment of the walls. J, In the subject report we will show how these problems have '~ been solved for the high-speed wind tunnel of the TU (Technical University) Berlin. ,j * Numbers in margin indicate pagination of original foreign text. '. ) "

Original page 7

I' , I 'y: (' 2. FUNDAMENTALS FOR THE DEVELOPMENT OF A CONTROL DEVICE L£ 2.1 Principle of the adaptive flexible walls The principle of the adaptive flexible walls rests on the .' controlled adjustment of the walls to coincide with s'tream lines. If the walls take on the shape of those stream surfaces which '; appear at this place during unlimited flow around the model, then the flow around the model is free of wall interference. Here we first assume for the sake of simplicity frictionless flow. Figure I shows a sketch of the flow field in which the walls were reshaped to streamline contours. Between the walls a real flow field is developed. One can imagine that this flow field is continued beyond the walls. In actuality this flow field does not exist; however, the assumption of a fictitious external flow field is necessary for the determination of the correct wall, contour. Under the assumption that the wall represents a stream surface equal pressure must exist on both sides of the wall since a stream line can take on no forces. Therefore, for the adaptive wind tunnel ," we have '. P = P. external lnternal. The internal pressure is measured, the external one can be calculated. It already becomes clear here that only the boundary values at the walls must be compared in order to obtain an adaptation of the walls. The problem to determine the correct wall contour is solved iteratively: as an example, starting with a straight wall we first measured the internal pressure distribution. The external pressure distribution is constant because of the straight wall. Therefore we have ,.J. p.lnternal I Pexternal - constant. - " 'llhe walls must now be deflected in such a way that the internal and ': 4 '>'/;'-. ,. .'

Original page 8

,e ,f the external pressure distribution approach each other, For this two possibilities must principally be taken into consideration Turning of the wall through trial and error (evolution strategy) Calculation of a wall contour based on the measured internal pressure distribution. ,- While in the first case many iteration steps are generally '~ necessary in order to obtain a streamline-type wall contour, this can be done in the 2. case by only a few iteration steps. seemed to be suitable I LQ The following scheme One forms a weighted average value between measured internal and calculated external pressure distribution and with it calculates a wall contour which would cause this pressure distribution in the fictitious external flow. The average value can be calculated as follows: ,J CPestim + = K . c + (l-K) • C Pinternal Pexternal (1 ) n l '. with CPestim' + = CPexternal + n l n n and O<K<l n l With this new estimated value the new wall contour can be calculated. The wall contour is then set as calculated and a renewed wall pressure measurement can follow. For a favorable selection of K it is possible to find an adaptive wall after two iteration steps. New theoretical works of Lo [5J show that a calculation of the ,- adaptive wall contour with only one iteration step is possible by the use of two measured flow quantities (u- and v-components) on a '. control surface. 5

Original page 9

" ., 2.2 Control algorithms As suggested in the previous section the factor K denotes the convergence behavior of the iteration resp. control.* An analytical determination of an optimal factor K, also denoted in the following " as control factor, is possible only for very simple llow cases; such as for Sears [6J. The problem here lies in the fact that the 'j entire flow field must be capable of analytical solution internally and externally which, however, is limited to a few special cases. Here we attempted in two ways to determine the control factor empirically. numerical simulation of the control with calculation of the internal- and external field (panel method, theory of small disturbances) experimental determination of the control factor. !.:L It is basically true that as the control factor K becomes ,J larger, a weighting of the measured CPinternal-component takes place at the expense of the CPestimated quantity. This can be '. interpreted as a larger back-coupling effect of the process quantity and, as a result, a smaller dampening resp. sensitivity increase of the entire control circuit. Simulaiion_on the cOmPuter Before the first experimental investigations were made, the control process was simulated on a computer. The thus determined values for K were intended to furnish the basis for furthe investigations for the determination of suitable control factors • ." For the simulation we had available, in addition to a program, (EXTER) * In the following we shall speak of control since the adaptive wind tunnel represents a back-coupled system. 6

Original page 10

for wall contour calculation, a program WINDKA which calculates the flow around a profile between two randomly shaped wind tunnel walls. It is based on the panel method and takes into account no viscous effects. Compressibility effects were taken into account by means of Krahn's transformation. The calculations were conducted " for an NACA 0012-Profile for various angles of incide'nce. Here the , ratios profile length to wind tunnel height resp. to wind tunnel " length of the actual ".rind tunnel were used as a basis. As an example, the controlled process for a control factor K = 0.5 and K = 0.35 is plotted in both figures, figure 2 and figure 3 0 for an a = 0 • It is shown that in general large control factors (K ~ 0.5) bring about a periodic, while small one (K ~ 0.25) bring about a monotone adaptation. It was found that K = 0.35 was a favorable value for the angle of incidence range from 0° to 10°. It is remarkable that 2-3 control steps were sufficient to obtain an adaptive wall shape. Ex£e£iment£l_d~termination ~ The experimental investigations were made within the framework ',. of a larger test series with an NACA 0012-Profile. Here it was shown that control factors in the range 0'.25 :5 K :5 0.35 produced a fast adaptation (2 - 3 control steps). Detailed information can be found in [7J. The optimal control factors found in the experimental investi- ~ gations are somewhat smaller than those determined during the numerical simulation. This differing result must be attributed primarily to the following facts: Compared to the calculated wind tunnel flow (program WINDKA) the actual wind tunnel flow is relatively sensitive toward changes of ,0 the wall contour, especially in the transsonic range. This ~ffect is shown clearly in the pressure distributions along the wall. A compensation of the greater sensitivity is attained by a smaller control factor. 7

Original page 11

" ,., An additional reason for the deviation of the control factors from the two mentioned methods rests in the fact that the program WINDKA does not take into account friction effects. 2.3 Theory for calculation of the external field .. The calculation of the external field makes possible the " determination of the wall contour based on the pressure distribution obtained with the aid of control algorithm. This calculation represents, in contrast to the generally conventional calculations in aerodynamics, an inverse computer method; it s designated as a design method. A solution for the design method is possible through the linearized potential equation [8J which opens up the possibility of relatively short computer time. Starting with the linearized potential equa:tion (1 - r4) ~xx + t/Jyy = 0 we obtain in the velocity components on the chord (y = 0) a ' weakly curved wall contour Yn = h(x); ,tt 1 dg 1 Jergdh x u (x.o} _= UQ) (I + it -g) '. dh Vex) = ax UQ) 0 In order to determine the wall contour h(x) resp. the slope ~, the above equation ust be interpreted as an integral equation. The solution was provided by Betz and Fuchs-Hopf as 1 L:l ~ =- f 1 E.t. (2) ~ = 0 Uoo ,0 '. 8 ............*'of" ; . ~., V ~ x(i-=-x) x-g '-

Original page 12

.. , whereby u/U denotes the disturbance velocity distribution on the oo chord. Between this disturbance velocity component and the pressure distribution there is here the relationship Cp = -2 u/U ' Since for oo the wall deformations we are dealing here with small deflections, one can set the velocity on a "chord" equal to that on the wall ,- surface. 'f Since equation 2 is to be used for a wind:tunnel wall of length le-Ia, we must first make a coordinate transformation with . ~ = '- la x x' - la und h = te=Ta le-Ia = H dg = ~ dx =-dx ' dh = dH 1= le-Ia I . l le dH -- I I u) l ":!::Tx'ox· = - -It I f/e-' f~'-,a~ I dg ' Ux'-Iate-x' . '=Ia co ... In order to take compressibility into account, one uses for dH dH KO f.1p I '. dx' = dx' VI-M 2' 00 From this we obtain le u(gl-Ia) dHKOMP = _VI - M' f -1- U"CO TX'-=hie-x' ~ d dx' · tt g' =la 2.4 Numerical method [(J.:f~ ~~'-Ia~' 1 , (3) Based on equation J we developed the program EXTER which carries out the numerical integration according to Simpson's Rule • .- For the discontinuity at the location g' = x' the Cauchy main value '. is also introduced: 9

Original page 13

,', 1"/{(((t?7'z Ilk X-!::a,1::' LZ2..V//777 ] ," The step width E· was determined empirically. For this we ill 'f selected a flow case which allowed an exact calculation of the v-component resp. the wall contour. During the numerical solution the step width was reduced until the error between the exact and the numerically determined v-component took on an order of magnitude of about 0.1. Here we obtained for'' a value of 0.01. As length of the wind tunnel wall we used here the length of the flexible portion of the walls which for the wind tunnel described here is 5.5 • 1 f.l with 1 fOl = 10 cm. The u-distribution pro ~ e pro ~ e . is determined from 2) pressure measurements per wall. Intermediate values for the numerical integration are formed with the aid of a ,. spline function. The calculated v-distribution and wall contour are also represented by 2) points. After an optimization of the '. program the calculation of the upper and lower wall shape acquires ca. 8 seconds. For this we must add the "input and output times. (In-out-spooling, output to printer etc). of ca. 2 - 15 sec. The indicated times were measured ona Hewlett-Packard Computer with a central unit of type 2ll7F and Floating-Point Processor. ). CONTROL DEVICE ).1 Total buildup In the following chapters we shall describe the electronic control device of the adaptive wind tunnel (figure 4). Sinc this device is composed primarily of computer resp. processors, no '. great meaning is attached to the software. 10

Original page 14

.., The block diagram in figure 5 presents an overview of the hardware buildup of the control device. The process computer constitutes the central unit. It coordinates the entire sequence of the control, carries out the data acquisition and calculation of the wall contour. A second computer (microprocessor-system) .' processes the wall contour data and releases them as ·analog values to the Servo-System which then adjusts the wall contour. Beyond that it controlled the motor current resp. torque in order to shut off the motors if these values are exceeded. With this device the walls are protected, on the one hand, from overloading, and on the other hand, the end of a readjustment phase is indicated. /11 3.2 Servo-System The Servo-System was constructed especially for the adjustment of the flexible walls. It consists of 2 x 8 individual Servo-Systems in accordance with the number of the positioning members. The requirement made of the entire system was, on the one hand, a fast adjustment of the walls, and on the other hand, an attempt was made to stress the flexible walls as little as possible during their adjustment. No exact statements could be made as yet concerning the adjustment accuracy during the development of the control device. From a size standpoint an accuracy of 1/10 mm was certain since the thickness tolerance of the flexible walls exhibited about the same magnitude. The displacement pickups selected in this way attained a resolution of 0.08 mm for a linearity error of 0.2 % referenced to 50 mm available stroke so that an adjustment tolerance of a maximum of 0.18 mm was attained. During most of the tests it was shown that this accuracy can be considered as sufficient. However, in the transsonic Mach number range an increase in accuracy by a factor 2 would be desirable. The wall adjustment speed is determined essentially by the motor size used. For mechanical reasons a motor with a power of 11 ..- ._-..----.-. ':.'.

Original page 15

12 watt could be selected which allows a maximum adjustment speed of 1 mm/sec. However, this value is attained only for large differences between should- and actual positions. For smaller values an rpm regulation reduces the adjusting speed in order to avoid, on the one hand, travel beyond the should- position, and, ,f on the other hand, to bring about a stress relief of 'the walls during the adjustment phase. Figure 6 shows the block diagram " of the Servo-System. The should-value of the position is given out by the microprocessor as 12-bit information and is converted in the digital/analog-converter (D/A-W) into an analog signal. The input amplifier forms the difference between the should- and actual position*. In accordance with the characteristic line of the rpm regulator (figure 7) the latter produces an output voltage as a function of the difference voltage which is led via the power amplifier to the electromotor. Its rpm is nearly proportional to the VOltage introduced. The motor adjusts the positioning member by means of a spindle. /12 ,. The instantaneous position is provided by a potentiometric displacement pickup. If the difference between the should- and the actual ~ position is large, then the circuit produces the highest possible voltag~.which corresponds to the rated motor voltage. The adjusting speed tnen is ca. 1 mm/sec. If the difference is smaller, the motor voltage decreases and thus also the adjusting speed. If the differece ~h 1. = 0.045 mm, the smallest adjusting speed of ca. '0.25 mm/sec m n. is obtained. Deviations smaller than ~H. produce no motor . m1n voltage. If in this range the motor voltage abruptly drops to zero, then because of the lower output resistance of the power amplifier, * The input amplifier is constructed as a sum amplifier; since should- and actual value voltages here have opposite signs, the difference is formed. 12 "

Original page 16

.. .', the rpm of the motor decreases greatly so that an overrun of the position is avoided. The adjustment speed approximately follows the following function: V - V • V = V. + max ml n •l::. H (4) l::.H l::.H s: l::.H · .' st mln l::.H·max -A Hmin ~ min max • Through this action it can become possible that, e.g., two position members which are required to adjust diffeent path lengths, reached their should position at about the same time so that the required low wall stress during the adjustment phase is obtained. The entire· electronics of the Servo-System, their current supply, and the controlled network portion of the electromotors are housed in 19"-slide-in units (figure 8). Above it the microprocessor system is located. It consists of the CPU-board, memory board, and 2 boards with 8 D/A-W each. 3.3 Microprocessor-process computer-coupling The Servo-System processes analog input valuf)3. '.Pherefore the . wall contour determined by the process computer must be converted to analog signals. First the digital wall positions are stored in a microprocessor and, coded appropriately, are paths to the li/A-H. The coupling of the two computers takes place on the process computer side via the process periphery* and on the microprocessoc Aide via a parallel input/output interface with interru.ption possibility. ----.......,'- .----- * Description of the process periphery (DEV systems 8.1 and 8.. 4) \laB also provided by the manufacturer. 13 ••_ ...... .d .",.. ....._ .,.... .\

Original page 17

Figure 9 shows the connection with the data flow direction. Ll1 The data are transmitted in 8-bit blocks whereby two blocks must be transmitted for one position. A cycle signal from the process computer provides the synchronization. After each cycle output the microprocessor sends out a occupied-signal which is led to the .' process computer via the input card of the process periphery . The process computer waits until the signal is returned by the " microprocessor (data have been processed) and then starts with a new output. Figure 10 shows the flow diagram. 3.4 Data acquisition device The data acquisition device determines the wall- and profile pressure values. In addition it must take on the control of the scanivalves and the interrogation of the pressure pickups. Figure 11 shows the components of this device. The DMS pressure pickups are supplied by the feed device with a highly accurate direct current voltage. It makes possible at the same time a displacement of the zero point. The direct current '. amplifiers amplify the low output voltage of the pressure pickups (~35 mV/+ 500 mbar) by a factor of IJO.This voltage is sUfficient in order to completely control the following A/D-W. Since the amplifiers possess an output with impressed current, a shunt resistance is necessary which produces a voltage decrease proportional to the current. In order to eliminate the line resistances between amplifier and analog/digital converter resp. multiplexer input, the shunt resistance was placed directly at the multiplexer inlet. Since the pressure pickups were connected to a common feed voltage source, a galvanic separation of the pressure pickup was necessary on the output side. This device prevented mutual ~ffects of the pressure pickup output signals and ground loops. The galvanic 14 ..................------

Original page 18

., separation is located between amplifier input and output. The amplified pressure pickup signals lie in the channels 0 and 1 of the multiplexer. One channel is selected per software and the instantaneous value is stored in Sample and Hold. The value if retained during the conversion time (8~s) of the A/D-W. The digital information of the A/D-W is recalled by the data acquisition program and is processed further in it. For the control of the scanivalves two signals are necessary: 114 step pulse and return pulse. The signals are generated via the software by setting resp. raising two output bits of the output card. From the short pulses the driver stage generates a 50 mslong step pulse resp. a 2.5 sec-long return pulse. At the same time an amplification of the pulses takes place for driving the scanivalves (24 V/4 A). 4. SOFTWARE FOR THE CONTROL DEVICE 4.1 Sequence of the control For the automation of the "adaptive wind tunnel" software was necessary for the process computer and for the microprocessor. While the programs of the process computer were written in FORTRAN, those for the microprocessor had to be written in Assembler. Since these programs communicate continuously with the process periphery, they had to take into account problems such as synchronization of the two computers, data coding, and interruption controls. Figure 12 presents a summary of the sequence of the control: The microprocessor can preselect an arbitrary initial wall c9ntour which is then adjusted by the Servo-System. After that the programs initiated in the process computer can be started by pushing a 15 ': ~----.... .'

Original page 19

button. The incident flow Mach number is indicated in 0.5 sec cycles by a luminous diode display. A Mach number correction can be made manually. A continuation of the programs for the wall- and profile pressure distribution measurement can be started by another push button. This is followed by a interrogation to see if the " control process is converging. At the time the judgment is still made by the service personnel. If no adaptation of the walls is achieved, a new wall contour is calculated and relayed to the microprocessor. This again causes the Servo-System to adjust the calculated wall contour. The process is repeated until an adaptation of the wall has been attained. 4.2 Microprocessor software L!..i The microprocessor system has the function to convert into analog values the wall contour data which are tr~nsmitted either by the process computer or are being input by the terminal. For that purpose 16 digital-analog converters are connected to the microprocessor bus. The program must control the input and output as well as assume the preparation of the wall contour data for the D/A-W. For the preparation of the wall cont'our data we must first explain the following relationships: The output signal of the displacement pickup lies in the range from -lOY to +lOY. The D/A-W is switched accordingly in order to attain the same voltage stroke. Figure 11 shows the displacement pickup voltage as a function of position. For a position HI the displacement pickup generates a voltage Uactual l' If this position is to be set, then the D/A-W must I' provide equally large voltage with reverse sign since UMotor = 0 if Ushould + Uactual = O. 16

Original page 20

With the aid of figure I) the following equation for the output voltage Uactual can be derived: Uist = tan 9 • H - U " o The slope tangent 0/ can be determined by a calibratio'n of the displacement pickup. The n/A-W must provide the following voltage: U = ~tan <j'. H + U soll o Based on the coding o'f the n/A-W listed in table 1 we must find at its inlet the digital information DAC = [- ta n Cf • H + U ] . Fl with Fl = ~ = 204.8 ~ o DAC = -Fl tan 'H + U . Flo_ ' _ , ____ , ,_ or nAC = -F . H + 0.2048 • U o • 3 (5) 10 ill The calculation of this equation is a part of the program REG2 described in the following. following tasks: It processes the additional 1. Input of the wall position by the terminal or by the process computer 2. Input of the values F and U o ). Lists of the wall position H, slope F and zero point U o 4 Output of the wall position to n/A-W. Figure 14 shows the program sequence of the program REG2. The program was written in such a way that the particular tasks can be fulfilled with a type of command input. The command 9an be found in the search- and jump table. " 17

Original page 21

•, After the start of the program* a double dot appears as an answer signal; a command input can now follow. The commandsST , OF ,KO serve for the input resp. change of the n n n values for positioning member height H, zero point displacement U ' o and transmition factor F of the displacement pickup. An optional parameter is n; it indicates. the positioning member number. If it is not given, the value of 1 is used. With the command LI we list for all 16 positioning members the positioning height,; zero point, and transmission factor. The command AG carries out the calculation in accordance with equation 5 and relays the value to the D/A-W whereby the wall contour H(x) is set immediately by the Servo-System. A process computer coupling is possible with the PR command. This coupling is necessary if wall contour data are to be transferred from the process computer to the microprocessor. After the PR command has been given, the microprocessor continuously interrogates the terminal and the interface to the process computer (bit 0 of the output card) for a signal. If the process computer (bit 0 set) signals the output of wall contour data, the microprocessor confirms by means of a ready-announcement- (bit 0 setting of the input card) its readiness to accept data. The data are made ready by means of the output cycle of the LlL process computer. The microprocessor interprets the output cycle as an interrupt and jumps into the interrupt program. The data are then acquired by it. At the same time an occupied-announcement is signalled to the process computer so that no additional data are sent. If the interrupt program has been processed, a return * Starting of programs is described in [9J. 18 .:0> ,"

Original page 22

,.., jump is made to the main program and the occupied-announcement is taken back. The process is continued until the wall positions have been transmitted for all 16 positioning members. When the transmission has been concluded, the wall contour, zero position, and transmission factors are listed similar to the case for the LI-command. If this wall contour is to be set, then·a "CR" signal must be input by the terminal. Otherwise another arbitrary signal ·can be input. After that the microprocessor waits for renewed transmission by the process computer. By means of an additional "CR" signal one can jump out of the command tlCR". The SP-command serves to terminate the program REG2. After its input the monitor announces itself. 4.2.1 Expansion of the microprocessor software At the beginning of chapter 3 it was mentioned that the microprocessor carries out control of the torque resp. motor current. This device is provided for the protection of the flexible walls and the motors against overloads. For this purpose it utilizes the nearly linear relation between torque and motor current. At the \0 moment this device has not yet been installed; however, the software has been made ready as an operating plan. Figure 15 shows the required operating sequence: The motor current of a motor is measured. If it is higher than the maximum value, an output line of the parallel input-output interface is set to 0 volts. In this way a continued processing (e.g. stopping of all motors) can take place. If the motor current should lie below the maximum current, the current is measured by the next motor. If all 16 motors have been checked, and if no excess current has been determined for any of them, an interrogation is made whether the sum of all motor currents (amount) is possibly equal to zero. If this is the case an announcement is given that the adjustment of the wall has been completed. After this the same control sequence starts anew. 19

Original page 23

" . This program is started about every 100 ms by a software- /18 interrupt from an interval-time pickup. The motor current is measured with a digital voltmeter with series-connected multiplexer and Sample and Hold which are tuned specificallYl t·o the hardware of the microprocessor system are carried out. 4.3 Process computer software Three tasks are essentially assigned to the process computer: calculation of the wall contours control of the pressure distribution measurement coordination of the control process. For this 3 programs were developed which will be described in the following. However, for an understanding knowledge is necessary concerning the operating system, the process per~phery, and their software. Detailed information concerning the running of the programs, the meaning of the variables, and the buildup of the data files can be found in the comments in the source programs. 4.3.1 Coordination program INTR Figure 16 shows the sequence of the coordination program INTR. It is a peculiarity of this program that it is called up by a hardware interrupt. During the rest of the time it is placed into a waiting series. Thus the process computer is free for other tasks such as servicing other terminals of other users. The program INTR is called up by the special process software (DEVSY, RTDEV)*. This first starts the program and enters its * Description of the process hardware is supplied by the manufacturer of the process periphery. 20

Original page 24

,I'· program name (INTR) as well as the process periphery unit (1) into an interrupt list. The program INTR enters the test parameters of several plate files (subroutine EXT). After this the program obtains a stop signal which is formed by a subroutine callup of the process software (XDEV). It is placed into a waiting condition and is not processed ~ further by the computer. A continuation of the program can only occur through an interrupt from the process periphery unit 1 (program Start-Scanner). This interrupt first starts a special program of the process software (RTDEV). It searches the interrupt list for entered programs for the unit 1. When the program name INTR has been found, the waiting program INTR is called up. It now operates further starting from the stop signal. First there is a continuous measurement of the Mach number and an indication on the large display. The Mach number can be adjusted during this time. When the "measurement" button is pushed, the program branches into wall- and profile measurements for the upper side. The program MESS is started for each measurement. In the '. subroutine pressure the C and u/U values are calculated. oo p When the upper side has been measured, the program (EXINT) is called up to calculate the upper wall contour. During this time the program INTR continues to run and starts the measurement of the pressures on the under side. During the measurement the program EXINT ends and records the result on a plate file (TEMPFI). After the measurement on the under side has been completed, the program EXINT is again started which now calculates the lower wall contour. In this case the program INTR waits for completion of EXINT. After the lower wall contour has been calculated, the wall contour is given to the waiting microprocessor. The output sequence is similar to the process 21 .. ~..

Original page 25

",,, computer-coupling described under microprocessor software. The program INTR can now be called off by the process software DEVSY so that it is cancelled from the interrupt list. Otherwise the program is again placed into a waiting state and can be started anew by a hardware interrupt. Figure 17 shows the time sequence of the 3 programs. In addition the sequence of the programs with the mutual data exchange is shown in figure 18. 4·3.2 Test program MESS &2 The program MESS allows the measurement of electrical voltages on the inlet channels 0 and 1 of the multiplexer with the aid of an A/D-W. For this the apparatuses must be plugged into the process periphery unit 1 (slot 3: A/D-W, slot 4: multiplexer with Sample and Hold). Beyond :that the program initiates the continuation of the switching of the scanivalves via the output card in slot 2. The program can only be called up by other programs through the EXEC calls*. During the call-up 5 parameters can be transmitted of Which, however, only one is r6quired here: 1. 4. 5. pause time in ms· 10 After the call-up the program first places itself into the waiting state for the duration of the pause time. This pause serves to equalize the time constant of the pneumatic system (pressure hoses, air chamber in the scanivalves etc.). Then follows the measurement of the voltage in mV for the channels 0 and 1. Finally the continuation of the switching of the scanivalves is initiated and the measured values are transmitted by means of an EXEC call to the calling program (e.g.:INTR). * See also RTE-IV Programmer's Reference Manual [llJ. 22

Original page 26

4.).) Computer program EXINT The wall contours are calculated by means of the program EXINT. It calculates a wall contour per program sequence. The start and the data transmission result from EXEC calls of the . , . calling program INTR. A parameter transmission during call-up controls the sequence of the program: 1. length of the data transmission-memory in words 2. number of the first u/U oo value ). number of the last u/U value oo 4. upper-!under side calculation 5· To 1) The part of the common block of the program INTR is made /21 available to the program EXINT. These are the field ranges resp. variables: X, U, UT, XHY, H, NWHU, NWHO, IWANU, IWANO, REFA Their meaning can be seen from the commentary of the source programs. The length of the data transmission memory is calculated in accordance with the following relation: length = number of the real variables . 2 + no. of the integer variables To 2) The numbering of the pressure-measurement holes in the and) walls and thus also the indexing of the measured values are made in the flow direction starting with the upper wallz pressure holes upper: pressure holes lowerz 1 -to 2) 24 to 46 From this it follows for the upper wall that the index number of the first u/U value is 1 and that of the last oo one 2). Correspondingly the numbers for the lower wall are 24 and 46. ': -ii:', ,- "~J" .. p... ...? j 2) ~. ;. " ~." ~ I' ~. ~.'l

Original page 27

" To 4) The calculation of the upper wall contour requires a 1, the calculation of the lower value a 2 as transmission value. The program EXINT first starts with the calcu~ation of the estimated values in accordance with the control algorithm equation 1 for one wall. With these estimated values the velocity distribution of the v-components in the subroutine VSTOR is calculated. The wall contour is calculated in the subroutine HKONT. The results are passed on to the printer resp. can, in order to guarantee a faster memory storage, be recorded via the spooler on a plate file. The results are transmitted to the calling program (INTR) in two ways: During the calculation of the upper wall contour ~ (5. parameter = 1) the results (part of the common block, figure 18) are recorded on a plate file (TEMPFI) which is read off at the appropriate time by the calling program. During the calculation of the lower wall contour (5. parameter = 2) the results are passed on by an EXEC call to the calling program. 5· SUMMARY ill The control device for the adaptive tunnel at the TU Berlin presented here has been in operation for one-half year. A series of profile investigations (CAST 7, NACA 0012) have already been conducted with it. Here we found a considerable reduction of the test times compared to those of earlier investigation with manual adjustment of the walls and value inputs into the computer. For a given test case the adjustment to an adaptive wall shape takes, as an example, 1 to 2 minutes. A reason for this short 24

Original page 28

.' . .\ control time can be found in the fast calculation of the wall contour which is determined by the computer type used and the optimization of the computer program. An additional reason can be found in the high adjusting speed of the flexible walls. A different measuring method for acquiring the wall ~ressure distribution would produce even greater reductions of the control time. Thus the use of an electronic multiple pressure measuring position switch would bring about a reduction of the measuring time from a present value of 50% of the test time to 1%. The adjusting accuracy of the Servo-System was considered to be entirely adequate during the profile investigations. On the other hand, the calculation method of the wall contour in the transsonic Mach number range is in need of expansion. As a first requirement boundary layer effects must be taken into account. The successful testing of the control device described here as well as its limited hardware expenditures for the adaptive wind tunnel must be evaluated as essential prerequisites for the carrying-over of this test technique to other, larger wind tunnels. 25 "', A. .'t.

Original page 29

,,' 6. REFERENCES /24 [lJ J.-P. Chevallier, Parois auto-correctrices pour soufflerie eme transsonique 12 colloque d'aerodynamique appliquee ENSMA/CEAT-Poitiers, Nov. 1975. [2J M.J. Goodyer, A low speed self streamlining wind tunnel AGARD-CP-174, 1975. [JJ W.R. Sears, R.J. Vidal, Interference-free wind tunnel flows by adaptive wall technology. ICAS-PAPER No. 76-02, Oct. 1976. [4J U. Ganzer, Wind tunnels with adaptive walls for the elimination of wall interferences. Zeitschrift fur Flugwissenschaften und Weltraumforschung Vol. J, 1979, Journal 2. [5J C.F. Lo, Tunnel Interference Assessment by Boundary Measurements AIAA Journal 1978 Vol. 16, No.4. [6J W.R. Sears, A note on adaptive-wall windtunnels. Journal of Applied Mathematics and Physics, ZAMP, Vol. 28, 1977. [7J J. Ziemann, The convergence behavior of the control of adaptive wind tunnel walls during profile investigations in the high angle of incidence range. ILR-Mitt. 66, TU Berlin, Jan. 1980. [8J H. Schlichting, E. Truckenbrodt, Aerodynamik des Flugzeuges (Aerodynamics of the Airplane) Springer Publishers Berlin/ Heidelberg/New York, 1967, 1. Volume. [9J Texas Instruments, TM 990/100 M Microcomputer User Manual Dec. 1977, 1. reworked edition, July 1978. [lol Hewlett Packard, Batch Spool Monitor. Reference Manual Part No. 92060-9001J. [llJ Hewlett Packard, RTE-IV Programmer's. Reference-Manual- Part No. 92067-90001 26 '- ·

Original page 30

-.

Original page 31

·' Table 1 Coding of the digital/analog converter D/A-W outlet voltage hexidecimal decimal - 10 V - 9.9951 V OV + 0.00488 V + 9.9951 V '. 28 input information 12 bit 800 -2048 801 :-2047 000 0 001 1 7FF 2047

Original page 32

.. - ....... -----.... ---~~... ... .. -.-.- ..:,- ..-.....-.-- - -..:••. - .. - ......._ . ..-.................6. ... __ .:... ...:..L.._"' .. --- ....__ ..__~ __ •.;....1&..-....,.. ------------------ --0--.B ------------- - " wall ---- I r t I -===-= " A internal field, actually existing flow B,B' fictitious outer field, :imagined flow ",---- -- - -" - . "start: given: straight walls h=constant l wall with wall oontour as l:x:>undary oondition uc=constant for field B,B' ......... - / r '. setting the walls to the contour h 1 I f I inverse flow field calculation I or .field B,B' e.g.: . 'I , v = f (u ) h = vdx , c I I . . ili . t distribution new u l.str ution as 1 rreasurarent of the pressUre in field A on the walls CPm' ~ .... , velocity distributi fu = Uc ? - ,.)---------------- A and B resp B' equal? );es end: ''\ free-flight conditions .~"..,...............,-.- .., "'"-... ._ ..._ 0.l:x:>undary oondition for field calculation IB ,B 'use u t1c I t oontrol algoritlm: estimate new u-distr;iJ:x.ttion e. g • i\n . t L u-= K + t1c (l-K) ~ I no _ . t * TKS: theory of small disturbances 29 ': Figure 1.

Original page 33

,- ,N. • * ;~ .. "...... '-,' / /'-,'. .--t4___ ----.......:.""'-:.. D. ~ / innm II I I / // 1/ / .. / I .i // ',- /' .// ,,/ '. // .,../ -100 o Carputer simulation of the control J 6 control steps f - r I I ~ .' , 4 _ t. (QUa .. i 777$" -. , control factor 0.. 5 • ~ \ \ \ i \ I \ \ \ " "" I. I ~ " " " " l ,!. " ............... i f -- i 200 300 I .r positioii referenced to profile nose, in mn .- 1 t . ,., Figure 2· , ..-' / #-,..----_. --.• - .. - .•-"",,,,,,..- -tift.; pc --..-.-.....-.--- .. ...,.. -.."""~... ---.....----... - ..--..---.--.----- .... --~ ..-. - ----.

Original page 34

I f I I ~ - • ( .:.=~~ wall deflection innm 2' I 1 i i ~ @ , -200 -100 a i Catputer simulation of the control -. :J . . " ¥ WN \ control factor 0.35 t ~ r I , • l !, t I i I , t 1 t t I I I 100 200 300 position referenced to profile nose innm . control steps . Figure 3 ,.;, 3 w t= .-- _. ___ ...__...__ ..____ ,_ --.- ---....._ .. ',",'9' "-""":Ic-,-.''''::.''-'''- --,........-..... - ....- ••••• __ '__ .<1. ... __ ._. - --- ----------....-._-

Original page 35

.. t;trl ............ , ...~ ·0 ..-'_ ...............,.... ..:.......... --....... ·..""··.··........ .~ ..."·.4. _•....__•..•.•.,. ____ ~&4. ~. ...... _. .' r-;·...r;.:;··'···"·r-·....,........,..···;'"7··1 . ·j,f.lw;;~~....-z.rz;·L.;:;:·?k· .......... .l- • •• • of . .1- 8 displaCE!IEI1t pickups J- . Vi" .. • •• • III j. '.'.) -:>. .~ "J" - electric rroto~ ;.. .. ' .•' I ," !:.t • . .:. i~'"'-' , ,.-. .;'W3 /CJll .•.m:...._......;J ! . .. , . • -. - "'. ~ •.• r .- "16:::·,. -,.. I')t· - positioning- rrember :- . ; / <,\1' . ;' ,...,.- .. .-;- I • • ., ...... ," ~ .A.i'.·.'-7.!,:::~~>._.: -flexible wall fr~~' :.fi I " .J '., t '1'-'. l .: t'" 1ci' . \ . ~ p " - - '"" ''''::~~~'!f;~ fl. profil~.l,lpWr:t .- . . ----- ''- I, J ::l L wall pressure-pickup holes .7mf"P-~~ ." .i. : I; , '. ·.l' ,,~.. ,".: ".".....'::.'.;';f' '.'. t.. ~ I' , ,I, :".,.J;' ., ~ I I !f' ,}. ~ .' "',,'-"'ii;- .... ,, .j "': ''tl • . -',:':.. ; I , : I:' t. . . ... .. t: ; .. L. .... ,.,1{!lL LJ "--"-"'.,-",! '~.:. I (:V. r ".. : " [' ~ :. . : a,;. , . " " . '--.--.' ! . __ ... .\ }_. .-t.,. .........-.-•.. ... t ...... ...- . r· . .:.. .".. r • I 'i :: ; , , • ~ H" , r.. .wii:.::l f I l i Test section of the ~aptive wind tunnel 32 ...........""..... -...._---- ~ .' ,Figure 4

Original page 36

! ,1 • I, ~ ~ ! l r tes~ value acquisition AID converteJ: i -. f.. .. I ~ f \ ~ ~ - r. - \ ~ process carq;>Uter displacement pickup IA I p;:>sitioning rranber Terminal 1M I . actual value --, I ____ . D/A converter --c-t ,:.' I froproces,sor 1·.--:-- -;(£Wd-~cont:rollr ' • I I STOP + . Block diagram of the test installation: adaptive win:i tunnel ~ w w ) ~ f Figure 5. ,i t t ~ f t ... -.-- _ - ..-- ~~". ,...-:""-·:-~l"·.- ....,.'.....\ '-.'"-1"- ...... .. .r-...._ .... ......'.........-- . ", ._.", ~ ....".... :;. " ....-......,...... '-,...,. ,,."""'..- .. :--:-J...... .,lo..· •.,.. )".""""- .....--.,.--...--------

Original page 37

-------- --- - ~ - - -- --- - - ---- ,;...,j t:=" displacerrent f -. r • ·1 I ! - - -- --------------- ---------- - -- '. \ \ t \ l .pickup '14t=== I 8 ~ o inlet-sum amplifier Re100kn Block diagram of the servo-systan 1./ -Uref UM flUMI q:m controller' ~ stage UM =!24V max 'max =4A i • ~ i· t ~ Figure 6 i ! .. ____ • _____• __ .. '"'_____ 'J",.......,................--:---..-:........t .. t,.:'tIJ''' ..... r .. .....,..'I;., I"f""o.".' ... .......-.."C ........ ,·_,_"""<-··....,!oIt...··.... ·(·'- ... "1'''-<.......... "_. ,"-,1""". ~ .....A..:"':; .....,./,J.'It'l".,.. -····-··..-~------

Original page 38

_ .... ... h •• ~ . • ......~••..h' .. ..........I .. ~~ __ . .._ .... ••• - ..... .... _ ... .-..I~., .......,, .. _ .•.,. i I I 1 i 1, j c 1 °E l, +-:I: > U) ·1 I <J 1 • > > o~+-... ~ U) :c N l .. <J I 1 , )( o E :r: .. <l ; I I . I \ \ , I • - . 0; 1 J - °0 1 i l!'\ (Y) ..... Q) ~ o~ IZ.! E , l........._ E Ll) E ~ 0 E d Ll) II II c °E ~ :C:C <J<l . ~ r-I r-I > ~ I ~ N § I ~ ~ II-l 0 o§ r-I on rn °i j !

Original page 39

" . .~•.• ,. ..... _ ......_ ..o..J ....!.•. _____ ......... "' ... ~~~............. ~'d ..... ----...;.-- ' ....... ~- , ... ' FI:"11\1'. .' ':....->•••• .,....,. ...- microprcx::esser .. ". "rJ-:-- ' . ·f·........... :t.'1 : 'l·.::.-..-.'.'.: r-',-;"".";:'. ..- -. -... •• . CPU-board ' . :--.-- ;.i .- '. --.. , I 'j'; ,":-'-' --i ,,"::::: '.. :: ::'. . rrarory board " • o ' .') .0. ..' . l' .' .'.. ' { ... _ 0 ,. !. } .L....-:..: ,I.;';.' .;. C -;.",', ,:1 :p digitci1./analCXJ conv;erter .. . -;:-: .L..IL.-'-."H~-:~~~: .' '..:..J,J; - • o r"''-='"""""-'-' '.. ""'''"::r'''''''-1''i" control unit -'in1etamplifi.' ;. I . I ' .. 0·.... :~~';--',<~-=.. :. o rpn controller) regler) . ~ I } ,i .' .J ~. .J' I ,t> l ~,I J' ~ i .- ....:.. ... .. , - L. t T . I' I I , , ! I f (lCMer anplifier f .,..J I J ··.···'w" Ji I . QJnerk parts I· '....''. 1'1 . . .,... r_"";''-- ." ,- - '. : • rI,~~I:......... . ,'. ! ., !" •... .,. I wi ........ ....;.,.........o,",,~~,I. .... : I I •• • or · .., ""~1'U''' ••• e'. ". ~ ...-'"T..,.....b'"Y'r"<":.;........!Q. •Tt\ .. ).·4'...-.,..-..... \ "J"''-:c '\ A ;:: ': ...-.t.;.,.~~ ! 0 .' " "' '....: I • I.iC 'i ,':- :'j"" ......,,...., ',I I ",j ~~~ ,:)1, ~ i t ':. __ ...., ~ ,4- ,;' '''' I .:'•.•:<; ",,;;:-(-:- t ':,. ~~:'I,·,J \e\ ' ........ • :1. .. . " I . :...., o '! 1"~ ..": ': -. . ': j-.:- ." ',~ '. '.::." I;' :j! t ... -.."'J',. ,r;L I '.:",-'. ." i... -f o! t f'- . 'r. . ~ •• ": .! ~ ,I ~~ ,,. •...'''..- ..,------.. .. .. f 'Ir'L .;... ;/I ;i .~~~.: ':"'' -,~~: .. , { .,$- _ \'-Iv -:' ...., (I. .J .;;; .. 'i' ~J.- . .l , ....... "0' ". ,. ~~ .• ' G' '.11) 2 ~ ~;«., ..;..........-J -.'1/ llower amplifier ~~~. ~ :! f ''';.;;;) ': .... control card I ! Electronics of the servo-systan 36 Figure 8

Original page 40

, . ........: ..!.I...... ~ ....... - .... ~~ ., !II.--"'_ J~ ..._ ....... -A..""'_":' _ ......._ ....._ ......c~.' ....~ ....' .•••.. ',... _ ......ao.- ...... ~ ... ..._. __ ......_. --.. "'"!~~•.• , " / .J I ; , : ~ . process periphery microprocessor • outplt card 16 bit 8 bit data :i cycle , 1bit input card ready-·: '2J. bit .. , ~ Process canputer-microprocessor coupling Figure 9 '. ., , ! , ~ i " '.'~ , . \ .; 11 I I I ! . ... 37

Original page 41

, . ~ #"".. l. ".~ ""p'M4) , "...-:. 't~ ........ -+6 .... _ ................_ ~~4~ ..........--............-. t- --. ~~ ,data process canputer , I { cycle. " ,t r 'I, l I ! : ! : I: I data valid I I ! t r i Signal flow diagram of the data transmission fran process canputermicroprocessor J8 -.;....-.. ...... , ........- ... .. -.. '. Figure 10 \ ~ \ , " r t. ! t I I, ~ I .

Original page 42

..~ .. ..-- ......- -../"., ..-=-=- .'..$ "d''''' ......-............ .... hW ....... ..... ••.....__... _ ... .. ..- .......... :... o.L&.o.-..Io_'- ... _ ...____ _ -:-.t ,- I ,l I i t I f t AID converter I IlUlltiplexer outJ?Ut card 'I process periphery 12 bit 16 charmels 16 bit of the , Sample & Hold pressure piCkup channel 1 profile " pressure picklip ~ 1r r1 i , feed device with zero-point:- I Idisplacement process canputer , :.s CJl iHI ~ driver stage I ~ . i I Scanivalve I, . '·1 I .! -I ,! , i ! , lEasured value acquisition Figure 11 '. -.......------ .....--- .....--,- ... ---~-- ....-- .- '-, . " I I. i • 39

Original page 43

................-. , . ~~CV"'" . ••• . ,.-' __.......01.-. --"....... _-"':..__.... _ ........____ ...-'-......._____......-. .. __ .. . _........ " ...... .. - t • I I , I (STAR-i) .1 ~ r oreseIection -of-rmtial ~ wall contour H(x) through inr:ut to microprocessor o adjustl'oont of the walls' 4 Control sequence plan 40 , ! ! '\ '( I I I f .1 .1 ( . i I I ; I t I f r Figure 12 :--; ....- ... -.~~~~.,......----.,.......---... - .- --_ ... ,.__.,,' '----:-"'"7!"'"

Original page 44

.....~ ..- __ 0'--"-- --.-. - .....- -..... - ......-----.:.-'--- --..: .-_...............-.--.--.--.--------..-.-----... 1 I ) 1 j t 1 I 1 t ~ ~ U' ~ om I > 1 S2 i " (J i I /' I .....';.. - ..--...--.------- .. -. -.- ..... --.. M '. r-I '. Q). .~ .~ 3:~ ~ J: ] > ~ CIl' I d a ..f' -g ro. 0 ::! iJ m ~ !::1 ..~ )( .~ 0 E J: 1 I .~ I r-I Iffi' .~ ., 0 --------------.-------.-----.- ---.--.-"--.----------.-----.-- .. --..-.. '. .j I.. -'

Original page 45

lo_ .. ,.,... . "') ..,-.,- --...--.-.......-.• ,.-.---..•.• ............ ,., ...2. ..."""_____.......:.1...-----..... ..-- l .S I Ala" .' REG 2 AQ inlilt: cciiitiaIrl., (I) search- and jl.IlTp table: ADR:A i ST n change position height H OF n change zero point U Such- und Sprungtabelle: I n: positioning ne{Iber ijo. o ST n Stellhohe H verandern fj) ROn change slope F orn Nullpunkt Uo verandern ' AG output wall contour- to- DAC On Steigung F veruncern PR process computer coupling AG Ausg'Jbe Wcndkontur all DAC PR Prozeflrechner- Kopplung . SP end of the program SP Ende des Programms ~ LI listing of S, Uo, F 1I listen' von S. Uo. F 1I 9 H.U()oF SPC( display all values . '1-16 S,U , F /.. rtum .j.; . 0 I L m rromtor I read new value \ r 'I -T ' , "r_--,/~ '.., jurrp to A: Rffi2 I ,. -1 jump to ADR -- I .-•. 1 I I P: te'" al i inoot tran :rnun I d:;;::: '. _ I « - -- I I ..".- . _ ..:." yes , outPU~ of all Vcus to OAC "CR" fran tenninal? , - L· - -- ,. T '"/' ~ jump to 'A j~ to ADR' I h-; r no L.' '.-1 - lo gna1-'{-rall process ccirputer?-' . ...... ..,/' ~~~. yes ·1 .. i I I trt -i6-'ues ----., \ pam) .~ J.~~.~ LI ADR;D inter&t, I - i I ready-armouncement . I .t 1 occupied announcement to process canputer ~ . to process ccrrplter . r .---.I I put f tenninal . I acquire data) I _ . 8 bit .. .: I interrupt program J yes f start ~ i. .r endof, .::-:.:-.:: addr-'.' >:-.-.,J" , jump to I return t I yes :interrupt program -Jump ess ADR:C ........--- .. ~ ,AGADR:E I \ jump to ADR ' . '. ....:.L E fl,.."'----' 1jump to IE. ,-'--.~..... I ! I £ I ': microprocessor program REl32 F.CJll!e 14 I Program sequence of the t""""-.- .. -. r-.-","'' .--.------.-.. -- ..... --- .. '-- ··-'r·-·.·.--...._-- ..........'!""""'-.. . .-. 42 ...... .;'- '.

Original page 46

, . . ;..--- ...,,' -"---""---,"-'~'" - .......--...-.- ... -'. ----...-.-.- .-....-.................. -.- ..... . .' c-_n , ready-announct no are all notcrs positionerl ([ IM(I)~O) ? r I l ,r I, t i rreasurarent of the notorcurrentby positioning rrember Noo1 I haVe all notors been yesxreasurerl (I=16)? 'Ibrque control .yes i I " , erase output bit I .. notors S'roP ...1 j i ! 1 .\ I j I • B I I l t j f 1 '1 I Figure 15 I I '- 43

Original page 47

• :.."\ .~~-... -. h" ..........._ ............._ ._ .....»._ ...... :............. _ ........___............ "' •• w _ •..••. •.•. OUJ.J.Luutine EX'r reading-in test pararteters i _--lI...1 IN1'R is put on hold by DEVSY tand versetzt [-- --InterruPt . .A <lnt by yes ........... slot 0 ? - >-- /" end IN1'R .tine Mach measure Mach number and display - interrogation bit 8 -. bit 8 erased?:--, (Mach number set?)1 yes initiation for wall and profile measurarent: '--- u side --I start of the program MESS: pause rreasuranent pressure pickup voltage switching on the scanivalves end of MESS, 1-- -ra I __ .. -_ .. ,....:...-.:-----I subroutine Druck (pressure) calculation of the pressure coefficients, disturbance speeds 1.0 r-l Q), ,no, ur:per wall an::l profile -g, side treaSured? .,J .r:i !il, ~ -caJ.lup of the program EXINl' for' I calculation of the upper wall contour i initiation for wall- -an::l profile measurenent: . lower side' I .....--- i start of MESS: '. pause IYeaSurenent pressure pickup voltage switching on the scanivalves L end of MESS I --=--~;-,-I---:- --calculation of the pressure coefficients ," I disturbance speed- I, ~ I yeq callup of the program I EXINl' -- I Cll I. _for calculation I :S of the lower wall contour ~ m' Ausgabe r-l ~ output of the wall Q). contours to microprocessor i ffi : yes II) -"t. ".&-d---=---r---" T J :"4 . ,

Original page 48

\ II • . , , .... . ~ , Programm I • . rrea.surarent of upper side measurarent of upper side calculationJ output, reading-m' - "'--.-12~~~'--"- ~and ction-o~- :ii-seC -·....i-.Of.,...·lower·sidel ::.lsec r- HI C 1 id H r:: It sec I I INTR I l H-------H H H-- ower s e--H I ( I II II I I I I I: II I I I I I. II I I ( I II II I I I I II II I I I I· II II I I I I II II I I I I I( II I I..! ( I II II I MESS VH H H Hl1 ti H t' 05 I I I I • sec I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I ( EXINT· ~ ~ f--t parn active I II I II I II I II I II ~ I I III I II I I ,,. · I II . H H ( ·I I I I ~ • I • I • I I I ! I t I · I-----------; t t ~ r · Tima sequence of the three programs for the process: adaptive wind tunnel -"__or-_ __.._" __ ' "'---,!.-.--'.'.' Figure 17. . .... ---. -'.- ••• _-" - .., ......-.-.......... · .... r. ' __ "

Original page 49

,. . ..--............... .• " • ""....,J4_,w,•....--.....'-.....~.-- ~...... .. . .....,•._ ....... - - ~ ..;. ...,. ' ~~ __ _ ..•:.-..J .... • ..:...-........ •••. __ .............' r Ji,~TR -. MESS DVM (1), DVM (2) Buffer 11" ,A DVM (1) DVM (2) INTR N' 1<23 Common-Block 468 rds ... Buffer f- 23 0. . ! EXINT MESS DVM (111\ DVM (2) v A INTR 1<46 J ,'l Common-Block ~ 1=46 A EXINT l'l Common - Block ~ ... INTR ". H TMS 9900 y ,A I'· U,H ~ t· file Buffer l. TEMPFI . I , ! i \ j Buffer I t j ~ i v ! 1 .1 f f I I I I Buffer I ~ y I I j I ! ! ! Figure 18 Data transmission - Organization of the three programs for the process: adaptive wind tUIU1el 46 .....,:, ".6''':-- ,...,...4_~1__........ ......... -: - '-

Original page 50

·. I J ~ . ...

Original page 51

, . Technical data for the individual systems of the electronic /A-l control device 1. Servo-System 1.1 Direct current motor with permanent magnetic field rated power 12 W starting moment Man 8.83 Ncm rated torque M 1. 91 Ncm min-l rated rpm 6000 power pickup 24 W type of operation 100% ED rated voltage 24 V 1.2 Spur gear drive, displacement spindle gear reduction 1 : 106 maximum torque 300 Ncm spindle M 10 x 1 1.3 Displacement pickup, potentiometric total resistance 2 K.Q non-linearity 0.2% resolution 0.08 mm available lift + 25 mm resistance material chrome-nickel-wire 1.4 Control part - power amplifier - electronics maximum outlet current 4 A minimum outlet voltage 6 V maximum outlet voltage 24 V switch threshold 22 mV limit device onset 1.1 V A-I 43

Original page 52

, ' \ ~ dead range displacement pickup feed 44 mV /A-2 + 12.5 V transmission factor displacement 32 mV/0.08 mm = 400 pickup* layout value voltage supply control 'part + 20 V voltage supply power section + 30 V 2. Microprocessor system 2.1 Microprocessor TMS 9900 (Texas Instrument Inc) data bus width 16 bit memory size system cycle input/output interfaces cassetteinterfabe (KCS-code) 2.2 Digital-/analog converter 5 KB 3 MHz 2 x in series 1 x parallel 16 bit with 16 interrupt lines resolution 12 bit calibration 4.8828125 mV/LSB input information via DMA 2' complement output + 5 rnA / + 10 V non-linearity .::!:. 0.01 % 3. Process computer 3.1 Central unit (HP 2117 F) data bus width 16 bit memory size 256 KB operating system Real-Time-Execute (RTE IV) plate size 2 x 9.8 MB * see also Table A 1 A-2 49

Original page 53

· ~ 3.2 Process periphery (DEV-system) lk.J. Slot 0 input card 24 bit with interrupt Slot 2 output card 16 bit Slot 3 analogidigital converter' resolution 12 bit inlet range + 5 V switching single ended 2' complement conversion time 8 ~s Slot 4 multiplexer with Sample and Hold no. of channels 16 accuracy 12 bit 4. Measuring device 4.1 Pressure pickup pressure range + 1013 mbar outlet voltage max. at + 30 mV 10 V feed ? 4.2 Amplifier, galvanic separation gain 130 inlet range maximum .±. 1 V outlet max. + 20 rnA shunt resistance 300 frequency range o to 10 Hz 4.3 Driver stage inlets 1 TTL-Last outlets 24 vi 4A return pulse duration 2.5 sec step pulse duration 50 ms A-3 50

Original page 54

·.' !. .. Table A 1 Displacement pickup - data /A-4 Positioning max. Uo/mV calibration transmission member lift/mm for straight factor F factor mV/mm No. walls 1 5.18 -1120 2 51.16 -1348 3 51.56 - 795 4 51.16 - 773 5 51.05 -1500 6 51.26 - 214 7 51.26 -1502 8 51.66 - 494 9 51.21 - 538 10 51·31 -1558 11 51.16 - 965 12 51·31 -1355 13 51.10 -1690 ~ 14 51.05 - 746 15 51.16 - 866 16 5·20 -3417 A-4 tan tp 330.89 1615·67 86.25 421.17 92.47 451·5 89.84 438.67 88.75 33 3 above 96.46 4471.00•1 87.07 425.17' 90.28 440.83 90.04 439.67 86.97 424.67 89.63 437.671 below 87.14 425.50 82.94 405·00 92.77 453.00 88.00 429.67 323.86 1581. 33 51 " , t " ",

Original page 55

...' ,. '-" . ' ". ?' t .-.;,................... ...,r ....~ 'ed?~ .'. ,-- -- _ .... _.Ao _-....-..JW.. ___ ....:....... -....·M N'" « .,..................,..........-...:......._____._ .. _. . - AS- , process periphery I ! I I i interrupt lIlIll.t . terrupt lrogram start) -- . lIl Bit 8 Lcaro .- ..-' .. micrpprocessor systm Slot 0 Masse Bit 0 input/ootput I start of m 'arent I Wclil and p .e I paralll I I . ! ,. .. I CP Bit 0+7 ~ Interface ! i -y IUtput card cycle Interru ptdecoder I Masse • I S' ve Com f / Slot 2 - driver Bit 14 - Bit 15 " VI> converter ,.' 12 bit /8 ps i Slot 3 ~~~." f- Ul ~ g R' 8·~ III multiplexer 16 chaimels __--'1 --I c:::J.-.... .. KO 4 ~ Sample & Hold T 1F 0300Q 300Q Slot' I Tl~ V Kl ~ c::J 10ks} output _card ..-- I Slot 5 Bit 0+7 )I Steo I staqe Home \ , C anplifier-wall' C mass C arrplifier profile • I I • !. I I ~ Display I SUrnnary of the entire hardware wildup 52 \ - ,;. &~ ;A;;\

Original page 56

.. ~ ,.,..'." ~ ( Terminal Plan for the Hardware Buildup /A-6 Apparatus meaning terminal designation input card bit 0 A " bit 8 1 ! " interrupt S " mass 15 output card ready- S announcement " bit 7 J " bit 6 H " bit 5 F " bit 4 E " bit .3 D " bit 2 C .J " bit 1 B " " bit 0 A " mass 14;15 " bit 14 7 " bit 15 8 multiplexer channel 0 C S & H channel 1 .3 mass 10 analog from N ~ control 13 signal * microprocessor A-6 terminal meaning apparatus designation 20 outlet PO. plug p4* start of remote control the meas. program " start mass " 8 Interrupt p4* INT5 40 high-bit P15 p4* .38 P14 p4* .36 Pl.3 p4* .34 P12 p4* .32 Pll p4* .30 PIO p4* 28 P 9 p4* 26 low-bit P 8 p4* 9 to .39 mass p4* step {driVer stage return step scanivalve 4 wall press. apparatus plug pickup 6 profile " pressure pickup 16/18 mass " L analog in A/D converter H control " signal 53

Original page 57

m :::J c.. o -h C o n s::: 3 tD :::J f"+