US2828070A - Electric computer - Google Patents

Description

Marph 25, 1958 D. M. BOYD, JR 25828070 ELECTRIC COMPUTER I Filed Dec. s1, 1954 David M Boyd Jr.

A TTOR/VEYS.'

ELECTRIC COMPUTER David M. Boyd, Sir., Clarendon Hills, Ill., assigner to Universal @il Products Company, Des Plaines, lll., a corporation of Delaware Application December 31, 1954, Serial No. 47 8,97 1

2 Claims. (Cl. 23S- 61) This invention relates to an electric computer for simultaneously solving inter-related equations and in particular to an electric computer especially adapted to making tray by tray temperature and composition calculations for designing fractionation or other multi-stage equipment.

`Many problems in engineering and research require the solution of long, complex series of inter-related mathematical equations in order to arrive at a proper design for equipment or to determine the course of a reaction. Frequently, in the solution of these equations each stage of solution depends upon the solution of the last previous stage so that an ultimate answer to the problem is dificult to obtain in a short period of time since the various parts of the problem cannot be worked on independently by several people. Typical of these problems are those associated with the design of rnulti-stage equipment such as fractionation columns, multiple effect evaporators, liquid-liquid or liquid-solid extraction processes, etc. The design of the equipment for these processes requires long hours of routine work which must be accomplished by highly trained men who could be better used for more creative problems. he laborious manual solution of complex mathematical equations also invites human error which frequently requires recalculations to be made or alternatively incorporates mistakes into the ultimate design. lt is an object of this invention to rapidly and accurately calculate such problems by electric means.

To accomplish this object it is an embodiment of this invention to provide an electric computer comprising in combination a constant voltage source connected to a plurality of identical parallel circuits, each of said circuits comprising a first voltage dividing means connected to a bank of parallel first voltage multiplying means which are adapted to be preset to speciiic multipliers, suitable programing means for alternately inserting pairs of consecutive lirst multiplying means into an interpolating bal-I ancing circuit which adds the voltages to similarly obtained voltages from the other parallel circuits and balances the circuit at said constant voltage when one of the sums of the consecutive pairs of iirst multiplying means causes one voltage greater and one voltage less than said constant voltage, recording means to record the position of balance of said interpolating circuit as the proportion` of the dilference that said constant voltage is from said greater voltage and said less voltage, second multiplying means adapted to be preset to specific multipliers connected one to a fixed voltage dividingmeans to multiply the iixed divided voltage and another to multiply the voltage obtained by one of the pair of the aforesaid lirst multiplying means, a second adding circuit connected to said second multiplying means whereby their voltages are added, balancing means connected to said second adding circuit and to a second voltage dividing means whereby the resultant sum of the voltages is balanced against said constant voltage to obtain a voltage equal to said sum from said second voltage dividing means, connecting said tates Patent" y second voltage dividing means to a bank of parallel riice fourth multiplying means which are adapted to be preset to specific multipliers, suitable programing means for alternately inserting consecutive pairs of fourth multiplying means into a second interpolating balancing circuit which adds the voltages to similarly obtained voltages from the other parallel circuits and balances the circuit at said constant voltage when one of the sums of the consecutive pairs of fourth multiplying means causes one voltage greater and one voltage less than said constant voltage, connecting said second interpolating circuit to said recording means to record the position of balance of said interpolating circuit as a proportion of the difference that said constant voltage is from said greater and said less voltage, fifth multiplying means adapted to be preset to specific multipliers connected one to said xed voltage dividing means to multiply the resultant fixed voltage and another to multiply the voltage obtained from one of the pairs of the aforesaid fourth multiplying means, a fourth adding circuit connected to said fifth multiplying means whereby the voltages are added, balancing means connected to said fourth adding circuit and to said rst voltage dividing means whereby the sum of the voltages is balanced against said constant voltage to obtain a voltage from said first voltage dividing means equal to said sum from said voltage dividing means.

Since the apparatus and method of this equation are particularly adapted to aiding in the design of fractionation columns it Will be described in relation thereto, however, it is to be understood that this choice of an example is not intended to be limiting on the apparatus or its function. of a series of identical parallel circuits for each component of a mixture to be fractionated. Therefore, a two component system will have two parallel circuits and a live component system will `have live Parallel circuits,

however, some systems contain hundreds of components,`

for example, a petroleum fraction, and in these cases a circuit will be used for each boiling range desired instead of each component. Each of the parallel circuits representing a component or a boiling range is comprised of a voltage dividing means for establishing a voltage proportional to the desired purityof that component and a series of parallel voltage multiplying means which are programmed to function as consecutive pairs until a proper pair is in the circuit. That is multiplying means 1 and 2 Will function as a pair and then multiplying means 2 and 3 will function as a pair, next 3 and 4, etc. There may be as many multiplying means in each of these banks as desired, the number being chosen in accordance with the accuracy desired.

Each parallel circuit also contains asecond group of multiplying means for correcting' an answer obtained in the first group for the influence of another stream.

Since the operation of this apparatus is rather conn plex it may be best described with reference to the ac,

companying drawing which illustrates an extremely simpley embodiment and is not to be limiting on this invention.

rEhe drawimy illustrates a circuit established to desifin 4,

fractionator for a two component system that is desired to produce a pure overhead from the mixture ot`v 'l` his column are condensed and returned to the top tray and one part is withdrawn as product. In the explanation, K is the constant (at a specic temperature and pressure) from equation Y=KX where Y is the mol percent of one component'in1 the .l vapor phase and X is the mol percent of that componentr Basically the apparatus of this invention consists in the liquid phase' from which the vapor came, or the liquid phase in equilibrium with that vapor. El is any base voltage and for the purposes of this explanation it will be considered 100 volts,V however, itis understood thatl any voltage may be suitable. imposed across the circuit by connecting` voltage` dividing means 1, 2, 3 and 4 to a suitable source of power and to the ground. Voltage dividing means l. and 3 are setto take a percent of the base voltage equivalent to the desired percent of light component in the product from the column. Therefore, in this example, since it is desired to have 95% purity of light material in the product, voltage dividing means 1 and 3 will be set to take 95% of the base voltage or in this case 95 volts. Thus, in lines 5 and 21 a potential of 95 volts exists. The function of line 21 will be considered later and will not be discussed further at this point.

Similarly, voltage dividing means 2 and 4 are set to establish a voltage proportional to the percent of heavy material in the overhead product and in this case, since theY percent of light material and heavy material must be 100, voltage dividing means 2 and 4 are regulated to place a voltage of 5 volts or 5% of El in lines l2 and 34. As with line 21 above described, a detailed discussion of line 34 will be deferred.

It is now desired to obtain the composition ofthe liquid on the top tray of the column from which the vapor containing 95% light material and 5% heavy materiai was obtained. Since K multiplied by the composition of the liquid will produce the. answer Y, which is the composition of the vapor, then the composition of the vapor Y multiplied by the reciprocal of K will produce the composition of the liquid in the light materialV from which the vapor was obtained. Therefore, voltage multiplying means 6 is established in series with line 5 and multiplies the voltage in line 5 by to produce in line 8 avoltage corresponding to the amount of light material in the liquid phase on the top tray. Similarly, the voltage in line 5 is transmitted to voltage multiplier 7 wherein it is multiplied by to produce in line 9 another voltage corresponding to the composition in light material of the liquid phase on the top tray.

Since the-true K value is unknown and the K values used are incorrect, the two voltages corresponding to the liquid composition on the top tray areV also not correct, however, this difficulty is overcome by the use of a series of K values and the interpolating circuit. Since it is known that the column will operate over a given temperature range and it is also known that the column will operate over a given composition range, the K value for the components of the mixture to be separated may be known for a given temperature and composition. Thus, for example, if the column operates over a 100 F. to 200 F. temperature range, K values may be set into the series of multiplying means corresponding to the K value of the components at 10 temperature increments over the temperature range of operation. The K values are arbitrarily chosen and will usually Vbe at convenient even increments for example, at each 10. the bank of multipliers represents a series of trials at obtaining the value K, which is a number' which quantitatively describes the difference in volatility between the two components in the mixture to be separated at the temperature in the stage being calculated.

Simultaneously with the above described multiplications, the voltage in line 12 which represents the composition Lof the vapor in heavier material is also multiplied by suitable The base voltage is,

In other words, t

Vvalue selecting 4 Kl etc. values by multipliers 13 and 14 respectively and whatever additional multipliers are in parallel with these two to establish in lines 15 and 16 respectively two values for the composition of the liquid ony the top tray in heavier component. The

the others must also have l0 values. It is also necessary that the e values in the circuit calculating the light material be diffferent by a proportional amount, which proportional amount represents the ratio of the K values of these two components at any temperature. Therefore, if the volatility of the light component as compared to the volatility of the heavy component were such Vthat. the K value for theV light component were twice the K value for the heavy component, the corresponding multipliers in the lparallel circuits would4 differ by aV factor of 100% in the magnitude, of their eiect upon Vthe input voltage to each of the cor.- respondirigmultipliers. By making the difference in K value proportional between the light material circuit and the heavy material circuit one assumes that there is a constant diiererice between the K Values lover the entire temperature range or, in other words, that the plot of K versus temperature for the two components would produce two lines whose distance apart on the graph paper would be substantially unchanged over the temperature range involved. Although this assumption is not strictly true'it is usually made in calculations of fractionating colurrms along with assumptions of equal molal boil-up and equal molal overflow. It has been found from experience that these assumptions are substantially true and devia tions fromthem are extremely small and ordinarily negligible compared with other quantities involved. YBy establishing Y .l 'values is automatically set. It is of course necessary that the mechanisms for the two parallel circuits coact so thatthe same pair of values is in the two circuits at the same time.

The voltages established in lines 8 and 1S, which represent the composition of the liquid on the top tray in light material Vand in heavy material are connected to adding circuitl which produces in line 17 thesulm of these voltages. VAs hereinbefore stated the amount of heavy material plus the amount of light material, percentage-wise, should equal or in this case 100 volts. However, since the K values are arbitrarily chosen the sum of the products will probably not be 100 volts or E1, but a voltage E3 which will probably be greater than or less than E1. It the value of E3 is equal to the vale of El, then of course the problem of the top tray liquid composition and temperature is solved and the arbitrarily chosen K value is correct, however, if it is not equal to E1 further calculations must be made and these are made by adding the voltages established in lines 9 and 16, which are equivalent to a second calculation of composition of heavy and light material in the liquid phase on the top tray. This addition is performed by adding circuit 11 which produces in line 18 a voltage E3 which will be different from E2.

It is desired that if E2 and E3 are neither equal to E1, one of them is higher than E1 and one of them is lower than El. When this situation exists E2 and E3 through lines 17 and 18 respectively, are placed at opposite ends of interpolating circuit 19 which establishes the distance between them that is proportional to El to obtain the fraction of the distance between them representing E1 and therefore an interpolation between these values for the proper value of K.

lf the arbitrarily chosen values for K are wrong in the same direction, that is if the value Vfor E2 and the value of E3 ultimately arrived at are both higher than or both lower than E1 the circuit will drop the multiplier for out of the circuit, place the multiplier into that position in the circuit, that is the voltage obtained by multiplier will be in line 5, and introduce into the circuit a new value which will connect to line 6 producing a new voltage therein. Simultaneously, the circuit calculating the heavier material will drop again be tested by the interpolating circuit` to see ifone is greater than and one is 4less than E1. This operation may be repeated by each succeeding pair of consecutive values in the bank until E2 and E3 values bridging E1 are obtained and the interpolation is made to obtain the proper value of Since these-valueswere established for a specific tem perature, the ultimately obtained value for resulting from the interpolation is equivalent to the temperature on the top tray as well as the composition of the liquid. Suitable recording means connected to the interpolating means, such as a teletype machine adapted to record a particular number at a specific rest position of a stepping relay, will be connected to the interpolating means so that the answer ultimately obtained may be preserved as a permanent record.

The apparatus of this invention hasto this point calculated the temperature and composition of the liquid on the top tray of the column and it is now desired to calculate the composition of the vapor coming from the second tray down in the column hereinafter called the second tray or, in other words, the vapor coming to the top tray of the column.

lt is readily seen that the liquid product from the column, or ,1/5 of the total boil-up from the top tray of the column must pass through the entire column from the feed point to the top tray in the vapor phase. `As hereinbefore described, assuming equal molal overtlow and equal molal boil-up on each tray, the top tray has, of its five parts of overhead or boil-up, one part lost as product and four parts returning to the tray as reflux.

The reilux, or` its thermal equivalent, goes to the trays below. It maybe seen from this description and from the assumption of equal molal boil-up and equal molal overow that the mols of liquid passing to any tray is different from the vapor passing from that tray by the number of mols of product.

Therefore, the number of mols of liquid passing from the'top tray, the composition of this liquid, and the number of mols of vapor coming from the second tray are all known. By material balance the composition of the vapor can be determined. The composition of the vapor on the second tray is the sum of the ratio of mols of product to mols of boil-up times the product composition plus the ratio of mols of reux to mols of boil-up times the composition of the liquidfrom the tray above.

To aid in this explanation, a mathematical analysis of the portion of the column above` the feed tray may be made. Making V the mols of vapor passing upwardly through the column, L the mols of liquid passing down, D Vthe mols of product, Y the mol fraction of light material in the vapor and X the mol fraction of light material in the liquid, the following relationships are true:

where subscripts l and 2 indicates the rst and `second trays. Also, by material balance, this equation, as here-v inbefore stated, indicates that the mols of vapor diier fromthe mols of liquid by the quantity of product. i

-the top tray, which is equivalent to the product from the column, to Vs of the composition of light material of the liquid phase on the next higher tray. It may be seen that if the rellux ratio were l to l the composition dierence between the liquid phase on a tray and the vapor phase on the next lower tray would be 1/2 ofthe total boil-up from the top tray and if the reflux ratio were 3to ,l it would be V4, etc. l v

i liguid passingV to the.` second tray,

7 Y In the operation of thisiin'yention, to'obtain the composition of the vapor ontlie second tray, a voltage di- Yiding means 3 is inserted into Vthe circuit and line 21V carries the voltage taken from this dividing means toY a multiplying means which multiplies it 'by 1,/5, since theY eQuXratio is 4 to 1, and' transmits this through line 28 toad'ding means 30. The voltage dividing means yfor obtaining this figure will be set to place a voltage in line 21 proportional tothe purityfof the light material in the product from the column or in this particular case 95 'i ill" liquid phase in lilies` 8" and. valu'finay be I p a inthe one of these lines which is closest to the true value. Therefore, double throwsinglepole switch 25V and line 26 areY operated byl relay or other means connecting to onthe iirst tray is between theV voltages 9 and a Vclose approximation to the true the interpolating circuit 19 so that eitherline 24 cpzo'nnec't-V ingto line 9` or line 23 connecting to line S will be used inthe circuit depending Yupon which voltagel is closer toV tha true. value of composition. When one or the other multiplicationV volts. The composition of light materialV made by using for this value the voltageY isultiinately selected'thevalue of that voltagefs trans-Y mittedto voltage multiplier V27 which multiplies it by a factor of valuepreviouslyobtained by multiplyingthe boil-up lon th? top tray by This addition, i lent of adding the amount ofV overhead. product tothe. resiilts ina voltage inY line 31"equivalent to the amounty oflightl material in the vapor'on the second tray.y This voltage is fed into` balancing circuit 32 which mechanically establishes` the same voltage in line 105 by adjusting the voltage dividing means 1431 to produce the calfcircuit the composition of "thevapor phase on'the second tray inheavier material is obtained. from-the sec;

' ond' parallel circuit by means of miltiply'ing the voltage in line l34 which is established by placing an initial voltage: of 5 volts from potentiometer 4 and multiplying the thus obtained voltage by Vs desir'edfvoltage.Y By an dentiby multiplying means lfand 4/sand transmits theimultiplied voltage to voltage adding vcircuit wherein its value nis `added to the 25' Whh is the equiva- Y' Y the stripping or. upper portion ofk the the calculations will be stopped and the apparatus reset to calculateffrom thefbottom' tray of they column to the @cdkey-U. Y. Y i VThe 'calculation of thebottom or stripping portion of the Acolumn isaccomplished"by setting, the multiplying means 22 'and 27 for the light materialcomposition and to .condltlonsin that portion of the column. These calculationsare made according to the equation: Y L1 W1 where is the mols of bottoms withdrawn. This equa- Y tion "is derived similarly to the previously used equation rate when-'these` values are determined they may readily 'beVv set 4into the apparatus'of this invention to calculate the bottom part of the column. Thefvalue corresponding to Ll/Lf--W being placed in multipliers 27 and 40 and the'value W/Lf-Wfbeing placed in multipliers 22 and 35'." Whenlthe vmultiplying means have been adjusted and the calculations started, the apparatus Yof this invention will calculate eachl tray up from the bottom until the feed tray is reached.

Any suitable known or commercially available parts `may be used in the assembly of the present invention.

The various voltage dividing means that may be used may include potentiometers of the slide Wire or stepping relay Y type, transformers, electronic circuits, etc. The preferred voltage dividing means are potentiometers since their accuracy,V simplicity and availability make then well adapted'to this function and especially the stepping relay Y type which are easily actuatedfby magnetic means. The

feeding that voltage into adding circuitV 43 through line 41.' This voltage is added to a voltage obtainedby multiplying thevoltage either in line 16 or in line 15, whichever is closer to the correct answer, through'ilinc 37 or lavand` selector switch 38 connected to line 39 and multiplying the thus Vobtained voltage by ,4/5 through multif plying circuit 40, adding the` thusl obtained voltage'in adding circuit 43 to'poduce la result equivalent to the composition of the vapor on` the second tray in heavy material. This voltage isv transmitted by balancing means 45 to'voltage dividing means 102 `andrproduces* in line 112la voltage equivalent` to that composition.

The voltages inlines 165. and,112` represent theA cornf positionf the vapor on theu second tray in light material and in heavy material respectively YYand these lines occupy the samepositionin an identicaljcircuit as arejoccu'pied;

by lines -`5 and 12 in the previously described circuit.- Therefore, attachingL a circuitto linesv 105 and 112 identical to vthose attachedto lines 5A and 12 will produce a calculation that will render the composition of the liquid on thel second tray and the4 composition. of the vapor on thepthird tray. The second circuitinthe gure is nurnbered so that each partin the circuit corresponds totheA pantin the-MQVionsly described circuit having a number l0 0 lass,Y `It, 'may bersevenf that the, balancing means 132andsuitable parts which ;"opevrat`eto Vadjust the voltage in lines 5. and 1.2. so thatv 'Y the circuit used to calculate thefiirst trayis also used. tof

calsulatetheV thirdl Vtray as. .Well as. the'ffth `trayr,ithe seventh tray, and the` ninth tray; etc. until Vthe feed tray is reachedat which time, thccalculations are stopped.

The,secondl setiof circuitsnumbered with the lGO) series Y less in light material than thejfee'df, as herrebefore'statedV When the compositioniof either the voltage multiplying meansY and voltage addingrirneans uste'drn'ay be 'commercially'availablefcircuits which arge usuallyV made up as units to be bodily included into vthe assembly.V These units are usually electronic or electromechanical circuits such as those shown in Figure l. 7(a), (b), V(c) and (d) of Electronic Analog. Computers, by GLA; and r. Kean, erst editi@ publishedV 1952 by McGraw-Hill Publishing Companyv and since they form lno part of this inventionrexcept for theirifunction they will not be described in detail herein. The voltage balancing means may be of any suitable type such as`a null detector coupled with a slide wire and serlvoLmoto'r combination or a steppingrelay Acpupled with incremental resistances across each stage. "Theprogramming switches may also be steppingy relays or other sequencing switches. Interpolating means employed may' be abalancing means installed to perform the particular function of balancing the circuit toY the lconstant voltage. For example, a slide wire and servomotor :combinationvmayhe used toY balance i y theY circuit between B2 and EstoY be equal tothe' voltage El. As hereinbefore stated ltheseVV parts are generally commercially available and it is contemplated that the assembly comprising thisV invention will be made of any perform thesel functions. Il'clai'm' as my invention:l Y 'i v Y* Y 12- An electric computer comprisingin combination a constant voltage source connected toa plurality of iden` tical parallel circuits, each of `said' circuits'comprising a firstivoltage dividing meansgladapted to take a portion of said constantyoltage, connected to a bank of electrij cally parallel irstjvoltage.. multiplying means which are j adaptedto. be preset',v to specific` multipliers, programing 70V adapted s toVV alternately insert pairs of consecutive lirst multiplying meansVV into an interpolating kcircuit which means connecting to said first voltage multiplying-rmeans contains addingv means` at opposite sides thereof to add jue-multiplied @ragga te munipned voltages from the other parallel circuits and bring the circuit to" rest Yat Y 35 and 0 .for` the heavy material composition to conformV said constant voltage when one of the sums of the consecutive pairs of rst multiplying means causes one voltage greater and one voltage less than said constant voltage, recording means connecting to said interpolating circuit whereby the position of rest of said interpolating circuit is recorded as the diierence that said constant voltage is from said greater voltage and said less voltage, second multiplying means adapted to be preset to specific multipliers connected one to a fixed voltage dividing means to multiply the xed divided voltage and another to multiply the voltage obtained by one of the pair of the aforesaid rst multiplying means, a second adding circuit connected to said second multiplying means whereby their voltages are added, balancing means connected to said second adding circuit and to a second the voltages is balanced against said constant voltage to obtain a voltage equal to said sum from said second voltage dividing means, means connecting said second voltage dividing means to a bank of parallel third multiplying means which are adapted to be preset to specific multipliers, programing means connected to said multiplying means adapted to alternately insert consecutive pairs of third multiplying means into a second interpolating circuit which contains adding means at opposite ends thereof to add the multiplied voltages to multiplied voltages from the other parallel circuits and bring the circuit to rest at said constant voltage when one of the sums of the consecutive pairs of third multiplying means causes voltage dividing means whereby the resultant sum of one voltage greater and one voltage less than said constant voltage, means connecting said second interpolating circuit to said recording means to record the position of rest `of said interpolating circuit as the difference that said constant voltage is from said greater and said less voltage, fourth multiplying means adapted to be preset to specific multipliers connected one to said fixed voltage dividing means to multiply the resultant fixed voltage and another to multiply the Voltage obtained from one of the pairs of the aforesaid third multiplying means, a fourth adding circuit connected to said fourth multiplying means whereby the voltages are added, balancing means connected to said fourth adding circuit and to said first voltage dividing means whereby the sum of the voltages is balanced against said constant voltage to ootain a voltage from each first voltage dividing means equal to said sum from said voltage dividing means.

2. The computer of claim l further characterized in that said banks of parallel multiplying means in each of said parallel circuits, each multiplying means is preset to a value differing from the value of the corresponding multiplying means in each other parallel circuit by an amount proportional to the difference in volatility of the substances to be separated.

Bubb May 5, 1953 Bubb June 19, 1955

Images