US2500997A - Electrical computing device - Google Patents

Description

March 21, 1950 T. D. MORGAN 2,500,997

ELECTRICAL COMPUTING DEVICE Filed July 13, 1944 INVENTOR THOMAS o. MORGAN ATTORNEY Patented Mar. 21, 1950 PATENT rice ELECTRICAL CQMPUTING DEVICE of Delaware Application July 13,

2 Shims.-

The present invention relates to a device for solving systems of simultaneous linear equations which are solvable by the well known method of iteration or successive approximations. Such systems of equations which are capable of practical solution by this method will be herein termed convergent systems of simultaneous equations.

Although the device may be utilized in solving any convergent system of n linear simultaneous equations of the type above mentioned it will be described in connection with the determination of percentages of the various constituents of a hydrocarbon mixture. Recently, infrared spectrophotometry has been used for the purpose of analyzing hydrocarbons, as is described in some detail in National Petroleum News for May 3, 1944, section 2, at page R274. Such analysis of hydrocarbons by means of infrared spectrophotometry ofiers many advantages over other analysis methods but requires a large amount of computation which the present invention eliminates. In analyzing petroleum hydrocarbons to determine the percentages of the constituents of a mixture the infrared spectrophotometer is utilized in the following manner, it being assumed that the analyst knows the wave lengths at which each constituent of the mixture has as large an extinction coefficient as possible when compared with the extinction coefiicients of the other constituents. The analyst then, by the use of the spectrophotometer, determine the total extinction produced by a sample of the mixture at each of the selected wave lengths, there being, of course, one wave length for each mixture constituent. For example, if a four component mixture be considered the procedure will be to first analyze a sample of this mixture at a wave length L1, this wave length being chosen to cause as great an extinction as possible with respect to the first component of the mixture as compared with the extinction caused by the remaining components thereof. In a similar manner, a second extinction value will be determined for the sample at a wave length L2 chosen to cause as great an extinction as possible due to the second component of the mixture as compared with the remaining components. In a similar manner, readings will be taken at wave lengths L3 and L4. Then, since the total extinction brought about by the mixture at any wave length must be 1944, Serial No. 5%,745

equal to the sum of the extinctions brought about by each of the components of the mixture a set of equations may be written as follows:

In these equations w, r, y, and a are the percents 1c of the components of the mixture; A1, A2, A3 and A4 are the extinction coefilcients for component w at wave lengths L1, L2, L3 and L4 respectively; B1, B2, B3 and B4 are the extinction coefficients for component :6 at wave lengths L1, L2, L3 and L4 respectively; C1, C2, C3 and C4 are the extinction coemcients for component y at Wave lengths L1, L2, L3 and L4; D1, D2, D3 and D4 are the extinction coefficients for component a at wave lengths L1, L2, L3 and L4 respectively; and K1, K2, K3 and K4 are the total extinction coefficients at wave lengths L1, L2, L3 and L4 respectively.

As has been stated, equations of this type may be solved by the method of systematically making successive approximations of the percentage of each component so that the approximations approach the correct solution to the problem and repeating the approximations until the difference between successive approximations becomes less than the experimental error of the original data,

at which time an answer will be at hand which is consistent in accuracy with the accuracy of the experimental data. However, this method requires a very considerable knowledge of mathematical procedure and considerable time. Further, even when reduced to a routine procedure the time element is still large and a knowledge of the effect of starting with a particular approximation is required in order to reduce the time to the minimum possible.

0 The present invention provides electrical circuits including potentiometers which are adjusted in accordance with the experimental data and through which the results may be read directly in a minimum of time.

It is an object of this invention to provide electrical means for solving simultaneous equations. It is another object of the invention to provide a device for solving such simultaneous equations resulting from the infrared spectrophotometer 0 analysis of petroleum hydrocarbons.

It is a further object of the invention to provide such an electrical computer which shall be simple in construction and easy to operate.

Other objects and features of the invention will appear when the following description is considered in connection with the annexed drawing, in which:

Figure 1 is a schematic wiring diagram of my invention; and

Figure 2 is a schematic circuit diagram showing the manner in which reversing switches are utilized to insert negative coeflicients.

Referring now to the drawing it will be seen that there are provided a plurality of potentiometers K1, K2, K3 and K4 which potentiometers are to be set in accordance with the readings of total extinction as mentioned hereinabove. It will be noted that each of these potentiometers is in a circuit which passes through additional potentiometers and may be closed through a galvanometer ill.

Thus, the circuit extending from potentiometer K1 proceeds from its adjustable contact to a potentiometer A1 and thence through potentiommeter m1 and its adjustable contact to potentiometer B1; thence through adjustable contact of potentiometer B1 to potentiometer x1 through its adjustable contact to potentiometer C1, thencethrough its contact to potentiometer yi, thence to potentiometer D1, potentiometer 21 and to the galvanometer it through the variable resistance H.

The circuits for potentiometers K2, K3 and K4 are similar to the one just above described and lead also to the galvanometer through the variable resistance li. Each of the potentiometers K1, &, K3 and K4 has in circuit therewith a battery respectively designated !2, l3, l4 and i5 and each of the potentiometers A1 through D4 has in its circuit a battery, the latter batteries (16 in number) being poled oppositely to the batteries 82, 33, i4 and 15. Further, each galvanometer circuit has in series therewith a switch respectively designated Ni, ii, l8 and i9, these switches being used to close the separate circuits through the galvanometer, one at a time in an order which will be described hereinafter. It should be noted that the potentiometers wi, wz, we and 1124 have their movable contacts mechanically interconnected and that these contacts are operated by a handle 20 which cooperates with a scale to indicate its setting. In like manner, the potentiometers x1, :22, 1:3 and :04 are mechanically interconnected and connected to an adjusting knob or handle and likewise the y potentiometers and the e potentiometers respectively are interconnected and are connected to operating handles y and 2. Further, the potentiometers A1, A2, A: and A4 D4 are provided with scales (not shown) so that their position may be adjusted.

In using the computer the potentiometer K1 is set to a position such that its scale indicates the measured extinction at wave length L1, potentiometer K2 is set so that its scale indicates the measured extinction at wave length L2 potentiometer K3 is set so that its scale indicates the measured extinction at wave length L3 and potentiometer K4 is set so that its scale indicates proportionately to the measured extinction at wave length L4.

Also potentiometers A1 through D1 are set to indications proportional to the extinctions brought about by pure substances w, :c, y and a respectively at wave length L1, potentiometers A2 through D2 are set to indications proportional to the extinctions brought about by pure substances w, x, y and 2 respectively at wave length L2, potentiometers A3 through D3 are set to indications proportional to the extinctions brought about by pure substances w, m, y and 2 respectively at wave length L3, and potentiometers A4 through D4.- are set to indications proportional to the extinctions brought about by pure substances w, m, y and 2 respectively at wave length L4.

Switch IE is now closed thereby completing a circuit through potentiometers K1, 101, B1, x1 C1, n, D1, 21, rheostate II and galvanometer Hi. Adjusting knob or handle 11; is then operated until the galvanometer reading is zero. This adjustment is equivalent to making the first approximation in the mathematical solution of the equations (and assuming that the total extinction brought about in the sample was due solely to the extinction owing to the pure substance w). It will be noted that the setting of the knob w has caused adjustment of potentiometers wz, ws and 204 as well as of 2121 thereby making the same assumption as to the remaining circuits as was made with respect to the one just considered.

In the next step of the procedure switch I6 is opened and switch i1 closed thereby completing a circuit through the potentiometers bearing subscripts 2 and through rheostat l I and galvanometer IS. The knob or handle :0 is now operated until the galvometer again reads zero. This amounts to setting up electric potentials such that the total extinction represented in potentiometer K2 is balanced out by the settings of the remaining potentiometers. It should be noted that in this as well as in the other three circuits the battery in the circuit of potentiometer K2 is poled oppositely to the batteries of the circuits of potentiometers A, B, C and D.

It may be well at this point to mention that reversing switches may be used in circuit of all the batteries in order that any negative terms in the data may be taken care of by merely operating the corresponding switch. The manner in which such a reversing switch is connected is shown in Figure 2 wherein a double pole double throw switch 29 has its blades connected to opposite terminals of the battery, and its two opposite sets of terminals connected to the fixed terminals of potentiometer D1 so that movement of the switch from one position to the other changes the polarity of the battery connection to the potentiometer. Also, switches may be put in series with all of the batteries to open the battery circuits and these switches may be mechanically interconnected so that when the device is not in use all batteries may be open circuited.

The drawing herein illustrates the complete operation of the device of my invention in its simplest form. It will be obvious to those skilled in the art that without departing from my invention suitable switches may be provided to permit the use of the same battery in place of the batteries i2, i3, M and I5. In a like manner, a second single battery may supply potent-iometers A1, A2, A3 and As successively. Also a third single battery may supply potentiometers B1, B2, B3 and B4 successively, and the same is true of C1, C2, C3, C4, etc. Moreover, a single potentiometer may be switched successively to various positions instead of using the series of ganged potentiometers w1, wz, 203, L04; m1, r2, m3 and :04, etc.

It will be seen that the adjustment of knob :12 has effected an adjustment of all the ac potentiometers and has thereby altered the current which flows through the K1 circuit. However, the K2 circuit although it was modified by setting of the w handle has been adjusted subsequently thereto, and consequently, the zero reading of the galvanometer persists at this moment.

Now switch I! is opened, switch is closed and handle 3 adjusted until the galvanometer again reads zero. As before, the adjustment of the y handle alters the setting of all potentiometers yr, 1112, ya and g4 and, consequently, the circuits in the K1 and K2 circuits would, if now tested, no longer have zero values.

Following this switch !8 is opened, switch 19 closed and handle C adjusted until the galvanometer again reads zero.

The operations above described are repeated until no readjustment of the adjusting handles or knobs w, r, y and z are necessary to give zero readings on galvanometer ill. When this occurs readings are taken from the scales associated with the handles w, x, y and z and these readings give directly (in percentage) the proportions of the constituents w, m, y and z of the original mixture.

From the above it will be seen that each variable, each constant and each coefiicient in the equations hereinabove set forth has its electrical counterpart in the diagram of the figure and that by inserting the values into these electrical counterparts and successively and repeatedly adjusting the circuits respresentative of one of the four simultaneous equations until the galvanometer reads zero, the equations are solved and the percentages of the constituents of the mixture are readable directly from the dials or scales of the instrument.

It Will of course be understood that while the above description and example has considered the analysis of a hydrocarbon mixture having four components and a device for solving equa tions resulting from such analysis it is entirely within the scope of this invention to extend the system to any number of components. For example, if a seven component mixture were to be analyzed the system would be extended to seven circuits each comprising a K potentiometer and seven of the coefiicient potentiometers (such as AD) and of the associated potentiometers such as w representing the seven components of the mixture. Furthermore, as has been indicated hereinabove, although the computer of my invention has been described in connection with solving simultaneous equations resulting from the analysis of hydrocarbon mixtures the system is applicable to various types of mixtures.

Consequently, I wish to be limited not by the foregoing specification, but solely by the appended claims.

What is claimed is:

1. In a device for solving convergent systems of linear simultaneous equations, in combination, a plurality of networks each representing an equation to be solved, each network including a cell representing a constant term, said cell including a potentiometer, a current source connected in series therewith, and a reversing switch for changing the polarity of the voltage drop pro duced across said potentiometer by the current source, and a cell representing each variable term, each such cell including a first potentiometer, a battery connected in series therewith, a reversing switch for changing the polarity of the voltage drop produced across said potentiometer by said current source, and a second potentiometer having one terminal connected to the slider of said first potentiometer and its other terminal connected to a fixed terminal of said first potentiometer, leads connecting the slider of each second potentiometer to a fixed potentiometer terminal of the next succeeding cell, a galvanometer, a variable resistance connected in series therewith, a plurality of switches for connecting said galvanometer and resistor in circuit selectively with the respective networks, and means for simultaneously adjusting said second potentiometers by sets, each set including one second potentiometer from each network.

2. A device for solving convergent systems of linear simultaneous equations as set forth in claim 1 in which the potentiometer of each cell representing a constant term has a fixed resistor connected in series therewith.

THOMAS D. MORGAN.

REFERENCES CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS Number Name Date 1,655,276 Lichtscheindl Jan. 3, 1923 1,893,009 Ward Jan. 3, 1933 2,408,081 Lovell Sept. 24, 1946 2,417,098 Wilcox Mar. 11, 1947 2,432,504 Boghosian Dec. 16, 1947 FOREIGN PATENTS Number Country Date 722,351 France Dec. 29, 1931

Images