US3853723A

US3853723A - Mercury cell anode short detection and current balancing

Info

Publication number: US3853723A
Application number: US00377993A
Authority: US
Inventors: R Mack
Original assignee: PPG Industries Inc
Current assignee: PPG Industries Inc
Priority date: 1973-07-10
Filing date: 1973-07-10
Publication date: 1974-12-10
Anticipated expiration: 1991-12-10
Also published as: DE2432691A1; NL164613B; FR2236969B1; FR2236969A1; NL164613C; NL7409047A; GB1479671A; BE817477A; IT1014489B; JPS5039298A

Abstract

Disclosed are a method and apparatus for conducting electrolysis in electrolytic cells having flowable mercury amalgam cathodes. Also disclosed are methods and apparatus for maintaining the current flow in each row of anodes of the cell within a predetermined range of the average current in all of the anode rows and within a predetermined range of a reference current. In this way, incipient short circuit conditions may be detected, short circuits prevented, the current load across the cell balanced, and the anode-cathode voltage drop reduced to a minimum practical voltage.

Description

Unite States atent 11 1 [in 3,853,723 Mack Dec. 10, 1974 [5 MERCURY CELL ANODE SHORT 78,557 12/1970 Germany 204/225 DETECTION AND CURRENT BALANCING 1,212,488 11/1970 Great Britain 204/225 75 Inventor: Robert Mack, Lake Charles, La. [73] Assignee: PPG Industries, Inc., Pittsburgh, Pa. [22] Filed: July 10, 1973 [2]] Appl. No.: 377,993

Primary Examiner-John H, Mack Assistant Examiner-Aaron Weisstuch Attorney, Agent, or Firm-Richard M. Goldman [5 7] ABSTRACT Disclosed are a method and apparatus for conducting [52] Cl 204/99 204/219 81 electrolysis in electrolytic cells having flowable mer- 51 I Cl 0 14 cury amalgam cathodes. Also disclosed are methods d 9 and apparatus for maintaining the current flow in each 1 le 0 earc 1 row of anodes of the cell within a predetermined 204/220 range of the average current in all of the anode rows and within a predetermined range of a reference cur- [56] References cued rent. In this way, incipient short circuit conditions may UNITED STATES PATENTS be detected, short circuits prevented, the current load 3,531,392 9/1970 Schmeiser 204/225 across the cell balanced, and the anode-cathode volt- FOREIGN PATENTS OR APPLICATIONS age drop reduced to a minimum practical voltage. 1,148,322 4/1969 Great Britain 204/99 20 Claims, 6 Drawing Figures Error Manipulated ir g r Set Point Signal Ge Variable G p Ou+pu+ f C troller Electrode Mercur- Cell) Anode mr oi e (on Spac1ng) y Curr-em) 4's Discrere Anode Currem G h= Conrinous Anode Current (Lock8.Hold) PATENTED 53 1 U 3. 853 .723

SHEET 1 OF 6 Fig.2

PATENTEB DEE I DISH SHEET 2 BF 6 MERCURY CELL ANODE SHORT DETECTION AND C f NT BALANCING BACKGROUND OF THE INVENTION According to one method of electrolyzing sodium chloride brine to yield caustic soda and chlorine, the electrolysis is carried out in a mercury cell. Mercury cells are characterized by the presence of a conducting surface inclined slightly from the horizontal in the longitudinal direction. A mercury amalgam film, typically from about $4; of an inch to about A of an inch or more in thickness flows across this plate in the direction of the inclination thereof. The flowing mercury amalgam film is the cathode.

Flowing on top of the amalgam is the electrolyte, that is, the aqueous sodium chloride solution. Typically, the electrolyte is of a thickness of from about A of an inch to about 2 inches or more at the inlet and as much as or 12 inches at the outlet.

Anodes, such as carbon anodes or metallic anodes are usually spaced about /8 of an inch to about 3/ 16 of an inch above the surface of the mercury-amalgam film. In this way, electrical current can flow from the anodes through the electrolyte to the flowing mercury amalgam cathode. Structurally, a group of anodes are mechanically supported by and commonly movable on a frame or structural member which is electrically insulated from the cell body. Several anodes connected to a single bus bar and arrayed laterally across the width of the cell are referred to as an anode row. A group of anode rows supported by a common frame or structural member are referred to as an anode bank.

In a typical electrolytic cell, there are a number of these anode banks arrayed along the longitudinal axis of the cell. For example, a typical cell may be 60 or 70 feet long and have anywhere from 12 to about 30 anode rows arrayed along the length of the cell in about 4 to about 8 banks.

In a typical cell circuit, a plurality of individual cells are arranged in series. Typically, two or more rows of cells are in side-by-side relationship with the positive terminal of a power source being connected to the anodes of the end cell of the first row of cells and the negative terminal of the power source being connected to the cathode of the end cell of the last row. Within the series, the cathodes of one cell are connected to the anodes of the next adjacent cell in the series. Typically, there may be from 30 to 80 cells in a cell circuit, although there may be more or less, and several rows of cells may be installed in a single circuit.

In a normally running mercury amalgam cathode electrolytic cell, chlorine is liberated at the anode and the sodium is liberated at the cathode, the sodium forming an amalgam with the mercury.

In the operation of such mercury amalgam cathode cells, where each cell has a plurality of vertically movable anode rows, it has been found that the current flow from an individual anode to the cathode exhibits variations relative to the total current flow through the cell due to changes in the physical and electrical properties of the amalgam, the formation of mercury compounds on the cell bottom, the build-up of cell butter, the depletion of electrolyte content in the brine, and the thermal expansion of the cell body and cell hardware. Additionally, the anode banks may work out of alignment due to various cell operations.

Due to the contribution of any or all of the above factors, an anode may touch the mercury film cathode, shorting the cell. This causes excessive current flow through the anode, which may damage the anode. Under short circuit conditions hydrogen may form and collect in the chlorine space, giving rise to an explosive hydrogen-chlorine mixture.

Additionally, when the anodes come out of alignment, the current across the cell is unbalanced. That is, one anode row may carry significantly more electrical current than an adjacent anode row, resulting in uneven and sporadic operation of the cell, and higher voltages than are actually necessary for cell operation.

Short circuits may be eliminated by raising an individual anode to a greater height above the cathode when the current through that anode reaches or exceeds a preset level such as disclosed in German Patent 2,107,305 assigned to Dyanmit Novel AG or as disclosed in Canadian Patent 909,720 to Schaffer. Other methods of coping with incipient short circuits in flowable mercury cathode electrolytic cells are shown in U.S. Pat. No. 3,574,073 to R. W. Ralston, in U.S. Pat. No. 3,644,190 to N. Weist, and in U.S. Pat. No. 3,689,398 to Caleffi. In all of these teachings of the prior art after discovery of an incipient short circuit by suitable measuring means, the anode is withdrawn from the cathode, but, without any provision for either returning the anode to a position of normal operation or subsequent balancing of the cell.

The prior art shows various methods of positioning anodes to balance the current load across the individual anodes at start-up of the cell. Thus, according to the prior art, all of the anodes may be adjusted such that they are lowered toward the cathode until the incipience of a short circuit, at which time they are slightly backed 0H, for example, as disclosed in U.S. Pat. No. 3,361,654 to DePrez, or U.S. Pat. No. 3,396,095 to Van Diest. Similarly, it is known to initially position the anodes in a mercury cathode cell by measuring the resistance or conductivity of the brine and initially adjusting the electrodes responsive thereto, as shown in U.S. Pat. No. 3,464,903 -to Shaw and U.S. Pat. No. 3,476,660 to Selwa.

It is also known according to the prior art to adjust individual anodes, independently of other anodes in the cell, such that a predetermined current flow is maintained through each anode, such as disclosed in U.S. Pat. No. 3,556,973 to Sensen and U.S. Pat. No. 3,627,666 to Bonfils.

SUMMARY OF THE INVENTION It has now surprisingly been found that the first indication of an incipient short circuit condition between an individual anode and a cathode is that the bus bar to the particular anode row in which the anode is located carries a gradually increasing current compared to the other anode rows, i.e., compared to the average current flow to the cell.

It has now surprisingly further been found that a flowing mercury amalgam cathode electrolytic cell may be operated at a lower average cell voltage if the individual current flow through each bank of adjustable anodes is substantially continuously measured and substantially continuously maintained within a preset range of the average of the individual current flows through each of the rows of anodes.

DETAILED DESCRIPTION OF THE INVENTION According to this invention, a flowing mercury amalgam cathode electrolytic cell circuit having a plurality of electrolytic cells is operated while continuously maintaining each of the individual current flows through each of the anode rows of a cell within a preset range of the average of the individual current flows through all of the anode rows of the cell over extended periods of electrolysis. In this way, by keeping the individual anode row currents within a preset range of the average anode row current, the anodes of the electrolytic cell are maintained in balance, incipient short circuit situations are detected and short circuits are avoided.

By continuously maintaining" is meant monitoring and controlling a particular anode-current flow with sufficient frequency as to detect and respond to changes in current before such current changes reach an undesirable magnitude. This may be accomplished by measuring and controlling each of the currents at a time interval of less than the time constant of the anode current flows. The time constant of an individual anode is the physical parameter represented by the constant Tau in the LaPlace transform of the anode transfer function, i.e., (1 [1 S 11"), A single row of anodes in a typical flowing mercury amalgam cathode cell has a time constant, for purposes of measuring and controlling the current flow through the anode row of from several minutes to several hours. Accordingly, if the in dividual current flow through a single anode row to the flowing mercury amalgam cathode is measured and controlled at intervals of less than once every 20 seconds, e.g., from about once every two seconds to about once every 20 seconds, then the current flow through an anode bank to the cathode is continuously controlled.

By an extended period of electrolysis, is meant a period of electrolysis such that the initial start-up phenomena associated with long-term operation of the cell, such as changes in power input to the plant, changes in concentration of the electrolyte and the like have had time to occur and have an effect on the cell, e.g., time periods in excess of more than 1 day, such as 3 days, 5 days, or several months.

According to this invention, short circuits may be prevented and the individual cells balanced by the use of a control element. Short circuits are characterized by direct contact of the anode with the flowing mercury cathode, resulting in a greatly decreased voltage and increased current. In one exemplification of this invention, where the input to the controller is the current flow in the anode row bus bars, the control element monitors the current flow in each individual anode row bus bar, increasing the anode-cathode gap if the current goes above a predetermined level or set-point and decreasing the anode-cathode gap if the current goes below a predetermined level or set-point. Typically the set-point is a function of the average current flow, i.e., the average current plus or minus a constant, or plus or minus a constant multiple of the average current. According to this invention, when an incipient short circuit is detected, e.g., by a high current flow, the anode row is moved vertically away from the cathode to a sufficient distance to return the current flow to the reference value, which may either be a value preset into the controller or a value related to the instantaneous operation of the cell such as the average current through each of the anode rows in the cell, or to a value below the reference.

The individual cells are also maintained in balance by the control element. A cell is deemed to be in balance when each of the individual anode to cathode voltage drops, or each of the individual current flows through each anode row, are approximately equal throughout the cell. That is, the individual anode to cathode volt age drop differs from the average anode to cathode voltage drop by less than a predetermined amount or in a preferred exemplification, the individual anode row current differs from the average anode row current by less than the predetermined amount. The amount by which an individual voltage differs from the average voltage or the preset voltage, or in a preferred exemplification, the amount by which an anode row current flow differs from the average current flow, i.e., the toI erance, indicates the fineness of the control.

While control of the circuit. prevention of short circuits, and balancing of the cell may be accomplished by using the anode to cathode voltage drop, it is preferred. according to this invention, to use the anode row bus bar current flow. This is because the voltage difference between the extremes of voltage encountered may be less than 25 percent of the total anode to cathode voltage measured, while the extremes of bus bar current measured are one hundred percent of the current measured.

One exemplification of apparatus useful in practicing this invention is shown in the figures.

FIG. I is a schematic diagram of an individual mercury cell.

FIG. 2 is an isometric partial cutaway view of a mercury cell.

FIG. 3 is a schematic diagram of a multiplexing circuit useful with this invention.

FIG. 4 is a schematic diagram of the switching elements of the multiplexing circuit shown in FIG. 3.

FIG. 5 is a schematic diagram of a magnetic latching matrix useful with this invention.

FIG. 6 is a signal flow control diagram for a control method of this invention.

According to one exemplification of this invention, the voltage drop across a homogeneous section of an anode bus bar is measured. As this voltage drop is proportional to the anode bus bar current flowing to the anode row, a first signal proportional to the current in the anode bus bar is generated. Thereafter, a second electrical signal, i.e., a set point signal is preset or generated. This second electrical signal is proportional to the current corresponding to an incipient short circuit condition. The first and second electrical signals are then compared to generate a third electrical signal which is an error signal, If the first electrical signal is greater than the second electrical signal i.e., if the anode bus bar current exceeds the maximum permissible current, i.e.. the control current, means for vertically moving the anode bank proportional to the error signal are actuated. Currents of greater than 2.0 times the average current are indicative of short-circuit con ditions. Generally the set point current is from about 1. l to about I .5 times the average current and preferably L1 to 1.2 times the set point current. Control cur rents ofless than 1.05 times the average current are uneconomical in terms of control effort. Simultaneously, a record of short circuit conditions may be made on a typewriter console. In this way incipient short circuits are detected and short circuits are prevented. Additionally, a cathode ray tube display or a typewriter console, or both, may be actuated, at the operators request to exhibit cell operation data, e.g., as an operating aid.

Once any short circuit conditions or incipient short circuit conditions have been corrected, the current to an individual cell can be balanced. In balancing an individual cell, the individual anode bus bar current flows are measured as described hereinabove. Thereafter, a signal proportional to the anode bus bar current in an individual anode bus bar is generated. This is done for each anode bus bar to the cell. An electrical signal is also generated which is proportional to the average anode bus bar current. This signal may be the actual sum of the individual anode bus bar signals divided by the total number of signals. Alternatively, it may be the current from the power supply to the cell circuit divided by the number of bus bars or anode rows in an individual cell. Whenever an average current is referred to, it is understood that such current may be either of the types described above.

The average of the anode bus bar electrical signals for an anode bank are compared with this average anode bus bar current signal and an error signal is generated for each anode bank. If the error signal exceeds a threshold value, means are actuated for vertically moving the anode bank in question.

After balancing the cell, the cell voltage may be adjusted. The minimum cell voltage is dependent on the total cell current and may be determined therefrom or preset. The average voltage drop across the cell may be determined by measurement and the minimum voltage drop compared thereto. If the average voltage drop exceeds the minimum voltage drop, the anode vertical adjustment means is actuated simultaneously for all of the anode banks in the cell. In this way reduced voltage drop may be obtained without disturbing current balance.

In control of a cell circuit according to this invention, a controller receives electrical signal inputs from the individual rows or banks of anodes of each of the electrolytic cells in the circuit, and generates electrical control signals back to the individual rows or banks of anodes.

This may be understood by references to FIGS. 1 and 2 where there is shown a mercury cell 1 having a flowing mercury amalgam cathode 72, flowing across a steel surface 70, and metal anodes, 74, with current fed to the anodes 74, by bus bars 80, which are connected to anode stems 76, which pass through a rubber of plastic or other pliable top 78. The anodes 74, and anode stems 76, are supported by frames 82. The frames 82, may be in the nature of l beams or more complex structures. The anode frame 82, and its rows of anodes 74, are vertically movable on threaded rods 84. This vertical movement may be accomplished through motor 88, and the associated drive gearing 86. According to this invention, the current flow in the bus bars 80, to a row of anodes 74, is measured, fed to a controller where it is compared to a control point, and an error signal generated which actuates the motor 88.

According to one exemplification of this invention, the control element is an analog computer. The individual inputs to the analog computer are the individual amperages measured on each anode row bus bar and 6 the individual voltages measured from each anode to the cathode. These individual currents and voltages are selectively, and either randomly or sequentially, transmitted to the analog computer through a multiplexer, and an amplifier.

The outputs from the analog computer are individual error signals corresponding to each individual anode row. The error signals are transmitted back to the vertical adjustment mechanism individual anode banks by way of a multiplexer.

According to a preferred method of this invention, the control element is a storable program digital computer. A storable program digital computer is preferred if there are more than about 450 individual anode bus bars being monitored, e.g., as many as 2100 or more individual bus bars being monitored.

The individual inputs to the digital computer are individual amperages measured on each anode row bus bar and individual voltages measured across each anode row and the cathode. These individual currents and voltages are selectively transmitted to the digital computer through a multiplexer and an amplifier. The outputs from the digital computer are individual error signals corresponding to each individual anode row. The error signals are transmitted back to the anode bank adjustment mechanisms by way of a multiplexer.

One such storable program digital computer controlled system is shown schematically in FIG. 3 for a single cell. As shown in FIG. 3, wires measure the voltage drop across a uniform section of the anode row bus bar 80 of a cell 1. When actuated by relay 241 contacts 221 through 232 are closed allowing the electrical signals to pass through the first level multiplexer 201, i.e., contacts 221 through 232, and to be available at the second level multiplexer, 301, i.e., contacts 321 through 332. The second level multiplexer 301 is controlled by the digital computer 431, such that two contacts may be selectively closed under control of the digital computer and any current or voltage signal converted to a digital value. In order to get the electrical signal into the digital computer in a digital form the low level signal is amplified through amplifier 411 and digitalized through the Analog-to-Digital Converter 421, from which the signal is transferred to the digital computer 431.

The digital computer 431 actuates the output matrix 450, including latching relays 451 through 463. An audible alarm and light 96 may be actuated and anode row motors 88 may be driven through the output matrix 450. The output matrix switch 464 controls relay 241 to select the cell that is to be monitored by closing contacts 221 through 232.

The input to the controller is a signal-proportional to the current flowing through individual anode bus bars. According to one exemplification, a shunt may be connected across a uniform section of homogeneousanode bus bars. Typically the shunt is of sufficient length in relation to current flow through the busbar to provide a voltage drop of about 30 to 40 millivolts across the shunt.

In order to provide the cell control described herein, it is not necessary to continuously monitor all of the individual cells in the electrolytic cell circuit. It is only necessary that the cells be monitored at time intervals of less than the time constant of the individual cell. This may be accomplished by multiplexing the input signals from each individual anode row of each cell to the controller.

There are several ways of multiplexing the anode row current signals and anode to cathode voltage signals into the controller. The signals may be inputed sequentially to the amplifier 411 through a single level of switching wherein all contacts are in parallel and on the same level.

A coded addressing system multiplexer can be used. In a coded addressing system multiplexer, a signal is generated, typically by the computer. This signal activates an input selection unit which closes the appropriate switch and the sensor or group of sensors having the address corresponding to the signal are read. Each bus bar input unit has its own n-digit address and an AND gate. If the address of the individual anode bus bar input unit is the one transmitted, the input unit then transmits its reading back to the controller. The bus bar input unit may be the shunt millivolt signal described above.

In the use of a coded address system multiplexer, when a bus bar signal is interrogated, its reading is transmitted back to the controller. This may be done electromechanically by reed relays. Alternatively, this may be accomplished by a solid state device such as a transistor or a field effect transistor. If a field effect transistor is used, it should be one characterized by a large on/off impedance ratio, the absence of a junction EMF, and gate isolation.

Alternatively, a rotary sequential scanning switch may beused. According to still another exemplification, a magnetic latching switch matrix may be used to sequentially close relays corresponding to each anode row shunt, and thereby cause each anode row current to be sequentially read by the controller.

A preferred multiplexing system is a multiple multilevel multiplexing system. In a two-level system, a first first-level multiplexer simultaneously reads all of the bus bar inputs from one cell. A first second-level multiplexer, which is part of the controller, reads all of the inputs from the first first-level multiplexer. Thereafter another first-level multiplexer is connected to another second-level multiplexer which sequentially reads all of the inputs from the second first-level multiplexer. Meanwhile the first of the second-level multiplexer drops the prior first level multiplexer and picks up another first-level multiplexer.

Thus, with reference to FIG. 4 in a typical cell circuit with 64 cells, where each of the cells has four individual anode bus bars per cell, the sensors are connected to the controller by a system containing 64 first-level multiplexers (201 through 208) and four second-level multiplexers (301 through 304). According to this exemplification, at any given time, l6 relays corresponding to 4 relays per cell times the number of second-level multiplexers, i.e., 4 second-level multiplexers, are simultaneously energized. The input from the first of the first multiplexers (201 i.e., the first 4 anodes bus bars inputs, are sequentially read by the second-level multiplexers (301). As soon as the input from the first 4 anode bus bars are read into the controller 431, the first first-level multiplexer 201 is de-energized. Then a second first-level multiplexer 205, is connected to that second-level multiplexer 301. During this time, the input from the next first-level multiplexer 202 is transmitted through the next second level multiplexer 302 to the controller 411 in the same manner. This process continues in sequence through the cells of the circuit.

In a multilevel method of multiplexing, the contacts are in parallel and series, with the second level of contacts, e.g., 321 through 332, being in series with the first level of contacts, e.g., 221 through 232. The second level of multiplexing need contain only enough contacts to read all current and voltage inputs from a single cell. The first level of multiplexing contains as many contacts as are required to read all inputs from a single cell times the number of cells. Through the output Contact matrix, a contact such as contact 464 energizes relay 241 to select all contacts 221 through 232 for a single cell in the first level multiplexer. The computer then reads the current and voltage signals of the cell selected in the first level multiplexer by individually selecting contacts 321 through 332. The computer may sequentially or randomly read the inputs. Once all inputs for a single cell have been read, all contacts 221 through 232 for that cell are opened by deactivating relay 241 through contact 464. After the contacts of the cell just read in the first level multiplexer have had time to open, contacts 221 through 232 for another cell in the first level multiplexer may be closed and the read procedure through contacts 321 through 332 repeated. Contacts in both the first-level multiplexer and secondlevel multiplexer may be solid state, reed relays, electromechanical relays or any combination of these.

The first level multiplexer may be mercury wetted reed relays, one relay per anode bus bar, and one group of relays per cell. The second level multiplexer may be part of the computer, and comprise mercury wetted relays also.

The preferred multiplexing system is a two-level multiplexer as shown in FIG. 3. The preferred contacts are mercury wetted reed relays. Even though FIG. 3 shows a single two-level multiplexing system a preferred scheme of input is to have multiple two-level multiplexers with a common output at the amplifier 411. By using multiple two-level multiplexers, the operation of the multiplexers may be overlapped to give faster operation by gaining time that is normally allowed for contact settle after switching. An example of multiple two-level multiplexers is shown in FIG. 4 in which two first-level multiplexers are shown connected to each of the second level multiplexers (e.g., first-

level multiplexers

201 and 205 to second-level multiplexer 301, first-

level multiplexers

202 and 206 to second-level multiplexer 302, first-

level multiplexers

203 and 207 to second-level multiplexer 303, and first-

level multiplexers

204 and 208 to second-level multiplexer 304). As shown in FIG. 4, all the cells in the circuit are connected in series with each cell having its own first-level multiplexer. Every fourth cell and therefore, every fourth first-level multiplexer is connected to a common second level multiplexer.

The two-level multiplexer system useful in this invention is shown in FIG. 4. As there shown, the circuit starts with all contacts 221 through 232 of each of the first-level multiplexers 201 through 208 open and all contacts 321 through 332 of the second level multiplexers 301 through 304 open. The contacts 221 through 232 inclusive of the first four first-level multiplexers 201 to 204 inclusive are closed, e.g., by coil 241. The outputs from each of the switches 221 through 232 of multiplexer 201 are connected to the switches 321 through 332 of the second-level multi- 9 plexer 301. The controller sequentially reads the electrical signals on each contact 321 through 332 of second-level multiplexer 301. At this point all of the contacts 221 through 232 of the first first-level multiplexer 201 are opened and all of the contacts in the first-level multiplexer 205 are closed. While the contacts in first-level multiplexer 201 are opened and the contacts in first-level multiplexer 205 are closed, the electrical signal on each contact of second-level multiplexer 302 coming from first-level multiplexer 202 is sequentially read by the controller 431. After reading all of the inputs from first-level multiplexer 202 through second-level multiplexer 302, all of the contacts in first-level multiplexer 202 are opened and simultaneously all of the contacts in first-level multiplexer 206 are closed. As this is occurring, the electrical signals on each contact of second-level multiplexer 303 are sequentially read by the controller 431. After reading all of the inputs through second-level multiplexer 303, all of the contacts of first-level 203 are opened and simultaneously all of the contacts in firstlevel multiplexer 207 are closed. As this is occurring the controller reads the electrical signals on each contact of first-level multiplexer 204 through secondlevel multiplexer 304. After reading all of the contacts of multiplexer 204 through second-level .multiplexer 304, all of the contacts in first-level multiplexer 204 are opened and simultaneously all of the contacts in second-level multiplexer 208 are closed. As the contacts in first-level multiplexer 204 are opened and the contacts in first level multiplexer 208 are closed, the electrical signal on each contact of multiplexer 205 is sequentially read through second-level multiplexer 301. After all of the contacts on multiplexer 205 have been read, first-level multiplexer 205 is opened and first-level multiplexer 209, not shown, but next in seqeuence, would be closed.

While the method of this invention is illustrated with multiplexer, or two second-level multiplexers with onehalf of the cells connected to each second-level multiplexer, or even it second-level multiplexers with one over ri of the cells connected tov each multiplexer. The number of multiplexers is determined by economic factors, and the number of second level switches available on the controller.

In an electrolytic cell control circuit of the type herein contemplated, a filter may be interposed between the input from the individual electrolytic cells and the controller. The filter serves to minimize the noise created by the process being monitored, e.g., the A.C. ripple on'the D.C. electrolysis power, and the noise created by the control circuit, especially the noise created by the multiplexer elements. The filter is interposed between the individual anode bus bars and the controller. In a multi-level multiplexer system, as a twolevel multiplexer system, locating the filters between a first level multiplexer and the second-level multiplexer requires a low settling time filter while locating the filter between the first-level multiplexer and the process requires higher initial cost for the increased number of filters. In one exemplification of an electrolytic cell circuit control of the type herein contemplated, the second level multiplexer has capacitors connected across Additionally, in a control circuit of the type herein contemplated, an amplifier may be required to raise the signal transmitted to a level that can be acted upon by the analog to digital converter.

The analog to digital converter converts the input signal to digital form and also encompasses what is commonly called a sample and hold element. This is shown on signal flow diagram FIG. 6 with the legend: le ")/s.

When the controller is a storable program digital computer, considerable flexibility is offered in the control of the process. Preferably the controller should be a storable program computer with in excess of 10,000 words of 16 bit per word memory, and preferably in excess of 12,000 words of 16bit per word memory.

With a storable program computer, the control characteristic may be proportional-integral, i.e., proportional-reset having a control equation of:

Pn K[e,,3,, At/rR e where'e is the error signal, n is the interval number or sample number, K, is the proportional gain, t is a sampling interval, 7;; is the reset or interval time constant and Pn is the output.

The sampling time or AT is on the order of from about 05m about seconds. Most commonly for an electrolytic cell circuit of the type herein contemplated AT is from about 4 to about 10 seconds. The longer sampling times refer to the approximate time constant of the drift of the average cellvoltage or average cell current level while the shorter time constant refers to the approximate. time constant of incipient short circuits and the drift of individual anodes.

The output from the controller may be a cathode ray tube display, typewriter messages, motor control signals, or a combination thereof. Cathode ray tube dis-' plays may indicate which anodes are most out of adjustment, the graphic cell balance, or the like,

The output from the controller also includes actuation of motors 88 to cause vertical movement of the anode rows 82. The electrical signal indicates both the magnitude and the direction of anode bank movement.

Since most digital computers are limited to the number of contact closures that can be incorporated and a mercury circuit short detection system requires a large number of contact closures, it is necessary to expand the contact closures available from the computer. One method of doing this is to have a matrix 450 of relays driven by the contact closures available in the computer. The coils of the relays are arranged in a matrix pattern as shown in FIG. 5, and in relation to the control circuit, as relays, in FIG. 3. The contact closures (701 through 710) are arranged to either drive a column or a row of the matrix. Each relay utilizes a close coil 0 and an open coil 0 arranged in magnetic flux opposition. If both the close coil and open coil are energized simultaneously, the magnetic flux generated by the close coil will override the magnetic flux generated by the open coil. The relays are magnetic latching such that once they are closed or opened they remain in that state even though voltage is removed from their coils. The relay located at the point of intersection of a column and row is the relay that is latched closed when the computer contacts for that column and row are closed. For example, if

computer contacts

703 and 709 are closed, relay 553 will latch and remain closed even after

computer contacts

703 and 709 are opened. In order to unlatch a relay in the matrix the computer contact in the row in which the relay is located must be closed and the computer contact in the column in which the relay is located must be opened, thus energizing the open coil without energizing the close coil. For example, in order to open relay S3 computer contact 703 is opened and computer contact 709 is closed. The contacts of the relays of the matrix may be used to energize anode adjustment motors, short detection lights, analog input relays, or energize a horn.

The latching relay matrix 450 contains contact closures 701 through 708. The contact closures 701 through 708 are arrayed in rows and columns to provide a matrix. Preferably they are arrayed in equal rows and columns to provide a latching matrix. Closing contact closures in any column and in any row serves to latch closed the latching relay at the intersection thereof. Latching open a contact in any row latches open all of the relays located in that row.

In maintaining cell balance, the anode bank is moved in relation to the error signal. A particular type of control which may be used with such a control circuit is one providing for a constant speed of anode movement but with a time of movement which is proportional to the magnitude of the error signal and less than the sampling time. If the anode bank has not obtained a desired degree of movement within one sampling interval, the current is measured, a new error signal generated, and the anode moved again during the next sampling interval. Single pole double throw center stable motor relays may be used on each individual anode bank motor. As to the actual raising or lowering of the anodes, several alternatives are possible. For example, upon an incipient short circuit condition occurring in any anode bank of the cell, that anode bank may be raised until the current is zero, or some fractional part of the average current, such as O.l times the average current, or 0.5 times the average current, or until the current drops below a selected alarm point. The anode bank may then be slowly lowered, as described above, to restore current balance to the cell, or all of the anode banks may be rapidly raised, then slowly lowered to restore current balance to the cell. The movement of the anodes may be continuous rate, full space, stepped rate, or incre' ment.

When restoring a condition of current balance to the cell or returning the cell to a minimum voltage the anode banks may be moved as great a distance as possible in one time interval or a constant distance or multiple thereof not greater than the maximum distance called for. Alternatively, the maximum distance the anode bank might be moved could be some fraction of the amount necessary for current balance, so that cell balance is gradually approached, and not overshot.

The following example is illustrative.

EXAMPLE A short circuit detection and cell balance system was installed on four cells of a 68 cell, 300,000 ampere, mercury cell circuit. The 68 cells were divided into four groups of 17 relays each such that there were four second-level multiplexers each connected to 17 first-level multiplexers of 30 relays.

Each of the mercury cells in the circuit had six anode frames with four anode rows per frame, providing a total of 24 anode rows. The measured inputs from the four cells were the anode bus bar current to each anode row, and the voltage across the cell measured from each anode frame to the cathode. The measured inputs from each cell were switched through 30 double poles, mercury wetted relay contacts which served as the firstlevel multiplexer, into the input of the second-level multiplexer.

The second-level of multiplexer was IBM analog input relays. These were divided into four groups of 30 relays each, corresponding to the number of first-level multiplexers. Also connected directly to the secondlevel multiplexers are a zero reference voltage, a calibration reference voltage, the total current flow across the cell circuit, the total voltage across the cell circuit, and output from a power supply used to drive the coils of the latching relays.

The output from the second level multiplexer went to a programmable, multi-range amplifier, to an analogto-digital converter, and to the IBM System 7 Computer.

The IBM System 7 Computer has a 12,000 word, 16 bits per word memory. The output from the computer drives a magnetic latching relay matrix with two coil, single contact relays. The latching relay matrix had the dual function of opening and closing the double pole contacts on the input from the anode row as well as activating the control action for incipient short circuit conditions and for cell balance.

In the operation of the test system, the 30 inputs from each first level multiplexer were alternatively latched into the second level multiplexer, and were read individually by the IBM System 7 Computer through the second level multiplexer. The alarm point was 1.2 times the average current, and the anode rows exceeding this value were indicated by a light, horn, and typewriter message. A cathode ray tube output was also used for data display.

The results obtained are shown in Table I.

TABLE I Average Anode Stem Average Anode Stem to Cathode Voltage to Cathode Voltage Days Drop. Cells Not Dro Cells Attached on Attached to To Computer Difference Line Controller (Volts) (Volts) (Volts) 3 4.25 4.02 0.23 7 4.33 4.08 0.25 14 4.31 4.04 0.27 I? 4. l9 4.02 O.l7 21 4.24 4.03 0.21 28 4.29 4.12 0.17

of vertically movable anode banks, and a current flow from the anodes in said anode banks to the cathode, and having a common control element the improvement comprising:

discretely measuring each of the individual current flows through the anode rows of a single cell at intervals sufficient to detect and respond to incipient changes therein;

electrically generating individual first electrical signals proportional to the individual current flows in each of the individual anode rows, simultaneously transmitting all of the said first electrical signals from a single cell to and through a first level of switches to a second level of switches;

individually transmitting each of said first electrical signals from said second level of switches to the common control element;

electrically generating a second electrical signal proportional to the average of the individual current flows through said anode rows; and

electrically generating individual anode row error signals proportional to the difference between said individual first electrical signals and said second electrical signal whereby to control said cell whereby to maintain the individual current flows within a preset range of the average of the individual current flows through the anode rows of said cell.

2. The method of claim 1 comprising moving all of said anode banks vertically away from said cathode whenever any of said first signals exceeds a value corresponding to short circuit incipience.

3. The method of claim 1 comprising moving each of said anode banks in relation to a function of said error signal whereby to maintain cell balance.

4. The method of claim 1 comprising measuring the total current flow to the cell circuit to generate a signal is proportional to the total current flow to the cell circuit.

5. The method of claim 4 comprising electrically dividing the said signal proportional to total current by the number of said first electrical signals whereby to generate said second electrical signal proportional to the average anode row current flow.

6. The method of claim 3 comprising measuring the current flow through each of the individual anode row bus bars whereby to generate the said first individual electrical signals.

7. The method of claim 1 comprising controlling a relay matrix output through said common control element.

8. The method of claim 7 comprising actuating contact closures whereby to actuate a latching relay matrix.

9. The method of claim 8 wherein the contact closures are arrayed in an equal number of columns and rows.

10. The method of claim 8 comprising closing a contact in a row and closing a contact in a column whereby to latch closed a relay at the intersection of said row and said column.

11. The method of claim 8 comprising closing a contact in a row whereby to latch open all of the relays located in the row.

12. The method of claim 1 comprising sequentially controlling said cells.

13. The method of claim 1 comprising randomly controlling said cells.

14. The method of claim 1 comprising:

simultaneously transmitting all of the said first electrical signals from a first single cell to and through a first group of first level of switches to a first group of second level switches;

simultaneously transmitting all of the said first electrical signals from a second single cell to and through a second group of said first level of switches;

individually transmitting each of said first electrical signals from the first group of said second level of switches to said common control element, electrically disconnecting said first group of said second level switches from said first group of said first level of switches, electrically connecting said first group of second level switches to a subsequent group of said first level of switches, and individually transmitting each of said first electrical signals from the second group of said first level of switches through said second group of said second level of switches to said common control element.

15. In a mercury cell circuit having a plurality of flowing mercury amalgam cathode electrolytic cells in series, each of said cells being electrically connected to the cells adjacent thereto by bus bars, and a controlcircuit having a storable program digital computer; the improvement comprising shunts responsive to current flow on each of said bus bars; and first level multiplexing means and second level multiplexing means interposed between said bus bars and said storable program digital computer.

16. The cell circuit of claim 15 wherein said first multiplexing means comprise one first level multiplexer per mercury cell.

17. The cell circuit of claim 16 wherein said second multiplexing means comprise second level multiplexers interposed between said first level multiplexers and said storable program digital computer.

18. The cell circuit of claim 17 wherein there is more than one first level multiplexer per second level multiplexer.

19. The cell circuit of claim 15 wherein the storable program digital computer includes output means for actuating said multiplexing means comprising a contact closure actuated latching relay.

20. The cell circuit of claim 19 wherein the contact closures are arrayed in an equal number of columns and rows.

Claims

1. IN A METHOD OF CONDUCTING ELECTRLYSIS IN AN ELECTRLYTIC CELL CIRCUIT HAVING A PLURALITY OF ELECTRLYTIC CELLS, EACH OF SAID CELLS HAVING A FLOWING MERCURRY AMALGAM CATHODE AND A PLURALITY OF ANODE ROWS IN A PLURALITY OF VERTICALLY MOVABLE ANODE BANKS, AND A CURRENT FLOW FROM THE ANODES IN SAID ANODE BANKS TO THE CATHODE, AND HAVING A CONNON CONTROL ELEMENT THE IMPROVEMENT COMPRISING: DISCRETELY MEASURING EACH OF THE INDIVIDUAL CURRENT FLOWS THROUGH THE ANODE ROWS OF A SINGLE CELL AT INTERVALS SUFFICIENT TO DETECT AND RESPOND TO INCIPIENT CHANGES THEREIN; ELECTRICALLY GENERATING INDIVIDUAL FIRST ELECTRICAL SIGNALS PROPORTIONAL TO THE INDIVIDUAL CURRENT FLOWS IN EACH OF THE INDIVIDUAL ANODE ROWS; SIMULTANEOUSLY TRANSMITTING ALL OF THE SAID FIRST ELECTRICAL SIGNALS FROM A SINGLE CELL TO AND THROUGH A FIRST LEVEL OF SWITCHES TO A SECOND LEVEL OF SWITCHES; INDVIDUALLY TRANSMITTING EACH OF EACH FIRST ELECTRICAL SIGNALS FROM SAID SECOND LEVEL OF SWITCHES TO THE COMMON CONTROL ELEMENT; ELECTRICALLY GENERATING A SECOND ELECTRICAL SIGNAL PROPORTIONAL TO THE AVERAGE OF THE INDIVIDUAL CURRENT FLOWS THROUGH SAID ANODE ROW ERROR SIGNALS ELECTRICALLY GENERATING INDIVIDUAL ANODE ROW ERROR SIGNALS PROPORTIONAL TO THE DIFFERENCE BETWEEN SAID INDIVIDUAL FIRST ELECTRICAL SIGNALS AND SAID SECOND ELECTRICAL SIGNAL WHEREBY TO CONTRL SAID CELL WHEREBY TO MAINTAIN THE INDIVIDUAL CURRENT FLOWS WITHIN A PRESET RANGE OF THE AVERAGE OF THE INDIVIDUAL CURRENT FLOWS THROUGH THE ANODE ROWS OF SAID CELL.

12. The method of claim 1 comprising sequentially controlling said cells.

13. The method of claim 1 comprising randomly controlling said cells.

14. The method of claim 1 comprising: simultaneously transmitting all of the said first electrical signals from a first single cell to and through a first group of first level of switches to a first group of second level switches; simultaneously transmitting all of the said first electrical signals from a second single cell to and through a second group of said first level of switches; individually transmitting each of said first electrical signals from the first group of said second level of switches to said common control element, electrically disconnecting said first group of said second level switches from said first group of said first level of switches, electrically connecting said first group of second level switches to a subsequent group of said first level of switches, and individually transmitting each of said first electrical signals from the second group of said first level of switches through said second group of said second level of switches to said common control element.

15. In a mercury cell circuit having a plurality of flowing mercury amalgam cathode electrolytic cells in series, each of said cells being electrically connected to the cells adjacent thereto by bus bars, and a control circuit having a storable program digital computer; the improvement comprising shunts responsive to current flow on each of said bus bars; and first level multiplexing means and second level multiplexing means interposed between said bus bars and said storable program digital computer.