GB1587563A - Method of and apparatus for monitoring for electrode displacement in the joule effect heating of heat softenable material - Google Patents

Description

(54) METHOD OF AND APPARATUS FOR MONITORING FOR ELECTRODE DISPLACEMENT IN THE JOULE EFFECT HEATING OF HEAT SOFTENABLE MATERIAL (71) We, OWENS-CORNING FInEaGLAS CORPORATION, a corporation organised and existing under the laws of the State of Dela ware, United States of America, of Fiber glas Tower, Toledo, Ohio, United States of America, do hereby declare the invention for which we pray that a patent may be granted to us and the method by which it is to be performed, to be particularly described in and by the following state ment This invention relates to the monitoring of processing apparatus and more particularly to methods of an apparatus for monitoring for conditions indicative of dis placement or slumping of an electrode in a molten body of heat softenable material in a heating and melting apparatus for Joule effect heated heat softenable material.

This slumping can result in electrode fail ure or refractory wall failure.

The electrical heating or heat softenable materials by Joule effect involves establish ing electrical current communication with the material through electrodes. A typical heat softenable material, glass, has been maintained at or above the unusual work ing temperatures, about 26000F, by immers ing one or more pairs of electrodes in the molten glass and passing controlled pulses of electrical power from the electrodes through the glass. Frequently these elec trodes have substantial portions of their glass contacting surfaces spaced from the walls of the container or furnace for the glass to reduce the heat imposed on those walls and extend wall life. One electrode form involves a right circular cylinder which is extended vertically through the bottom of the furnace. Although a number of expediants are employed to protect the electrode and the bottom refractory, it has been found that there is a tendency for electrodes to erode in the region of the wall through which they pass and to slump from a vertical orientation.

In the case of the cylindrical electrodes cooling jackets have been mounted around the electrodes beneath the furnace bottom to reduce the electrode temperature external of the furnace to a level at which it is not subject to attack by constituents of the atmosphere. Further, it is not uncommon to maintain an inert atmosphere within the jacket and around the electrode so it is protected in the portion of its length in which the temperature approaches that within the furnace. Even with such precautions electrodes tend to erode, usually in the region of and slightly below the furnace floormolten glass interface. Such erosion can result in the breaking off of the electrode in the melt.

Upon breaking an electrode, the furnace wall in the vicinity of the remaining stub portion is attacked and will develop leaks unless corrective action is taken promptly.

In Patent Specification No. 1,527,082 changes in the cross sectional area of an electrode as by a necking down of the electrode and the creation of fractures in the necked region were thought to result in characteristic changes in the electrical parameters for the electrodes which could be monitored and, at threshold levels, actuate indicators.

Recently, it has been observed that Joule effect heated glass melting tanks can be punctured by the slumping of electrodes mounted to extend vertically through the furnace bottom. These punctures are attributed to electrical, physical and thermal erosion of the refractory furnace bottom wall by the slumped electrodes. The nature of the slumping influences the detrimental effects on the furnace operation since in some conditions adjacent electrodes can slump to produce an effective short circuit between them. It has also been observed that the electrodes which are continued in operation after they have slumped tend to be eroded to a cutt-off state in the vicinity of the bend from the desired virtical orientation. Thus it is desirable to detect the slumping of electrodes in the opaque mass of the thermally soften able material, or more generally the displacement of the electrode from its preferred orientation, in order to establish corrective procedures or avoid the undesired consequences of continuing the mode of operation then in effect.

According to the invention a method monitoring the state of electrodes immersed in a mass of heat softenable material for Joule effect heating of the material to determine changes in resistance due to displacement of an electrode from its preferred orientation comprises ascertaining resistance values of the electrode and the material in proximity to the electrode when the electrode is initially utilized in the material; monitoring resistance values of the electrode and the material in proximity to the electrode during the useful life of the electrode; providing a signal indicative of electrode failure if the difference between the monitored value and the initial value of the resistance amounts to the decrease in resistance of from l5to to SOWo of the initial resistance or to an increase in resistance which is more than 20o and up to 50% of the initial resistance.

On receipt of the signal indicating electrode failure, which may actuate alarm indicators, instruction print out and/or process controls, corrective measures can be undertaken.

The change in monitored resistance values is indicative of changes in the geometrical relationship between an electrode and the heat softenable material and other electrodes which could eventually cause failure of the system if corrective action is not taken. The resistance change may occur over weeks, months or even years of utilization.

Plural electrodes and electrode pairs may be monitored by multiplexing to provide a repetitive cyclic scan of the resistance between pairs.

The electrical energy for Joule effect heating may be applied to the electrode as pulses of opposite polarity of a nonsinusoidal waveform and the resistance values may be derived from the r.m.s.

values of current and voltage.

A circuit can be employed with a comparator and which develops signal levels which represent values closely approximating true r.m.s. values of current and voltage for detecting the onset of critical current voltage ratio or resistance levels.

According to another aspect of the invention, there is provided apparatus for heating a heat softenable material by the Joule effect heating of the material using electrodes immersed in the material and means for monitoring the state of the electrodes immersed in the material to determine changes in resistance due to displacement of an electrode from its preferred orientation, the monitoring means comprising means for ascertaining resistance values of the electrodes and the material in proximity to the electrodes when the electrodes are initially utilized in the material, means for monitoring the resistance values of the electrodes and the material in proximity to the electrodes during the useful lives of the electrodes, and means for providing a signal indicative of electrode failure if the difference between the monitored values and initial values of resistance amounts to the decrease in resistance of from 150/0 to 50% of the initial resistance or to an increase in resistance which is more than 20% and up to 50% of the initial resistance.

An electrode pair indicator may be provided to identify the pair of the plurality from which the failure signal originated.

Some embodiments of the invention will now be described by way of example with reference to the accompanying drawings, in which: Figure 1 is a schematic and block diagram of a monitoring circuit for a typical electrode pair wherein voltages representing approximate r.m.s. values of current and voltage applied to an electrode pair for Joule effect heating of material in the vicinity of those electrodes is monitored as a ratio compared to a set point value to ascertain either short term deviations in the ratio or deviations from an actual or extrapolated original ratio for long term deviations; Figures 2, 3, 4, 5 and 6 represent variants of the connection of electrodes and the slumping patterns for those electrodes for which resistance models have been developed to aid in the evaluation of observed resistance changes taken from a fragmentary cross section of Figure 1 at lines 2-2 for Figures 2, 3, 4 and 6 and at line 5-5 for Figure 5; and Figure 7 is a schematic representation of a plan of a glass melter and typical electrode firing patterns which result in a relatively uniform resistance increase for any electrode slump direction.

It has been discovered that incipient failure of an electrode immersed in a fluid mass of heat softenable material to which it communicates electrical current for Joule effect heating is indicated by an increase in the resistance between electrode pairs carrying such current. This phenomenon is believed to be due to erosion of the electrode to reduce its cross section and to the development of fractures in the necked down region of erosion.

Since a significant increase in resistance between electrode pairs has been observed as much as twenty-four hours in advance of an electrode failure, it has been deduced that the onset of the process ultimately resulting in failure is a decrease in cross sectional area along the electrode length. This decrease or necking down has most frequently been observed in the region of the electrode in the vicinity of its emergence from the refractory wall of the container for the molten material. Characteristically, the rate of increase of resistance increases with time. It has been found that a twenty percent (20%) increase is a reliable indicator of incipient trouble, hence such an increase has been employed as the criterion for actuating a trouble indicator in British Patent Specification No. 1,527,082.

The present invention, on the other hand, is concerned with monitoring for changes in resistance due to displacement of an electrode from its preferred orientation normally over a long period.

Both incipient and long term resistance changes between Joule effect electrodes employed in heat softenable materials are significant in indicating the tradition of the apparatus for melting and refining the material and enabling the life of the apparatus to be extended. Changes in resistance of current-voltage relationships between immersed electrodes from the initial values observed when the electrodes are first placed in operation to values deviating at least 15'0/0 up to 50% from the initial values can indicate electrode displacement. Normally such changes occur as long term effects, e.g. over periods of at least weeks and more frequently months or even years.

Vertically mounted electrodes upstanding in a body of molten glass will normally alter their paired electrode resistance as they slump from the vertical orientation from their initial value as a gradual change which over a long period will amount to increases and/or decreases from 15% to 50% as opposed to the relatively abrupt or short term change due, for example, to breaking off of an electrode. These long term changes are therefore measured against the initial values of resistance rather than the periodically reset set points employed to detect incipient electrode failure due to short term effects.

The system of British Patent No. 1,527, 082 can be modified to monitor for electrode slumping by comparing actual values of resistance or current voltage ratio with the original value of zone resistance of resistance between paired electrodes and when that change is observed to be gradual, over a period of at least weeks, and is in excess of 15'0/0 of the original value to actuate an indicator. However, a preferred system is described hereafter.

Electrode slump indication has been found to be significant in regard to the direction of slump where interpair electrical parameter changes are available for more than one electrode in the region in which a slumped electrode is indicated.

Slumping of one electrode has been found to affect the current-voltage ratios of adjacent electrode pairs especially where certain phase relationships exist between the power supplied to the various electrode pairs.

As shown in Figure 1 the plan pattern of electrodes in a typical tank 101 typically is six ranks of four electrodes each spaced generally evenly along the longitudinal axis of the rectangular tank. More particularly the electrodes are arranged to be energized in working pairs either as adjacent electrodes 106 and 107 in a rank, or alternate or interleaved paired electrodes 102 and 104, and 103 and 105 in a rank, and pairs in different ranks either parallel to the tank longitudinal axis as 102 and 106 or skewed to that axis as 102 and 107. A frequently employed arrangement is that shown in Figure 1. where the principal currents are between alternate or interleaved electrodes across a rank as shown by the power supplied from the secondary 108 of transformer 109 through back-to-back controlled rectifiers 111 and 112 subject to firing control 113 applied to their gate electrodes. In a typical interleaved firing control for a tank supplied with polyphase power the electrodes of a rank are connected to a common phase and successive ranks of electrodes are connected to different phases.

However, where inter rank connections are made for the supplying transformers and controlled rectifiers, as between paired electrodes 102 and 106 in the suggested alternate, this common phase per rank is not the case. Further, as shown in the electrode connections adjacent the submerged throat 114 and bottom channel 115, adjacent electrodes may be coupled through their principal power supplies from a common phase as paired electrodes 116 and 117 and paired electrodes 118 and 119.

Adjacent paired electrodes 116 and 117 and 118 and 119 are supplied and controlled in the manner of typical electrode pair 103105 as by means of a phase controlled firing circuit 121 for back-to-back controlled rectifiers 122 and 123 connected in series with a transformer secondary 124.

Resistance change as represented as a change in the ratio of the current to voltage imposed on the heat softenable material through the electrode for Joule effect heating of that material can be ascertained by a number of techniques. As discussed in British Patent No. 1,527,082, the waveform resulting from the phase controlled firing of controlled rectifiers resistance is most accurately indicated as a ratio of current to voltage from the root means square values of those quantities in view of the waveform distortions introduced by the phase controlled firing of the controlled rectifiers in the power source circuits to the Joule effect electrodes and the interphase currents within the mass of heat softened material. However, the impedance of the circuits for Joule effect heating is primarily resistive hence other current and voltage values can be employed, particularly where it is the change in their ratio which is significant.

Thus the instantaneous values might be taken after firing of the controlled rectifiers, or the maximum values, or the average values could be employed. While a preferred measurement of voltage and current is through r.m.s. converters, as described in British Patent 1,527,082, for the purposes of ascertaining resistance changes in the sense of changes in the ratio of current-tovoltage alternative means can be employed.

Advantageously, these alternative value sensing means should indicate signal values which approximate r.m.s. values since those values can then be employed in other measurements than the change in ratio to be monitored for the present evaluation and detection method. One type of circuit whch affords a close proportionality to r.m.s.

values is the network shown in Figure 3 employed for the sensing of current and voltage. It comprises a voltage divider which develops voltage across a resistorcapacitor combination which is closely proportional to the true r.m.s. values A similar circuit is employed for each of the current and voltage sensing circuits. A scanning means connected as indicated at 153, and which may be a decoder as described in British Patent No. 1,527,082 shifts the voltage and current sensing circuit between different electrode pairs.

Voltage is picked up across the leads 125 and 126 to electrodes 103 and 105 by means of transformer 127 having a secondary 128 connected between ground and the r.m.s.

proportioning circuit. The simulating circuit comprises a dual time constant charging circuit which develops a signal at lead 129 which results from a sum of the signals representing time constants of resistor 131 and capacitor 132 through diode 133 and resistor 134 and capacitor 132 through diode 135 where the currents are divided in accordance with the values of resistors 131, 136, 134 and 137. In one example, resistance values of 390 ohms for resistor 131, 1000 ohms for resistor 136, 560 ohms for resistor 134 and 300 ohms for resistor 137 were employed with a 20 micro-farad capacitor 132 to produce an output on lead 129 which closely approximated voltage values having a constant proportionality to the actual r.m.s. voltage values across leads 125 and 126. By proper calibration of voltmeter 138 a direct reading of the voltage across the electrodes is provided.

Current sensing signals proportional to r.m.s. current can be achieved in a circuit 139 from a signal picked up by a current coil 141 associated with a lead to the Joule effect electrodes, as lead 126, through transformer 142 having one end of its secondary 143 grounded and a resistance 144 connected across its secondary. An ammeter 145 is calibrated to indicate the proportioned r.m.s. current on output lead 146.

A resistance value or ratio of current to voltage is derived from analog divider 147 to which the current and voltage signals are applied from leads 146 and 129 so that an output signal on lead 148 is passed to a comparator 149. A set point potentiometer 151 is associated with the comparator to enable the establishment of standard values against which the ratio signal is compared.

When that standard is reached by the out put signal on 148 an alarm 150 is actuated.

The system can be arranged to scan a plurality or all of the electrode pairs as by a relay selector system of the type shown in British Patent No. 1,527,082 and can be provided with a visual indicator (not shown) to identify the out of range pair. The visual indicator in one form is a lamp array with a lock up control for holding the indication of an out of limits pair even as scanning is continued. A manual override of the scanning (not shown) is employed to set the selection of the electrode pair for which the alarm was actuated.

It is advantageous to employ the same proportioning circuits for each electrode pair to avoid variations in values indicated due to variations between circuits. Where different resistance standard values are to be monitored as where the electrodes are immersed in portions of the tank 101 having different temperature characteristics the set point for comparator 149 are differ ent for the different base resistance values from which the given deviation is to be effective. One means for shifting set points with the scan of the electrode pairs is by switching to various tap positions (not shown) on potentiometer 151. This can be done by switching to separate taps each having an appropriate setting by means not shown or by switching to separate potentiometers each having its appropriate setting (by means not shown).

The simplified proportioning circuits while providing a truer indication of the current-voltage ratio or resistance for the range of values derived from a phase controlled firing at different phase settings than for averaging circuits is less accurate over the range phase angle setting than a true r.m.s. converter as described in British Patent No. 1,527,082. However, it is sufficiently accurate to detect deviations within a few percent of true r.m.s. values and thus to enable the monitoring of long term resistance changes which indicate electrode slumping. In such instances, since the reference value of the ratio or resistance is that originally obtained at the start up of the tank compaign or mounting of the electrode pair in question, no need to reset set points is required after the limiting signal deviations have been established for the comparator 149.

A correlation between long term resistance changes and the displacement of electrodes within the heat softenable mass has been observed. In certain glass melting tanks some Joule effect electrodes which were installed extending vertically through the tank bottom have slumped toward or to the bottom of the tank thereby changing the operating parameters of the system.

Molten glass in a cold top Joule effect melting operation exhibits a thermal gradient through the depth of the batch cover, the batch molten glass interface and the depth of the molten glass. Glass at the glass-tank bottom interface is hundreds of degrees cooler than that at and near the top of the Joule effect electrodes. In view of the steep negative temperature coefficient of resistance of molten glass, the slumping of an electrode tends to orient it in high resistivity material. Countering this effect is the presence of the electrode and thus the Joule effect heating of the glass, which is most intense in the immediate vicinity of the electrodes, tends to concentrate around the electrode and, to the extent it is not inhibited by the higher resistivity of the cooler molten glass tends to heat the glass adjacent the bottom of the tank. Volumetric effects must also be considered where one or both of two paired electrodes have slumped in the change of resistance of the circuit including paired Joule effect electrodes and the intervening molten glass, since as the electrode approaches the tank bottom it tends to lose its radial flux paths into the cylinder of the molten glass mass around the electrode and to have only the volume of molten glass above them in which to conduct current. In effect the volume through which current is effectively conducted is reduced by the proximity of the -slumped electrode to the furnace bottom.

It is desirable in the Joule effect heating of glass to detect the slumping or displacement of electrodes from their preferred orientation because that slumping disrupts the thermal patterns in the molten glass and thus the melting and refining parameters. Further, the refractory employed in the walls of glass tanks erodes at a relatively high rate when subjected to electrical potentials which cause current to flow therein. As an electrode slumps toward the bottom wall the rate of both thermal and electrical erosion of that wall increases substantially, even to the degree of puncturing the wall and thus prematurely ending the campaign of the tank. Early detection of electrode slumping enables the tank attendants to take corrective measures which maintain processing parameters and extend the useful life of the equipment. Typically, the slumped electrode can be removed from the operation of the tank as by disconnecting it from its source of electrical power or replacing is by pushing it through the furnace bottom in advance of a replacement electrode without interruption of tank operation. In certain instances the slumping can be accommodated by a reconnection of the electrodes and their power supplies. Accordingly, it is desirable to ascertain the nature of the slumping.

Figures 2 through 6 illustrate various slumping patterns for which resistance change characteristics have been measured and various connection of sources of power, represented by the dashed lines between electrode pairs. Figures 2, 3, 4 and 6 represent interleaved firing across the section 2-2 of Figure 1 such that the dashed lines 152 and 153 represent the connection of the controlled sources of Joule effect heating power supplied to electrode pairs 105-103 and 104-102 respectively. The molten glassto-tank floor interface is represented as line 154 and the cross sections are drawn as having the molten glass depth greater than the height of upstanding electrodes 104 and 102 of Figure 2 from the interface 154.

When a pair of firing eletcrodes in an interleaved firing rank slump toward each other as in Figure 2 electrode pair 105-103, the resistance of the zone between those electrodes will increase. If the electrodes slump rapidly, that is in a period of hours as opposed to the more usual weeks or months, the resistance will increase about 30%. This is attributed to the thermal inertia of the molten glass mass. The greatest current density and greatest Joule effect heating occurs in a thin sheath of the glass around the electrodes. In the case of three inch diameter molybdum electrodes a proponderance of the energy dissipation occurs within a sheath of a wall thickness of about three electrode diameters with most dissipation occurring in the hotter material above the tank bottom.

It is theorized that when an electrode slumps rapidly it tends to slump out of its high temperature sheath and thereby be partially or totally surrounded with cooler glass of higher resistivity which increases the resistance of the glass between the electrodes.

In those instances where the electrodes of Figure 2 slump toward each other slowly, an increase of about 20% will be observed in the resistance between the slumped electrodes. However, when the other pair of electrodes 104-102 remains in its preferred orientation, vertical, the resistance between them will decrease by about 25%. This decrease is attributed to reduction of the effective length-to-area of current flux lines between the electrodes since the opposing electrical poles of the slumped pair are removed from the major conduction region.

Another electrode slumping pattern is illustrated in Figure 3 where the electrodes of both of the electrode pairs 105-103 and 104-102 slump toward each other. In a sequential slumping of the pairs, when the second to slump pair of electrodes slumps toward each other its zone resistance increases from the initial value by about 15% and the first to slump pair has its zone resistance decreased by about 5% to about 15% of its initial value, that is it decreases from its initial 20% increase relative to its initial reference value by 5% of that initial reference value.

In Figure 3 electrodes 104 and 103 are represented to have slumped toward each other without contacting each other. It is to be recognized that slumping of electrodes can occur in any direction and that the examples showing slumping in the general plane of a cross section have been chosen for convenience of illustration and discussion. Where electrodes 104 and 103 slump into actual or near engagement with each other and are effectively electrically connected by a low resistance path, as illustrated in Figure 1, they connect the secondaries of their supply transformers in series aiding across the outer electrodes 102 and 105 and the preponderant current path will be across the outer electrodes 105 and 102.

The resistance of each of the two zones will decrease significantly to of the order of one-half their initial values.

If a pair of electrodes slumps in opposite directions, away from each other, as shown for the electrodes of Figure 6 the resistance of each zone increases about 33 above the initial zone resistance. This is attributable to the greater effective electrode spacing in the molten glass, the location of the electrodes in a cooler portion of the melt, and the volumetric effect of the reduced area for effective conduction in the melt due to the proximity of the refractory floor to the slumped electrodes. Little effect on the zone resistance of an adjacent zone in the rank when the second to slump pair slump away from each other, as its resistance also increases about 33% from its original value. It is believed that the slumping of the first pair to slump away from each other has little effect on the resistance of the zone of the pair which remains in its vertical preferred orientation because like electrical polarities are close together as 105 to 104 and 103 to 102 negating the cell constant change in the melt.

Outward slumping of electrode pairs connected for side-by-side electrode firing as illustrated for electrode pairs 116-117 and 118-119 supplied along dotted lines 110 and 120 in Figure 5 has been observed to increase the zone resistance about 30% of its original resistance in the preferred electrode orientation. As in the case of paired electrodes 102 and 104 where only voltage sensing transformer 152 and current pick up coil 153 are shown for association with the r.m.s. approximating voltage and current circuits in a manner corresponding to transformer 127 and coil 141, the voltages and currents in the circuits to paired electrodes 116-117 and 118-119 are ascertained through transformers 154 and 155 and pick up coils 156 and 157 respectively for connection to the r.m.s. approximating circuits.

An illustration of an electrode array and electrode firing pattern which exhibits uniform zone or paired electr the melt 162, and in the illustrated array have about a twenty-five percent greater resistance in the diagonal zones represented as 163 through 166 than in the same geometry with interleaved electrode pair firing as disclosed in Figure 3. It is theorized that this increased zone resistance increases the volume of the melt proximate to the electrodes in which the preponderance of the Joule effect heating occurs thereby reducing volumetric effects on the zone resistance change experienced when an electrode slumps. Further, the absence of electrodes of other pairs in the volume of melt between each electrode pair reduces the cell effect of displacement of such electrode on the change in resistance of the zone. In practice it has been found that a zone having slumped electrode in a system of the type shown in Figure 9 exhibits an increase over its initial condition with both electrodes vertical and upstanding from the bottom of tank 161 of about thirty per cent for any direction of slumping.

As shown in Figure 7 a single phase, phase A, of the power source is coupled across the first two ranks of electrodes with all the electrodes of each rank having like instantaneous polarities imposed at the supply transformer secondaries. Similarly another phase is individual to electrode ranks 177 and 178 and still another phase to electrode ranks 179 and 181. Power is supplied to each zone as previously described.

Zone 163 between electrode 171 of rank 167 and electrode 184 of rank 186 is supplied from secondary 187 of transformer 188 through back-to-back parallel connected controlled rectifiers 191 and 192 controlled by a firing control 193. Each zone is monitored for its electrical parameters as described above employing a current pick-up coil 194 on a cable to an electrode and a voltage sensing transformer 195 connected across the cables to the electrodes of a zone.

Current sensing means 196 and voltage sensing means 197 transpose the signals from the current pick up coils, as 194, and voltage sensing transformers, as 195, respectively for application to analog divider which issues a ratio or resistance signal to comparator 199 all as previously described.

The initial set poit for the comparator 199 is established based upon the initial resistance of the associated electrode pair and its zone is set on potentiometer 201 and an alarm 202 responds to the preset deviation of resistance from the original value. Only a typical block diagram and the current pick up coils and voltage coils for the first bank of zones is illustrated. It is to be understood that scanning means for associating the block diagram elements for ascertaining resistance and limits thereon with each zone of the bank as well as other banks is contemplated e.g. by using a relaydecoder as described above by reference to British Patent No. 1,527,082 wherein only one ascertaining circuit is employed and that zone identification and set point association can be coordinated with the scan function by the decoder. Further, where desired, both increased and decreased resistances relative to original zone resistance values can be accommodated by the scanning sequence and set point selection means.

The magnitudes of the long term changes in zone resistance caused by the displacement of electrodes from the preferred orientation are a function of many factors apart from the aforementioned directions of slumping, speed of slumping, slumping relationships between electrodes of a pair providing the principal zone current, and the slumping relationships or lack of slumping betwen electrodes of several pairs.

Particularly significant effects upon the resistance change are the firing configuration of the electrodes, their relative instantaneous polarity and their phasing. In any arrangement a long term change in resistance of a zone in the range of about 150/0 to about 50% of the original zone resistance is indicative of the slumping of an electrode, the displacement of an electrode from its preferred orientation and in many instances the magnitude of the resistance change and the relationships of the changes is resistance in nearby electrode pair zones provides a rather precise indication of the nature and direction of the electrode displacement from the preferred orientation.

However on increase in resistance of up to 20% can also indicate thinning of electrodes as described and claimed in British Patent 1,527,082.

WHAT WE CLAIM IS: 1. The method of monitoring the state of electrodes immersed in a mass of heat softenable material for Joule effect heating of the material to determine changes in resistance due to displacement of an electrode from its preferred orientation which comprises ascertaining resistance values of the electrode and the material in proximity to the electrode when the electrode is initially utilized in the material; monitoring resistance values of the electrode and the material in proximity to the electrode during the useful life of the electrode; and providing a signal indicative of electrode failure if the difference between the monitored value and the initial value of the resistance amounts to the decrease in resistance of from 15% to 50% of the initial resistance or to an increase in resistance which is more than 20% and up to

**WARNING** end of DESC field may overlap start of CLMS **.

Claims (10)

**WARNING** start of CLMS field may overlap end of DESC **. the melt 162, and in the illustrated array have about a twenty-five percent greater resistance in the diagonal zones represented as 163 through 166 than in the same geometry with interleaved electrode pair firing as disclosed in Figure 3. It is theorized that this increased zone resistance increases the volume of the melt proximate to the electrodes in which the preponderance of the Joule effect heating occurs thereby reducing volumetric effects on the zone resistance change experienced when an electrode slumps. Further, the absence of electrodes of other pairs in the volume of melt between each electrode pair reduces the cell effect of displacement of such electrode on the change in resistance of the zone. In practice it has been found that a zone having slumped electrode in a system of the type shown in Figure 9 exhibits an increase over its initial condition with both electrodes vertical and upstanding from the bottom of tank 161 of about thirty per cent for any direction of slumping. As shown in Figure 7 a single phase, phase A, of the power source is coupled across the first two ranks of electrodes with all the electrodes of each rank having like instantaneous polarities imposed at the supply transformer secondaries. Similarly another phase is individual to electrode ranks 177 and 178 and still another phase to electrode ranks 179 and 181. Power is supplied to each zone as previously described. Zone 163 between electrode 171 of rank 167 and electrode 184 of rank 186 is supplied from secondary 187 of transformer 188 through back-to-back parallel connected controlled rectifiers 191 and 192 controlled by a firing control 193. Each zone is monitored for its electrical parameters as described above employing a current pick-up coil 194 on a cable to an electrode and a voltage sensing transformer 195 connected across the cables to the electrodes of a zone. Current sensing means 196 and voltage sensing means 197 transpose the signals from the current pick up coils, as 194, and voltage sensing transformers, as 195, respectively for application to analog divider which issues a ratio or resistance signal to comparator 199 all as previously described. The initial set poit for the comparator 199 is established based upon the initial resistance of the associated electrode pair and its zone is set on potentiometer 201 and an alarm 202 responds to the preset deviation of resistance from the original value. Only a typical block diagram and the current pick up coils and voltage coils for the first bank of zones is illustrated. It is to be understood that scanning means for associating the block diagram elements for ascertaining resistance and limits thereon with each zone of the bank as well as other banks is contemplated e.g. by using a relaydecoder as described above by reference to British Patent No. 1,527,082 wherein only one ascertaining circuit is employed and that zone identification and set point association can be coordinated with the scan function by the decoder. Further, where desired, both increased and decreased resistances relative to original zone resistance values can be accommodated by the scanning sequence and set point selection means. The magnitudes of the long term changes in zone resistance caused by the displacement of electrodes from the preferred orientation are a function of many factors apart from the aforementioned directions of slumping, speed of slumping, slumping relationships between electrodes of a pair providing the principal zone current, and the slumping relationships or lack of slumping betwen electrodes of several pairs. Particularly significant effects upon the resistance change are the firing configuration of the electrodes, their relative instantaneous polarity and their phasing. In any arrangement a long term change in resistance of a zone in the range of about 150/0 to about 50% of the original zone resistance is indicative of the slumping of an electrode, the displacement of an electrode from its preferred orientation and in many instances the magnitude of the resistance change and the relationships of the changes is resistance in nearby electrode pair zones provides a rather precise indication of the nature and direction of the electrode displacement from the preferred orientation. However on increase in resistance of up to 20% can also indicate thinning of electrodes as described and claimed in British Patent 1,527,082. WHAT WE CLAIM IS:

1. The method of monitoring the state of electrodes immersed in a mass of heat softenable material for Joule effect heating of the material to determine changes in resistance due to displacement of an electrode from its preferred orientation which comprises ascertaining resistance values of the electrode and the material in proximity to the electrode when the electrode is initially utilized in the material; monitoring resistance values of the electrode and the material in proximity to the electrode during the useful life of the electrode; and providing a signal indicative of electrode failure if the difference between the monitored value and the initial value of the resistance amounts to the decrease in resistance of from 15% to 50% of the initial resistance or to an increase in resistance which is more than 20% and up to

50% of the initial resistance.

2. The method according to claim 1, wherein the electrical energy for Joule effect heating is applied through the electrode as pulsations of alternately opposed polarity of a sinusoidal waveform; and the resistance values are derived from current and voltage which are at least approximate r.m.s. values.

3. The method according to claim 1 or claim 2 wherein the resistance is ascertained between two electrodes initially immersed in a body of heat softenable material and upstanding vertically therein and wherein a long term change in ascertained resistance over the initially ascertained resistance to actuate the indicator is indicative of the slumping of an electrode from its upstanding vertical orientation.

4. The method according to claim 3, wherein the heat softenable material is heated in a tank having a longitudinal axis and a bottom; wherein the electrodes are mounted as vertical elongate conductors upstanding from the tank bottom, are arranged in ranks of at least two pairs of electrodes in each of a plurality of ranks arrayed transverse of the longitudinal axis of the tank; wherein the paired electrodes are interleaved to establish principal conductive paths between paired electrodes within the heat softenable material having at least one electrode of another pair between each electrode pair; and wherein a decrease in resistance between first paired electrodes from the initial resistance between those electrodes to actuate the indicator is indicative of the slumping of an electrode of another pair located between the first paired electrodes.

5. The method according to claim 3, wherein the heat softenable material is heated in a tank having a longitudinal axis and a bottom; wherein the electrodes are mounted as vertical elongate conductors upstanding from the tank bottom are arranged in ranks of at least two pairs of electrodes in each of a plurality of ranks arrayed transverse of the longitudinal axis of the tank, wherein the paired electrodes are interleaved and the principal conductive path with the heat soften able material is between paired electrodes in a rank separated by an electrode in said rank; and wherein an increase in resistance over the initial resistance between other paired electrodes in the rank of the first paired electrodes and including an electrode between said first paired electrodes is indicative of a slumping of the electrode of said first paired electrodes between said second mentioned pair.

6. The method according to claim 5, wherein the increase in resistance between the first paired electrodes is between twenty and thirty percent of the initial resistance between those electrodes and the decrease of resistance between the other paired eleccent of trodes is about twenty-five percent of the initial resistance between those electrodes.

7. The method according to claim 3, wherein the heat softenable material is heated in a tank having a longitudinal axis and a bottom, wherein the electrodes are mounted as vertical elongate conductiors upstanding from the tank bottom, arranged in ranks of a least two pairs the electrodes of which are interleaved in each of a plurality of ranks arrayed transverse of the longitudinal axis of the tank; and wherein an increase in resistance over the initial resistance between paired electrodes of about fifteen percent over the initial resistance of these electrodes indicates a slumping of at least one of those electrodes or each pair toward its paired electrode.

8. The method according to claim 3, wherein the heat softenable material is heated in a tank having a longitudinal axis and a bottom wherein the electrodes are mounted as vertical elongate conductors upstanding from the tank bottom arranged in ranks of at least two pairs the electrodes of which are interleaved in each of a plurality of ranks arrayed transverse of the longitudinal axis of the tank; wherein adjacent electrodes of different pairs are supplied with voltages of different polarities a preponderance of the time; and wherein a decrease in resistance between paired electrodes of interleaved pairs of the order of fifty percent of their respective initial resistance values is indicative of a slumping of the adjacent electrodes of different pairs into close proximity to each other.

9. The method according to claim 3, wherein an increase of resistance between an electrode pair of about thirty-three percent over the initial resistance between that pair is indicative of the slumping of the electrodes of that pair away from each other.

10. A combination of apparatus for heating a heat softenable material by the Joule effect heating of the material using electrodes immersed in the material and means for monitoring the state of the electrodes immersed in the material to determine changes in resistance due to displacement of an electrode from its preferred orientation, the monitoring means comprising means for ascertaining resistance values of the electrodes and the material in proximity to the electrodes when the electrodes are initially utilized in the material, means for monitoring the resistance values of the electrodes and the material in proximity to the electrodes during the useful lives of