compensation

Technical field of the invention The invention relates to the field of fused salt electrolysis, and more precisely to electrolytic cells suitable for the Hall-Heroult process for making aluminium by fused salt electrolysis. In particular, the invention relates to a particular arrangement of the busbar system in an electrolysis plant in which electrolytic cells are arranged side by side, forming compensation loops that are capable of counterbalancing the vertical magnetic fields generated by adjacent rows of cells.

Introduction

The Hall-Heroult process is the only continuous industrial process for producing metallic aluminium from aluminium oxide. Aluminium oxide (Al ₂ 0 ₃ ) is dissolved in molten cryolite (Na ₃ AIF ₆ ), and the resulting mixture (typically at a temperature comprised between 940 °C and 970 °C) acts as a liquid electrolyte in an electrolytic cell. An electrolytic cell (also called "pot") used for the Hall-Heroult process typically comprises a steel shell (so-called pot shell), a lining (comprising refractory bricks protecting said steel shell against heat, and cathode blocks usually made from graphite, anthracite or a mixture of both), and a plurality of anodes (usually made from carbon) that plunge into the liquid electrolyte. Anodes and cathodes are connected to external busbars. An electrical current is passed through the cell (typically at a voltage between 3.5 V and 5 V) which electrochemically reduces the aluminium oxide, split by the electrolyte into aluminium and oxygen ions, into aluminium at the cathode and oxygen at the anode; said oxygen reacting with the carbon of the anode to form carbon dioxide. The resulting metallic aluminium is not miscible with the liquid electrolyte, has a higher density than the liquid electrolyte and will thus accumulate as a liquid metal pad on the cathode surface from where it needs to be removed from time to time, usually by suction into a crucible.

The electrical energy is a major operational cost in the Hall-Heroult process. Capital cost is an important issue, too. Ever since the invention of the process at the end of the 19 ^th century much effort has been undertaken to improve the energy efficiency (expressed in kWh per kg or ton of aluminium), and there has also been a trend to increase the size of the pots and the current intensity at which they are operated in order to increase the plant productivity and bring down the capital cost per unit of aluminium produced in the plant.

Industrial electrolytic cells used for the Hall-Heroult process are generally rectangular in shape and connected electrically in series, the ends of the series being connected to the positive and negative poles of an electrical rectification and control substation. The general outline of these cells is known to a person skilled in the art and will not be repeated here in detail. They have a length usually comprised between 8 and 25 meters and a width usually comprised between 3 and 5 meters. The cells (also called "pots") are always operated in series of several tens (up to several hundreds) of pots (such a series being also called a "potline"); within each series DC currents flow from one cell to the neighbouring cell. For protection the cells are arranged in a building, with the cells arranged in rows either side-by- side, that is to say that the long side of each cell is perpendicular to the axis of the series, or end-to-end, that is to say that the long side of each cell is parallel to the axis of the series. It is customary to designate the sides for side-by-side cells (or ends for end-to end cells) of the cells by the terms "upstream" and "downstream" with reference to the current orientation in the series. The current enters upstream and exits downstream of the cell. The electrical currents in most modern electrolytic cells using the Hall-Heroult process exceed 200 kA and can reach 400 kA, 500 kA or even more; in these potlines the pots are arranged side by side. Most newly installed pots operate at a current comprised between about 350 kA and 600 kA, and more often in the order of 400 kA to 500 kA. Pots operating at still higher current are being developed and operated by several companies. These enormous electrical DC currents flow through various conductors, such as electrolyte, liquid metal, anodes, cathode, connecting conductors, where they generate heat with ohmic voltage drops and where they generate significant magnetic fields. As mentioned above, electrolysis according to the Hall- Heroult process is a continuous process driven by the flow of electric current across the electrolyte, whereby said electric current reduces the aluminium atoms that are bounded in the alumina present in the molten electrolyte. Four equilibria define the optimum cell operation window, leading to the high current efficiency: electrical equilibrium, magnetic equilibrium, thermal equilibrium and chemical equilibrium. These equilibria are partly determined by the cell design, and can also be acted upon in the process of cell operation.

Conditions of electrical equilibrium of the cell are attained when the distribution of the current is as uniform as possible throughout the electrolyte; the thickness of electrolyte between the anode and the cathode (inter-electrode spacing) in a typical Hall-Heroult cell is of the order of about two to five centimeters. The main operating parameter by which the operator can act on the thermal and electrical equilibria is the inter-electrode spacing, which modifies the Ohmic resistance of the electrolyte and thereby modifies both the thermal equilibrium (through the heat generated by this Ohmic resistance) and the electrical equilibrium (through the change in cell voltage related to this Ohmic resistance). The main permanent perturbation factor of the electrical equilibrium is the current path in the liquid metal, as will be explained below; this factor is determined by the cell design and cell operation. Discontinuous events during cell operation such as anode change and so-called anode effects will also perturb the electrical equilibrium.

The magnetic equilibrium of the cell is determined to a large extent by the cell design and in particular by the busbar structure. Perturbation factors are mainly related to electrical currents arising from conductors outside of the cell.

The thermal equilibrium is determined to a significant extent by cell design, and in particular by the choice and thickness of materials and components, and by the lining; main operational parameters that influence on the thermal equilibrium are the thickness of the anode coverage and the inter-electrode spacing. Perturbation factors are mainly related to specific discontinuous operations (anode change, metal tapping, adding of electrolyte) or to so-called anode effects (this term and the phenomenon that it designates are known to a person skilled in the art and need not to be explained here).

The chemical equilibrium is determined by the chemical composition of the electrolytic bath; addition of alumina and aluminium fluoride are the principal operational parameters.

The present invention is related to the magnetic equilibrium of an electrolytic Hall-Heroult cell. Such cells are of rectangular shape, and as such they are symmetric by construction. Asymmetry arises from asymmetry of the electric current flow in the cell. Electrical current enters the cell through anodes which cover a large part of the surface of the cell, crosses the electrolyte and the liquid metal pad, and is collected by the cathode which forms the whole surface of the cell bottom. The cathode is made from a carbonaceous material and contains steel collector bars which enable an electrical contact to be established with the cathode busbar. However, the electrical conductivity of both the cathode and the steel cathode bar is much lower than that of the liquid metal pad. As a consequence the current lines in the liquid metal are not vertical but have horizontal components, interacting with the vertical magnetic field and leading to magnetohydrodynamic (MHD) perturbations. Furthermore, at the downstream side of the cell, the cathode busbar is linked to a small number of connecting elements called anode risers, through which the current is fed into the anode beam of the downstream cell. Further perturbation of the magnetic field in the cell is due to boundary effects, the cell not having infinite dimensions.

As a consequence, the magnetic field in the cell locally has a spatial distribution which, combined with electrical currents in the cell, creates Laplace forces; these induce movement of liquid conductors (electrolyte and metal) and deform the metal-bath interface hydrostatically. Laplace forces may also induce metal-bath interface oscillations. The resulting unevenness of the metal surface leads to a local variation in anode-to-metal pad distance across the length of the pot, which is represented by small fluctuations of the overall cell voltage signal; this may even lead to a short-circuit between the anode and the cathode. These oscillations of the metal-bath interface are called magnetohydrodynamic (MHD) instabilities which are detrimental to the performance of the process; they require the distance between anode and cathode to be increased and this counter measure increases the electrical resistance of the cell, leading to ohmic losses and eventually to an increase in energy consumption. The MHD instabilities are specifically the result of the vertical magnetic field component which tends to increase with the size of the pots and with the cell current.

Certain perturbative events (adding alumina, anode change, metal tapping, anode effects) may increase these instabilities and metal and bath velocities. The perturbative effects of an event will be higher if the vertical component of the magnetic field, in particular in the upstream corners of the cell, is high. It is therefore desirable, in order to reduce these magnetohydrodynamic instabilities, to decrease as far as possible the vertical component of the magnetic field (designated as B _z , z being the coordinate running upwards from the bottom to the top of the cell) in the liquid metal; a root-mean square average value of about one millitesla is a usual upper limit. Moreover, the horizontal components B _x and B _y (x being the longitudinal axis of the cell) should be anti-symmetrical with respect to the longitudinal and transverse axis, respectively. A required property of B _z is also the anti-symmetry with respect to the cell centre, i.e. equal and opposite values in each corner of the cell.

The magneto-hydrodynamic (MHD) stability of an electrolysis cell depends on two fundamental factors, one internal, the other external. The main factor (internal) is the busbar design of the cell itself. A good busbar design is capable to ensure a magnetically stable cell operation. Several patents have been published which present a design in which the magnetic fields created by the various parts of the cell and the connecting conductors compensate one another, thus decreasing magnetohydrodynamic instabilities in the cell. The targeted result is a cell having a magnetic field in the liquid metal pad which is symmetric or anti-symmetric with respect to the cell axes or the cell centre as explained above. In particular, it is desirable to reduce the vertical magnetic field in the two upstream corners of the cell. This is most often achieved by appropriate busbar design and/or compensation conductors. An example for a compensation loop for the symmetrisation of the vertical component of the magnetic field in an electrolysis cell is shown in GB 2 041 409 (assigned to Aluminium Pechiney). The magnetic equilibrium of a cell may be perturbed by neighboring current conducting systems capable of inducing a magnetic field bias in the operating cells which influence the magnetic stability of the cell; this is the second factor (external) that governs the MHD stability of a cell. Such neighboring current conducting systems can be neighboring cells (as cells are usually arranged side-by-side in series of up to several hundred cells and divided in at least two potrooms), neighboring potrooms, potlines, crossovers, and rectifiers.

The present invention addresses the second factor, in a peculiar situation, as will be explained below.

The perturbation by external factors leads to local variations of the vertical component of the magnetic field B _z , (z being the coordinate running upwards from the bottom to the top of the cell) in the liquid metal pad which destroy the anti-symmetry of B _z with respect to cell centre, required for good MHD stability of the cell. More precisely, the value of the vertical magnetic field should be zero in the geometrical center of the liquid metal pad, but the contribution from the adjacent rows gives a bias that can be greater than one millitesla, depending on potline current and distance to the adjacent rows of cells. The magnetic effect of neighboring cells can be decreased or compensated by an appropriate design of the potline, and prior art describes certain designs intended to resolve this problem. As an early example of such design, US 4,090,930 (assigned to Aluminium Pechiney) discloses the use of a compensation loop which produces an additional magnetic field substantially equal to that created by the adjacent row and opposite to it, by diverting a portion of the current from the upstream conductor from the cathode of the downstream cell, passing the diverted current below the cell and rejoining the diverted current with the outer upstream conductor after being passed below the cell. US 4,169,034 (assigned to Aluminium Pechiney) discloses the use of two compensation loops which produce an additional compensating magnetic field substantially equal to that created by the adjacent rows. US 4,683,047 (assigned to Alcan International Ltd.) achieves the same goal with asymmetric conductors below the cell. GB 2 027 056 (assigned to Swiss Aluminium Ltd.) discloses the use of an asymmetric cathode busbar system. A smaller perturbation factor of a cell operating under conditions of magnetic equilibrium is due to magnetic fields generated by more distant electric conductors carrying high currents. These conductors belong to the rectifiers and power station feeding the cell line, and to other lines of electric cells present in the same plant. Such perturbation induces a slight magnetic dissymmetry (so-called "bias") of the cell: the local variations of the vertical component of the magnetic field are no longer symmetric over the longitudinal direction of the cell, as they would be in the absence of an external magnetic field. In particular, the magnetic fields generated by adjacent rows of cells lead to a perturbation of the magnetic equilibrium of the row; this may lead to a decrease in overall efficiency of the plant.

It is customary to designate as the "adjacent row" the row closest to the row in question, while the "field of the adjacent row" is the resultant of the fields of all the rows other than the row in question.

Interaction between magnetic fields of adjacent rows can of course be decreased easily by increasing the center-to center distance between adjacent rows. With 200 kA pots this minimum distance was usually considered to be of the order of 50 m, but with 450 kA pots it should now be increased to 110 m or 120 m. However, this option may be expensive if the high consumption of land is an economic issue; the additional land may be unavailable when adding a new potline to an existing plant. Additional cost of linkage conductors may be a drawback, too. It is therefore desirable to be able to decrease or to compensate this effect by using an appropriate design of electrical conductors in the plant. Prior art claims to offer certain remedies to the perturbations described above, but patent documents do not always clearly address which of the abovementioned problems they are dealing with. US 4,072,597 (assigned to Aluminium Pechiney) discloses a busbar system that is asymmetric with respect to a vertical plane orthogonal to the long axis of the pot and passing through the center of the pot; this patent applies to a cell with end risers. US 4,169,034 (mentioned above) discloses the use of an auxiliary conductor at each end of the cell, parallel to the long axis of the pot and situated in the plane of the bath/metal interface as near as possible to the pot shell; a direct current of appropriate direction and intensity (of the order of 10 to 30 kA for a cell operating at 175 kA) is passed through this auxiliary conductor. US 4,713, 161 (assigned to Aluminium Pechiney) discloses the use of two correcting conductors per potline, one on each side of the pot, that extend parallel to the axis of the potline; the total current in these corrective conductors is up to 70 % of the current through individual cells. This seems to give satisfactory results for potlines operating at very high amperage. However, as explained below, the use of external compensation loops leads to locally high magnetic fields on both sides of the potline, which may increase the interaction between the magnetic fields generated by adjacent rows; as a consequence, the distance between two rows needs to be increased when building a new plant or adding a new row to an existing plant. As considerable progress has been achieved in recent years in the control, the stability and the homogeneity of high amperage electrolysis cells, the tolerance to perturbative effects of magnetic fields generated by adjacent rows is decreasing, knowing that the magnitude of these perturbative effects increases with the amperage. Furthermore, the investment cost of a potline is so high that it is desirable not to increase, and if possible, to decrease, the distance between two adjacent rows of electrolytic cells when designing a new plant or potline comprising cells of increased amperage.

WO 2015/017924 (assigned to Rio Tinto Alcan) discloses a compensation circuit parallel to the axis of the series; said circuit extends underneath the potline. This circuit comprises a set of three to ten parallel conductors which carry a current of between 50 % and 150 % of the electrolysis current. It is claimed that this allows decreasing the distance between two rows of cells to less than 40 m. Such a compensation circuit carrying such a high current requires a significant amount of metal, and ohmic losses will occur. The main disadvantage of this solution seems to be however the high cost of the power supply station needed for supplying current to the compensation circuit. WO 2017/064547 (assigned to Rio Tinto Alcan) discloses various designs of compensation circuits that extend either between two adjacent rows or at the opposite side (i.e. outside the rows). This solution is simple but not very efficient.

The problem that the present invention endeavors to resolve is therefore to decrease the effect of interaction between magnetic fields generated by two neighboring rows belonging to the same series of electrolytic cells when decreasing the distance between the rows of cells and/or increasing the current in the cells, without using long lines of compensation conductors carrying high current that is not needed for the electrolysis process and that requires independent power supply stations. Generally, there are a few well known solutions used to magnetically compensate the unbalance of vertical magnetic field (bias) of cells that are meant to operate in regions with high magnetic fields. For end pots, a special arrangement is used in the linkage busbars to compensate for the bias introduced by the crossover busbar in close proximity. For a more drastic arrangement where the full potroom/potline is expected to experience high magnetic field bias, a compensation loop is used. This involves additional current carrying busbars running around the potrooms with a dedicated power source (rectifiers).

However, there are a few draw-backs and complications with using dedicated compensation loops. An investment representing capital cost is required for a dedicated power supply and a permanent energy and operational cost for the magnetic compensation loop has to be sustained. From a safety perspective, there is the issue of properly isolating the compensation loop electrical system from the potline electrical system, as there is a huge electrical potential difference between the two. Another safety issue is related to an accidental power outage in the potline: in this case, due to the change in magnetic flux, the amperage drop between the potline and its compensation may induce very high currents/voltages in the compensation loop, potentially damaging the system.

Compensation loops are usually designed and installed upon the initial design and construction of the potline. They can be adapted when additional pots are added to the potline at a later stage. A peculiar problem arises when pots designed to be operated at a significantly higher current are added to an existing potline. This problem may arise when the capacity of an existing potline is increased by adding pots using a different technology.

When new pots are added to an existing potline, this usually takes place at the end of the potline. For example, in an existing potline comprising several hundreds of pots operating at 460 kA it may be decided to add a dozen of pots using a different technology operating at 600 kA or 700 kA at the end of the potline. These additional pots may perturb the magnetic equilibrium of the existing pots, or may be perturbed by the adjacent return current of the added pots if the pots are added at the end of one potroom only. On the other hand, if pots operating at higher current are added at the end of both potrooms, their distance will be too small for the magnetic perturbation to be negligible, and appropriate means are required to minimize this perturbation.

One of the problems that the present invention addresses is the compensation of the influence of the return current of added pots operating at a higher current than the existing pots on their magnetic equilibrium. Objects of the invention

The general idea behind a compensation system/loop is to intentionally produce a vertical magnetic field in the electrolysis cells which opposes and neutralizes, or compensates, the existing vertical magnetic field bias due to neighboring potrooms, potlines and crossovers or return busbars when there is no return potroom.

According to the invention this is achieved by deviating a portion of the potline current through a set of conductors that extend around a group of cells, thereby producing an opposing vertical magnetic field to the one generated by the neighboring rows of cells. According to the invention, the current is deviated from the potline busbars (in particular from the existing linkage busbar system) in the passageways between groups of cells and led back into the potline busbars after completing its way around the group of cells. The amount of current deviated into the compensation loop depends on the required magnetic compensation. This is achieved by creating a parallel circuit of busbars and adjusting resistances to provide the required currents in each path (see figure 5). Most of the potline current continues its normal path along the passageway. The portion of current in the self- regulating compensation loop running around the cells is adjusted during design phase of the potline layout to provide sufficient magnetic field compensation.

According to the invention the problem is solved by a system which requires no dedicated power source, and significantly reduces electrical hazard related to high electrical potential differences between the magnetic compensation loop and the potline.

More precisely, according to the invention, the problem is solved by an electrolysis plant, in particular for the Hall-Heroult process, comprising a plurality of electrolysis cells connected in series, each two neighboring cells being connected by a set of electrical conductors carrying a current of high intensity l _H , and said cells being arranged along at least two parallel rows forming a potline or along several rows in several potlines, the direction of said current of high intensity l _H in the cells arranged along two neighboring rows of the same potline being opposite to each other, said cells forming at least a first group of cells and a second group of cells, said electrolysis plant being characterized in that it further comprises a set of conductors (so-called "compensation loop") for at least partially compensating the vertical magnetic field induced by a first conductor carrying a first current of high intensity l _H1 upon a second group of cells (so-called "target group") connected by second conductors carrying a second current of high intensity l _H2 , wherein said current intensities l _H1 and / _H2 can be equal or different from each other, said compensation loop surrounding at least partially said target group of cells and having at least one portion that extends parallel to a line, wherein said compensation loop carries a current of low intensity l _c , with l _c < IH, such that the vertical magnetic field induced by said first current of high intensity l _H1 upon said target group of cells through which flows a second current of high intensity l _H2 s at least partially compensated.

Said compensation loop is connected in parallel to said current of high intensity l _H ; this will render in many cases the compensation loop self-regulating, i.e. a change in the high intensity current l _H will lead to a corresponding change in the current l _c in the compensation loop which will adapt the compensation of the vertical magnetic field to the new (increased or decreased) value of the high intensity current l _H .

This electrolysis plant forms the first object of the present invention.

In an advantageous embodiment said first and second currents are flowing in opposite directions. In another advantageous embodiment said compensation current l _c is derived from a current conductor carrying a current of high intensity l _H . Said compensation current l _c can be derived from a current conductor carrying said first current of high intensity l _H and/or from a current conductor carrying a current of high intensity bridging the last cell of an upstream group of cells to the first cell of the next downstream group of cells. In an advantageous embodiment, said compensation current l _c is selected by taking into account the distance between neighboring rows or lined of cells and the high intensity current l _H .

Typically said compensation current l _c is comprised between about 0.5 % and about 50 %, preferably between about 0.5 % and about 25 %, still more preferably between about 1 % and about 12 %, and most preferably between about 2 % and about 8 %, and most preferably between 2 % and 7 % of said current of high intensity l _H from which it is derived.

In one embodiment said target group of cells is operated at a higher current intensity l _H∑ than the other groups of cells.

Said first current of high intensity l _H1 can be generated, at least in part, by a booster rectifier station (BRS). In an advantageous embodiment of the electrolysis plant according to the invention the distance between said portion of the compensation loop that extends parallel to a line (L1 ,L2) and said target group of cells is smaller than the distance between said first conductor and said target group. In another advantageous embodiment said compensation loop is surrounding said target group of cells over at least three adjacent sides, preferably including the two longest sides, and preferably over at least 90% of the length of adjacent sides, including the two longest sides.

All the various embodiments recited above in relation with the electrolysis plant can be combined with one another.

The system according to the invention requires no dedicated power supplies for the compensation loop current. The compensation current will be derived from a booster rectifier station only if such a station is used for boosting the current supplied to a group of cells in the potline. It also requires only negligible operational cost in terms of energy consumption, provided that the investment in additional busbar mass (i.e. busbars of sufficient conductive cross-section) in the magnetic compensation loops is sufficient. The system is self-regulating for increased amperage: by changing the potline amperage, the loop current changes accordingly and automatically adjusts the required compensating magnetic field, whereas prior art systems (such as the one described in US 4,713, 161), in which the compensation loop is not connected in parallel to the potline current, would require an adjustment of the current in order to keep the compensation effect at the same level.

In terms of safety, the system according to the invention significantly reduces the electrical hazards generally faced with compensation loops according to prior art. The electrical potential difference between the loop busbars and the cells is limited to the number of cells in one section (of the order of a few tens or a few hundreds of volts) only, which is much lower than the full potline voltage (of the order of one or two thousands of volts). Since the potline current is not a closed loop within the self-regulating magnetic compensation loop there is no risk of induced currents/voltages in the self-regulating magnetic compensation loops when there is a sudden potline power outage. Also, with the system according to the invention there is no need for a special arrangement of linkage busbars for end pots to account for the influence of crossovers in close proximity. This is taken into account using the existing loop system. According to the invention, the conductors of the compensation circuit are approximately in the same plane as the pots. In an advantageous embodiment they are situated substantially at the horizontal level of the liquid aluminium pad in the cells.

Another objects of the invention is a method to extend the capacity of an existing electrolysis plant comprising a plurality of electrolysis cells connected in series, each two neighboring cells being connected by a set of electrical conductors carrying a current of high intensity l _H , and said cells being arranged along at least two parallel rows forming a potline or along several rows in several potlines, the direction of said current of high intensity l _H in the cells arranged along two neighboring rows of the same potline being opposite to each other, said cells forming at least a first group of cells and a second group of cells, said method comprising the addition of a group of cells operating at higher current intensity as the other cells, or the replacement of a group of cells operating at a given current intensity by a group of cells operating at a higher current intensity, or the replacement of a group of cells operating at a given current intensity by a group of cells operating at a higher intensity, and said method further comprising the addition of a set of conductors surrounding at least in part said group of cells operating at higher intensity (so-called "target group"), said set of conductors forming a compensation loop connected in parallel to said current of high intensity l _H and being capable of at least partially compensating the vertical magnetic field induced by a first conductor carrying a high current (which can in particular be the conductors carrying a potline current of intensity l _H or the conductors of a booster rectifier station) upon said target group.

Said cells operating at a higher intensity are preferably adjacent cells.

In said method, and in accordance with the specific features of the first object of the present invention:

Said compensation loop carries a compensation current l _c that is derived from a current conductor carrying a current of high intensity l _H .

Said compensation current l _c is advantageously selected by taking into account the distance between neighboring rows or lined of cells and the high intensity current l _H .

Said compensation current l _c is advantageously comprised between about 0.5 % and about 50 %, preferably between about 0.5 % and about 25 %, still more preferably between about 1 % and about 12 %, and most preferably between about 2 % and about 8 % of said current of high intensity l _H from which it is derived.

Said compensation loop is surrounding said target group of cells over at least three adjacent sides, including the two longest sides. Said compensation loop is surrounding said target group of cells over at least 90% of the length of adjacent sides, including the two longest sides.

All the various embodiments recited above in relation with method to extend the capacity of an electrolysis plant can be combined with one another; and with any of the embodiment recited above in relation with the electrolysis plant itself..

Figures

Figures 1 to 3 illustrate general features of an electrolysis cell and electrolysis plant using the Hall-Heroult process. Figures 4 to 9 illustrate embodiments of the invention.

Figures 1 and 3(a) show schematic views of a Hall-Heroult cell. Figure 1 schematically shows a vertical cross section of a typical Hall-Heroult electrolysis cell. It illustrates the current flow from the anode through the electrolyte to the cathode.

Figure 3(b) shows a vertical cross sections of three neighboring electrolytic Hall-Heroult cells connected in series, showing in particular the series connection of the cells that allow to feed the cathodic current of a cell into the anode beam of the neighboring downstream cell through anodic risers.

Figure 2 is a schematic view, showing the global arrangement of an electrolysis plant using the hall-Heroult process.

Figure 4 schematically shows a simplified electrical diagram of a so-called "series" of pots or "potline" according to the invention.

Figures 5 to 7 and 9 show each a specific embodiment of the invention, derived from the structure shown on figure 4.

Figures 5(a) and 5(b) show two different representations of a same, first embodiment of the invention. The representation of figure 5(a) shows the individual cells that form a sector of a potline, while the representation of figure 5(b) shows only the sectors as a block. In figure 5(b) the potline current busbars are represented in full lines, the compensation loop in dotted lines.

Figures 6(a) und 6(b) show two different representations of a same, second embodiment of the invention. Same remark concerning the figures (a) and (b) as for figure 5. Figures 7 and 8 relate a special case where pots operating at higher amperage are added to an existing potline ; there is no representation of the individual cells forming the group of the potline. Figures 7 (a), 7(b) and 7 (c) illustrate one embodiment of the invention where the booster circuit and the main line adjacent row are on the same side of the added cells. Figures 8 (a) to 8 (d) show another embodiment where the booster circuit and the adjacent row of the main line are on the opposite sides of the added cells.

Figures 9(a) and 9(b) show two variants of an embodiment for a potline with four groups of cells. The variant of Figure 9(a) is similar to that of figure 5.

Detailed description

The present invention relates to the Hall-Heroult electrolysis process, which is carried out in electrolysis cells called "pots" of substantial rectangular shape. The Hall-Heroult process as such, the general structure of a Hall-Heroult electrolysis pot, the way to operate the latter, as well as the cell arrangement are known to a person skilled in the art and will not be described here in more detail.

In the present description, the terms "upper" and "lower" refer to mechanical elements in use, with respect to a horizontal ground surface. Moreover, unless otherwise specifically mentioned, "conductive" means "electrically conductive". The expression "extends in parallel" has a geometrical meaning, whereas the expression "connected in parallel" refers to electrical circuitry. The expressions "upstream" and "downstream" refer to the direction of electrical current flow.

As the Hall-Heroult process is known as such, it is sufficient to explain (as done here in particular in relation with Figures 1 and 3), that the pot 1 typically comprises a potshell 2 usually made from steel, and a lining comprising a carbonaceous cathode formed from individual, parallel cathode blocks 3 and a side lining 4, said lining defining a volume for the liquid electrolyte 5 and the liquid metal pad 6 produced by the electrolysis. Current is fed into the anode busbar 16 (also called "anode beam"), flows from the anode beam 16 to the anode rod 7 (said anode rod 7 being connected to the carbon anode 8 by means of anode yoke 20) and to the carbon anode 8 in contact with the liquid electrolyte 5 where the electrolytic reaction takes place, crosses the liquid metal pad 6 resulting from the electrolysis process and eventually will be collected at the cathode block 3. As cathode blocks are symmetric and have collector bar 9 ends coming out on each side, in side by side arrangements of electrolytic cells half of the current collected by the collector bars 9 of the cathode blocks 3 will flow directly to the downstream longitudinal part 10 of the cathode busbar system, while the other half flows to the upstream longitudinal part 11. Flexible connectors 12 are used to connect the ends of the cathode collector bars 9 to the cathode busbar 10, 11. Said collector bars 9 can be full bars, as in figure 1 , or half bars 9a, 9b, as in figure 3b. Conductor 13 carries the current collected at the upstream part 11 of the cathodic busbar system to the anode risers 18 of the downstream pot. The current collected at the downstream part 10 of the cathodic busbar system directly feeds the anode risers 18. A Hall-Heroult cell further comprises an alumina feeding system 14 (usually located inside the carcass of the superstructure 19) through which alumina powder is fed from time to time into the cell volume (see arrow on figure 3b). The air space above the cell is closed by a set of covers or hood panels 15 that can be removed for maintenance and anode change; the anode rods 7 are adjustably fixed to the anode beam 16 using anode clamps 17 that allow to adjust the anode heights in order to keep the inter-electrode spacing constant as the anode is consumed. The present invention is directed to certain features related to the global arrangement of a plant, or aluminium smelter, using the Hall-Heroult process. As schematically shown on Figure 2, an aluminium smelter comprises a plurality of electrolytic cells Ci, C ₂ , ... , C _n-1 , C _n , arranged the one behind the other (side by side or end to end) along two parallel lines L1 and L2, each of which comprises n/2, i.e. m cells. These cells are electrically connected in series by means of conductors, which are not shown on Figure 1. The number of cells in a series is typically comprised between 50 and 500. The electrolysis current therefore passes from one cell to the next, along arrow DC. The cells are arranged transversally in reference of main direction D1 or D2 (axis of the row) of the line L1 or L2. In other words the main dimension, or length, of each cell is substantially orthogonal to the main direction of a respective line, i.e. the circulation direction D1 , D2 of current. Figure 2 depicts a typical "clockwise" current orientation; the present invention is explained below for a counter-clockwise arrangement. This arrangement of n cells connected in series along two parallel lines is called a "series" of pots or "potline". The cells depicted on figure 2 are arranged side-by-side; end-to-end arrangements are generally not used in newly built plant, but the present invention could be applied to end-to-end arrangements, too.

As can be seen on Figures 3a and 3b, anode risers 18 are provided to carry the current collected at the downstream part 10 and at the upstream part 1 1 of the cathode busbar system of the upstream pot (noted here C _n -i) to the anode beam 16 of the neighbouring downstream cell (noted here C _n ). Figure 4 schematically shows a simplified electrical diagram of a so-called "series" of pots or '"potline" according to the invention. The electrolytic cells are arranged side-by-side in two parallel lines L1 , L2. The cells are denoted by the letter C, as in figure 2. They form groups of cells; Figure 4 shows four such groups sectors G1 , G2, G3, G4. Each group can comprise an equal number or a different number of pots.

The grouping into groups is according to the order of the pots in the potline: the first group G1 is formed by the first n pots (C _1:1 , C _1i2 , C _1in-1 , C _1in ) in the potline (only four pots per group are shown on figure 4); while the last group (G4 on figure 4) is formed by the last p pots (C _4, i, C ₄ ,2, C _4,q- i, C _4,q ) in the potline. In the example of figure 4, when the potline is in normal operation, electrical current is supplied from a rectifier substation RS to the first pot C _1:1 of the first group G1 of the first potroom D1 of the potline. This current flows through the conductors of pots Ci _, i, Ci,2 ... Ci _,n- i, Ci _,n of the first group G1 and flows then through the conductors of pots C ₂ ,i, C ₂ ,2 ... C ₂ , _o -i, Ci, _o of the second group G2, and flows then through the conductors of pots C _3, i, C ₃ ,2 ... C _3,p- i, C _3, p of the third group G3 _h and flows then through the conductors of pots C _4, i, C ₄ ,2 ... C _4,q- i, C _4,q of the fourth and last group G4.

As mentioned in relation with figure 1 , the current crosses a pot by entering from the anode busbar (anode beam), to which it is fed, through the anode rod 7 (shown on figure 1) into the anode 8, crosses the molten electrolyte 5, where alumina is reduced into aluminium and oxygen, crosses the underlying molten aluminium pad 6 and enters into the cathode 3 where it is collected by the cathode collector bar 9 and carried to the cathode busbar system 10,11. From there it is fed into the anode risers of the neighbouring downstream cell (not shown in figure 1 ). The conductors through which the current is fed from the cathode busbar of the upstream pot into the anode busbar of the downstream pot are referenced in figure 4 with the reference numbers 100 to 108. They comprise so-called anodic risers 18 (shown on figure 3). In the example of figure 4 four such conductors are schematically represented; their number is of no importance in relation with the present invention.

When leaving the last pot C _2,0 of the last group (here G2) of the first potroom, the current is collected by a crossover busbar system 104 which conducts the current back to the second potroom D2 where it enters the first group of the second potroom (here group G3) through the first pot C _3, i. Then the current flows from the cathode busbars of pot C _3i1 to the anode risers of pot C ₃ ,2, and flows then though the conductors of pots C _3,p-1 C _3,p and so on, and eventually leaves the potline by the last pot C _4,q of the last group (here G4) of the second potroom where it is collected by the rectifier substation RS. Typically two adjacent groups in each potroom are separated by a passageway Pi for workers and trucks (shown on figure 4 as a dotted line).

The invention is based on the idea to use a small part of the potline current itself for the compensation loops. Figure 5 shows a first embodiment of the invention. Figures 5(a) and (b) show the same circuit, but figure 5(a) allows relating this circuit to the structure of a plant as shown on figure 4: the circuit of figure 5 is a partial representation of a circuit of an electrolysis plant similar to the one shown on figure 4, the groups of cells corresponding to the direction D2 (namely groups G3 and G4) not being shown. The difference between the circuit on figure 4 and on figure 5 is the presence of the compensation loop 540 on figure 5. Said compensation loop 540 is surrounding the target group G2 (current direction D1) and compensates the vertical component of the magnetical field induced upon said target group by the conductors of group G3 in which the current flows in the direction D2 opposite to direction D1.

The main busbar system is represented on figure 5(a) by four packs of parallel conductors 100, 101 , 102, 103, each of which is represented here by four parallel conductors. The exact number of parallel conductors does not matter in the framework of the present invention. What does matter is their electrical resistance, as will be explained below. Concerning the compensation circuit, using a plurality of parallel conductors gives some leeway for modifying the total resistance of the whole compensation circuit by cutting in or out one or more of these parallel conductors.

In figure 5(b) the compensation circuit 540 is represented as a dotted line. It starts at connection point 541 located downstream with respect to the last pot Ci _,n of group G1 , fully surrounds the group of pots G2 and goes back into the potline busbar at a connection point 542 located downstream with respect to point 541 but upstream with respect to the first pot C _2, i of group G2. The current in the compensation circuit, called here "compensation current" l _c , is a small fraction of the potline current of high intensity l _P from which it is derived.

According to the invention, the conductors of the compensation circuit are approximately in the same plane as the pots. In an advantageous embodiment they are situated substantially at the horizontal level of the liquid aluminium pad 6 in the cells. Figure 6 shows a second embodiment of the invention. Figures 6(a) and (b) show the same circuit, but figure 6(a) allows relating this circuit to the structure of a plant as shown on figure 4. It refers to an electrolysis plant having at least eight groups of cells connected in series, four of which are aligned along current direction D1 (only three of them, namely G2, G3 and G4, are shown on figure 6, whereas the four other ones, in which the current direction D2 is opposite to direction D1 , are not shown.

Figures 6 and 5 apply the same principle of a compensation circuit, but in the embodiment of figure 6 two target groups of pots, namely G3 and G4, have a compensation circuit, designed to compensate the vertical component of the magnetical field induced upon each of said target groups G3, G4 by the conductors of groups G5 and G6, respectively, in which the current flows in the direction D2 opposite to direction D1.

The compensation circuit 640 for the group of pots G4 starts at connection point 641 , located downstream with respect to the last pot of the preceding (upstream) group G3, fully surrounds the pots of group G4 and goes back into the potline busbar at a connection point 642 located downstream with respect to connection point 641 but upstream with respect to the first pot C ₄ ,i of group G4. The compensation circuit 630 for the group of pots G3 starts at connection point 631 , located downstream with respect to the last pot of the preceding (upstream) group G2, fully surrounds the of pots of group G3 and goes back into the potline busbar at a connection point 632 located downstream with respect to connection point 631 but upstream with respect to the first pot C _3i1 of group G3.

In the framework of the present invention, the current in the compensation circuit, called here "compensation current" l _c , is always a small fraction of the current of high intensity l _H (which can be, as will be explained below, a potline current l _P , or a booster rectifier current l _B ); this small fraction is typically between about 0.5 % and about 15 %, preferably between about 1 % and about 12 %, more preferably between about 2 % and about 8 % of the current of high intensity l _P . The exact percentage required depends on the distance between the rows of the main potline and on the distance of the compensation loop from the cells to be compensated.

As an example, for a potline current l _P of 450 kA the compensation current l _c is advantageously comprised between 5 kA and 50 kA, preferably between 10 kA and 40 kA, and still more preferably between 10 kA and 30 kA.

Figures 7 and 8 relate to a third embodiment of the invention. Figure 7 shows a potline comprising two potrooms PR1 and PR2, each comprising a large number (for example 200) pots operating at about 460 kA, divided into a plurality of groups (the number of which is not relevant here, and for this reason the target group is designated as GX, wherein X is representing an integer designating the number of the target group). The centre lines of the potrooms are separated by a distance d ₅ that should be sufficient to avoid interaction between magnetic fields magnetic interaction of adjacent potlines or other conductors. It is known to decrease this interaction by increasing the center-to center distance between adjacent rows. With 200 kA pots this minimum distance d ₅ was usually considered to be of the order of 50 m, but with 450 kA pots it should be increased to 1 10 m or 120 m.

Only one pot for each potroom is identified on figure 8 in order to clarify the terminology. Figures 7 and 8 show an existing potline with two potrooms each representing one of the main directions D1 , D2, in which the current enters the first pot C _1:1 of the first group G1 (not shown on the figure) and leaves the last pot Ci _,n of the first group G1 (not shown on the figure) is modified at its geographic end (i.e. at the end of the potrooms, between the last group G2 of the first potroom and the first group G3 of the second potroom) by adding a group GX comprising a rather small number of pots.

In the example of figure 7 (a) the potline is to be extended by adding a small number of pots (for example 12) forming a group GX at the end of potroom PR1. These added pots operate at a much higher current (for example 700 kA), and additional current is supplied by a booster rectifier stations BRS. The problem is now to minimize the magnetic interaction on the target group of cells GX from the conductors 911 feeding the return current into the cells of potroom PR2 and from conductors 911' of BRS which are additive because they are on the same side of GX cells. The magnetic interaction between these conductors on the pots of target group GX is larger than that between the conductors of potrooms PR1 and PR2, because the current in GX is much higher than in conductor 911. In figure 7 (a), the distance d ₆ between conductor 911' and group GX is smaller than the distance d ₅ : while the distance d ₅ is sufficient for having only negligible magnetic interaction between potrooms PR1 and PR2, this is no longer the case for conductor 911' and group GX. A partial solution to this problem is shown on figure 7 (b) : The distance d ₇ between group GX and conductor 911' has been increased to a much higher value (for example 170 m) in order to decrease their magnetic interaction on the target group GX. This may not be feasible if the land is not available and will anyway result in increased capital cost for the additional conductors and operating cost for ohmic losses in said additional conductors.

The complete solution to this problem according to this invention is a self-regulating compensation loop shown in Figure 7 (c): Part of the BRS current of high intensity l _B is deviated into a compensation loop which runs around the target group of cells GX. The compensation loop length is determined by the number of cells in the target group GX and the cross-section is determined so that the loop electrical resistance gives the desired current in the loop. The current of low intensity l _c in the compensation loop flows in the opposite sense to the current in the BRS circuit. The compensation current in the loop is self- regulating for current increase in the BRS loop; If the booster current is increased to increase the current in GX cells, the compensation loop current also increases.

Figures 8 (a) to 8 (d) show another situation where the BRS circuit and the main potline current of the row adjacent to the target group of cells GX are on the opposite sides of the group of cells GX. In this case the magnetic field from the return potline current partially or completely compensates the the vertical component of the magnetic field induced by the BRS circuit. Partial compensation will be achieved if the return potline current remains in the original distance d ₅ , extends in a straight line parallel to GX cells. In this case, shown in Figures 8 (a) and (8b), the remaining magnetic field bias is compensated according to the invention by a compensation loop, which feeds from the main circuit (Figure 8 (a)) or from the booster circuit (Figure 8 (b)). The loop in Figure 8 (a) is self-compensating for current increase in the main potline; the loop in Figure 8 (b) is self-compensating for current increase in the BRS circuit.

In another, preferred, embodiment of the invention the extension of the return row D1 of the main potline is shifted towards the target group of GX cells to such a distance that the vertical magnetic field from BRS circuit is completely compensated for a given current of GX cells. Figures 8 (c) and 8 (d) show the case where a considerable current increase is planned in GX cells. The position of the adjacent row of busbars is chosen so that GX cells will be compensated at maximum current. At less than maximum current, the GX cells will be overcompensated; in this case, the invention uses a magnetic compensation loop as shown in Figures 8 (c) and 8 (d). Note that in this case the loop current flows in the opposite direction than in Figures 8 (a) and 8 (b), i.e., the loop is intended to compensate the magnetic field of the current in the return row of the main potline circuit. Again the loop current can be tapped from the main potline circuit (Figure 8 (c)) or from the BRS circuit (Figure 8 (d)). The loop in Figure 8 (c) is self-compensating for current increase in the main potline; the compensation loop in Figure 8 (d) is self-compensating for current increase in the BRS circuit.

In other words: In figures 8(a) and 8(b) the aim is to compensate the vertical magnetic field induced by the booster current of high intensity l _B upon the potline current of high intensity l _P flowing through added pots GX. This is achieved by having in a portion of the compensation loop that is close to the target group GX and parallel to a line L1 said compensation current flowing in a direction D2 opposite to the direction D1 of said booster current; said compensation current of low intensity l _c can be derivated from the main potline current of high intensity l _P as in the embodiment of figure 8(a), or from the booster current of high intensity l _B in the embodiment of figure 8(b).

In figures 8(c) and 8(d) the aim is to compensate the vertical magnetic field induced by the main potline current of high intensity l _P upon the potline current of high intensity l _P flowing through added pots GX. This is achieved having in a portion of the compensation loop that is close to the target group GX and parallel to a line L1 said compensation current flowing in a direction D2 opposite the direction D1 of the potline current; said compensation current of low intensity l _c can be derivated from the main potline current of high intensity l _p as in the embodiment of figure 8(c), or from the booster current of high intensity l _B in the embodiment of figure 8(d). If the number of added pots in GX is too high such a modification might not be possible due to the limit of the existing rectifier substation. Furthermore, the number of pots that can be added in GX can be limited by several other factors such as availability of land, the limits of the existing electrical components which would need a complete redesign of the existing electrical system, or it can be economically unreasonable compared to the construction of a new potline, in particular due to peripherics like building and cranes.

In the present case, the original potline (only part of which is depicted on figures 8 (a) to 8 (d), namely groups G2 and G3) operates at about 460 kA; this is the current flowing through the pots of groups G2 and G3. The added pots in group GX operate at a significantly higher current (700 kA) than the existing ones: their perturbative effect on neighbouring pots and potlines is therefore higher than the perturbative effect of the existing pots of groups G2 and G3. This additional current is supplied by a booster rectifier substation BRS that is installed next to group GX. As shown in figures 8 (a) and 8 (b), this additional current of 240 kA enters into the circuit at connection point 723 located between groups G2 and GX, and leaves the circuit at connection point 724 located between Group GX and G3. According to the invention, group GX is compensated by a compensation circuit 700 that starts at connection point 721 located upstream of the first pot of group GX, surrounds group GX and then goes back into the main busbar at a connection point 722 located upstream with respect to the first pot of group GX but downstream with respect to connection points 721 and 723. In the embodiment of the present invention, the compensation loop 8 (a) is self-regulated for current increase in the main potline and the compensation loop in 8 (b) for current increase in BRS circuit. This means that the compensation current \ _c is a constant fraction of the potline current l _p from which it is derived. When the potline current l _p is modified, for instance by modifying one or more pots of the upstream potlines, then the compensation current \ _c will adjust. This gives a satisfactory adjustment for rather small variations of the potline current l _p , such variations being typically of the order of ± 10 to 15 %.

In the example of figure 8 (a) and 8 (c) the compensation loop is fed by the potline current of the pre-existing potline, i.e. the current that leaves group G2. It would be within the scope of the present invention that the compensation loop is fed by the booster current coming out of the booster rectifier station BRS; this variant is shown figures 8 (b) and 8 (d).

In Figures 7 and 8, the distance d ₅ between the pot centre lines of the main current is in general fixed as explained above in function of the amperage and of the land available. The distance d-, between the cells GX and the booster BRS is general determined by space available. The distances d ₂ and d ₃ between the cells GX and the compensation loop are near the cells GX at a distance which compensates well the vertical magnetic field bias for a given current in the loop. The distances d ₂ and d ₃ can be equal or unequal. The distance d ₄ is determined so that this part of the main potline current compensates well the the magnetic field of BRS circuit in GX cells. The cells of group GX do not need to be arranged along on of the lines L1 or L2 (as it is the actually case in the embodiments shown on figures7 and 8) but should be parallel to said lines.

Figure 9(a) shows an embodiment for a potline with four groups of cells. This embodiment is similar to that of figure 5: group G2 is surrounded by a compensation loop 800 that derives from the potline current busbar system at connection point 801 , located between group G1 and group G2, surrounds group G2 and goes back into the potline current busbar system at connection point 802 located between connection point 801 and the first cell C ₂ ,i of group G2.

Figure 9(b) shows a variant of the embodiment of figure 9(a) in which the compensation loops surrounds group G2 : the compensation loop 805 derives from the potline current busbar system at connection point 803 located between group G2 and group G3, surrounds group G2, and goes back into the potline current busbar system at connection point 804 located between group G2 and group G3, downstream with respect to connection point 803. As can be seen from all the embodiments and variants shown on figures 5, 6, 8 and 9, in the portion of the compensation that is located between both potrooms (each of which represent one of the directions D1 or D2) the direction of the compensation current \ _c is flowing in a direction that is opposite to the direction of the potline current l _p against which the group(s) surrounded by the compensation loop is to be shielded. As an example, in figure 9(a) the direction of the compensation current in the portion of the compensation circuit that is located between group G2 in the potroom representing direction D1 and the corresponding group G3 in the potroom representing direction D2 is opposite to the direction of the potline current in the potroom representing the direction D2. Likewise, in figure 9(b) the direction of the compensation current in the portion of the compensation circuit that is located between the potroom representing direction D1 and the potroom representing direction D2 is opposite to the potline current direction D2. The same arrangement can be seen on figure 8. It should be noted that in figures 5 and 6 the groups of pots in the second potroom are not shown, but a person skilled in the art can see easily that the current in the compensation circuit derived from the potroom representing direction D1 is opposite to the potline current direction D2.

As can be seen from figures 5, 6, 7(a), 7(c), 8, and 9, the compensation loop can surround the target group of cells over at least three adjacent sides, including the two longest sides. In the embodiments of figures 5, 6, 8(a), 8(d) and 9(a), the compensation loop is surrounding said target group of cells over all of the length of its adjacent sides.

Example 1

In a potline having 36 pots per group the self-compensation had to be made on one group of pots adjacent to the crossover. The adjacent row of pots was at 1 10 m distance. A magnetic field bias of 1 mT had to be reduced to 0.5 mT, to be equal to the bias in the mid-sections of the potline. The current was taken out at the exit from the end pot at point 803 of Figure 9(b) and brought back from around the group to the end of the crossover as at point 804. The line current was about 460 kA, the loop current was 15 kA. The distance from point 803 to point 804 between two connections of the loop to the linkage busbars of the passageway was 125 m. The distance between the center lines of two neighboring pots was 6 m. Eight busbars of section 200 mm x 800 mm were used for linkage between points 803 and 804 with a resistance of 3.878 μΩ. The self-compensating loop had 2 busbars of section 100 mm x 800 mm and a length of 601 m with a mass of 260 tonnes and a resistance of 1 14.768 μΩ. Example 2

In the example of figure 7 (a) the potline of 260 kA was extended by adding five pots forming a group GX at the end of potroom PR1. These added pots operate at a much higher current of 460 kA, and additional current is supplied by a booster rectifier stations BRS which supplies 200 kA. The adjacent row of pots is at 65 m distance. The booster rectifier station BRS is at 53 m distance from the centre of the group of 5 pots. The magnetic field bias of 2.2 mT had to be reduced to less than 0.5 mT. The current was taken out near the exit from the BRS as shown in figure 7 (a) led around the extended group of 5 pots and brought back from around the group to the entrance of the whole current into the first pot of the extended group. The line current of the extended group was 460 kA, the loop current was about 42 kA. The self-compensating loop was composed of three busbars of cross section 72 mm x 700 mm, passed at about 1 1 m from the centerline of the pots, had a length of 134 m and a mass of 55.5 tonnes. The self-compensating loop resistance was 33.17 μΩ. The resistance of the main booster busbar between the outgoing self compensating loop and the main potline was 4.8 μΩ and 2.6 μΩ in the potline with full current up to the point where the the self- compensating loop joins the potline.