In particular, the present invention relates to a method for controlling power or voltage drop at a variable load.

The present invention also relates to methods, devices and a system for limiting a current in a variable load.

The present invention also relates to a method and a system for voltage regulation of an ac circuit with a capacitive load or a load dominated by capacitive elements, and where the capacitance of the load may be more or less (stochastically) variable (within certain bounds).

The present invention also relates to a power supply control system and a method for controlling a power supply system with a variable resistive load. Here, the term "resistive load"shall be interpreted as a pure resistive load or a load with dominating resistive elements. The resistive load can also comprise a bucking voltage.

BACKGROUND OF THE INVENTION Inductive components, i. e. inductors and transformers with permeability control are described e. g. in the applicant's PCT/NO01/00217 and PCT/N002/00435 which are hereby incorporated by reference. These components provide permeability by means of substantially perpendicular flux.

Other types of controllable inductors are also known. US 5.684. 678 shows a resonant converter with a controlled inductor, where the controlled inductor is used to vary the effective capacitance of the resonant capacitor in the converter. The inductor comprises a power winding wound around a core and a control winding.

The control winding creates a controlled flux in the core, which biases the core to an operating region on its permeability curve. One example of the controlled inductor uses orthogonal control flux to vary core permeability. This inductor yields the least surrounding flux leakage field. Moreover, it is stated that it has a very limited range of control of inductance, which makes it unsuitable for the present applications.

Transformers are usually employed as an intermediary between a power supply and a load. The power supply is connected on the primary side of the transformer while

the load is connected to the secondary side. A load impedance reduction leads to an, increased current in the secondary and the primary side and a lower secondary voltage. This increased current can damage parts of the transformer. On the other hand, an increase of the load impedance will lead to an increase in voltage. The output is not controlled, but given by the primary source and the transformer plus the load. In an extreme case the load impedance is reduced to zero, leading to a secondary short circuit.

Variable load: control of power or voltage drop A first object of the invention is the to provide a method and a device which permits control of power or voltage drop in a variable load, to be able to ensure optimal working conditions for the load and the rest of the circuit at all times.

Reduced load impedance: short circuit situation While the one embodiment of the present invention will be explained with reference to a special case of load impedance reduction, a short circuit situation, it can obviously be used for limiting secondary current in other cases of load impedance reduction.

When a transformer is shorted, several serious effects might arise: personnel injuries, excessive mechanical forces, possibility for fire, decreased lifetime for the device itself and for interconnected equipment. Fuses will normally prevent or limit the consequences of such results, but one will still have degradation of equipment.

Further, it is an object of the invention to provide methods, devices and systems to control a short circuit current (or a current caused by a variable load) by controlling the mutual and/or leakage flux in a transformer, and hence limit the current to a safe level.

Variable capacitive load: coalescer The invention will also be illustrated in detail by means of an example related to voltage control in a circuit comprising an electrostatic coalescer. An electrostatic coalescer is an apparatus for separating water from oil, and represents a variable capacitive load. The method and system according to the invention will be applicable for voltage regulation of any capacitive load, as it addresses the main problems relating to this issue.

Today, such variable capacitive loads could be fed by a variac via a transformer when voltage regulation is wanted.

The variac solution is not a good solution for continuous and automatic regulation, as it requires mechanical intervention, which makes it especially ill suited for subsea applications. Furthermore, the capacitive load may come into resonance with inductive elements in the transformer, creating overvoltages that may destroy

components and represent danger for the surroundings. It is therefore an object of the present invention also to overcome these drawbacks related to prior art.

Variable resistive load : DC-supply Power supply for large loads is nowadays performed by means of transformers and rectifier circuits. Rectifier circuit control is today, as an example, performed with the use of thyristors in different configurations, e. g. configurations in six pulse bridge coupling, six pulse star coupling or in double three pulse star coupling with interphase transformer, depending on the current needs on the DC side. These rectifiers have some disadvantages, i. e. noise generation to the voltage on the DC side, voltage distortion at the point of common coupling, line notching, current distortion. Such rectifiers can have a complex structure, and an additional transformer, coils and filter elements may be needed for some applications.

It is a further object of the invention to achieve an improved power supply control system and a method for controlling a power supply system with variable resistive load.

SUMMARY OF THE INVENTION Example 1 and 2: Controlling power in a load To achieve the above mentioned objects, the invention in one aspect is related to a method for controlling power supplied to a load, by means of a transformer with a primary winding connected to a power source, a secondary winding connected to the load, and a magnetic flux path, comprising: - dividing the flux path in a common flux path for the primary and the secondary winding and at least one bypass flux path, - upon, variations in the load impedance adapting the power transferred to the load to the new load impedance or maintaining the voltage drop in the load impedance substantially unchanged, by - changing the effective permeability of the common flux path to change the common flux, or - changing the effective permeability of the at least one bypass flux path to change the leakage flux, or - changing both the effective permeability of the common flux path and the effective permeability of the bypass flux path to change the common flux and the leakage flux.

The load can be a constant load or a variable load. Further, the load can be resistive, inductive or capacitive, or be dominated by resistive, inductive or capacitive elements. The control of the power can comprise many types of control based on

different measurements and/or estimates of reference values, such as limiting a current through the load, voltage regulation of the load voltage etc.

In an embodiment of the invention the bypass flux path is linked essentially only by the primary winding. In another embodiment the bypass flux path is linked essentially only by the secondary winding.

According to a preferred embodiment of the invention the effective permeability is changed by means of a magnetic field creating a flux substantially perpendicular to the common flux and/or the bypass flux respectively.

Variable voltage There are several possibilities for regulation, depending on the type of load (motor, electrostatic filter, lamps, rectifiers, ovens/heaters, etc) and the method according to the invention is adapted to the different applications, which will implement a variable voltage source, which comprises: a) upon decrease in the load impedance decreasing the effective permeability of the common flux path and/or increasing the permeability of the bypass flux path to reduce power (or keep constant power, or keep constant current) supplied to the load, or b) upon increase in the load impedance increasing the effective permeability of the common flux path and/or decreasing the permeability of the bypass flux path to increase power (or keep constant power, or keep constant current) supplied to the load.

Constant voltage In other applications it is vital to ensure constant voltage drop in a load, that is to implement a constant voltage source. The method according to the invention will the comprise: a) upon decrease in the load impedance increasing the effective permeability of the common flux path and/or decreasing the permeability of the bypass flux path or b) upon increase in the load impedance decreasing the effective permeability of the common flux path and/or increasing the permeability of the bypass flux path to decrease power supplied to the load, to maintain voltage drop at the load impedance unchanged.

In an embodiment of the invention, the power supplied to the load is reduced by limiting current in the load.

The invention also comprises the use of a method according to the methods above, for limiting a short circuit current.

The invention comprises also a device to perform the methods related to the invention. The device for controlling power supplied to a load comprises: - a transformer with a primary winding, a secondary winding for connection to the load, and a core establishing a magnetic flux path, wherein: - the transformer's core comprises a common member for establishing a common flux path for the primary and the secondary winding and at least one bypass member for establishing a bypass flux path, and for adapting the power transferred to the load to the new load impedance, or maintaining the voltage drop in the load impedance substantially unchanged, - at least one permeability control device, where a) the permeability control device is connected to the common member for, upon variation in the load impedance, changing the effective permeability of the common flux path to change the common flux, or b) the permeability control device is connected to the bypass member for, upon decrease in the load impedance, decreasing the effective permeability of the bypass flux path to increase the common flux, or c) a first permeability control device is connected to the common member and a second permeability control device is connected to the bypass member for, upon decrease in the load impedance, increasing the effective permeability of the common flux path and decreasing the effective permeability of the bypass flux path to decrease the leakage flux.

In a preferred embodiment, the primary winding is adapted to link essentially all flux in the bypass member.

In another preferred embodiment, the secondary winding is adapted to link essentially all flux in the bypass member.

In a preferred embodiment, the permeability control device comprises a control winding adapted to create a flux substantially perpendicular to the common flux and/or the leakage flux respectively.

The device according to the invention comprises a transformer. This transformer can be the"power supply"transformer mentioned above, that is a transformer connected more or less directly between the load and the power supply. It can also be an additional device connected to the load circuit with substantially a protection function, or be a part of the power supply (that is with power electronics upstream and downstream).

The transformer can also be implemented as a combination of a parallel and a series inductor.

The system further comprises a system for controlling power supplied to a load comprising: - a measurement unit for measuring load current and/or load voltage, -an interface for input of set point values, - a device according to the above mentioned devices, - a processor connected to the measurement unit and the at least one permeability control device for controlling the transformer's permeability based on the current measurements.

Example 3 and 4: Voltage regulation of a circuit with a capacitive or resistive load The invention further relates to a method for voltage regulation of a circuit with a load and a series connected controllable reactor provided in a magnetic controlling device, comprising: - measuring or estimating voltage drop and/or current through the load, - based on these measured/estimated values and on reference values for voltage drop across and/or current through the load, varying the effective permeability of the reactor's magnetic core, and thus the reactor's reactance to obtain a voltage drop across the reactor leading to the reference value of voltage drop (Vref) across the load.

In a preferred embodiment, the method comprises varying the effective permeability in the series reactor's magnetic core by means of a control current (control) which creates a field substantially perpendicular to the main magnetic field in the series reactor's magnetic core.

In a preferred embodiment, the method comprises varying the effective permeability in the parallel reactor's magnetic core.

The invention further comprises a system for voltage regulation of a circuit with a load, comprising: - a series reactor provided in a magnetic controlling device for connection in series with the load, - a measuring and/or estimating unit for measuring/estimating voltage drop across and/or current through the load, - an input device for inputting reference values for voltage drop and/or current through across the load, - a permeability control device for regulating the series reactor's reactance, - a processing unit connected to the measuring/estimating unit, the input device and the permeability control device for, based on the measured and/or estimated voltage and/or current and on the reference values, controlling the output of the permeability control device to obtain a voltage drop across the reactor leading to the reference value of voltage drop across the load.

Preferably, the permeability control device comprises a control current source which creates a field substantially perpendicular to the existing flux in the series reactor's magnetic core.

Capacitive load As is well known, when a circuit comprises inductors and capacitors, there is a risk of resonance.

When the load is capacitive, the method in a preferred embodiment further comprises : - measuring or estimating a range of impedance values for the capacitive load, and deriving a range of series resonance values for the series reactor, - based on the measured/estimated voltage drop and/or current values and on established values calculating a desired series impedance for the series reactor which will lead to the reference voltage drop across the capacitive load, - comparing the desired impedance value with the range of resonance values for the series reactor, and - if the desired impedance value belongs to the range of resonance values or differs from this range by a predetermined magnitude, providing an overvoltage protection device for the capacitive load.

When the load is capacitive, the system in a preferred embodiment further comprises that the measuring and/or estimating unit for measuring/estimating voltage drop across and/or current through the capacitive load also measures or estimates a range of impedance values for the capacitive load, and the processing unit connected to the measuring/estimating unit, the input device and the permeability control device performs the following steps: - based on the range of impedance values for the capacitive load, deriving a range of series resonance values for the controllable reactor, - based on the measured/estimated voltage drop and/or current values and on established values calculating a desired series impedance for the controllable reactor which will lead to the reference voltage drop across the capacitive load, - comparing the desired impedance value with the range of resonance values for the reactor, and - if the desired impedance value belongs to the range of resonance values or differs from this range by a predetermined magnitude, controlling use of an overvoltage protection device for the capacitive load, - based on the measured/estimated values and on reference values for voltage drop across and/or current through the capacitive load controlling the output of the permeability control device permeability control device to obtain a voltage drop across the reactor leading to the reference value of voltage drop across the load.

The overvoltage protection device can comprise a second reactor connected in parallel to the capacitive load. In this case the processing unit can be connected to a second permeability control device for varying the second reactor's core effective permeability. This is done in a preferred embodiment of the invention by a system where the permeability control device comprises a control current (Icontrol) source which creates a field substantially perpendicular to the existing flux in the reactor's magnetic core.

Connection of the load to a power supply will be triggered by the processing unit in an embodiment of the invention.

The overvoltage protection device can be a known device, such as surge arresters and varistors possibly in combination with fuses, and/or protective power electronics. These devices may lead to unwanted disturbances in the process itself since they essentially short-circuit across the capacitive load, such that e. g. the actual oil/water separation process is disturbed. Furthermore varistors degrade over time and may have to be changed often when overvoltages occur frequently.

Frequent replacement of components is not acceptable e. g. in subsea applications. A preferred solution is that the overvoltage protection device is able to prevent the overvoltage from occurring in the first place. In a preferred embodiment, the overvoltage protection device comprises a second reactor connected in parallel with the capacitive load.

In a preferred embodiment, the series and parallel reactors comprise the leakage and the magnetizing impedance in a transformer with a primary winding connected to a power source, a secondary winding connected to the load, and a magnetic flux path, wherein: - the transformer's core comprises a common member for establishing a common flux path for the primary and the secondary winding and at least one bypass member for establishing a leakage flux path, and - at least one permeability control device connected to the common member for, upon variation in the load impedance, changing the effective permeability of the common flux path to change the common flux and thus the power transferred to the load.

In a preferred embodiment, the method comprises implementing the series and parallel reactor devices by means of the leakage and/or the magnetizing impedance in a transformer with a primary winding connected to a power source, a secondary winding connected to the capacitive load, and a magnetic flux path, where: - the flux path is divided in a common flux path for the primary and the secondary winding and at least one bypass flux path, - the effective permeability of the reactor's magnetic core is varied by changing the effective permeability of the common flux path to change the resonance frequency of the total impedance (load plus controllable reactor (s)).

In a preferred embodiment, the method comprises increasing or decreasing the effective permeability of the series reactor's magnetic core to obtain a reactor voltage drop leading to reference voltage drop across the load.

Resonance occurs at a frequency which depends on the capacitance and inductance values of the circuit. Thus, resonance can be avoided by varying the frequency of the power supply and keeping this frequency at a value outside the resonance frequency range defined by the variable capacitances and inductances in the circuit.

The power supply frequency can e. g. be varied by means of frequency converter.

If the overvoltage protection device is a surge arrester or a varistor, connection will be triggered automatically by the protection device, and it will not be necessary to attend to this by means of the processing unit.

Resistive load In a preferred embodiment, where the load is resistive, the above mentioned resonance is not an important issue. Consequently, if the load is resistive the magnetic controlling device can comprise a second reactor connected in parallel with the resistive load.

In a preferred embodiment, the input device and/or the permeability control device (s) is integrated with the processing unit.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described by means of examples illustrated in the drawings. In a first example, a load impedance is reduced and power transference to the load is increased to keep the voltage constant. In a second example the current in a variable load is limited by means of the invention. In a third example, the voltage in a circuit with a capacitive load is controlled by means of the invention. In a fourth example the power supply of a resistive load is controlled by means of the invention. The drawings show : Fig. 1 illustrates a first embodiment of the invention.

Fig. 2 shows an equivalent circuit for the device in figure 1.

Fig. 3 shows a second embodiment of the invention.

Fig. 4 shows an equivalent circuit for the device in figure 3.

Fig. 5 shows an embodiment of the device according to the invention.

Fig. 6 shows how the windings are positioned in the device illustrated in fig. 5.

Fig. 7 shows permeability variation for different values of control current.

Fig. 8 shows schematically a control system according to the first example.

Fig. 9 shows schematically a control system according to the second example.

Fig. 10 shows a basic circuit model according to the third example.

Fig. 11 shows the voltage across the coalescer.

Fig. 12 shows the effect of changing the value Xc in fig. 10.

Fig. 13 shows a basic circuit model with overvoltage protection.

Fig. 14 shows a basic circuit model with a controllable reactor also for the parallel branch.

Fig. 15 shows the voltage over the electrostatic coalescer with inductive overvoltage protection.

Fig. 16 shows a basic circuit comprising a voltage regulator, an overvoltage protection and a step-up transformer.

Fig. 17 shows a control circuit according to the third example.

Fig. 18 shows an aluminum furnace comprising a power supply control system related to the fourth example.

Fig. 19 shows a block diagram of the power supply control system according to the fourth example of the invention.

Fig. 20a shows a simplified sketch of the magnetic controlling and transformer device in a first embodiment of the power supply control system.

Fig. 20b shows a simplified sketch of the magnetic controlling and transformer device in a second embodiment of the power supply control system.

Fig. 20c shows a simplified sketch of the magnetic controlling and transformer device in a third embodiment of the power supply control system.

Fig. 21 shows a simplified sketch of a magnetic controlling device (controlled reactor type).

Detailed description of the invention General description of examples according to the invention Figure 1 illustrates an embodiment of the invention. In this figure, the following nomenclature is used: gl effective permeability of core 1 R2 effective permeability of core 2 In control current for core 1 Ic2 control current for core 2 Vp Primary voltage Vs Secondary voltage PW Primary winding SW Secondary winding.

The method according to the invention is implemented by means of a transformer T with a primary winding PW, a secondary winding SW for connection to a variable load VL and a magnetic flux path. In this figure, transformer T is shown as comprising a common member CM and a bypass member BM, which provide first and second flux paths 1 and 2 respectively. 1 denotes a common flux path for the primary winding PW and the secondary winding SW and 2 is a bypass flux path for leakage flux. Leakage flux in the context of the present specification is flux which is not linked by both windings but which is comprised in a magnetic path (leakage flux in air is thus not considered).

At least one of the flux paths has controllable relative permeability. Although the figure shows the different flux paths implemented by means of separate, adjacent magnetic members or cores, it is possible to use separate members or cores substantially separated from one another (as long as the primary and/or the secondary winding links common flux and leakage flux), and-also cores which provide a composite geometry.

It will also be possible to implement the invention by means of a series and a parallel reactor, where one or both can be controlled.

The primary voltage Vp generates a flux, in accordance with Faraday's law, which in the embodiment of the invention shown in figure 1, flows in the two paths. If the permeability and geometric dimensions are identical for the two paths, the flux divides evenly between them due to equal reluctances. When the permeability of common flux path 1 decreases, the flux"sees"an increased reluctance in this path, and relatively more flux flows then in bypass path 2. As the total flux is unchanged, altering the permeabilities of the path (s) causes less flux in common path 1 and

hence lower induced voltage at the secondary terminal, Vs. For a given load impedance VL on terminal Vs, the current is reduced.

In the embodiment shown in figure 1 the bypass flux path 2 is linked essentially only by the primary winding PW, but it is possible to envisage an embodiment where the secondary winding links the leakage flux.

Fig. 2 shows an equivalent circuit for the device in figure 1.

In this figure, LI is related to leakage flux, while Lm is related to common flux. N1 and N2 represent the transformer effect (where a primary voltage/current is transformed to a secondary current/voltage following transformer ratio N1/N2). In this circuit, Ll is variable. Lm will also be variable if it is possible to control permeability in both paths as will be discussed later.

The described embodiment of the invention comprises control of the flux in a single path (common path 1), but, depending on the application this control method can have disadvantages. In a case where limited secondary current is the purpose the permeability of leakage path 2 can be held constant during decreasing of permeability of common path 1, one will then achieve limited secondary current.

But the present drawbacks will be present: 1) As the total flux (common flux plus leakage flux) is given solely by the primary voltage, the primary current will increase even if the secondary current is reduced.

This is due to a reduced equivalent"magnetizing"inductance Lm which represents the common flux between primary and secondary windings. As the primary voltage and permeability of path 2 is assumed constant, decreasing permeability of common path 1 will give lower total impedance level seen from the source (connected to the primary winding, not shown in the figure), which increases the primary (reactive) current.

2) When the primary current increases, the power factor decreases at the same time, due to more inductive load. One ends up with a high reactive current.

3) During normal operation, one will have a rather high series inductance in the circuit caused by leakage path 2 in parallel to the common path 1 if the permeability of bypass path 2 is not reduced.

Based in the above mentioned, and depending on the application, one should have one or more controllable cores, where the common-member CM (path 1) should have as high permeability as possible during normal operation, while bypass member BM (path 2) should have as low permeability as possible. During a short circuit case, the situation is opposite, that is high permeability for leakage path 2 and low for common path 1.

The examples and figures so far have been related to two controllable paths. It is, however, possible to divide the flux in more than two paths. In figure 1 one can for example place an additional controllable leakage path 3 that is enveloped by the secondary winding together with path 1. This situation is shown in figures 3 and 4.

Then one has three adjustable parameters (a common flux path 1 and two leakage paths 2 and 3) to get appropriate behaviour.

Figure 4 shows the equivalent circuit of this embodiment, where equal reference numbers refer to the same parts as in figure 2, and where leakage flux path in bypass member BM1 (secondary bypass) is represented by inductor L2. The three possibilities for variation mentioned before correspond to three variable inductors LI, L2 and Lm.

It is also possible to let some of the turns in a winding envelop one core only, while the rest is enveloping two or more cores.

Figures 5 and 6 shows a device according to one embodiment of the invention, with cylindrical parts for accommodating primary and secondary windings and through holes for control windings. It comprises a transformer built with two sets of controllable cores (one corresponding to common member CM and the other corresponding to bypass member BM). Together with a control system one can make a system with galvanic isolation and controllable current (/voltage).

The controllable members can be made the same way as the cores shown in PCT/NO01/00217. In the case of an inductor, the core consists of substantially 3 parts: Inner core, outer core and couplers. A transformer built on this concept uses the same parts, only twice as many, where one unit (inner, outer core + couplers, these are not shown) is the common flux path for primary and secondary windings, the other one makes up the bypass flux path. Those two units are concentric mounted, which means that one unit has to have smaller diameter than the other.

Some of the above mentioned elements will be further described in relation to the following examples.

Example 1: Power supply control method and device From the above description, a method for controlling power supplied to a load is provided, by means of a transformer with a primary winding connected to a power source, a secondary winding connected to the load, and a magnetic flux path, comprising: - dividing the flux path in a common flux path for the primary and the secondary winding and at least one bypass flux path, - upon variations in the load impedance adapting the power transferred to the load

to the new load impedance or maintaining the voltage drop in the load impedance substantially unchanged, by - changing the effective permeability of the common flux path to change the common flux, or - changing the effective permeability of the at least one bypass flux path to change the leakage flux, or - changing both the effective permeability of the common flux path and the effective permeability of the bypass flux path to change the common flux and the leakage flux.

The invention comprises also a control system which, based on current values in the load, controls permeability in the flux path (s).

A control system according to the invention will then comprise: - a measurement unit MU for measuring load current and/or voltage, - an interface I for input of set point values, -a device T according to the invention, a processor P connected to the measurement unit and the at least one permeability control device for controlling the common member's and/or the bypass member's permeability based on the current measurements on the current/voltage measurements.

In one embodiment of the system (figure 8), it will comprise: - Two independent control-current (Isl, Is2) sources which can provide the wanted level for permeabilities for the paths.

- A measurement unit, which monitors the load current (or the primary current) - A processing unit which controls control-current according to load impedance changes.

The circuits that deliver the two control currents Isl and Is2 can for example be realized by a thyristor rectifier. Or, if faster response is required, one can use a diode rectifier together with a PWM bridge of IGBT's or MOSFET's or other forced commutated devices.

Regarding the measuring unit, as the load impedance is fed by alternating current, the simplest and cheapest way of measuring the current is by using a current

transformer. It is, however, possible to use s current shunt together with an amplifier circuit that preferably is galvanic isolated from the rest of the control electronics. The converted signals must then be transferred to the processing unit by either an optical device (optocoupler) or a signal transformer. A Hall element device can also be utilized.

The processing unit can be very simple or rather complex depending on the application. In a simple embodiment it can be an open loop with no measurement unit. The control system takes one or two user signals and adjusts proportionally one or two control currents that lead to adjusted output voltage. This system relies on the user only. The output is therefore manually adjusted. This can be regarded as an electronically controlled variac.

A more advanced system can comprise signal conditioner for measurements of secondary and/or primary voltages and currents. These measured signals, together with user input signal (s), are used to calculate necessary levels of control currents to achieve wanted output voltage and/or current (and hence, power). The system can also include temperature measurements either via measuring the resistance in the control and/or working coil (s) of the perpendicular field transformer device. The temperature measurements can be used to impose limitation to the control currents in order to prevent overheating (worst case shut the system down), or opposite, allow overload during shorter periods as the components have robustness to transient overload situations. The control system can also be adapted to calculate the power factor of the load as well as for the system itself. It is also possible to calculate energy delivered to the load (kWh), efficiency, used reactive power with more.

If several system is running in parallel, the control system should be able to run in master and slave modes, which makes it possible to incorporate load sharing and/or redundant systems. Then the control system needs to have a kind of bus system (field busses like CAN, Profibus, DeviceNet,... or ordinary Ethernet (TCP/IP) ) to be able to communicate with other systems. Wireless solution can also be considered (Bluetooth, IR, GSM, GPRS, UMTS, other radio solutions).

The system should also be able to log system and load parameters.

To be able to perform such operation, the control system has included a micro controller or processor. The system also needs a user input device, to get necessary commands and settings. It is also necessary to have a user output device to present system messages. This user input and/or output can be an integrated part of the processing unit, or it can be a separate devise communicating with the processing unit through the busses or wireless solutions mentioned above.

As mentioned before, the effective permeability for the flux paths will be controlled in one embodiment of the invention by a flux which is perpendicular to the common and leakage flux. This control flux is generated by means of a control current (Isl, Is2 in figure 7) in a control winding. The relation between different values of control current and permeability is empiric and is shown in figure 8. This shows different B-H curves depending on the control current.

Example 2: Control methods, devices and a system for limiting a current in a variable load As mentioned in the introduction above, the current is reduced for a given increase in load impedance VL on terminal Vs. This gives one possible mechanism for limiting the current at the secondary side. In a short circuit situation, a maximum current will be established, and it will not be necessary to interrupt power supply to the circuit. In this case, if all flux is"moved"to the bypass path, making L1 large and Lm small (figure 2), the power source will"see"the transformer plus the load impedance as a large inductance. If this equivalent inductance is sufficiently large, it is possible to avoid interrupting power supply.

The invention also comprises a control system which, based on current values in the load, controls permeability in the flux path (s). A control system according to the invention will then comprise : - a measurement unit MU for measuring load current and/or voltage, - a current limiting device T according to the invention, - a processor P connected to the measurement unit and to at least one permeability control device for controlling the common member's and/or the bypass member's permeability based on the current measurements.

In one embodiment of the system (figure 9), it will comprise: - Two independent control-current (Isl, Is2) sources which can provide the wanted level for permeabilities for the paths.

- A measurement unit, which monitors the load current (or the primary current) - A processing unit which triggers the protection system if the current reaches a predetermined level.

The circuits that deliver the two control currents Isl and Is2 can for example be realized by a thyristor rectifier. Or, if faster response is required, one can use a diode rectifier together with a PWM bridge of IGBT's or MOSFET's or other forced commutated devices.

Regarding the measuring unit, as the load impedance is fed by alternating current, the simplest and cheapest way of measuring the current is by using a current transformer. It is, however, possible to use s current shunt together with an amplifier circuit that preferably is galvanic isolated from the rest of the control electronics. The converted signals must then be transferred to the processing unit by either an optical device (optocoupler) or a signal transformer. A Hall element device can also be utilized.

The processing unit will be very simple if the system is designed for short circuit protection only, that is, if another system is taking care of ordinary operation control. For example it can be realized by some analogue input filters to eliminate noise in the measured signal before the signal trigger a comparator. This comparator then activates the control current sources in accordance to the behaviour needed (for example maximum control current to Lm and minimum to Ll in figure 1.

As mentioned before, the effective permeability for the flux paths will be controlled in one embodiment of the invention by a flux which is perpendicular to the common and leakage flux. This control flux is generated by means of a control current (Isl, Is2 in figure 8) in a control winding. The relation between different values of control current and permeability is empiric and is shown in figure 7. This shows different B-H curves depending on the control current.

If the time constants of the control circuits are too high compared to a fault situation, there is some inherent protection due to some series inductance present. It is possible to design the value of the series'inductance to any value wanted, so that one can get a good compromise to inherent protection versus low voltage drop during normal operation.

One can also have a kind of"fail to safe"property, as a power loss in the control circuits bring the series inductance to its highest level and hence limit the output voltage and current.

Example 3: Voltage regulation of a circuit with a capacitive load The third example relates voltage regulation of an ac circuit with a capacitive load or a load dominated by capacitive elements, and where the capacitance of the load may be more or less (stochastically) variable (within certain bounds).

Fig. 10 shows a basic circuit model corresponding to the principle for voltage regulation according to the invention.

An electrical system comprising a variable capacitive impedance can favourably be analysed on a per unit basis in order to understand principles. When a per unit system is used and all but the basic circuit elements are ignored, including the transformer reactances if applicable, the system consists of a capacitor in parallel with a resistor, fed by an AC voltage source. A controllable reactance is placed in series with this load to achieve a variable voltage in the method according to the invention. This circuit solution is shown in figure 10.

When the input voltage equals 1, the voltage across the variable capacitive load is given by: The (capacitive) reactance of the load X, depends, of course, on the value of the capacitance, and can have a relatively large variation. This variation will be discussed later, but for the time being, X, is set equal to 1. R is set equal to 10 so that the load impedance is predominantly capacitive. The voltage across the load is shown figure 11 as a function of X-a controllable reactance according to the invention.

As one can see: - For a value of X = X the circuit is in resonance, and the capacitive load voltage is 10 times as high as the input voltage.. With a larger R, the damping in the circuit would be even less, and the resonance overvoltage would be higher.

- On the left flank of the curve, when X < Xc, the capacitive load voltage is higher than the input voltage, but cannot be lower than the input voltage.

This opens, in principle, for the possibility of having a low input voltage, while using the resonance curve to transform the capacitive load voltage up to the desired level.

- On the right flank of the curve, when X > Xc, the capacitive load voltage can be both higher and lower than the input voltage. With large enough X, the capacitive load voltage can be zero.

Regarding resonance, a case where the capacitance C of the capacitive load varies will be analysed. A variation in C from 0.2 to 5 gives a ratio of 25 between the

highest and the lowest value. Such a variation corresponds to a variation of X,, in the basic circuit of figure 10, in the range of 5 to 0.2. (Xc = 5 corresponds to C equal to 0.2 and vice versa.) In figure 11, where Xc = 1, let us say that we are operating in the point X = 2. If C increases, Xc decreases. Then the resonance curve shown will have a left-shift.

Keeping X constant (equal to 2), the voltage over the capacitive load decreases.

After a short while, a closed loop control system will reduce X in order to bring the capacitive load voltage back to its original value. A new stable operating point is thereby achieved.

If the opposite happens, i. e. C decreases, the resonance curve shifts to the right.

Consequently the voltage over the capacitive load increases. In the worst case the decrease in C is just the right amount to correspond to the top of the resonance curve, and the capacitive load voltage increases dramatically. Due to a certain time lag in the closed loop control system, this can be destructive. Preventive measures against such overvoltages are therefore essential. Figure 12 illustrates the above discussion.

Variations in the impedance of the capacitive load give a risk of resonance, as shown above. Variations in the given range don't represent a limitation concerning stationary operation, i. e. to achieve the wanted voltage when the capacitance of the capacitive load doesn't have fast variation.

However, the system so far discussed, can give unwanted resonance overvoltages due to impedance variations. The answer to key issue f is therefore that a great deal of caution should be shown towards resonance, and measures to solve this problem must be used. A discussion of possible methods to solve this problem follows: Methods maintaining the basic circuit The smallest possible C gives the largest possible Xc, in the basic discussion set equal to 5. If the smallest possible X (controllable reactance) is chosen larger than this value, one is ensured to always have right flank operation, thereby always avoiding resonance. The disadvantage of this solution is clear when C is larger than the minimum value. Then Xc is lower, and a smaller X is needed to sustain the capacitive load voltage. Since the smallest X was set (by design, for example) to a value higher than the one now needed, it is clear that the maximum voltage cannot be sustained when the capacitive load has a relatively high C.

However, if C almost always has its lowest value, and higher values correspond to unwanted operating modes where a decrease in voltage is actually wanted, the solution of setting the smallest possible X larger than the highest possible Xc, could be a good answer to the resonance problem.

If, on the other side C shows a kind of stochastic behaviour, with a typical value somewhat above its minimum value, the above solution is not recommended. A sudden decrease in C would lead to a shift of the resonance curve to the right, as shown in 12, with the danger of an overvoltage occurring. In this case, one could instead rely on some kind of overvoltage protection that short-circuits over the capacitive load once the voltage over it reaches a threshold voltage.

Another way to avoid resonance at all times is to set a"bleeding resistor"in parallel with the electrostatic capacitive load. The disadvantage of this method is the power dissipation in the resistor.

According to one embodiment of the invention, the resonance problem is addressed by means of a second controllable reactance, in parallel with the electrostatic capacitive load. The basic circuit is then modified, so that it looks like in figure 13.

The reactance of the new component (second controllable reactor) is denoted Xp.

The impedance of a parallel connection of an L and a C is: It is clear that if Xp < Xc, the source sees an inductive load. In the frequency domain, the resonant (angular) frequency of the load (that now includes the parallel inductor, disregarding R) is: <BR> <BR> I<BR> @0 0-<BR> lac Keeping the (angular) frequency <u of the source smaller than) at all times ensures that the circuit never comes into resonance. This can be done either by a tuning the parallel inductance (second controllable reactor) so that the minimum o that can occur, always will be larger than the angular frequency of the source; or by actually changing the frequency of the source to an appropriate value; or both.

Thus, a solution to the resonance issue according to the invention is to place also a frequency converter between the source and the first controllable reactor, and to use a controllable reactor also for the parallel branch. This is shown in figure 14, where also some stray inductances and resistances in the circuit are included.

To address other issues regarding the invention, reference is made to figure 13. In figure 15, the voltage over the capacitor is shown as a function of X, setting Xp equal to 0.5 times the minimum Xc. Two curves are shown; one for Xc = 1, and one for Xi = 5

As can be seen, in theory the voltage can now be regulated in the range of 1 to 0. In practice, however, a range from about 0.8 to 0 is more realistic, as designing the controllable reactance for almost zero minimum inductance makes the component more expensive and heavy.

This should in practice not impose a large limitation as a step-up transformer is in most cases used in front of the load. This transformer should then be designed with the above in mind. In some cases, an additional step-up transformer would have to be used if full voltage is needed. Figure 16 shows the basic circuit with also the transformer included.

The controllable reactances in the method and system according to the invention operate on the principle discussed for example in the above mentioned PCT/NO01/00217 and PCT/NO02/00435. A basic concept regarding the possibility of varying effective permeability of a circuit by means of a magnetic flux that is substantially perpendicular to the operative flux in a magnetic circuit is illustrated in figure 17, where different values of a control current lead to different B-H curves. The linearized BH curves give the effective permeability for a given control current (between 0.2 and 4A in the figure).

Example 4: Power supply control system of a circuit with a resistive load The following example will be described with reference to fig. 1,5, 6 and fig. 18- fig. 21. It shall be noted that there are several similarities with the method and system described in example 3.

The present invention can advantageously be used as a power supply control system 1 for an aluminum furnace or electrolytic cell in a melting plant as shown in fig. 18.

The melting plant of fig. 18 comprises a plurality of electrolytic cells or aluminum furnaces 104,107. A power supply control system 101 supplies a first group of electrolytic cells 104. A thyristor booster 103 supplies a second group of electrolytic cells 107. The melting plant further comprises a number of electrolytic cells 106, where the power supply system (s) supplying these cells 106 are not shown in fig. 18 for clarity. The melting plant further comprises a system control center 108 for monitoring and controlling the aluminum furnace processes and power supply systems 101,103.

Each power supply control system 101,103 is supplied with electric power from a mains supply 2 with a typical voltage of 6 or 12 kV. Power supply control systems 1,3 typically transform the above mentioned power to 40 V and 5000A DC current, which is distributed to the respective groups of cells 104,107. An electrolytic cell is

usually a resistive load, or a load with dominant resistive elements. The electrolytic cell may comprise a bucking voltage.

The power supply control system 101 according to the invention will now be explained more generally while referring to fig. 19. A power source 109 is supplying electric energy to the power system. Power source 109 corresponds preferably to the mains supply 102 in fig. 18, as described above. Power source 109 is connected to a magnetic controlling and transformer device 110, which transforms the electric energy from an input voltage level to a desired output voltage level. The magnetic controlling and transformer device 110 will be described in further detail below.

The magnetic controlling and transformer device 110 is connected to a rectifying device 113. The rectifying device 113 is preferably a diode rectifier, but other types of rectifiers can be used. Finally, the rectifying device 113 is connected to a load 114. The load 114 corresponds to one electrolytic cell or a group of electrolytic cells 104,107 in fig. 18.

As shown in fig. 19, the power supply control system 1 according to the invention further comprises a measurement-/estimation unit 115, a processing unit 116 and an input value unit 117 (fig. 19). The measurement-/estimation unit 115 measures or estimates a value representing the voltage drop over and/or current through the load 114. The measurement-/estimation unit 115 can measure voltage drop over and/or the current through the resistive load 114 directly, but can also measure the voltage over and/or current through the magnetic controlling and transformer device 110 or the rectifying device 113 or other appropriate places.

The input value unit 117 provides a reference value for the voltage drop over and/or current through the load 14. The reference value is preferably a digital signal, which is set manually or automatically, for example based on desirable process parameters. Consequently, the input value unit 117 is preferably connected to said system control center 108.

The measured/estimated value is converted to a digital signal by means of an A/D- converter (not shown). The digital measured/estimated value and the reference value are inputs to the processing unit 116.

The processing unit 116 produces a control signal to the magnetic controlling and transformer device 110 (fig. 19), based on the reference value from the input value unit 117 and the measured/estimated value for the voltage drop over and/or current through the load 114, in such a way that the voltage drop over and/or current through the load 114 is altered towards the reference value. The control signal is converted to an analog signal by means of a D/A-converter (not shown).

Three embodiments of the magnetic controlling and transformer device 110 will now be described with reference to fig. 20a, 20b and 20c. Here, the power source 109 is represented by three-phase AC voltage Va, Vb and Vc. VL is the voltage over the load 114.

First and second embodiment (fig. 20a and 20b): reluctance controlled reactor In the first embodiment, shown in fig. 20a, the magnetic controlling and transformer device 110 comprises a magnetic controlling device 111 connected to a transformer device 112. Here, the magnetic controlling device 11 is connected to the power source 109, while the transformer device 112 is connected to the rectifying device 113.

In the second embodiment, shown in fig. 20b, the magnetic controlling and transformer device 110 also comprises a magnetic controlling device 111 connected to a transformer device 112. Here, the magnetic controlling device 111 is connected to the rectifying device 113, while the transformer device 112 is connected to the power source 109.

In fig. 20a and 20b the magnetic controlling device 111 is represented by three controllable inductors, La, Lb and Lc, one for each phase respectively. In the following description, each such controllable inductor is denoted reluctance controlled reactor.

The principle behind a controlled reactor is shown in fig. 21. The controlled reactor 119 comprises a core 120 of a magnetizable material. A main winding 121 is wound around the core 120. The main winding 121 is adapted for connection to the power source 109 and the transformer device 112 in fig. 20a and 20b. A control winding 122 with control current Ic is also wound around the core 120. The control current Ic is regulated by means of the control signal from the processing unit 116, as described above.

Main winding 121 and control winding 122 are wound around core 120 in such way that the flux generated by control winding 122 is non-parallel to the flux generated by main winding 121, so that the fluxes will not influence each other substantially.

The main magnetic coupling between main winding 121 and control winding 122 will occur in the magnetizable material as a change in the relative permeability Flr A change in Str will cause a change in reactance and consequently in inductor L. In the power supply control system according to the invention, the inductor L of the reluctance controlled reactor (fig. 21) corresponds to one of the controllable inductors La, Lb and Le in fig. 20a and 20b.

In the embodiments of the invention shown in figs 3a and 20b the inductance of the reluctance controlled reactor will be used to regulate the voltage output of the magnetic controlling and transformer device 10 of fig. 20a and 20b.

Third embodiment (fig. 20c): reluctance controlled transformer The third embodiment of the invention will be described with reference to fig. 20c.

In fig. 20c the magnetic controlling device 111 and the transformer device 112 of the first and second embodiment are embedded in one device, denoted as reluctance controlled transformer.

The, reluctance controlled transformer preferably comprises two controllable inductors La) and La2, Lb and Lb2 and Lcl and L, 2 per phase, respectively. In fig. 20c the reluctance controlled transformer is connected as a Y-Y-connection.

The principle of a reluctance controlled transformer 129 is shown in fig. 1 and 6, which will be briefly described related to this example. Here, the reluctance controlled transformer comprises a first core 1,130 and a second core 2, 131. A primary winding PW, 132 is wound around the cores 130,131 and is provided with a primary voltage Vp. A secondary winding SW, 133 is wound around the core 130.

The voltage of the secondary side is denoted Vs. The primary winding is connected to the power source 109 and the secondary winding is connected to the rectifying device 113 in fig. 20c. The voltage Vp in fig. 1 is corresponding to the voltage between Vc and the neutral point on the primary side in fig. 20c.

In the embodiment shown in fig. 5 and 6, each of the cores 130,131 comprises two cylinders with an annular space 136,137 between each of the two cylinders. The diameter of core 131 is smaller than the diameter of core 130, and the core 131 is located inside core 130.

The reluctance controlled transformer further preferably comprises two control windings 134,135 (fig. 5). A first control winding 134 with control current Ic, (fig.

1) is located in the annular space 136 between the two cylinders of core 130, and a second control winding 135 with control current IC2 is located in the annular space 137 between the two cylinders of core 131.

Control currents are produced by control signals from the processing unit 116, as described above. A change in the control currents will provide a change in the effective permeability ti, 92 of the respective cores. Consequently, the inductance of the controlled transformer will change.

The windings PW (132), SW (133), together with the control windings provide a first flux path in the core 131 and a second, common flux path in the core 130.

Leakage flux in the context of the present specification is flux which is not linked

by both windings but is comprised in a magnetic path (leakage flux in air is thus not considered).

Although the figure shows the different flux paths implemented by means of separate, adjacent magnetic members or cores, it is possible to use separate members or cores substantially separated from one another (as long as the primary and/or the secondary winding links common flux and leakage flux), and also cores which provide a composite geometry.

The primary voltage Vp generates a flux, in accordance with Faraday's law, which in the embodiment of the invention shown in fig. 1, flows in the two paths. If the permeability and geometric dimensions are identical for the two paths, the flux divides evenly between them due to equal reluctances. When the permeability of common flux path in core 2,131 decreases, the flux"sees"an increased reluctance in this path, and consequently more flux then flows in flux path of core 2, 131. As the total flux is unchanged, altering the permeability of the path (s) causes less flux in common path in core 1,130 and hence lower induced voltage at the secondary terminal, Vs.

In the embodiment shown in fig. 1 the flux path in core 2,131 is essentially only linked by the primary winding PW, but it is possible to provide an embodiment where the secondary winding links the common flux.

The secondary winding SW is placed around one or more of these cores. Since the flux now can be controlled between the cores, the ampere turn balance between the primary winding and secondary winding will also be controllable. The magnetic circuit therefore comprises parallel paths that are fed from the ampere turns of the primary winding. Because the reluctances are parallel, the inductances in the T- equivalent to the transformer will be in series with the input voltage and cause an inductive voltage drop in the transformer.

In this embodiment the inductance of the reluctance controlled transformer device will be used to regulate the voltage output of the magnetic controlling and transformer device 110 (block denoted 111, 112 in fig. 20c).

Among the advantages that can be achieved with the three embodiments according to the present invention are that the power supply control system has adjustable DC voltage where the noise is avoided compared to a corresponding thyristor rectifier.

Further, it is possible to implement diode rectification through a large voltage range. Moreover, by regulating the reactance on the primary side, the rectifier can be protected against short circuit.

The transformer of the magnetic controlling and transformer device 110 can be dimensioned for a considerably lower harmonic load than with a thyristor rectifier, since there will be less harmonics.

Several modifications of the invention is possible within the scope of the invention as defined in the claims. Preferably, other parameters such as temperature, overcurrent, input voltage etc. are measured because of security issues. The voltage and/or current between the magnetic controlling and transformer device 110 and the rectifying device 113 can also measured and converted to an input signal to the processing unit.

The reluctance controlled transformer in fig. 20c can be connected as a Y-A- connection or a A-A-connection. Moreover, there can be one control winding per phase instead of two.