《高电压工程》课程教学资源（参考书籍）《High Voltage Engineering Test》Chapter 2 Generation of high voltages Chapter 3 Measurement of high voltages.pdf_P16-P20

Generation of high voltages 23 of the transformers.In addition to this,to every primary and secondary winding a capacitor Cp is connected in parallel,which provides an overcompensation of the magnetizing currents.The whole chain of cascaded transformers is loaded by a terminating resistor;thus the network acts similarly to a terminated transmission line along which the a.c.voltage remains nearly constant and has a phase shift between input (high-frequency power supply)and output (termination).The transformers,therefore,are not used to increase the a.c. voltage. It is now possible to connect to every stage indicated as usual Cockcroft-Walton cascade circuit,with only a small input voltage (some kV),producing,however,output voltages of some 10kV per stage.The storage columns of these Cockcroft-Walton cascades are then directly series connected,providing the high d.c.output voltage for the whole cascade transformer HVDC generator unit.Typically up to about 25 stages can be used, every stage being modular constructed.As these modules are quite small,they can be stacked in a cylindrical unit which is then insulated by SF6.Not shown in Fig.2.7 is the voltage regulation system,which is controlled by a parallel mixed R-C voltage divider and a high-frequency oscillator,whose frequency ranges from 50 to 100 kHz.As for these high frequencies the capacitors within the Cockcroft-Walton circuits can be very small,and the energy stored is accordingly low;regulation due to load variations or power voltage supply variations is very fast(response time typically about 1 msec).The small ripple factor is not only provided by the storage capacitor,but also by the phase- shifted input voltages of the cascade circuits.Amongst the disadvantages is the procedure to change polarity,as all modules have to be reversed. Summary and concluding remarks to 2.1.1 It has been shown that all a.c.to d.c.voltage conversion systems could be classed between the circuits of Figs 2.1 and 2.3,if single-phase a.c.voltages are converted into d.c.voltages.A high d.c.to a.c.voltage ratio can only be gained with a high product of a.c.frequency and energy stored in the smoothing capacitors,as they have to store electrical energy within each cycle, during which the a.c.power is oscillating.If,therefore,the d.c.output should be very stable and continuous,a high product (fC)is necessary.A reduction of stored energy is possible if the a.c.power is not only provided at ground potential,this means if a.c.power is injected into the circuits at different potential levels.The savings,therefore,can be made either on the a.c.or d.c. side.The large variety of possible circuits and technical expenditure is always strongly related to the 'quality'of the d.c.power needed,this means to the stability and the ripple of the output voltage

Generation of high voltages 23 of the transformers. In addition to this, to every primary and secondary winding a capacitor Cp is connected in parallel, which provides an overcompensation of the magnetizing currents. The whole chain of cascaded transformers is loaded by a terminating resistor; thus the network acts similarly to a terminated transmission line along which the a.c. voltage remains nearly constant and has a phase shift between input (high-frequency power supply) and output (termination). The transformers, therefore, are not used to increase the a.c. voltage. It is now possible to connect to every stage indicated as usual Cockcroft–Walton cascade circuit, with only a small input voltage (some kV), producing, however, output voltages of some 10 kV per stage. The storage columns of these Cockcroft–Walton cascades are then directly series connected, providing the high d.c. output voltage for the whole cascade transformer HVDC generator unit. Typically up to about 25 stages can be used, every stage being modular constructed. As these modules are quite small, they can be stacked in a cylindrical unit which is then insulated by SF6. Not shown in Fig. 2.7 is the voltage regulation system, which is controlled by a parallel mixed R-C voltage divider and a high-frequency oscillator, whose frequency ranges from 50 to 100 kHz. As for these high frequencies the capacitors within the Cockcroft–Walton circuits can be very small, and the energy stored is accordingly low; regulation due to load variations or power voltage supply variations is very fast (response time typically about 1 msec). The small ripple factor is not only provided by the storage capacitor, but also by the phaseshifted input voltages of the cascade circuits. Amongst the disadvantages is the procedure to change polarity, as all modules have to be reversed. Summary and concluding remarks to 2.1.1 It has been shown that all a.c. to d.c. voltage conversion systems could be classed between the circuits of Figs 2.1 and 2.3, if single-phase a.c. voltages are converted into d.c. voltages. A high d.c. to a.c. voltage ratio can only be gained with a high product of a.c. frequency and energy stored in the smoothing capacitors, as they have to store electrical energy within each cycle, during which the a.c. power is oscillating. If, therefore, the d.c. output should be very stable and continuous, a high product (fC) is necessary. A reduction of stored energy is possible if the a.c. power is not only provided at ground potential, this means if a.c. power is injected into the circuits at different potential levels. The savings, therefore, can be made either on the a.c. or d.c. side. The large variety of possible circuits and technical expenditure is always strongly related to the ‘quality’ of the d.c. power needed, this means to the stability and the ripple of the output voltage

Generation of high voltages 25 the belt passes.The entire equipment is usually enclosed in an earthed metal tank filled with insulating compressed gases of good performance such as air, mixtures of N2-CO2,Freon 12(CCl2,F2)or SF6.For simple applications the metal tank can be omitted,so that the insulation is provided by atmospheric air only. The potential of the h.v.terminal at any instant is V=O/C above earth, where o is the charge stored and C is the capacitance of the h.v.electrode to ground.The potential of the terminal rises at a rate given by dV/dt=I/C, where I=Sbv (2.13) is the net charging current to the terminal.In this equation,S is the charge density at the belt in coulombs/m2,b its width in m,and v the belt speed in m/sec.In practice,dV/dt may reach a value of 1 MV/sec and it appears that the final potential of the h.v.electrode would be infinite in the absence of any mechanism of charge loss.Equilibrium is in practice established at a potential such that the charging current equals the discharge current which includes load currents-also due to voltage dividers,leakage currents and corona losses,if present-and by voltage regulating systems which are based on voltage measurement and the controllable spray unit. While the h.v.terminal electrode can easily be shaped in such a way that local discharges are eliminated from its surface,the field distribution between this electrode and earth along the fast moving belt is of greatest importance. The belt,therefore,is placed within properly shaped field grading rings,the grading of which is provided by resistors and sometimes additional corona discharge elements. The lower spray unit,shown in Fig.2.8,may consist of a number of needles connected to the controllable d.c.source so that the discharge between the points and the belt is maintained.The collector needle system is placed near the point where the belt enters the h.v.terminal. A self-inducing arrangement is commonly used for spraying on the down- going belt charges of polarity opposite to that of the h.v.terminal.The rate of charging of the terminal,for a given speed of the belt,is therefore doubled.To obtain a self-charging system,the upper pulley is connected to the collector needle and is therefore maintained at a potential higher than that of the h.v. terminal.The device includes another system of points(shown as upper spray points in Fig.2.8)which is connected to the inside of the h.v.terminal and is directed towards the pulley at the position shown.As the pulley is at a higher positive potential,the negative charges of the corona at the upper spray points are collected by the belt.This neutralizes any remaining positive charges on the belt and leaves any excess negative charges which travel down with it and are neutralized at the lower spray points

Generation of high voltages 25 the belt passes. The entire equipment is usually enclosed in an earthed metal tank filled with insulating compressed gases of good performance such as air, mixtures of N2 –CO2, Freon 12 (CCl2, F2) or SF6. For simple applications the metal tank can be omitted, so that the insulation is provided by atmospheric air only. The potential of the h.v. terminal at any instant is V D Q/C above earth, where Q is the charge stored and C is the capacitance of the h.v. electrode to ground. The potential of the terminal rises at a rate given by dV/dt D I/C, where I D Sb O 2.13 is the net charging current to the terminal. In this equation, SO is the charge density at the belt in coulombs/m2, b its width in m, and v the belt speed in m/sec. In practice, dV/dt may reach a value of 1 MV/sec and it appears that the final potential of the h.v. electrode would be infinite in the absence of any mechanism of charge loss. Equilibrium is in practice established at a potential such that the charging current equals the discharge current which includes load currents – also due to voltage dividers, leakage currents and corona losses, if present – and by voltage regulating systems which are based on voltage measurement and the controllable spray unit. While the h.v. terminal electrode can easily be shaped in such a way that local discharges are eliminated from its surface, the field distribution between this electrode and earth along the fast moving belt is of greatest importance. The belt, therefore, is placed within properly shaped field grading rings, the grading of which is provided by resistors and sometimes additional corona discharge elements. The lower spray unit, shown in Fig. 2.8, may consist of a number of needles connected to the controllable d.c. source so that the discharge between the points and the belt is maintained. The collector needle system is placed near the point where the belt enters the h.v. terminal. A self-inducing arrangement is commonly used for spraying on the downgoing belt charges of polarity opposite to that of the h.v. terminal. The rate of charging of the terminal, for a given speed of the belt, is therefore doubled. To obtain a self-charging system, the upper pulley is connected to the collector needle and is therefore maintained at a potential higher than that of the h.v. terminal. The device includes another system of points (shown as upper spray points in Fig. 2.8) which is connected to the inside of the h.v. terminal and is directed towards the pulley at the position shown. As the pulley is at a higher positive potential, the negative charges of the corona at the upper spray points are collected by the belt. This neutralizes any remaining positive charges on the belt and leaves any excess negative charges which travel down with it and are neutralized at the lower spray points.

26 High Voltage Engineering:Fundamentals For a rough estimation of the current I which can be provided by such generators,we may assume a homogeneous electrical field E normal to the belt running between the lower spray points and the grounded lower pulley. As E=D/so =S/so,D being the flux density,so the permittivity and S the charge density according to eqn(3.13)deposited at the belt,with so 8.85 x 10-12 As/Vm,the charge density cannot be larger than about 2.7 x 10-5 As/m2 if E=30kV/cm.For a typical case the belt speed might be v= 20 m/sec and its width b=1 m.The charging current according to eqn (2.13) is then I540uA.Although with sandwiched belts the output current might be increased as well as with self-inducing arrangements mentioned above,the actual short-circuit currents are limited to not more than a few mA with the biggest generators. The main advantages of belt-driven electrostatic generators are the high d.c. voltages which can easily be reached,the lack of any fundamental ripple,and the precision and flexibility,though any stability of the voltage can only be achieved by suitable stabilizing devices.Then voltage fluctuations and voltage stability may be in the order down to 10-5. The shortcomings of these generators are the limited current output,as mentioned above,the limitations in belt velocity and its tendency for vibra- tions,which aggravates an accurate grading of the electrical fields,and the maintenance necessary due to the mechanically stressed parts. The largest generator of this type was set into operation at Oak Ridge National Laboratory.(15)A view of this tandem-type heavy ion accelerator is shown in Fig.2.9.This generator operates with 25 MV,and was tested up to internal flashovers with about 31 MV. For h.v.testing purposes only a limited amount of generators are in use due to the limited current output.A very interesting construction,however, comprising the Van de Graaff generator as well as a coaxial test arrangement for testing of gases,is used at MIT(6 by Cooke.This generator,with an output of about 4 MV,may be controlled to provide even very low frequency a.c.voltages. The disadvantages of the belt-driven generators led Felici to develop elec- trostatic machines with insulating cylindrical rotors which can sustain perfectly stable movement even at high speeds.The schematic diagram of such a machine(17)is shown in Fig.2.10.To ensure a constant narrow air gap,the stator is also made in the form of a cylinder.If the stator is a perfect insulator, ions are deposited on its surface which tend to weaken the field.In order to avoid such ion screening,a slight conductivity has to be provided for the stator and resistivities in the range 1011-1013/cm have been found satisfactory. The overall efficiency of the machine is higher than 90 per cent and the life expectancies are only limited by mechanical wearing of the bearings,provided the charge density on the rotor surface is kept within limits which depend upon the insulating material employed.Epoxy cylinders have a practically unlimited

26 High Voltage Engineering: Fundamentals For a rough estimation of the current I which can be provided by such generators, we may assume a homogeneous electrical field E normal to the belt running between the lower spray points and the grounded lower pulley. As E D D/ε0 D S/ε O 0, D being the flux density, ε0 the permittivity and SO the charge density according to eqn (3.13) deposited at the belt, with ε0 D 8.85 ð 1012 As/Vm, the charge density cannot be larger than about 2.7 ð 105 As/m2 if E D 30 kV/cm. For a typical case the belt speed might be v D 20 m/sec and its width b D 1 m. The charging current according to eqn (2.13) is then I ¾D 540 µA. Although with sandwiched belts the output current might be increased as well as with self-inducing arrangements mentioned above, the actual short-circuit currents are limited to not more than a few mA with the biggest generators. The main advantages of belt-driven electrostatic generators are the high d.c. voltages which can easily be reached, the lack of any fundamental ripple, and the precision and flexibility, though any stability of the voltage can only be achieved by suitable stabilizing devices. Then voltage fluctuations and voltage stability may be in the order down to 105. The shortcomings of these generators are the limited current output, as mentioned above, the limitations in belt velocity and its tendency for vibrations, which aggravates an accurate grading of the electrical fields, and the maintenance necessary due to the mechanically stressed parts. The largest generator of this type was set into operation at Oak Ridge National Laboratory.15 A view of this tandem-type heavy ion accelerator is shown in Fig. 2.9. This generator operates with 25 MV, and was tested up to internal flashovers with about 31 MV. For h.v. testing purposes only a limited amount of generators are in use due to the limited current output. A very interesting construction, however, comprising the Van de Graaff generator as well as a coaxial test arrangement for testing of gases, is used at MIT16 by Cooke. This generator, with an output of about 4 MV, may be controlled to provide even very low frequency a.c. voltages. The disadvantages of the belt-driven generators led Felici to develop electrostatic machines with insulating cylindrical rotors which can sustain perfectly stable movement even at high speeds. The schematic diagram of such a machine17 is shown in Fig. 2.10. To ensure a constant narrow air gap, the stator is also made in the form of a cylinder. If the stator is a perfect insulator, ions are deposited on its surface which tend to weaken the field. In order to avoid such ion screening, a slight conductivity has to be provided for the stator and resistivities in the range 1011 –1013"/cm have been found satisfactory. The overall efficiency of the machine is higher than 90 per cent and the life expectancies are only limited by mechanical wearing of the bearings, provided the charge density on the rotor surface is kept within limits which depend upon the insulating material employed. Epoxy cylinders have a practically unlimited