FI121986B - Output coil arrangement in an inverter and associated procedure - Google Patents

Description

Method for inverter output choke

BACKGROUND OF THE INVENTION The present invention relates to a method according to the preamble of independent claim 1 in connection with inverter output chokes.

An inverter is a device typically used to control motors in an economical and reliable manner. The inverter, or inverter, produces a voltage or current at its output that has a controllable frequency, whereby the motor can be optimally controlled according to the desired frequency of the output. The voltage-to-frequency inverter, such as that shown in FIG. 7, exemplifies, by the inverter rectifier section 73, its output with short voltage pulses with respect to the base frequency, the duration of which is adjusted to provide the desired output voltage. The DC link frequency converter has a DC link DC link Udc, the positive and negative voltages of which are connected as desired in pulses to phase outputs. Thus, in one phase there is an upper 71 and a lower 72 semiconductor switch between which a phase output voltage is obtained. Figure 7 further shows a frequency converter controlled motor M and an input transformer 74.

These short and fast voltage pulses are produced with high current and semiconductor tolerant power semiconductors. Nowadays, IGB transistors (insulated gate bipolar transistors) are widely used because of its good properties. The IGBT is capable of breaking up to hundreds of amps and, accordingly, the component has a voltage rating of thousands of volts.

However, in high current inverters, IGBT components must be connected in parallel to achieve the required current. The IGBT components are often packaged in modules, whereby one module has a plurality of IBGT switches and their zero diodes. In this case, the natural alternative to parallel switching of the inverter output switches is to connect the switches of one module o in parallel. Applications where multiple components are required The 30 parallel connections are typically three-phase. At each stage, there are thus parallel connected components, and these parallel connection parts are generally referred to as branches. One phase of the inverter output thus consists of a number of branches corresponding to the number of upper and lower switch components g connected in parallel.

35 In voltage-to-frequency inverters, the output voltage is formed in the inversion section by voltage pulses which have very high rise speeds when using current switch components. High voltage rise speeds cause voltage oscillations in the cable between the inverter and the rotating machine to stress the insulation of the machine. In addition to the voltage rise rate, the amplitude and frequency of the voltage oscillations are determined by the cable length and the aal-5 operating impedance, the machine wave impedance, and other electrical interfaces between the inverter and the machine. The critical length of a cable is called the minimum cable length at which maximum reflection in the machine terminals occurs.

The choke at the inverter output can extend the critical length of the ka-10 game described above. The desired voltage rise rate determines the magnitude of the inductance of the inductor. The magnitude of the inductance required can be reduced by the RC circuit as shown in Figure 2. However, the use of the RC circuit increases the operating losses and may make it difficult for some inverters to adjust.

Conventionally, the voltage rise rate has been limited by reverse chokes, which are either single-phase or three-phase. Figure 1 illustrates the basic construction of both a single-phase and a three-phase choke. The choke consists of a winding 1 which is wound around the core 2. The heart is generally rectangular and includes at least one air gap to prevent satiety. In addition, in some cases, various voltage cutters have been used to reduce the maximum voltages, as shown, for example, in Figure 3. Here, a diode bridge 31 is connected to the inverter output, which cuts the output voltage as it rises above the intermediate circuit voltage Udc.

25 The high voltage rise rate also increases the circulating current flowing through the bearings. Roughly, the circulating bearing current is proportional to the common-mode voltage of the motor and the resulting current through the stray capacitances. Output voltage of inverter

LO

Decreasing the 9 climb speeds will reduce this current. A special bearing current filter 30, such as that shown in Fig. 8, comprises, in addition to the inverter output choke L, a common-mode choke Lcomm, e.g.

CL

the capacitance CE in the intermediate circuit or inverter output and between the star point of the RC circuit and the ground plane.

LO

§ Regardless of whether the inverter has IGBT parallel connections, the 35 output choke is designed to reduce the voltage rise speed and amplitude to such a low level that voltage vibrations do not damage the rotating machine insulation and that the current flowing through the bearing does not cause bearing damage.

In the prior art, in the case of parallel-connected IGB transistors, the outputs of the IGB transistors are connected directly in parallel to form a phase output, and a choke is connected to a common current path of this parallel connection to limit the voltage rise rate. In order to prevent saturation of the throttle core and, on the other hand, to obtain the necessary inductance, the core must have a sufficiently large cross-sectional area. Because of the high current allowed by the parallel-connected components, the required choke becomes a very large and heavy-duty terminal.

The problem with the direct parallel connection of the IGBT switches is the difference in current between the parallel branches in normal switching situations. Current differences between the branches are due to differences in IGBT switch steepness and lattice capacitances, as well as gate driver factors such as lattice voltage and lattice resistance differences. Due to the non-ideality of the parallel coupled components and the circuits controlling them, the phase output current is not evenly distributed between the components during the switching situations.

For example, in the case of switch-off, when one switch starts-20 before switching off other switches in the same phase, part of the current is transferred to the other switch components of the phase. An increase of this current relative to the rated switch-specific phase current has been found to be up to tens of percent. This potential increase in current should be taken into account when designing the inverter. Differences between branch currents reduce the load capacity of the inverter because the output current must be limited to such a level that the instantaneous current duration of the switch components is not exceeded even in the branch with the highest current. Thus, the switches cannot be dimensioned solely on the basis of the phase current, since this dimensioning would cause the switching currents in the switching situations to cause switching.

LO

9 component failure.

o 30 The problem with existing rigidly parallel branched branches £ is also due to a malfunctioning single branch power switch, such as

CL

for example, the IGBT switch cannot be recognized.

BRIEF DESCRIPTION OF THE INVENTION It is an object of the present invention to provide a method which avoids the above drawbacks and enables the detection of a particular switch failure condition which has previously been impossible to detect. This object 4 is achieved by the method according to the invention, which is characterized in what is stated in the characterizing part of independent claim 1.

The method according to the invention is based on utilizing output chokes. By the method of the invention, the output semiconductor switch, which is misconducted or left conductive, can be incorrectly determined and, if necessary, the control of the device can be terminated to avoid any major damage to the inverter itself or the device controlled by it.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be further described in connection with preferred embodiments, with reference to the accompanying drawings, in which:

Figure 1 illustrates conventional single and three-phase ballast designs; Figure 2 shows an LCR filter at inverter output;

Fig. 3 shows a voltage cutter at the inverter output; Fig. 4 shows branch-specific inductors at inverter output;

Figure 5 shows a break-through in the output phase of an inverter with the legs rigidly connected in parallel;

Figure 6 shows a break-through in the output phase of the inverter with branch-specific chokes; Fig. 7 shows a voltage intermediate circuit frequency converter;

Figure 8 shows a bearing current filter,

Figure 9 shows an output choke unit having a row structure; and

Fig. 10 shows an output choke assembly having chokes symmetrical to one another.

DETAILED DESCRIPTION OF THE INVENTION g Figure 4 illustrates the structure of a choke arrangement in connection with an inverter oo. In the structure of Fig. 4, the output of each phase U, V, W comprises a plurality of x me branches, i.e. three parallel switch pairs, which are simultaneously controlled * 30 to provide the required output current. Fig. 4 shows how the switches of each S phase are formed by modules consisting of three pairs of switches "j> 41, 42, 43 connected to a DC intermediate circuit Udc. This arrangement comprises o branch-specific single-phase chokes LU1, LU2, LU3, LV1, LV2, LV3, LW1, LW2, LW3, the first ends of the chokes of which are adapted to be coupled to the outputs UU1, UU2, UU3, UV, UV3, UW1, UW2, UW3. These branch-specific outputs are normally made up of points between the upper and lower power switches. The potentials of these points are controlled alternately by the power switches to the positive and negative potentials of the voltage intermediate circuit.

Further, the other ends of these single-phase chokes are arranged to be interconnected, thereby providing the inverter phase output U, V, W, which is further coupled to the load to control it.

The operation of this arrangement will now be described with reference to Figures 5 and 6. Figure 5 illustrates prior art direct parallel branching of branches 10 to increase current capacity. The outputs of the power semiconductor components, which are preferably IGBT switches or similar fast components, are then connected in parallel. In this case, the points between all the upper V1, V3, V5 and lower V2, V4, V6 power semiconductors in the connection are at exactly the same potential.

15 Since the switch components and the control circuits that control the switch components always differ to some extent, the components tend to turn off and on at different times and at different speeds despite the same control. This results in a situation where, for example, when a component is shut down, one of the switches tries to switch off before the other. In shutdown condition 20, for example, in an IGBT switch, the voltage across the component tends to rise before the current decreases. However, since the voltage of all branches is the same, the voltage cannot rise in the desired manner, and current is first transferred from the onset component to the components of the other branches. Although typical delays are not higher than the nanosecond, this causes significant current transitions to other components.

Figure 6 shows a single-phase connection with three branches. Chokes are connected at each point between the coupling pairs, and

o

™ the other ends of these chokes are still connected to each other via a phase output

LO

o to give. In this solution, the above-mentioned non-idealities of the IGBTs and the gate controllers 30 cause instantaneous voltage differences between the output branches of the IGBT switches, which act across the chokes and the currents of the branches change

CL

voltage and the inductance of the inductors. By using the inductors in this way, the other switches move

LO

g current can be significantly reduced, allowing the switches to be dimensioned more precisely for load current purposes only.

6

The above arrangement enables the method of the invention to be implemented. This method can be used to determine a particular type of fault with parallel connected power semiconductors. The method according to the invention will be further described with reference to Figures 5 and 6.

5 In failure situations where only one of the parallel IGBT switches V2 is misconducted or remains conductive when the other V4 and V6 go out, while the opposite branch power switches V1, V3, and V5 are controlled or normally conductive, the entire fault current2 passes through this single switch. Such a situation is illustrated in Fig. 5 in connection with the conventional direct branching of the straight branches. This causes the current to grow so high that the IGBT V2 in question goes into an unsaturated state, and limits the fault current to its gate voltage. Figure 5 shows how the fault current also passes through switches V3 and V5 and through switch V2 to the negative rail of the DC link.

If the output branches of the IGBT switches are rigidly coupled in parallel, the opposite branch will carry the same fault current, increased or decreased by the load current Im, through a plurality of parallel IGBTs V1, V3, and V5, correspondingly lower current. In this case, too, the output voltage of these IGBTs remains only a few volts, i.e., only deviates from the output voltage at normal load current, which makes it difficult to detect the fault.

The fault current flow can be influenced by the illustrated arrangement of connecting a choke to each output branch of an IGBT switch pair, and connecting the branches in parallel only after the chokes, as shown in Figure 6. This does not substantially affect the magnitude of the fault current, but affects the current flow so that in the branch where all the parallel power switches V1, V3 and V5 are guided, the fault current initially flows through only one power switch V1, gradually transmitting to the others. for parallel switches V3 and V5 at the speed determined by the differences between the inductances of the inductors and the output voltages of the switches, g 30 In this case, the output x voltage of the power switch V1 through which the entire fault current initially travels is substantially higher than the corresponding fault points. The output voltage of this IGBT also decreases as the current of σ> V gradually passes to the parallel IGBTs V3 and V5, but with the appropriate design of the gauges, the output voltage dwells sufficiently high so that the fault condition of the IGBT collector can be detected. based on measurement. The design of the inductor is intended to obtain, for example, a discharge voltage of about 7 to 10 volts higher than the discharge voltage at normal load current.

Thus, the method of the invention operates by predetermining the voltage limit and the time limit used to indicate a fault condition. The voltage limit is the voltage above which the component's output voltage must be in order to reliably determine the fault condition. The time limit is used to prevent malfunction indication in normal switching situations, whereby the IGBT output voltage is briefly above the above voltage limit. Further, according to the invention, control information of the power semiconductor components 10 of the branches is determined. For example, component control information is obtained directly from the processor performing the control.

According to the invention, the collector voltage of the power semiconductor components of the branches is further determined in those power semiconductor components that are controlled to a conductive state. Measurement of collector voltage, or emission voltage, can be accomplished in a normal manner by utilizing measurement circuits typically provided in controller circuits or gate controllers. According to the invention, a fault condition is detected in the power semiconductor component with which the collector voltage of the serially connected power semiconductor component exceeds a predetermined voltage limit for a period longer than a predetermined time limit. Thus, by determining the collector voltage of the controlled switches, it can be deduced whether the switch above or below the switch in question has been accidentally left or is in a conductive state.

According to a preferred embodiment of the invention, the power switches are turned off after a fault has been detected, whereby the fault current is interrupted.

Alternatively, or in addition, an alarm signal can be generated from the failure and transmitted to the inverter or process operator of the inverter. According to a preferred embodiment of the choke arrangement, the chokes o are formed from single-phase, non-oscillating chokes. In other words, the winding of the throttle g 30 is formed around a non-closing core x Figure 1 shows a conventional single- and three-phase throttle, in which the magnetic core material is eigenmic, i.e., σ formed on the core material. the magnetic flux is directed to close along the core, g Figures 9 and 10 show the choke constructions for use in a gag 35. Figure 9 shows the structure of a silent single-stage damper, although the figure shows an entire throttle unit in transverse and vertical projection. The figure shows how a single inductor is formed by winding coil 92 around a columnar core 91. The choke terminals 94 are provided at one end of the choke and a metal shield 93 is formed around the winding to prevent the spreading fields 5 from spreading. The choke assembly shown in Figure 9 comprises a total of nine single-phase chokes. Such a number of chokes are required for the output of a three-phase system if each phase comprises three separate branches, as in the example of Figure 4.

The mechanical design of the choke must be such that there is no significant magnetic coupling between the chokes per branch, or that there is no magnetic coupling between the chokes of different output phases and the magnetic coupling between the parallel branches of the same phase is small and symmetrical. Such an implementation is achieved, for example, by the row-like structure shown in Fig. 9, in which all individual chokes are protected by a metal-15 guard 93.

Another possibility for avoiding unwanted magnetic coupling is the symmetrical structure shown in Fig. 10 as a vertical projection. The figure shows a core 101, a coil 102 wrapped around a core, and a metal shield 103. In this solution, the parallel branches of the parallel branches ku-20 are connected in a symmetrical configuration. The output choke assembly of Figure 10 is for an inverter having three parallel arms in each of the three steps. In this solution, the magnetic coupling between the chokes of the various phases is prevented by a metal shield, that is to say, within the single metal shield 103, the single-phase chokes 25 are symmetrically arranged relative to one another.

The use of non-conductive inductors is made possible by the fact that with conventional yoke inductors the high frequency magnetic flux does not block through the yoke. Thus, an inductance-free inductance at high frequencies is achieved with a silent-less choke. A significant advantage of a non-conductive choke is its small size and lightweight construction compared to previous embodiments. Thus, the manufacturing costs are also significantly lower than before.

The invention has been described above with respect to inverters having three parallel branches. However, it is clear that the invention can be applied to g also in a context where the number of branches differs from three, g 35 It is further apparent to one skilled in the art that the basic idea of the invention can be implemented in many different ways. The invention and its embodiments are thus not limited to the examples described above, but may vary within the scope of the claims.

δ

(M

i tn o i oo o

X

I do not

CL

sj- cn sj- m sj- o o

(M

Claims (3)

10 A method in conjunction with inverter output chokes, wherein the inverter phase output (U, V, W) is adapted to consist of two 5 or more parallel branches, each comprising an output of an upper and a lower power semiconductor component, the upper (V1, V3, V5) and the lower (V2, V4 and V6) to the negative voltage with the components in series, whereby the output of each branch (UU1, UU2, UU3, UV1, UV2, UV3, UW1, UW2, UW3) is comprised of 10 component points, the output choke comprising inductors (1), the first ends of the inductors of the inductors being connected to the outputs of the branches and the other ends connected together to form the inverter phase output (U, V, W), characterized in that the method comprises the steps of predetermining the voltage limit and Indiko determining the control power of the branch power semiconductor components, determining the collector voltage of the branch power semiconductor components in the power semiconductor components controlled to a conductive state, detecting a failure condition in A method according to claim 1, characterized in that the method further comprises the step of controlling the fault current by switching off the power semiconductor components. o The method according to claim 1 or 2, characterized in that the method further comprises the step of generating an alarm signal § responsive to the detection of a fault. X en CL Sj- cn sj- m sj- o o {M 11