DE19544410A1 - Real=time calculation of pulse pattern for three=phase inverters - has three binary switching signals calculated for each modulation cycle from set points for phase-to-phase voltages - Google Patents

Abstract

A procedure for real-time calculation of the pulse pattern for three-phase inverters where the switching state is determined by three binary signals which are calculated for each modulation cycle from set points for the voltages between the phase terminals. For each of the three signals which should come into effect for the next modulation cycle, the following parameters are calculated separately:- the value of the signal at the beginning of the cycle, the value at the end of the cycle, the switching time between the beginning and end of the cycle. The set points for the phase-to-phase voltages and the mean direct voltage at the terminals are normalised to half this direct voltage.

Description

The invention relates to a method for real-time calculation of the pulse pattern of three-phase two-point inverters, in which the switching state by the passage of time is determined by three binary switching signals, which for each modulation game from setpoints for the voltages between the three-phase connections.

The connection points on the AC side of a two-point inverter are connected to either the positive pole or the negative pole of the supplying DC voltage. The basic task of a pulse modulation method is to determine the time spans for the two possible voltage states at a connection point (= strand of the three-phase system) identified by an index .με (a; b; c) within a so-called modulation game with the duration T _m , that the mean value of the connection voltage formed via the modulation game coincides with a predetermined target value, the connection voltage being defined as the potential difference compared to the mean value of the potentials of the two DC voltage connections.

Fig. 2 shows an example of the circuit principle of a three-phase two-point inverter. Switching devices U _a , U _b1 , U _c (converter valves) can be used to connect the connections of a three-phase load either with the positive or negative intermediate circuit voltage (half intermediate circuit voltage = E _d ). The switching signals required to control the switching devices for the strings µ = a or µ = b or µ = c of the inverter are S _a , S _b , S _c (generally Sµ). The phase currents or stator currents are i _a , i _b , i _c . The voltages between the center potential point M 'of the input _DC voltage and the strands are e _aM' , e _{bM '} , e _cM' (generally e _{µM '} ).

As is known, the point in time for the potential change at the connection µ can be very easily determined by the switching variable S _µ (switching signal) within a modulation game by comparing a time-linear auxiliary function v (t) (comparison function) with the mean value of the standardized connection voltage desired for the modulation game under consideration ( = Pulse control value). Fig. 3 shows the relationships for two modulation games, if you normalize the voltage to half the value of the supplying DC voltage (DC link voltage of the inverter) 2 Ed and the time spans to half the modulation cycle time T _m .

Is used for the normalized time Xν of the ν-th modulation game

With
T _m = modulation cycle time
ν = ordinal number of the modulation games (ν = 1; 2; 3;...)
t = continuous time

this gives the time to be determined for the potential change at connection µ the simple relationship

With

The setpoints for the pulse control values

are usually calculated by the substructure of the signal processing for controlling and regulating the load connected to the inverter for each modulation game so long before the start time (ν-1) · T _{m of} the game that a causal sequence is ensured. The two-valued switching variables S _µ with µ = (a; b; c) determine that the connection is connected to the positive pole of the input DC voltage if S _µ = 1 applies and is connected to the negative pole for S _µ = 0. The switching variables are determined according to the following rule:

Inequality (3) is e.g. B. met for the entire modulation game if

applies. For

is the inequality in the whole modulation game not fulfilled. In both cases there is no potential change in the ν-th modulation game at the µ-th connection instead of for

there is always a potential change at ε0. The desired mean values of the internal connection voltages (pulse control values) can in principle be implemented without errors using pulse modulation, as long as the following two conditions e.g. B. for the (ν + 1) -th modulation game are fulfilled:

With

In order to indicate that the standardized mean values of the internal connection voltages required by the regulation and control are setpoints, the feasibility of which still has to be checked with regard to inequalities (4), an exclamation mark is added to the formula symbol. By fulfilling condition A, it is ensured that no average values v are obtained for the internal connection voltages, the amounts of which exceed the maximum realizable value E _d . For various reasons, a so-called minimum switching state time T _min must be observed between two potential changes at one connection of the inverter. Under steady-state modulation conditions, this means that the magnitude of the quantities is the first limit

must not exceed. By fulfilling condition B, the present invention ensures that dynamic, d. H. For

the time Δt between two potential changes does not become less than T _min . Even if one of the conditions (4) for the pulse control setpoints is not met, the mean value of the supply voltage required by the control and regulation system cannot be realized for the modulation game under consideration. The task of a so-called overmodulation method is to form uniquely realizable values from non-realizable target values for the mean voltage values of the connection voltages. The associated prior art is such. B. described in: "Holtz, J .; Lotzkat, W .; Khambadkune, A .: On Continuous Control of PWM Inverters in the Overmodulation Range Including the Six-Step Mode. IEEE Trans: O. Power Electr. Vol. 8 No 4 Oct. 1993, p. 546-553 ".

The previous solutions have the following shortcomings: the "preprocessor" for forming realizable control variables from the non-realizable setpoints for the space pointer of the connection voltages processes complicated equations which, in the form described, also only apply to fundamental oscillation variables, ie to steady-state operating states. Furthermore, no measures are described which are necessary to solve the "T _min " problem.

The invention has for its object a method for real-time calculation of Pulse pattern in the overmodulation range of three-phase two-point inverters, preferably for three-phase two-point inverters to indicate that no complicated Must process equations and this also applies to non-stationary operating states.

This object is achieved in conjunction with the preamble of claim 1 in that for each of the three switching signals S _a ; S _b ; S _c , which are to come into effect in the next modulation game, the following parameters are calculated separately:

• - the value S _{µ1 of} the switching signal required at the beginning of the modulation _game ,
• - the value S _{µ2 of} the switching signal required at the end of the modulation _game ,
• the time of switching from the value S _µ1 required at the beginning of the modulation game to the value S _{µ2 required} at the end of the modulation _game ,

where µ ε a; b; c = names of the three strands of a three-phase system.

An embodiment of the method is characterized in that, for simple calculation of the switching signal values and the switchover time, the target values for the voltages between the three-phase connections (a; b; c) and the mean value of the potentials of the two direct voltage connections to half the value of the direct voltage 2E _d between them Connections are standardized and so are the pulse control setpoints

form and that for each modulation game a time variable X _ν normalized to half the modulation game time T _{m is} formed, which runs from -1 to +1, where ν is the ordinal number of the modulation game.

It is provided that from the pulse control setpoints

three possibly corrected, new pulse control setpoints are formed in such a way that, for none of the required connection voltages, the amount is greater than the maximum possible value of half of the input DC voltage E _d and that, in the case of stationary modulation operation, between two potential changes on the same connection, at least a fixed time period T _min elapses, which corresponds to the minimum value of the duration of permissible switching states of the inverter.

A further measure based on this is characterized in that in modulation games, in which the amount of a pulse control setpoint is less than a first one limit

the unchanged value of the original pulse control setpoint is used for the assigned corrected pulse control setpoint, T _{m being} the modulation cycle time and T _min the minimum value of the duration of permissible switching states of the inverter.

It is useful in modulation games in which the amount of a pulse control setpoint is considered a second limit that is between 1 and the first limit lies for the assigned corrected pulse control setpoint the amount 1 and Sign of the original pulse control setpoint used.

The second limit value becomes the first limit value using the formula

determines the first limit and the second limit represent and preferably chosen a value between 0.5 and 0.39 for the factor k becomes.

In a further embodiment, in modulation games in which the amount of a pulse control setpoint lies between the two limit values and, for the assigned corrected pulse control setpoint the amount of the first limit and that Sign of the original pulse control setpoint used.  

Furthermore, the final pulse control values are determined from the corrected pulse control setpoints so that the time period Δt between two potential changes of the same connection does not fall below a specified minimum value of the duration of permissible switching states of the inverter T _min if the corrected pulse control setpoints for the next modulation game are not achieved with the ones Pulse control values of the previous modulation game match.

A further development of the method is characterized in that in the cases in which a pulse control value

of the previous modulation game and the associated new corrected pulse control setpoint both have the value 1 and the same sign, unnecessary and impermissible potential changes at the assigned connection µ are avoided by the value S _µ2 required at the end of the ν-th modulation game unchanged for both new values of the assigned switching variables S _µ1 and S _{µ2 of} the switching variables of the next (ν + 1) -th modulation game is taken over, which is expediently implemented whenever a check of the computational time interval Δt between two potential changes on the same connection gives a value which is between zero and the Half of the minimum value T _min , preferably at a quarter of the minimum value of the duration of permissible switching states of the inverter. This rule very reliably avoids accuracy problems in the numerical calculation.

In cases where a pulse control value

of the previous modulation game and the associated new corrected pulse control setpoint do not both have the amount 1 and the same sign, it is checked whether the calculated time interval Δt between two potential changes at the same connection is greater than the minimum value T _{min of} the duration of permissible switching states of the inverter, the assigned corrected pulse control setpoint is determined as the final pulse control value for the new modulation game when this condition is met.

In cases in which the arithmetical time interval Δt between two potential changes on the same connection is not greater than the minimum value T _{min of} the duration of permissible switching states of the inverter, the value that results from the value shown in Fig. 1 is used as the final pulse control value instead of the corrected pulse control setpoint 3 apparent formula.

for the time period Δt _µ between two potential changes on the same connection depending on the pulse control value

of the previous modulation game, if the minimum value T _{min of} the duration of permissible switching states of the inverter is selected for the time period Δt _µ between two potential changes at the same connection, the formula normalizes to T _m / 2, the X variables according to equation (2) the variables are replaced and the result is multiplied by Sign.

This means that equivalent relationships are used to check time periods which are defined by the minimum value T _{min of} the duration of permissible switching states of the inverter or fractions of this minimum value, namely between the first limit value, the pulse control value

the previous modulation game, the new, corrected pulse control setpoint, the new pulse control value and a sign size sign _n , which is assigned the value +1 and even the value -1 for even ordinal numbers ν of the modulation games.

The point in time at which the potential in the next modulation game at connection µ becomes is to be changed, determined according to the following equation:

With

A further development of the method is characterized in that in the cases in which a pulse control value

of the previous modulation game and the associated new corrected pulse control value do not both have the amount 1 and the same sign, which is determined as follows at the beginning of a modulation game of the switching variable S _µ1 :

(with X _{( ν +1)} = standardized time variable for the ν + 1) th modulation game) and that after the switching _delay X _{( ν +1) Sµ} , after which the potential changes at the connection µ, for the assigned switching variable S _µ1 the inverted binary value is assigned

The advantages that can be achieved with the invention consist in particular in that only elementary arithmetic operations with numerical values known before the start of a modulation game are required for determining the time profile of the switching signals S _μ .

According to the time course of a switching signal z. B. for the (ν + 1) th modulation game described by the value S _{µ1 of} the switching variable S _µ to be realized at the beginning at X _{( ν +1)} = -1 and by the value X _{( ν +1) Sµ of} the variable X _{( ν +1)} which determines at what point in time the potential at connection µ changes. From the value X _{( ν +1) Sµ} a comparison _value can be clearly determined in a known manner, which is constantly compared with the status of a counter that counts the clock pulses generated by a clock generator at constant time intervals since the beginning of the modulation game. If the comparison value and counter reading are the same, the value of the switching signal changes from S _µ1 to the value S _µ2 , which _according to the invention is not automatically set to the inverted value of the signal S _µ1 in all cases, but is determined separately under certain conditions in order to avoid unnecessary and to avoid impermissible switching of the inverter.

With high demands on the dynamics, from signal processing to control and rules for each modulation game the target values for the mean values of the normalized inner Connection voltages, called pulse control setpoints for short, as orthogonal coordinates  a complex space pointer recalculated; e.g. B. for the (ν + 1) th modulation game in the shape

With

The orthogonal coordinates and become the three inverter strings with connections a, b and c assigned to zero-system-free string coordinates e.g. B. calculated as follows:

Without the mean values of the voltage differences between the connections a, b and c to change, a zero system value can be added to the non-zero system variables, z. B. the usable for pulse control without overmodulation big as physically possible. As in "Weinhold, M .: Appropriate Pulse Width Modulation for a three-phase PWM AC-to converter. EPE Journ. Vol. 1 no. 2, Oct. 1991; p.  139-148 ", this is achieved, for example, by the method of" symmetrization " In this case, half of that of the three strand sizes becomes the zero system chosen that has the smallest amount:

This results in the new symmetrical setpoints

According to the structure diagram shown in FIG. 1, the next step is to check whether conditions corresponding to inequalities (4) are met for these three quantities. These checks are carried out individually and independently for each connection. The method according to the invention is therefore only described below using the example of connection a.

First, the pulse control values become a set of intermediate values (Intermediate setpoints or corrected pulse control setpoints) determined, although with stationary Modulatinsbetrieb fulfill the conditions (4), but are still checked thereupon whether this is also the case when required changes in the modulation operating state the case is.

If the inequality

is fulfilled, the following applies to the intermediate variable (corrected pulse control setpoint)

For each connection in which the associated inequality (9μ) is not met, the invention is used carried out another review; e.g. B. if this is for connection a the inequality is:

The second limit value determines whether a required, not realizable Pulse control value better by a realizable with the amount 1 or one with the amount (= first limit) is approximated. When setting the second limit with the factor k = 0.39, there are particularly small errors in stationary operation the realized fundamental voltage oscillations.

If inequality (11a) is satisfied, the intermediate variable (corrected Pulse control setpoint)

For each connection where the associated inequality (11µ) is not fulfilled, the formation of the pulse control setpoint multiplied by the sign of the pulse control setpoint, e.g. B. in the case of connection a:

In the case of symmetrization, it is sufficient to use only two of the three inequalities for the three Check connections, because the mathematically largest and the smallest pulse control value then always have the same amounts and the result for the unchecked connection is then already determined.

The corrected pulse control setpoints determined in this way meet under stationary modulation conditions all the inequalities (4). To ensure that this is also dynamic   the case is, a further condition is checked for each connection, e.g. B. for the Connection A:

In each strand, in which the assigned condition is fulfilled, a double could be formally Potential changes with Δt = 0 are required. According to rule (3), this is instead If no potential change in the strand concerned is caused, so that for the final pulse control values in the affected strands the intermediate value as the final one Pulse control value is set.

For the entire (ν + 1) -th modulation game, z. B. from the switching variable S _a1

(generally S _µ1 = switching signal required at the beginning of the pulse period)

This means that in this special case this also applies formally to the variable S _a2 , regardless of the value of X _{( ν +1) Sa} to be calculated

S _a2 = S _a1 (16a)

(general: S _µ2 = switching signal required at the end of the pulse period)

For all connections for which the assigned condition (14µ) is not met, the (ν + 1) th modulation game a potential change takes place, therefore in these cases:

The last thing to check is whether the T _min condition with the corrected pulse control setpoints is being met. B. for connection a with the inequality

If this inequality is not satisfied, then the final pulse control value for the ( _n +1) th modulation game z. B. determined for the connection a from the following equation:

In all other cases, z. B. for connection a:

This means that all pulse control values for the (ν + 1) th modulation game are clear and free of contradictions certainly. To determine the times of the potential change can be analogous to Eq. (2) be formulated uniformly:

The relationships (9µ). . . (22µ) can be calculated digitally in parallel for connections a, b and c if there are three processors. With just one processor, relationships can for the three connections can be processed one after the other without problems because they are completely independent of each other.

Based on the analog representation of the quantities in FIG. 2, the quantity _ν can be explained as the slope of the comparison function. In the method according to the invention, this comparison function is explicitly not required. You can therefore z. B. also define:
if ν is an even number, then

sign _ν = +1

otherwise

sign _ν = -1

or also recursively:

sign _{ν = 0} = +1
sign _{( ν +1)} = -sign _ν

The method described results for any desired space pointer Pulse control setpoints always clearly and without contradiction an assigned, realizable Space pointer of the pulse control values, but it makes sense to the area of application on

to limit, because there are already good procedures for target space pointers with larger amounts, such as B. the method of web speed modulation, as described in "Springmeier, F .: Direct stator size control of induction machines on the three-point inverter. Progress Ber. VDI series 21, No. 133, 1993, pp. 116-123 "for three-point inverters and the semiconductor switch at a sufficiently high permissible switching frequency can be easily transferred to two-point inverters.

The method is therefore preferably only for the modulation range of the inverter applied, in which the angle argument of the complex spatial pointer of the Connection voltages can be set almost continuously to any value, d. H. if stationary the amount of the pulse control variables in the range

lies.  

List of the formula symbols used

Claims (16)

1.Procedure for real-time calculation of the pulse pattern of three-phase two-point inverters, in which the switching state is determined by the time course of three binary switching signals (S _a ; S _b ; S _c ), which for each modulation game from setpoints for the voltages between the three-phase connections ( a; b; c) are calculated, characterized in that the following parameters are calculated separately for each of the three switching signals (S _a ; S _b ; S _c ) which are to come into effect in the next modulation game:

• - the value (S _µ1 ) of the switching signal required at the beginning of the modulation game,
• - the value (S _µ2 ) of the switching signal required at the end of the modulation game,
• the time of switching from the value required at the beginning of the modulation game (S _µ1 ) to the value required at the end of the modulation game (S _µ2 ),

where µ ε a; b; c = names of the three strands of a three-phase system. 2. The method according to claim 1, characterized in that for simple calculation of the switching signal values and the switching time, the target values for the voltages between the three-phase connections (a; b; c) and the mean value of the potentials of the two DC voltage connections to half the value of the DC voltage (2E _d ) are standardized between these connections and thus the pulse control setpoints form and that for each modulation game a time variable (X _ν ) normalized to half the modulation game duration (T _m ) is formed, which runs from -1 to +1, where ν is the ordinal number of the modulation game. 3. The method according to claim 2, characterized in that from the pulse control setpoints three possibly corrected, new pulse control setpoints are formed in such a way that for none of the required connection voltages the amount is greater than the maximum possible value (E _d ) of half the DC input voltage and that in the case of stationary modulation operation between two potential changes on the same connection, at least a defined period of time (T _min ) elapses, which corresponds to the minimum value of the duration of permissible switching states of the inverter. 4. The method according to claim 3, characterized in that in modulation games in which the amount of a pulse control setpoint is less than a first limit the unchanged value of the original pulse control setpoint is used for the assigned corrected pulse control setpoint, where (T _m ) is the duration of the modulation and (T _min ) is the minimum value of the duration of permissible switching states of the inverter. 5. The method according to claim 4, characterized in that that in modulation games, in which the amount of a pulse control setpoint is greater than a second one Limit value that lies between 1 and the first limit value for the assigned corrected pulse control setpoint the amount 1 and the sign of the original Pulse control setpoint is used. 6. The method according to claim 5, characterized in that the second limit value from the first limit value by the formula is determined, the first limit value and the second limit value being represented, and a value between 0.5 and 0.39 preferably being selected for the factor k. 7. The method according to claim 5 or 6, characterized in that in modulation games, in which the amount of a pulse control set point between the two Limits and lies for the assigned corrected pulse control setpoint  the amount of the first limit and the sign of the original Pulse control setpoint is used. 8. The method according to any one of claims 4 to 7, characterized in that the final pulse control values are determined from the corrected pulse control setpoints so that the time period Δt between two potential changes of the same connection also then a fixed minimum value of the duration of permissible switching states of the inverter (T _min ) does not fall below if the corrected pulse control setpoints for the next modulation game do not match the realized pulse control values of the previous modulation game. 9. The method according to any one of claims 3 to 8, characterized in that in the cases in which a pulse control value of the previous modulation game and the associated new corrected pulse control setpoint both have the value 1 and the same sign, unnecessary and impermissible potential changes at the assigned connection (μ) can be avoided by the value (S _µ2 ) of the assigned switching variables required at the end of the previous modulation game unchanged for both new values (S _µ1 and S _µ2 ) of the switching variables of the next modulation game is realized, which always happens when a check of the calculated time interval (Δt _µ ) between two potential changes on the same connection results in a value that is between zero and half of the Minimum value (T _min ), preferably decided by whether Δt _{μ is} greater or less than a quarter of the minimum value of the duration of permissible switching states of the inverter. 10. The method according to any one of claims 3 to 8, characterized in that in the cases in which a pulse control value of the previous modulation game and the associated new corrected pulse control setpoint not both have the amount 1 and the same sign, it is checked whether the calculated time span (Δt) between two potential changes on the same connection is greater than the minimum value (T _min ) of the duration of permissible switching states of the inverter is, wherein the assigned corrected pulse control setpoint is determined as the final pulse control value for the new modulation game when this condition is met. 11. The method according to any one of claims 9 and / or 10, characterized in that equivalent relationships are used for checking periods which are defined by the minimum value (T _min ) of the duration of permissible switching states of the inverter or fractions of this minimum value, and between the first limit value, the pulse control value the previous modulation game, the new, corrected pulse control setpoint, the new pulse control value and a sign size (sign _ν ) which is assigned the value +1 and even the value -1 for even ordinal numbers ν of the modulation games. 12. The method according to any one of claims 3 to 8, characterized in that in the cases in which the calculated time period (Δt) between two potential changes at the same connection is not greater than the minimum value (T _min ) of the duration of permissible switching states of the inverter , instead of the corrected pulse control setpoint, the value derived from the formula is used as the final pulse control value for the time span (Δt _µ ) between two potential changes on the same connection depending on the pulse control value of the previous modulation game results if the minimum value (T _min ) of the duration of permissible switching states of the inverter is selected for the time period (Δt _μ ) between two potential changes on the same connection, so that applies 13. The method according to claim 12, characterized in that the point in time at which the potential is to be changed at the connection µ in the next modulation game is determined according to the following equation: With 14. The method according to any one of claims 3 to 13, characterized in that in the cases in which a pulse control value of the previous modulation game and the associated new corrected pulse control value do not both have the amount 1 and the same sign, which is determined as follows at the beginning of a modulation game of the switching variable S _µ1 : With
X _{( ν +1)} = standardized time variable for the (ν + 1) th modulation game and that after the switching _delay X _{( ν +1) Sµ} , after which the potential changes at connection µ, the _inverted for the assigned switching variable S µ2 Binary value is assigned 15. The method according to any one of claims 3 to 13, characterized in that in the cases in which a pulse control value of the previous modulation game and the associated new corrected pulse control value both have the amount 1 and the same sign, which is determined as follows for the value of the switching variable S _µ1 to be realized at the beginning of a modulation game: and that there is then no change in the switching variables within the modulation game, so that applies