Method and apparatus for correction of a detected faulty actual valve of a signal affected by harmonic vibrations

#The present invention relates to a method for correction of an actual value of a signal affected by harmonics, the actual value determined by means of a first integrating actual value acquisition device, which value is affected by errors, and the signal is generated from a harmonic-free setting signal by means of a modulation function which approximates a harmonic-free setting signal, and a control system.

#In electrical drive technology, analog measurement variables, for example converter current, machine voltage, etc., are frequently converted into digital measurement values using integrating measurement value acquisition devices. With the use of an integrating measurement value acquisition device, measurement noise is filtered out and this measurement device can be structured in a very cost-effective manner. A typical structure of an integrating measurement value acquisition device is explained in greater detail in reference to FIG. 3.

#Harmonics of a measurement variable are only filtered out, in this connection, if they have the mean value of zero during the measurement interval (gate time). This is brought about, for example, in space vector modulation, by approximation, in that the gate time is made exactly as large as the duration of a switching state sequence, which approximates a reference vector.

#FIG. 1 shows a block schematic of a circuit arrangement for implementation of the method according to the present invention,

#FIG. 2 represents a pulse-width-modulated-signal with five control angles .alpha..sub.1 to .alpha..sub.5 per quarter period in a diagram over the angle of rotation .omega.t,

#FIG. 3 shows a block schematic of a known integrating actual value acquisition device, and

#FIG. 1 represents a block schematic of a circuit arrangement for implementation of the method according to the invention, for correction of an actual value u.sub.ixRF affected by errors, which is determined by means of a first integrating actual value acquisition device. This circuit arrangement consists of a modulator 4, a control system, including of a converter 6 and an asynchronous machine 8, two comparators 10 and 12, an equivalent model 14 of the control system, a first and second actual value acquisition device 2 and 16, and a pulse generator 20 for the gate time of the first and second actual value acquisition device 2 and 16. As the converter-powered control system, a line power converter can also be used; it is described in more detail in the journal Siemens Forsch.- und Entwickl.-Ber., Volume 16, 1987, No. 2, pages 55 to 59. A harmonic-free setting signal u.sub.yR is passed to the modulator 4 on the input side, which signal is converted to a pulse-width-modulated setting signal u.sub.yRF by means of a pulse pattern .alpha..sub.1, ..., .alpha..sub.5. This setting signal u.sub.yRF, which is affected by harmonics, shown in greater detail in FIG. 2, in a diagram above the angle of rotation .omega.t, controls the converter 6 which powers the asynchronous machine 8. As the result of the a pulse-width-modulated converter output voltage, a signal i.sub.xRF is provided which is affected by harmonics (also called the control signal or control variable). This control signal i.sub.xRF which is affected by harmonics is acquired by means of a measurement transducer 18 and converted into a digital actual value u.sub.ixRF using the first actual value acquisition device 2. Since the control signal i.sub.xRF is affected by harmonics, the measurement value u.sub.ixRF of the first actual value acquisition device is also affected by harmonics, which means that only an actual value u.sub.ixRF affected by errors can be determined, because the mean value is not equal to zero Within the gate time interval T.sub.AB.

#The harmonics are excited by the modulation process, i.e. by the pulse pattern .alpha..sub.1, ..., .alpha..sub.5 stored in memory. In a pulse width modulated signal u.sub.yRF, shown in greater detail in FIG. 2, in a diagram above the angle of rotation .omega.t, five control times .alpha..sub.1, ..., .alpha..sub.5 are established for every quarter period, which are stored in a memory of the modulator 4. The memory contains several sets of control times .alpha..sub.1, ..., .alpha..sub.5, one set for every amplitude of the setting signal U.sub.YR that can be realized. The number of sets depends on the resolution of the invertor control ratio. Since the converter 6 powers a three-phase asynchronous machine 8, a three-phase signal system u.sub.yR, u.sub.yS and u.sub.yT is preset as a setting voltage system, of which only the signal u.sub.yR will be considered here. The signal system u.sub.yR, u.sub.yS and u.sub.yT can be determined by a reference value transducer or by a control and regulation device. In the simplest case, a comparator can be provided for each phase, which are controlled as a function of the control times .alpha..sub.1, ..., .alpha..sub.5 stored in memory in each case. These control times .alpha..sub.1, ..., .alpha..sub.5 are read out from a memory arranged in the modulator 4, by means of amplitude values and phase positions of the setting voltages u.sub.yR, u.sub.yS and u.sub.yT which are present at the modulator 4. Precalculated, optimized pulse patterns are known, for example from the journal etz-A, Volume 94, 1973, Issue 5, pages 299 to 301. The application of precalculated, optimized pulse patterns for pulse-width-modulated converters with an intermediate voltage circuit to power three-phase asynchronous machines is also well-known. By the use of precalculated pulse patterns, a control signal i.sub.xRF affected by harmonics is formed, which is converted into an actual value u.sub.ixRF affected by errors, when a first integrating actual value acquisition device 2 is used.

#In order to eliminate the error caused by the harmonics, the harmonic signal u.sub.f which causes the harmonics must first be determined. For this purpose, the harmonic-free setting signal u.sub.yR is passed to the comparator 10 at the inverting input, and the setting signal u.sub.yRF affected by harmonics, generated by the modulator 4, is passed to the non-inverting input. At its output, the harmonic signal u.sub.F is then present. This harmonic signal u.sub.F is converted into a model signal u.sub.iF by means of the equivalent model 14 of the control system 6, 8, which model signal corresponds to the superimposed harmonic signal of the control signal i.sub.xRF As an equivalent model 14 for the control system 6, 8, consisting of a converter-powered asynchronous machine 8, an inductance is provided in the simplest case. The value of the inductance corresponds to the leakage inductance L.sub..sigma. of the asynchronous machine 8. The model signal u.sub.iF is then passed to the second actual value acquisition device 16, which determines mean values u.sub.iF in response thereto. In this connection, each mean value u.sub.iF is always determined within a gate time interval T.sub.AB of the first actual value acquisition device 2. At the end of each gate time interval T.sub.AB, a digital value is present, which counts as a measure of the error. These continuously determined correction values u.sub.iF are passed to the inverting input of the comparator 12. At the non-inverting input of this comparator 12, the actual values u.sub.ixRF of the continuous gate time intervals T.sub.AB, which are affected by errors, are present. At the output, a corrected actual value u.sub.ixR is then present, which can be processed further, for example in a current regulation, a flux model or a coordinate transformation circuit for determination of current components which form torque and flux.