DE202012000091U1 - Device for controlling an induction machine on a direct converter up to zero speed - Google Patents

Abstract

Device for controlling a three-phase machine (2) on a direct converter (1) to zero speed, the associated field-oriented control device (3) and (9) performing an observer (6) and a machine model (8) as well as speed adaptation (5) which has the following process steps: 1.) Reading in the actual machine values 2.) Calculating the correction values for these 3.) Calculating the model flow (φ) and a model current (I) from the corrected actual voltage values and the adapted speed (ṅ) 4. ) Calculating the complex deviation in (5) whereby the current differences are formed with the measured and predicted currents and multiplied by the model flow. The resulting deviations are fed to the controller with PI characteristics. The output is a speed adaptation variable proportional to the deviation. 5.) Depending on the machine type, the model is modified, but it always provides projectable values for the speed adaptation.

Description

The application relates to a device for controlling an induction machine to a direct converter to zero speed, the associated control device has a machine observer and a machine model which is adapted to the design of the machine. The models / observers are supplied with actual values by corresponding measuring devices which are explained below.

The need for a speed sensor is associated with considerable costs and risks for many industrial applications of drive technology. It is desirable to eliminate these sensors for cost reduction and increase in availability.

There are some prior art methods for field-controlled rotating field machines to operate and operate. However, restrictions arise due to system characteristics of the converters and machines used. The derivative of the flow occurs at a certain speed threshold from the measured stator currents and voltages. Since the measurement becomes less accurate with decreasing speed and thus decreasing voltage, this information alone can no longer be regarded as evaluable.

From the EP 2 019 482 A1 and the DE 10 2008 058 739 A1 principles are known which can be applied specifically to the application of voltage DC link converters with asynchronous machines and permanent magnet synchronous machines. These methods, however, are based on models that have to make do without measuring the real machine voltage.

This method can be improved in the use of a cyclo-converter as known in this application by the presence of voltage measuring devices and is thus more universally applicable to a variety of induction machines.

Furthermore, from the DE 10 2007 003 874 A1 is a method for sensorless operation of a converter-fed induction machine with a test signal known. By means of this test signal, which makes evaluable in the direction of a suspected flow axis, the estimated flow axis can be corrected. However, this requires, on the one hand, a corresponding signal source and, on the other hand, a machine is expected which clearly reflects the test signals and makes them detectable in another measured variable.

The physical structure of the direct converter makes the provision of the test signal impossible and, secondly, experience has shown that not all machines reflect this test signal due to their magnetic structure.

It is therefore obvious that there must be a specific and universal solution for the direct converter for the economical operation of drives of this type.

From the known control devices for field-oriented operation, models are known which deliver accurate estimated values from the stator measured variables up to a rotational speed of about 5... 10% of the nominal rotational speed alone. At speeds below are only machine observers such. B. by the Scriptures Kubota and Matsuse's "DSP-based speed adaptive flux observer" in April 1993 or by "Speed Control of the Asynchronous Motor Using an Observer with Low Parameter Sensitivity" by Zägelein (April 1984) known.

These observers allow a much smaller speed range to be approached, which can also result in zero speed or zero speed on passive loads when a useful pilot size signal is available.

The solution approaches DE 10 2008 058 739 A1 is calculated using the desired voltages, which can be improved in this application by the measurability of the machine voltage advantageous. Since the actual voltage values can be noisy despite the low harmonic components in the case of cycloconverters, according to the method a correction of the measured values by a state variable filter and a subsequent evaluation with a predictor device are improved.

The approach with the use of the setpoint values may be advantageous within certain limits in the case of the PWM converter due to the high harmonic components, but an exact model with the knowledge of the non-linear conditions with a low modulation factor is necessary. For this purpose, locking times and semiconductor properties for linearization must be known for each power converter that has been used.

This costly method can be eliminated in the direct converter and described measuring device advantageous and provides more accurate results even at very low frequencies.

If the control device is used flexibly with regard to the depicted motor model, a universal solution for almost all known rotary field machines can be created.

Because of the unknown previously described, in many applications a so-called "controlled operation" was introduced for starting or reversing through zero speed. This operating mode requires a precise knowledge of the necessary start-up torque that is necessary for many machine types to approach. If this moment is not known, the drive can either go through or "tip over". Both can cause damage to the driven system and must be avoided.

Since the control information for the converter is derived from the main flow angle in the case of a cyclo-converter, the position of this angle must be smooth and jump-free. Therefore, the tracking control ( 1 ; Pos. 5 ) used to equalize the target speed to the actual speed. A direct derivation of this angle from the derivatives of the EMF is not possible here because the necessary integrators tend to instability at low frequencies and thus the control information would be incorrectly routed to the converter, which can lead to damage.

For further explanation of the invention will now reference to 1 taken:
The drawing 1 shows an advantageous embodiment of the described observer structure with corresponding measuring device according to claim 1.

In the 1 are with 3 and 9 a field-oriented control device, with 4 a load-side converter of a direct converter, not shown, with 2 a encoderless induction machine and with 8th an observer. This field-oriented control device has a speed control loop 3 and a current loop 9 on, with the current loop 9 the speed control loop 3 is subordinate. That is, the output of the speed control loop 3 is a target value of a torque-forming current component Isq * of a machine current space vector. The output side is the current control loop 9 with the load-side converter 1 , (According to a direct converter) linked, the input side has a control device. In this illustration is the real load-side converter 1 represented as a converter. By means of at least two measuring devices not shown in more detail, at least two machine currents and voltages are measured, which represent a machine current room pointer I and machine voltage room vector U. These two measured machine sizes become the preparation of the correction device 4 and the predictor facility 7 to hand over. The current loop 9 The field-forming current setpoint Isd * is also transferred from a higher-level flow control device. The actual flow value can be the model output 8th or the observed river 6 serve. In this illustration, the associated machine-stream space vector I is shown.

Parallel to the observer 6 the controller has a machine model 8th , a speed adaptation 5 and an actual speed error compensation over n *. At the first entrance of the machine model 8th is the determined voltage value U and at the second input of 8th is the current value at. This machine model 8th has the equations of state of the corresponding induction machine. The outputs φ and I of this machine model 8th are with the speed adaptation 5 connected. The two outputs of the observer 6 are with the other inputs of the speed adaptation 5 connected. The output signal from 5 (ṅ) forms an input quantity of the actual speed error compensation via the setpoint n * and further in 7 and to 4 linked where the corrected speed ñ serves to correct the measured values. Also the speed controller 9 the speed ñ is used as the actual value. Advantageously, outside this representation, the speed ñ is used to form the flow angle.

By means of the device according to the invention the error of a faulty speed prediction is compensated by means of a model speed. This results in a robust field-oriented control device that can be used for various machine types by replacing the machine model. Thus, it is now possible to operate the fully regulated operation with active speed control loop up to the frequency zero point and to a standstill, the controlled operation can be omitted.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• EP 2019482 A1 [0004]
• DE 102008058739 A1 [0004, 0011]
• DE 102007003874 A1 [0006]

Cited non-patent literature

• Kubota and Matsuse's "DSP-Based Speed Adaptive Flux Observer" in April 1993 [0009]
• "Asynchronous motor speed control using a low parameter sensitivity observer" by Zägelein (April 1984) [0009]

Claims (4)

Device for controlling an induction machine ( 2 ) on a direct converter ( 1 ) to zero speed, with the associated field-oriented control device ( 3 ) and ( 9 ) an observer ( 6 ) and a machine model ( 8th ) and a speed adaptation ( 5 ), which comprises the following method steps: 1.) reading in the machine actual values 2.) calculating the correction values for these 3.) calculating the model flow (φ) and a model current (I) from the corrected voltage actual values and the adapted rotational speed (ṅ) 4.) Compute the complex deviation in ( 5 ) where the current differences are formed with the measured and anticipated currents and multiplied by the model flux. The resulting deviations are fed to the controller with PI characteristic. Output is a deviation proportional speed adaptation quantity. 5.) Depending on the type of machine, the model is modified, but always provides projectable sizes for speed adaptation. The object of the invention is achieved by a device having all the features of claim 1. Device according to claim 1, characterized in that the machine models of the control are preferably designed as rotor models. Device according to claim 1, characterized in that the machine models of the control are calculated from corrected actual values.

Images