SINAMICS S120 / SIMOTION D

Transcription

1 Cover Servo Drive Optimization Guide SINAMICS S120 / SIMOTION D Application November 2012 Applikationen & Tools Answers for industry.

2 Siemens Industry Online Support This article is taken from the Siemens Industry Online Support. The following link takes you directly to the download page of this document: Caution The functions and solutions described in this article confine themselves to the realization of the automation task predominantly. Please take into account furthermore that corresponding protective measures have to be taken up in the context of Industrial Security when connecting your equipment to other parts of the plant, the enterprise network or the Internet. Further information can be found under the Item- ID You can also actively use our Technical Forum from the Siemens Industry Online Support regarding this subject. Add your questions, suggestions and problems and discuss them together in our strong forum community: 2 V1.0, Item-ID:

3 SIEMENS Measurement of the mechanical system 1 Current Controller optimization 2 Speed controller optimization 3 Servo Drive Optimization Guide SINAMICS S120 / SIMOTION D Position controller optimization (SIMOTION) 4 Coordinated axes 5 Summary 6 Additional Information 7 Glossary 8 Related literature 9 Contact 10 History 11 V1.0, Item-ID:

4 Warranty and liability Warranty and liability Note The Application Examples are not binding and do not claim to be complete regarding the circuits shown, equipping and any eventuality. The Application Examples do not represent customer-specific solutions. They are only intended to provide support for typical applications. You are responsible for ensuring that the described products are used correctly. These application examples do not relieve you of the responsibility to use safe practices in application, installation, operation and maintenance. When using these Application Examples, you recognize that we cannot be made liable for any damage/claims beyond the liability clause described. We reserve the right to make changes to these Application Examples at any time without prior notice. If there are any deviations between the recommendations provided in these application examples and other Siemens publications e.g. Catalogs the contents of the other documents have priority. We do not accept any liability for the information contained in this document. Any claims against us based on whatever legal reason resulting from the use of the examples, information, programs, engineering and performance data etc., described in this Application Example shall be excluded. Such an exclusion shall not apply in the case of mandatory liability, e.g. under the German Product Liability Act ( Produkthaftungsgesetz ), in case of intent, gross negligence, or injury of life, body or health, guarantee for the quality of a product, fraudulent concealment of a deficiency or breach of a condition which goes to the root of the contract ( wesentliche Vertragspflichten ). The damages for a breach of a substantial contractual obligation are, however, limited to the foreseeable damage, typical for the type of contract, except in the event of intent or gross negligence or injury to life, body or health. The above provisions do not imply a change of the burden of proof to your detriment. Any form of duplication or distribution of these Application Examples or excerpts hereof is prohibited without the expressed consent of Siemens Industry Sector. 4 V1.0, Item-ID:

5 Preface Preface Objective of the Application This document offers a basic guide for servo drive optimization with SINAMICS S120 and SIMOTION D. Main Contents of this Application Note Chapter 1 Practice of measuring the mechanical system Chapter 2-4 Closed Loop optimization of a single axis Chapter 5 Outlook about optimization methods for coordinated axes Chapter 6 Short summary and important know-how Chapter 7 Alternative and optional practices Validity Tested with: SINAMICS FW (2.6.2) SIMOTION V4.2 SCOUT respectively STARTER 4.2 V1.0, Item-ID:

6 Table of contents Table of contents Warranty and liability... 4 Preface Measurement of the mechanical system Speed controlled system frequency response Mechanics frequency response Identification of inertia and torsion stiffness from the frequency response Current Controller optimization Siemens-drives Third-party-drives&motors Current controller setpoint frequency response Speed controller optimization Stability criteria Optimization in frequency domain Proportional gain factor (Kp) Integral time (Tn) Current setpoint filters in conjunction with multi-mass-systems Reference model Automatic controller setting (speed controller) Position controller optimization (SIMOTION) DSC (Dynamic Servo Control) Servo gain factor (Kv) unit Speed precontrol and balancing filter Position controller parameter Servo gain factor (Kv) optimization (frequency domain) Determination of balancing time vtc (time domain) Automatic controller setting (position controller) Coordinated axes Summary Additional Information Current controller setpoint jump Speed controller setpoint jump Filtering locked rotor frequency Glossary Related literature Bibliography Internet link specifications Contact V1.0, Item-ID:

7 1 Measurement of the mechanical system 1 Measurement of the mechanical system to determine the number and location of locked rotor frequencies and resonance frequencies. Note Comment all measured traces (type and parameter) to be able to evaluate the different tested filter effects later. 1.1 Speed controlled system frequency response Set the following settings before starting the measurement: deactivate the position controller gain factor: Kv = 0 precontrol factor: Kpc = 0% break speed controller feedback loop increase integral time Tn by factor 100 proportional gain factor Kp = 1/3 Kp_default The frequency response of the speed controlled system is traced with the "Measuring function" integrated into STARTER/SCOUT. To open the Measuring function press the following button (online-mode): Figure 1-1 toolbar in SIMOTION SCOUT Figure 1-2 measuring function Speed controlled system frequency response Choose SINAMICS_Integrated in the device selection. V1.0, Item-ID:

8 1 Measurement of the mechanical system Measuring function: Amplitude: Offset: Ramp-up time: Measuring periods: Bandwidth: Speed controlled system (excitation after current setpoint filter) ca. 3% - 5% of reference torque Offset in rpm > Amplitude in rpm (Take out the checkmark Values in % ) Offset to avoid reverse rotation duration till offset-speed is reached to start the measuring out of movement. ca. 20 affects the measuring time 4000Hz The required signals are pre-allocated. Figure 1-3 speed controlled system frequency response of a two-mass-system with 4kHz bandwidth The measurement provides only accurate values in the middle range. The lower 50% oft the trace window and the upper 50% of the bandwidth should be neglected (approximate values). In the lower frequency range the sampling is not good enough. In the upper part, amplitude und phase could be displayed in a wrong way due to aliasing effects. To increase the resolution in the lower frequency range, repeat the measurement with reduced bandwidth (ca. 600Hz). Both measurements can be displayed in a superimposed mode. By reduction of the bandwidth you achieve a dispersion of the measuring points in a smaller area. This way the resolution is increased. Reduce the number of measuring periods to keep the previous measuring time. 8 V1.0, Item-ID:

9 1 Measurement of the mechanical system 1-4 speed controlled system frequency response with reduced bandwidth (600Hz) Attention Also in this measurement the middle range is significant only! The locked rotor frequency is located in the significant area. Both frequency responses are congruent. The resolution could be improved in the lower frequency range. The following figure shows a measurement with once more reduced bandwidth: Figure 1-5 speed controlled system frequency response with reduced bandwidth (400Hz) V1.0, Item-ID:

10 1 Measurement of the mechanical system The locked rotor frequency is located on the verge of the significant area. The frequency responses are not congruent anymore. Note Choose the bandwidth twice as big as the frequency you want to have correctly measured/displayed! 1.2 Mechanics frequency response Attention The trace is possible only if encoder on the load side exists! The mechanics frequency response can t be found as a built-in measuring function. But you can create the frequency response manually by using the mathematics function. Trace the motor-encoder s and load-encoder s actual speed [r61] and link them by a transfer function: Transfer = ([r61]_motor ; [r61]_load). For that the measuring function speed controlled system could be extended as follows: Figure 1-6 mathematics function 1) Open the Mathematics function 2) New Formula 3) Bode diagram (transmission function) 4) Choose Signals: motor.r61, load.r61 5) Apply 10 V1.0, Item-ID:

11 1 Measurement of the mechanical system Note Name/label the new measuring functions in the comment field. Note With Firmware V4.3 and higher encoders can be read in with the SINAMICS drive object encoder. The evaluation occurs by SIMOTION with the TO External Encoders. Note Change the direction of rotation of the load motor to Counter-clockwise. Expert-List: p1821 Otherwise the mechanics frequency response phase will not be 0 in the front area. Figure 1-7 mechanics frequency response (red) superimposed speed controlled system frequency response (blue) Attention Pay attention to the scaling! Adapt the scaling of superimposed curves. (right click on the trace window Scaling: As curve x.x ) V1.0, Item-ID:

12 1 Measurement of the mechanical system 1.3 Identification of inertia and torsion stiffness from the frequency response In most cases the motor inertia is known from data sheets. In contrary the load inertia is often unknown. With the speed controlled system frequency response inertia components and torsion stiffness can be calculated. Figure 1-8 identification of inertia and torsion stiffness from the frequency response -20dB/decade JMot Jtotal Formula to calculate the total inertia (J_total): J total f T f R f 1 f 2 db ² f This reflects a pair of values consisting of amplitude [db] and frequency [Hz] in a gradient area of -20 db/decade! Example: The pair of values 21.67dB and f_1 = 50Hz results in the total inertia of the system J_total = kgm². The pair of values 21.91dB and f_2 = 1500Hz results in the motor inertia J_M = kgm². To calculate the torsion stiffness between to masses, the locked rotor frequency, the resonance frequency and the inertia of the masses has to be known. Depending on the known values you can use the following equations: c T 4 ² J M ( f R ² ft ²) c T ( f 1 J R L 2 ) 1 J M 2 c T ( f 2 ) T 2 J L 12 V1.0, Item-ID:

13 2 Current Controller optimization 2 Current Controller optimization Two methods are available: Current controller setpoint frequency response (frequency domain) Current controller setpoint jump (time domain) [Chp. 7.1] 2.1 Siemens-drives Using Siemens-drives&motors the controller settings (proportional gain factor Kp and integral time Tn) are preset automatically. This is done by the autoconfiguration of DRIVE-CLiQ drives or by entering the motor code number. The proportional gain factor depends on the inductance of the motor winding, the integral time on the current controller clock cycle: 62.5us Tn = 1ms 125us Tn = 2ms 250us Tn = 4ms Normally the current controller optimization of Siemens-drives&motors is not necessary! If there are problems with the behavior of the overall control system, start the fault analysis with a check of the current controller setpoint frequency response. 2.2 Third-party-drives&motors Two possibilities to set the controller characteristic parameter of third-partydrives&motors: By entering the motor nominal data, a calculation of the motor model starts to determine proportional gain and integral time. Entering the motor model data manually. Attention Pay attention to all physical units of the motor data! Manufacturer's specification can be different (Units, frequency or angular frequency etc.)! Note Because of this it s recommended to always check the current controller setpoint frequency response when third-party-drives&motors are used. V1.0, Item-ID:

14 2 Current Controller optimization 2.3 Current controller setpoint frequency response Use the default settings of the measuring function to trace the current controller setpoint frequency response: Figure 2-1 measuring function Current controller setpoint frequency response Device selection: Measuring function: Amplitude: Measuring periods: Bandwidth : SINAMICS_Integrated Current controller setpoint frequency response (after current setpoint filter) -default- ca. 20 affects the measuring time 4000 Hz Do not make any changes at the current controller parameters proportional gain factor (Kp) and integral time (Tn). Note Activated current setpoint filter(s) have no effects on the measurement. 14 V1.0, Item-ID:

15 2 Current Controller optimization Figure 2-2 current controller setpoint frequency response (Kp = ; Tn = 2.0ms (defaultvalues)) Note Note The amplitude should stay on the 0dB line till phase tilt With CU firmware V4.4 and higher parts of the current controller are calculated inside the motor module. This improves the amplitude and phase characteristics. Example trace above was recorded with SINAMICS V2.6.2 With this FW the small tolerance to 0dB is explainable. V1.0, Item-ID:

16 2 Current Controller optimization The following trace shows the result of a too low proportional gain factor: Figure 2-3 current controller proportional gain factor undersized (Kp = 15, Tn = 10ms) Note Because of the too low proportional gain factor, the magnitude is well below 0dB. Raise Kp! The integral time results from the current controller sample time. Because of this the integral time shouldn t be adapted manually. Note Always document all your parameter settings to reproduce changes later. 16 V1.0, Item-ID:

17 3 Speed controller optimization 3 Speed controller optimization Premise: optimized current controller Two methods are available: Speed controller setpoint frequency response (frequency domain) Speed controller setpoint jump (time domain) [Chap. 7.2] 3.1 Stability criteria Attention The correct stability hast to be evaluated with the open speed control loop! In SINAMICS trace there is a measuring function for the closed speed control loop. The open speed control loop has to be created manually with the mathematics function. To create the open speed control loop, the speed controller setpoint frequency response has to be extended with the transfer function. Transfer = system deviation [r64] ; actual speed [r61] (For more information about mathematics function see chapter 1.2) The coherence between the closed control loop and the open control loop is characterized by the following formula: F cl Fo 1 Fo If the denominator equal 0 (Fo = -1), the closed control loop is instable. This occurs if the actual value Fo = 1 and the phase amounts (in logarithmic description: 0dB and -180 phase) Reverse: To guarantee stability a adequate phase margin _R is required while amplitude amounts 0dB. If the phase amounts equal -180, a distance to 0dB is required (amplitude margin). Note Amplitude margin and phase margin have to be evaluated in the open control loop response, the resonance magnification in the closed control loop response. V1.0, Item-ID:

18 3 Speed controller optimization Figure 3-1 stability criteria resonance magnification < 3dB amplitude margin characteristic angular frequency f(-3db) phase margin Table 3-1 Resonance magnification: Characteristic angular frequency: Amplitude margin: Phase margin: Critical points: The following criteria should be observed: Amplitude amplification in the lower frequency area. Amplitude amplification max 3dB tolerable. corresponds to an overshoot of 43% (time domain). Related to a short integral time. Frequency of -3dB in the amplitude response. The characteristic angular frequency should be as high as possible to reach a high dynamic and short rise time. Distance to 0dB in the amplitude response. Determines the damping behavior of the control loop (along with phase margin). Distance to -180 in the phase response. Determines the damping behavior of the control loop (along with amplitude margin). If the phase reaches -180 the amplitude margin should be at least 10dB. If the amplitude margin is smaller than 10dB the phase margin should be at least V1.0, Item-ID:

19 3 Speed controller optimization 3.2 Optimization in frequency domain Proportional gain factor (Kp) One typical goal is to set the proportional gain factor high enough to obtain the required machine dynamic. But at the same time the stability of the control loop must be guaranteed. The reachable Kp factor is related to the total system inertia (motor and load) The bigger the total inertia the higher the reachable Kp. Note Optimize the speed controller with coupled load only! By inserting a new drive to a project the Kp factor gets pre-assigned depending on the motor inertia and an additional factor of 1/3: Kp 1000 JM The factor 1/3 is a preventive reserve to avoid mechanical oscillating at first use of the machine. The maximum reachable Kp factor without coupled load can be approximated by the following equation: Kp 1000 JM 1 3 Depending on additional load inertia the ideal Kp factor can be determined by following measurements. Attention For speed controller proportional gain factor optimization the integral component of the PI-controller has to be deactivated. In STARTER/SCOUT you can do this by setting Tn = 0. (The integral time is endless high) Attention It is possible that setting Tn = 0 on other systems effects a real integral time of 0ms! The integral time will be optimized after the ideal proportional gain factor was found. V1.0, Item-ID:

20 3 Speed controller optimization To get a first reference of the speed control loop start a measurement with default Kp setting. For that you the measuring function Speed controller setpoint frequency response (closed control loop) can be used. It is recommended to add the transfer function of the open speed control loop to evaluate the stability criteria! (see chapter 3.1) Figure 3-2 speed controller setpoint frequency response Device selection: Measuring function: Amplitude: Offset: Ramp-up time: Measuring periods: Bandwidth: SINAMICS_Integrated Speed controller setpoint frequency response (after speed setpoint filter) + transfer function of open speed control loop -default- Offset in rpm > Amplitude in rpm (Take out the checkmark Values in % ) Offset to avoid reverse rotation duration till offset-speed is reached to start the measuring out of movement. ca. 20 affects the measuring time 4000Hz Proportional gain: Kp = 1000 x J M x 1/3 Integral time: Tn = 0 Note If the Kp is increased much during optimization the amplitude must be reduced. 20 V1.0, Item-ID:

21 3 Speed controller optimization Figure 3-3 measurement 1: speed control loop (Kp = 0.031) resonance frequency locked rotor frequency (red curve: closed speed control loop, blue curve: open speed control loop) In this trace both curves are congruent, because of the low proportional gain (feedback loop is interrupted) The locked rotor frequency has no effect on the stability of the control loop, but it affects the reachable system dynamic. The excitation with resonance frequency in conjunction with a high proportional gain (peak near 0dB or above) and low phase margin can lead into an unstable condition. The low proportional gain is recognizable by the course of the magnitude right through to the locked rotor frequency. The closed loop magnitude should stay on the 0dB line till the locked rotor frequency. The big amplitude margin above the resonance frequency shows the possibility to raise the proportional gain factor. Rule of thumb for Kp increase: Factor 1.4 (40%) corresponds to 3dB amplitude amplification Note Document all your parameter settings to reproduce changes later. V1.0, Item-ID:

22 3 Speed controller optimization Figure 3-4 conversion logarithmic <-> linear factor 1 Important values: 0 db 1 3 db db db 10-3 db db db db Figure 3-5 measurement 2: speed control loop (Kp = 0.15) The open control loop magnitude reaches almost the 0dB line. At this point there is no phase margin left. The speed controller can t compensate the resonance by itself. In this case the proportional gain factor has to be reduced. Alternatively the resonance spot can be damped by a filter. In this way the Kp factor can be kept. Perhaps another Kp raise is possible. Use the measuring cursor to determine the exact frequency of the resonance spot. 22 V1.0, Item-ID:

23 3 Speed controller optimization Attention It s not necessary to use a filter at all resonance frequencies. See chapter 3.3 Current setpoint filters in conjunction with multi-masssystems There are four current setpoint filters to damp disturbing oscillations/resonances. One of these filters is active by default to mask out the encoder noise. It s configured as a low-pass filter with the characteristic frequency of 2000Hz and a damping of 0.7 (70%). Figure 3-6 current setpoint filters By clicking on the buttons the filters can be parameterized: Figure 3-7 settings of the current setpoint filter(s) V1.0, Item-ID:

24 3 Speed controller optimization To activate a filter check the checkbox filter effective At the beginning the filter type must be chosen. Select between: low-pass filter, band-stop, low-pass filter with reduction and general filter 2nd order. Note While low-pass filters have a better filter efficiency, band stop filters causes less phase tilt. This has advantages relating to the dynamics of the speed loop. For single resonance frequencies a band-stop filter is recommended. If there are several closely spaced resonance frequencies, a single low-pass filter could be better. Further filter-parameters: Notch frequency: Bandwidth: Notch depth: Reduction: determined measuring cursor frequency (pole) 1:1-2:1 notch frequency ca. 50% of magnitude difference between pole and zero Set notch depth as deep as necessary only. Notch depth causes phase tilt! Effects a permanent reduction of the magnitude after the notch. The characteristic of the parameterized filter is shown in the subjacent frequency response. With the button Superimposition all activated filter will be summed up and displayed in one frequency response to have a look at the total filter characteristics. To have a look on the filter results repeat the last measuring (no changes at the controller characteristic parameter) after setting the filter. Figure 3-8 measurement 3: speed control loop (Kp = 0.15; filter active) 24 V1.0, Item-ID:

25 3 Speed controller optimization Because of the filter the resonance frequency is damped and the trace shows an adequate amplitude margin. Now the proportional gain factor Kp can be raised. Note In this example a knowingly deeper notch depth was set. This is to be prepared for an anticipated variance of the resonance frequency in the course of time. Figure 3-9 measurement 4: speed control loop (Kp = 0.7; filter bandwidth: 1:1) With this Kp factor stability limit is reached (amplitude margin too small). The characteristic angular frequency amounts f(-3db) = 50Hz. In this case the speed controller proportional gain has to be reduced to guarantee an adequate amplitude margin (ca. 10dB). It can alternatively be tried to achieve the amplitude margin by adjusting the current setpoint filter (e.g. raise the filter bandwidth and set a filter reduction). V1.0, Item-ID:

26 3 Speed controller optimization Figure 3-10 measurement 5: speed control loop (Kp = 1.0; filter bandwidth: 2:1; reduction:-5 db) Because of expanding the filter bandwidth a Kp raise from 0.7 to 1.0 was possible. Furthermore there is still enough amplitude margin. The characteristic angular frequency amounts f(-3db) 60Hz. The amplitude amplification always amounts less then 3dB. At magnitude -3dB the phase margin is adequate. At a phase of -180 Phase the amplitude margin is adequate. The stability criteria are complied. 26 V1.0, Item-ID:

27 3 Speed controller optimization Integral time (Tn) After the speed controller proportional gain factor Kp is optimized, the integral component of the PI controller can be activated with the integral time Tn. Reducing the integral time affects a shortening of the integral-action time. As a result of this the amplification in the lower frequency domain of the closed speed control loop rises. This can be seen at the amplitude increase of the closed speed control loop (red) in the front area of the frequency response. Attention Optimization with symmetric optimum, the resonance magnification should not be higher then 3dB! Figure 3-11 measurement 6: speed control loop (Kp = 1.0; Tn = 10ms) Because of reducing the integral time the controller bandwidth got higher again (characteristic angular frequency f(-3db) = 82Hz). The stability criteria are still complied. V1.0, Item-ID:

28 3 Speed controller optimization 3.3 Current setpoint filters in conjunction with multi-masssystems Besides the desired damping effects of the current setpoint filters there are also downside effects: Depending on the filter type and the adjusted parameter a phase tilt is the negative result. This impairs the dynamic and the stability of the control loop. Be careful with the number of filters and the choice of the filter parameters. Not every resonance frequency has to be damped actively by a current setpoint filter. Resonances in the space of the controller bandwidth of the speed controller (in this area the controller reacts dynamically) are responded by the controller them self. The controller bandwidth defines the area in which a sufficient (>= 60 ) phase margin is available. The following closed loop speed controller frequency responses shows four resonance spots at which - at first sight - current setpoint filters are required: Figure 3-12 simulated speed control loop (Kp = 4 Nms/rad) First three resonance frequencies are located in the space of the controller bandwidth (noticeable at the phase margin). At the fourth resonance frequency the phase margin equals 0. But the amplitude margin is still enough. 28 V1.0, Item-ID:

29 3 Speed controller optimization After raising the proportional gain factor only the fourth resonance spot gets over 0dB. In this area there is no phase margin left. Figure 3-13 simulated speed control loop (Kp = 15 Nms/rad) Warning: Oscillation! The amplitude amplification above 0dB indicates an instable condition of the control loop. Here a current setpoint filter is recommended. There is no amplitude amplification over 0dB at the rest of the resonance frequencies noticeable. Because of the phase margin these resonances are damped by the speed control loop them self. No current setpoint filters necessary. V1.0, Item-ID:

30 3 Speed controller optimization 3.4 Reference model Note Optional improvement for applications with dynamic setpoint changes. The reference model is located between the speed command value and the speed actual value of the speed controller s integral component (see block diagram). In this way a separation of setpoint controlling and disturbance controlling is reached. The reference model delays the command actual value deviation for the speed controller s integral component. In this way the settling process can be shortened. The integral component will be deactivated if the speed setpoint is reached. Figure 3-14 block diagram: reference model PI-Controller n set M set n act n set n act PI-Controller M set Reference model n act The typical goal of the reference model is to reduce the amplitude amplification (caused of a short integral time) in the lower frequency domain of the closed speed control loop. (The previously adjusted short integral time Tn ensures a good disturbance control.) In time domain the reference model effect a reduction of the actual value overshoot but retains the integral component. 30 V1.0, Item-ID:

31 3 Speed controller optimization To activate the reference model set the select menu Reference model in the speed controller menu to ON. By clicking on the now appeared button Reference model the following parameter can be adjusted: Natural frequency [p1433]: Start value: frequency at which the phase reaches -90 for the first time. Following: Reference model adjustment by evaluating the closed speed control loop. Damping [p1434]: default-value; adjust if necessary Dead time factor [91435]: default-value; adjust if necessary Figure 3- reference model OFF Amplification because of low integral time recognizable Figure 3-15 reference model ON, natural frequency: 80 Hz Startwert: 80Hz bei ca. -90 natural frequency too low: magnitude reduction recognizable V1.0, Item-ID:

32 3 Speed controller optimization Because of raising natural frequency the reference model time constant decreases. The command value reaches the integrator earlier. Figure 3-16 reference model ON, natural frequency: 140 Hz desired settings: magnitude consistently at 0 db. 32 V1.0, Item-ID:

33 3 Speed controller optimization 3.5 Automatic controller setting (speed controller) Besides the manual optimization of the speed controller an automatic control loop optimization is possible. Press the button to start the automatic controller setting. (You can find it next to the measuring function button (see chapter 1.1)) Features of the automatic controller setting: System identification using FFT analysis Automatic setting of current setpoint filters, e.g. for damping resonances Automatic setting of the controller (gain factor Kp, integral time Tn) (Quote from SINAMICS S120 Function Manual) Attention Note The automatic controller stetting sets current setpoint filters with endless notch depth. This way the phase tilt is much bigger compared to a finite notch depth of manual tuning. Therefore it is advisable to adjust the notch depth manually after the automatic optimization! To improve the result of the automatic controller setting fit the notch depth to a proper (endless) value. V1.0, Item-ID:

34 4 Position controller optimization (SIMOTION) 4 Position controller optimization (SIMO- TION) Premise: optimized current controller and speed controller Compared to the PI current- and speed controller the position controller is realized as a P-controller. For hydraulic axis a PID-controller is prevalent. Optimizing servo drives the option speed precontrol is recommended. The most important position controller parameters: Kv position controller proportional gain factor Kpc speed precontrol weighting factor vtc balancing time (velocity Time Constant) Chapter 4.1 and the option: DSC Dynamic Servo Control Figure 4-1 menu closed-loop control in SIMOTION Chapter 4.4 Chapter 4.5 Chapter 5.0 Chapter V1.0, Item-ID:

35 4 Position controller optimization (SIMOTION) 4.1 DSC (Dynamic Servo Control) The function "Dynamic Servo Control" (DSC) is a controller structure calculated in the speed controller sample time. With the DSC, the dynamically active component of the position controller in the drive is executed with 125us. Advantages of DSC: Higher servo gain factor Kv (position controller gain) possible Larger bandwidth -> higher dynamic response Shorter response times for disturbance characteristic Use DSC especially for dynamic applications! For a position axis with position control and an assigned SINAMICS drive, the system sets DCS by default. Deactivation only in offline mode. Note To use the DSC function, the position controller must be set as P controller. You can see in the chart below that the benefit of DSC (higher reachable Kv factor) is even more the higher the minimum locked rotor frequency of the systems is. Figure 4-2 Kv-factor relating to the locked rotor frequency 1666,66 with DSC (0,125 ms position controller clock) reachable Kv-factor [1/s] 166,66 16,66 1 ms position controller clock 4 ms position controller clock 1, minimum locked rotor frequency [Hz] Attention In practice only ca. 2/3 of the Kv factor shown in the chart can be reached! This is because of the damping which isn t considered in this chart. V1.0, Item-ID:

36 4 Position controller optimization (SIMOTION) 4.2 Servo gain factor (Kv) unit Attention There is a difference between the servo gain factor unit of the SIMOTION position controller and the E-Pos (basic positioner) in SINAMICS! This unit differs also in the machine tool control system SINUMERIK. Basically the servo gain factor Kv is defined by the formula: Kv linear speed following error The difference of the Kv factor unit results on different units of above mentioned parameter: 1 Kv( MC ) s SIMOTION: E-Pos / SINUMERIK: Kv The result is a conversion factor of 16.6: Kv ( EPos / MT ) ( MC ) 16, 6 Kv( EPos / MT ) 1000 min There is a relation between the reachable servo gain factor and the least locked rotor frequency of the mechanical system. Kv Kv ( MC ) ( EPos / MT ) ft 10 min ft min 1 s 1000 min (see also: chart, chapter 4.1) 36 V1.0, Item-ID:

37 4 Position controller optimization (SIMOTION) 4.3 Speed precontrol and balancing filter The speed precontrol is used to reduce the following error during a positioning process. In this way the position controller gets more dynamic. The speed command value is added up directly on the position controller output. The speed controller input is faster available, as the known speed setpoint can bypass the position controller. Figure 4-3 block diagram - speed precontrol Note The additional command value is scaled by thy speed precontrol weighting factor Kpc. The factor Kpc should be 100% generally! The remaining function of the position controller using speed precontrol is the disturbance rejection. To activate the speed precontrol set the checkbox Precontrol On in the closedloop control menu. Note Activate the balancing filter when using speed precontrol. In the closed-loop control menu there are three balancing filter modes available: Balancing filter not active [balancefiltermode=off] Balancing filter active (a PT1 filter is used as the balancing filter) [balancefiltermode=mode_1] Extended balancing filter active [balancefiltermode=mode_2] (digital FIR (finite-duration impulse response) filter) V1.0, Item-ID:

38 4 Position controller optimization (SIMOTION) Note The extended balancing filter (balancefiltermode=mode_2) is recommended. The balancing filter is a simplified model of the speed control loop. It is used to prevent the position controller from overriding the manipulated velocity variable during the acceleration and deceleration phases. This is accomplished by delaying the position command value of the position controller by the balancing time with reference to the velocity precontrol. To adjust the balancing filter, go to the register Dynamic controller data. (Activate the checkbox Expert mode in the closed-loop control menu to see the dynamic controller data.) There are three time constants available: ttc Current control loop equivalent time (here not relevant) vtc Speed control loop equivalent time (for balancing filter) ptc Position control loop equivalent time (here not relevant) vtc too short: Position command value delay not sufficient. The position controller compensates the following error in addition to the speed precontrol. The result is an overshoot of the speed command value during acceleration and deceleration. vtc too long: Delay to strong. System deviation between command value and actual value negative (position exceeded). The position controller reacts against the speed precontrol. In the measuring a creeping in into the final position is noticeable! vtc ideal: The system deviation equals 0. Consequently the controller output without precontrol [r60] equals 0 as well. Attention If speed precontrol is deactivated (Kpc = 0%) set vtc = 0ms! 38 V1.0, Item-ID:

39 4 Position controller optimization (SIMOTION) 4.4 Position controller parameter NOTE In this document the position controller optimization is described for SIMOTION Servo gain factor (Kv) optimization (frequency domain) deactivate precontrol: Kpc = 0% balancing time: vtc = 0ms Figure 4-4 position controller setpoint frequency response Device selection: Measuring function: Amplitude: Offset: Measuring periods: D435 Position control setpoint frequency response -default- Offset in /s > Amplitude in /s Offset to avoid reverse rotation ca. 3 affects the measuring time There is a relation between the reachable servo gain factor and the minimum locked rotor frequency of the mechanical system. Use the chart from chapter 4.1 to estimate the servo gain factor in preparation of the optimization. Attention In practice only about 2/3 of the Kv factor shown in the chart can be reached! This is because of the damping which isn t considered in this chart. If the damping is good a higher Kv is possible. V1.0, Item-ID:

40 4 Position controller optimization (SIMOTION) Figure 4-5 position control setpoint frequency response (Kv too low) Figure 4-6 position control setpoint frequency response (Kv ideal) Amplitude response should stay near 0dB until phase tilt. No amplitude amplification tolerable! 40 V1.0, Item-ID:

41 4 Position controller optimization (SIMOTION) Determination of balancing time vtc (time domain) Set the following parameter at the beginning of the optimization: precontrol Kpc = 100% balancing time vtc = 0ms (start value) servo gain factor Kv = optimized value Attention Do not use the measuring functions position control ramp and final controlling element jump for balancing time optimization of oscillatory systems, as speed set point jumps are included in these measuring functions. The optimization could be affected negatively because of the infinite accelerations. Specify a trapezoidal or a smooth velocity profile instead with the control panel. The significant signals have to be traced with the device trace function generator. Figure 4-7 control panel Adjust the velocity profile by clicking on the button Position-controlled traversing of the axis Register Parameter : Setting of the movement speed Entry in /s 1/ min [ / s ] s Register Dynamic response : Choose the velocity profile: trapezoidal with an endless jerk or smooth with the possibility of jerk limiting. V1.0, Item-ID:

42 4 Position controller optimization (SIMOTION) Note The jerk is the derivative of the acceleration and defines the change of acceleration respectively deceleration. Decreasing the jerk the corners of the acceleration profile get rounded off. Attention Pay attention to torque limitation (negative/positive) at the following measurements. The torque limit violation is caused of too high acceleration or jerk values. To look up the actual torque limit go to open-loop/closed-loop control torque limitation. Specify the velocity profile in consideration of torque limitations before starting the measurement for balancing time optimization Note The torque actual value cannot be recorded straight away with the SIMOTION trace. Because of this the balancing time optimization is described for the SIN- AMICS trace in this document. Without consideration of the torque actual value all signals can also be recorded in the SIMOTION trace. Trace the signals torque actual value [r80] and actual speed motor encoder [r61] with the device trace function generator (device: SINAMICS_Integrated): Figure 4-8 device trace function generator: torque actual value Note: Recording: isochronous recording endless trace 42 V1.0, Item-ID:

43 4 Position controller optimization (SIMOTION) To start/stop the recording press the Play/Stop button at the top of the device trace function generator mask. During the recording the specified velocity profile has to be traversed with the control panel. (The function generator is not used!) Approximate the torque limits by adapting the parameters velocity, acceleration and jerk. Note It s recommended to keep distance (ca. 10%) to the adjusted torque limit! Figure 4-9 torque actual value (blue) In the test configuration the torque limit equals 5Nm. A maximum value of 4.5Nm is tolerable. After defining the velocity profile the measurement for balancing time optimization can be performed. Significant parameters are: [r60] Speed setpoint before the setpoint filter [rpm] [r62] Speed setpoint after the filter [rpm] There are two criteria to optimize the balancing time: quick reach of speed setpoint [r62] without overshooting speed setpoint, without precontrol value [r60], close to 0 V1.0, Item-ID:

44 4 Position controller optimization (SIMOTION) Trace the signals [r60] [r62] with the device trace function generator: Figure 4-10 device trace function generator: speed setpoint Note: Recording: isochronous recording endless trace Figure 4-11 measurement 1: vtc = 0ms clearly too short The blue curve shows the position controller output including the precontrol value (signal [r62]). The trace shows a distinct overshoot of the speed command value. Proportional to the speed overshoot there is an overshoot of the position setpoint. However use the speed command value to detect any overshoots because the evaluating by position is more difficult. The red curve shows the position controller output before adding the precontrol value (signal [r60]). If speed precontrol is inactive (Kpc = 0%) the signals [r60] and [r62] are congruent. If speed precontrol is active (Kpc = 100%) signal [r60] should equal to 0. (The position controller serves as disturbance controller only) 44 V1.0, Item-ID:

45 4 Position controller optimization (SIMOTION) If the time constant is not chosen optimally, the position controller "helps" at the beginning of the acceleration and delay phases (additionally to the speed precontrol). Because of this the deviation of the red curve of the value 0 when accelerating and decelerating. With the help of both criteria the short balancing time is noticeable. Figure 4-12 measurement 2: vtc = 1ms too short The too short balancing time can be also recognized with vtc = 1ms (overshoot and deviation). Figure 4-13 measurement 3: vtc = 5ms too long If vtc is too long (extreme example) the command value is reached very slowly (blue curve, zoom). Because of negative system deviation between command value and actual value the position controller reacts against the speed precontrol (red curve). V1.0, Item-ID:

46 4 Position controller optimization (SIMOTION) Figure 4-14 measurement 4: vtc = 2.1ms ideal If the balancing time is ideal the speed command value is reached quickly without overshooting. Because of the jogging mode by using the control panel there is no possibility to superimpose several measurement (different timings of power-on and power-off). In consequence of this the evaluation of the speed command value is possibly difficult despite the zoom (the threshold between quickly and creeping reach of the final position is difficult to identify). Because of this the parameter [r60] is most suitable for fine adjustment of the balancing time! Measurement 4 shows the position controller output is nearly 0. Approximate the ideal balancing time step by step using signal [r60]. 46 V1.0, Item-ID:

47 4 Position controller optimization (SIMOTION) The following trace makes clear, even minimal change of balancing time has consequences: Figure 4-15 measurement 5: vtc = 2.2ms minimal too long It s hard to identify an overshoot of the speed setpoint [r62] (blue) caused by the minimal too big balancing time. The position controller output without precontrol [r60] (red) shows a larger deviation compared to measurement 4 with vtc = 2.1ms. V1.0, Item-ID:

48 4 Position controller optimization (SIMOTION) 4.5 Automatic controller setting (position controller) Just as automatic controller setting of the speed controller, there is the possibility to optimize the position controller automatically. To start the automatic controller setting press the button in the closed-loop control menu of a specific axis. Attention Only the proportional gain factor Kv will be optimized! Speed precontrol and balancing time remain unconsidered. 48 V1.0, Item-ID:

49 5 Coordinated axes 5 Coordinated axes Moving several axes in combination (e.g. circular path or 3D path) the individual axis must offer the same dynamic e.g. to be able to follow the interpolated setpoints of a motion controller within specification. For this the dynamic behavior of the axes must be adapted to each other. This can be realized with the function dynamic compensation. Background: Axes which have to perform coordinated motion often drive different mechatronic systems (mechanical components and motor & drive) in the machine. Therefore the speed controller and position controller optimization of these axes can be quite different. The result of this: One axis is more dynamic as another one. To activate the dynamic compensation use the expert list of a specific axis (Configuration data: DynamicComp ) or set the checkbox Dynamic response filter in the closed-loop control menu. By setting a reaction time you can adjust the axis on each other. To simulate the effects use the circularity test of the measuring function. V1.0, Item-ID:

50 6 Summary 6 Summary 1) Backup the existing configuration before you start making any changes 2) Measurement of the mechanical system Primarily: speed controlled system frequency response Measurement significant in the middle area only! The lower 50% oft the trace window and the upper 50% of the measurement should be neglected (approximate values). Comment all traces! (Meaningful name e.g. trace type and adjusted parameters) 3) Check up current controller setpoint frequency response Watch on consistent scaling of superimposed curves! Current controller parameters for Siemens-drives are preset sufficiently! Amplitude response near 0dB until phase tilt 4) Speed controller setpoint frequency response Evaluate stability and set current setpoint filters based on the open speed control loop! Measuring of the open speed control loop: Mathematics function: Transfer = system deviation [r64] ; actual speed [r61] Deactivate integral component: Tn = 0 (or increase by factor 100) Gain factor start value: Kp 1000 JM 1 3 Increase proportional gain factor (Kp) set current setpoint filter Increase proportional gain factor (Kp) watch on stability criteria! Kp-increase (Rule of thumb): by factors of 1.4 3dB 50 V1.0, Item-ID:

51 6 Summary Reduce integral time (Tn): Effects amplification in the lower frequency domain and a higher controller bandwidth. Closed loop amplitude amplification maximum 3dB! 5) Reference model (optional at a dynamic setpoint change) Intentions: Separation of setpoint controlling and disturbance controlling. Reduces the amplitude amplification caused by a short integral time (The integral component persists). Natural frequency: start value: f(-90 ) Optimization with closed control loop! 6) Position controller Activate Dynamic Servo Control (DSC) For position controller optimization: Speed precontrol factor Kpc = 0% Balancing time vtc = 0ms Note: Different units of position controller proportional gain factor (Kv)! SIMOTION: [Kv] = 1/s Basic positioner: [Kv] = 1000/min Raise Kv step by step Amplitude amplification not bigger then 0dB! Activate speed precontrol: Kpc = 100% Optimization of balancing time vtc: Start value: vtc = 0ms Optimization with velocity profile trapezoidal or smooth Control panel Dynamic Response Balancing time optimization with parameter [r60], [r62] and following criteria: quick reach of speed setpoint [r62] without overshooting speed setpoint without precontrol value [r60] as possible 0 V1.0, Item-ID:

52 7 Additional Information 7 Additional Information Note This chapter is for alternative and optional proceedings. 7.1 Current controller setpoint jump Alternative to the Current controller setpoint frequency response the current controller can be inspected/optimized by a setpoint jump (time domain). Measuring function: Current controller setpoint jump (after current setpoint filter) Figure 7-1 current controller setpoint jump (Kp=15; Tn=10ms) As you can see in the figure, there is a very high rise time. Furthermore the system deviation is too high. raise proportional gain factor Kp Attention Pay attention to the same scaling of the curves! 52 V1.0, Item-ID:

53 7 Additional Information Figure 7-2 current controller setpoint jump (Kp = ; Tn = 10ms) The result of raising Kp is a smaller system deviation and a quicker rise time. For elimination of the persistent system deviation, the integral component of the control loop hast to be activated. reduce integral time Figure 7-3 current controller setpoint jump (Kp = ; Tn = 2ms (default-values)); Measuring time: 10ms No more persistent system deviation. V1.0, Item-ID:

54 7 Additional Information Figure 7-4 measuring time: 100ms If you continue raising the proportional gain factor, the current controller reacts faster. At the same time the overshoot increases. Figure 7-5 current controller setpoint jump (Kp = 43, TN = 2ms); measuring time: 10ms Attention Overshoot not bigger then 10%! Note Current controller setpoint jump also can be used for Kp-adaption with third-party drives. 54 V1.0, Item-ID:

55 7 Additional Information 7.2 Speed controller setpoint jump Alternative to the speed controller frequency response the speed controller can be optimized with a setpoint jump (time domain). The goal is to achieve a short rise time with no or small overshooting. Figure 7-6 speed controller setpoint jump Measuring function: Settling time: Amplitude: Offset: Ramp-up time: Speed controller setpoint jump (after speed setpoint filter) to reduce oscillations caused by ramp-up defines the height of the jump basic-speed (otherwise jump from standstill) ramp-up time to the adjusted offset The following trace shows the result of an undersized proportional gain factor: Figure 7-7 speed controller setpoint jump (Kp = 0.1; Tn = 1000ms) The rise time and the system deviation are too high. V1.0, Item-ID:

56 7 Additional Information The red curve shows the speed setpoint [r62], the blue curve the actual speed [r61]. The green curve represents the torque actual value [r80]. Attention Pay attention to the same scaling of the curves! Tracing the torque actual value is not required mandatorily, however it should be noticed. The parameters r1538 and r1539 in the expert list show the positive and negative torque limit. Do not exceed these values at any time. Note Keep a reserve of ca. 10% to the torque limit values to consider machine run in in the course of time. raise proportional gain factor Kp Figure 7-8 speed controller setpoint jump (Kp = 1.0; Tn = 1000ms) The result of raising Kp is a smaller system deviation and a quicker rise time. For elimination of the persistent system deviation, the integral component of the control loop hast to be activated. reduce integral time 56 V1.0, Item-ID:

57 7 Additional Information Figure 7-9 optimized speed controller (Kp = 1.0; Tn = 10ms) Attention Symmetric optimum: One-time overshoot up to 43% Not in all applications an overshoot of the actual speed is acceptable! V1.0, Item-ID:

58 7 Additional Information In the following trace the speed controller proportional gain factor is too high: Figure 7-10 Kp too high oscillating actual value The oscillating actual speed is an indication of a too high proportional gain factor. The control loop is unstable. reduce Kp! 58 V1.0, Item-ID:

59 7 Additional Information 7.3 Filtering locked rotor frequency Attention The following described methods should only be applied by experienced users! The shown effect can t be guaranteed! Setting a filter on the locked rotor frequency which causes an amplitude- and phase amplification, perhaps the characteristic angular frequency and the therefore the dynamic of the control loop can be improved. Figure 7-11Filter settings To parameterize such a filter choose filter type band-stop first. The notch frequency is the value of the locked rotor frequency. Bandwidth and notch depth corresponding to characteristics of locked rotor frequency. After setting these values change the filter type to General filter 2nd order. The new filter parameters are taken automatically. To complete the parameterization raise the numerator damping till vertical exaggeration of magnitude and phase is noticeable. If required, numerator frequency, denominator frequency and denominator can also be adapted. V1.0, Item-ID: