Electro-mechanical interaction:
Introduction

Electric variable frequency drives provide lot of benefits in terms of energy savings, controllability and flexibility. But the VFD is just means to control the speed or torque of electric motors. The useful work is provided by the driven load such as e.g. pump, fan, compressor or mill. Therefore, the seamless integration of the electric and mechanical (sub)systems is very important. Vibration impacts the quality of the output product and the lifetime of the mechanical components. Excessive torsional stress might be a root cause of fatigue. Many critical applications are equipped with instrumentation for supervision. The protection system would trip the drive if vibration exceeds specific threshold. This post and the whole series will discuss various topics related electro-mechanical interaction, its mitigation and and reliable operation.

Key point in this context is the electro-mechanical interaction. When not considered properly, it can lead to :

– operational restrictions (restricted speed ranges)

– increased vibrations and mechanical wear

– reduced maintenance intervals

– increased audible noise

– failure of a shaft component (worst case)

– consequent downtime and loss of production

Electro-mechanical interaction
Figure 1: Electro-mechanical interactions in variable frequency drive systems

As obvious from the title, electro-mechanical interaction is an interdisciplinary subject. Deep understanding requires knowledge from electrical (power electronics, drives and motors), mechanical (rotordynamics) and control engineering. The challenge is nicely formulated in following quote:

“Electrical engineers see electrical problems. Mechanical engineers see mechanical. But many problems lie in the boundary between. An electromechanical perspective reveals the entire picture: the electrical, the mechanical and the complex interactions between the two.”
Source:
Lloyd’s Register Consulting

Essential part is the understanding of system damping and the components the damping consists of. The damping determines how large the torque response will be and how quickly it decays once the excitation is removed. Damping of a closed-loop system (with VFD) can be very different to a damping of fix speed system (fix speed, direct on-line motor). When a resonant system is excited with force or torque of the resonance frequency, an oscillation appears. Once the excitation is removed, the system oscillations will decay. Poor damped system (damping D << 1) will take a long time until oscillation disappears (low damping = low losses in the system). Turbomachinery (steam and gas turbines, compressors, high-speed pumps etc) inherently have low damping. These systems are therefore sensitive to torsional vibration.

Figure 2: Supercritical (top), critical (middle) and subcritical (bottom) damping

In rotordynamics we distinguish two main types of vibration: Lateral and Torsional.

Lateral vibration

– Also called radial or transverse vibration

– Complex models including foundations etc.

– Measurable with state of the art instrumentation

– Adjustable alarm and trip limits

Torsional vibration

– “Torsional twist”

– Concern for multi-mass (elastic) machine trains

– Relatively simple models

– Limited observability

– Usually not instrumented → may not be recognized

– Dangerous (problems can grow until failure)

Other types of vibration are axial and disk/blade vibration.

“Lateral vibration is easy to measure and difficult to predict. Torsional vibration is difficult to measure and easy to predict.”
Malcolm Leader
President - Applied Machinery Dynamics
Shaft string_ONGC
Figure 3a: Variable frequency drive system of a compressor with speed-increasing gear
Figure 3b: Shaft string of gas turbine and electric motor drive for multi-stage compressors (gearless)

Torsional analysis is often mandatory. The objectives of torsional analysis are:

– calculate torsional natural frequencies and corresponding mode shapes

– verify the electro-mechanical integrity

– identify critical speeds, torsional response and potential issues

– dimension shaft components such as couplings

The analysis is applied since several decades and the problematic is well-known for fix speed motors. However, in variable frequency drive systems additional effects come to play. These effects might be less known or less obvious, especially for mechanically oriented professionals. When such effects are neglected, the torsional analysis may not predict correctly the torsional behavior of the system.

electro-mechanical interaction in VSDS
Figure 4: Simplified block diagram of VFD speed controlled motor with elastic mechanics

In this post series following points will be covered:

Origin of electro-mechanical interaction in systems with variable frequency drives

Steady state torsional excitation due to:

– Torque ripple (pulsating torque) and its spectrum

This mechanism is easy to imagine. Every VFD (incl. topologies with output sine filter) has certain torque ripple when observed in time domain. Analyzing the frequency spectrum of the torque several characteristic and non-characteristic components can be found. Most of these frequency components have certain relation to the fundamental output frequency (in figure 4 we can see characteristic components of 6x, 12x and 18x the fundamental electric frequency). In other words the frequency spectrum of torque changes with motor speed. At specific motor speed one of the frequency components in torque would match one of the torsional natural frequencies of the shaft string. We talk about resonances and critical speeds. The system damping is crucial and influences how damped or undamped the system is and what the torque response will be.

air gap torque ripple - electro-mechanical interaction
Figure 5a: Motor air gap torque of 6.5 MW / 1'500 rpm induction motor (time domain)
Figure 5b: Motor air gap torque of 6.5 MW / 1'500 rpm induction motor (frequency domain)

– Closed-loop interaction (feedback controlled systems with latencies)

Closed loop interactions are usually less obvious than torque ripple. The mechanism depends on several factors, such as controller settings (proportional gains and integral time constants), time delays, available damping functions etc. Improper control design and software parameterization lead to torsional amplification inside the feedback control loop. This mechanism is mainly linked to speed control. It is much less pronounced in torque control loop. 

Figure 6: Closed loop effects (torsional amplification) in speed controlled VFD system

Transient torsional excitation due to:

There are several sources of transient excitation. Besides external disturbances (load disturbance, process control etc) the most severe transient excitation comes from:

– Transient fault torques (short circuit torques)

– Direct on-line starting (for VFDs with manual bypass)

Figure 7: Transient short circuit torques of 5.5 MW induction motor (36 Hz is min. operating frequency, 63 Hz maximum continuous frequency)

Mitigation of  torsional oscillations and vibration

The negative effects of electro-mechanical interaction can be minimized by performing below steps in design, engineering and commissioning phase.

– Design considerations

– Best practice / experience

– Software parameterization

– Torsional stability

– Advanced software functions

– Virtual commissioning

Functions for enhanced damping

Increase of natural mechanical damping is seldom possible since higher damping would mean higher losses, lower efficiency and increased material cost. Moreover, it would not be practical, either. Consequently, the aim is to increase the system damping by advanced control algorithms. When properly done, it is the most effective method. We distinguish two groups of damping functions based on their principle:

– Passive damping functions

– Active damping functions

Figure 8: Torque on the electric motor (EM) coupling when crossing through torsional resonance. Top - without damping, Bottom - with damping algorithm

Diagnostics and verification

The VFD includes a powerful signal processor computing the quantities in real time (sampling time of 25 µs or similar). Lot of signals can be determined using appropriate observers. Note that e.g. every DTC based VFD contains a sophisticated mathematical motor model as key part of its control. This model calculates motor actual torque with fairly high precision. Smart processing of motor torque signal reveals lot of information for diagnostic purpose. This comes almost “for free” – no special torque sensor is required.

Few ideas how diagnostics can be used:

– Torque measurement and/or estimation

– Site testing and diagnostics

– Verification of parameters from design stage

– Improvements / adjustment of parameters (if necessary)

Figure 9: Waterfall diagram of motor speed (postprocessed signal from motor model) - visible first tosional natural frequency @ 17.5 Hz

Do you want to learn more about electro-mechanical interaction, torque oscillations and torsional vibration? Then just follow this series. We will explain the topic step by step in dedicated posts.

References for electro-mechanical interaction

See also post dedicated to literature on VFD systems: click HERE to access.

[1] API 684 Rotordynamic Tutorial, Lateral critical speeds, unbalance response, stability, train torsionals and rotor balancing, second ed., Washington: API Publishing Services, 2005.  

[2] T. Holopainen, J. Niiranen, P. Jörg and D. Andreo, “Electric motors and drives in torsional vibration analysis and design,” in Proceedings of the Forty-Second Turbomachinery Symposium, Houston, Texas, 2013.  

[3] P. Rotondo, D. Andreo, S. Falomi, P. Jörg, A. Lenzi, T. Hattenbach, D. Fioravanti and S. De Franciscis, “Combined torsional and electromechanical analysis of an LNG compression train with variable speed drive system,” in Proceedings of the thirty-eighth Turbomachinery Symposium, 2009. 

[4] S. Del Puglia, S. De Franciscis, S. Van de moortel, P. Jörg, T. Hattenbach, D. Sgro, L. Antonelli and S. Falomi, “Torsional interaction optimization in a LNG train with a load commutated inverter,” in 8th IFToMM International Conference on Rotordynamics, Seoul, Korea, 2010.

[5] M. Tallfors, Parameter estimation and model based control design of drive train systems, Stockholm: Kungliga Tekniska Högskolan (KTH), 2005. 

[6] K. Tanaka, A. Adachi, N. Takahashi and Y. Fukushima, “Torsional-lateral coupled vibration of centrifugal compressor system at interharmonic frequencies related to control loop frequencies in voltage source inverter,” in 38th Turbomachinery Symposium, Houston, TX, 2009. 

[7] S. E. Saarakkala and M. Hinkkanen, “State-space speed control of two-mass mechanical systems: Analytical tuning and experimental evaluation,” IEEE Trans. on Industry Applications, vol. 50, no. 5, pp. 3428-3437, September/October 2014. 

[8] V. Hütten, C. Beer, T. Krause, V. A. Ganesan and S. Demmig, “VSDS motor inverter design concept for compressor trains avoiding interharmonics in operating speed range and verification,” in Proceedings of the Forty-Second Turbomachinery Symposium, Houston, Texas, 2013.

[9] K. Szabat and T. Orlowska-Kowalska, “Vibration suppression in a two-mass drive system using PI speed controller and additional feedbacks – Comparative study,” IEEE Trans. on Industrial Electronics, vol. 54, no. 2, pp. 1193 – 1206, April 2007. 

[10] F. C. Nelson, “Rotor dynamics without equations,” vol. 10, no. 3, pp. 2 – 10, 2007.

[11] A. Muszynska, Rotordynamics, Boca Raton, Florida: CRC Press, Taylor & Francis group, 2005.

[12] A. Tabesh and R. Iravani, “Frequency response analysis of torsional dynamics,” IEEE Trans. on Power System, vol. 19, no. 3, pp. 1430 – 1437, 2004. 

[13] A. Tabesh and R. Iravani, “On the application of the complex torque coefficients method to the analysis of torsional dynamics,” IEEE Trans. on Energy Conversion, vol. 20, no. 2, pp. 268 – 275, June 2005.

[14] Advanced drive services, https://new.abb.com/drives/services/advanced-services