Steel Mill in the Neighborhood: Part 2 (In Your Backyard)


The steel industry is just giddy on new and upgrades to steel mills.  Google the subject and you’ll find — the Minnesota Steel & Iron project in Itasca County, MN, the upgrade of the Pacific Steel Casting mills in Berkeley, CA, and Brazilian steel company CSN’s plans to build a steel-rolling mill in Kentucky, among others, in recent news.  For whatever the reason, mill developers see an increased demand for their product providing impetus for increased capacity.  A key economic factor for additional milling capacity is the availability of steady, low-cost power supplies for their mills.

This bodes well for the power industry in at least one respect — mills represent a high load factor customer, not a lot of MW installed required for the annual energy consumption.  However, they do pose new challenges to maintaining reliability and power quality.


Electric aspects of mill technology have advanced significantly in a number of key areas, including:  improvements rolling processes, faster reactive controls on static var compensators [4], being able to interconnect to weak grid nodes using DC arcing [6].  Economies of scale applied to mill design have resulted in larger mill sizes.  Mills have become quite good citizens on the power grid, cooperating with operators, controlling, if not reducing, their power quality impact and operating efficiently and reliably.

Technical due diligence requires that we address the technical concerns.  The failures at Comanche station (power spikes from a neighboring mill) [8] and at Port Washington (torsional interactions with a steel plant) [7] were not too long ago to be forgotten.  New torsional relays offer protection against recurrence.  But how reliable and effective will these devices be?  Also, we have yet to see standard generator protection schemes cover flicker and harmonics.

To clarify these issues seemed like a good idea for a second article on this subject.  For the first Techblog on steel mills, please click on this link.

More Quality Issues

Modern steel mills comprise of two arc furnaces of 100 MW or more each, a rolling mill of about 175 MW, a static var compensator (SVC) and harmonic filters [1].  The demand MWs vary in accordance with perhaps up to 5 different load cycles that can lead to an average, cyclic load impact of 150 MW and extreme load impacts of 200 MW.  The cyclic load introduces flicker and each impact load generates harmonic currents.  In addition, the SVC controls introduce both flicker and harmonics, and capacitors modify frequency response.

To control flicker, mills are typically located where the fault levels are high (or where the system is relatively stiff).  This means locating close to a power plant or to a strong nexus of high voltage transmission lines.  Generators, due to their low impedance to ground, act as a harmonic sponge, absorbing harmonic currents from the steel mill plant.  So having them close to a mill is beneficial, since they would generally reduce total harmonic distortion in the system.  However, an undesired effect of this proximity is the heating that the harmonic currents cause in generators.  An even more ominous impact of proximity is the potential excitation of turbine shaft resonant modes that can lead to torsional stress and shaft failure.

Recently, solutions have been developed to allow interconnection even at weak nodes of the grid.  If effective, this may lead to siting of mills away from generators, leaving the power quality mitigation on the hands of SVCs (with fast controls) and well-designed harmonic filters, and avoiding the potentially hazardous side effects on the generators.

Case Study

In January 2004, the PSCo Comanche 2 station experienced a catastrophic failure.  The initial indication was that the bindings at the end of the generator rotor that support the end turns of the copper windings broke free and hit the stator.  This type of failure is normally associated to corrosion.  However, in this particular case, there was steel mill close by …

The steel mill included two DC arc furnaces for which voltage flicker control was implemented through a static var compensator (SVC).   The SVC produced a variety of exciting frequencies from the combined which varied as a function of the changing short circuit ratio (SCR) at the point of common coupling with the network.  The SCR changes as the external load changes.  The SVC operation itself varied as the steel mill’s load cycle, resulting in a cyclically varying set of excitation frequencies.

In the prior prior to the failure, one of the exciting frequencies coincided with a resonant frequency in the Comanche 2’s shaft.  Somehow, this did not cause the same issue with Unit 1, and the conjecture is that modifications made to Unit 2, also changed its resonant frequencies.  The repeated excitation in Unit 2 caused torsional stress that eventually led to the failure.

Mitigation measures applied to Unit 1 included modification of the SVC controls to change the exciting currents, modification to the shaft to eliminate resonance close to frequencies developed by the steel mill, and the continuous monitoring of shaft vibrations.

Treatment options

Treating some of the issues one at a time then:

  1. Flicker.  Comes from the mill operations and SVC [2].  Tend to be worst at light load, when fault levels are lowest.  Standards such as IEEE 519 provide basis for acceptance criteria.  Impacts primarily customers with lamps, noting that the tolerance level from incandescent and fluorescent lamps may differ.  Is mitigated by increasing the SVC range and/or locating close to a large power plant, or using DC arcing.  May also be mitigated operationally by reducing power level of the mill during certain system conditions or shutting down the mill altogether.

  2. Harmonics.  Comes from the mill operations, SVC and capacitor banks.  Analyzed with a frequency scan at various operating configurations [5]. Typical levels by harmonic number for various components of the mill may be applied to estimate the currents introduced into the grid.  When the mill is already in service, field measurements would refine the harmonic contributions.   Tend to be worst at peak load when the harmonic modulation is applied to higher current amplitudes, and when the harmonic is close to a resonant frequency.  Impacts primarily equipment that are sensitive to voltage spikes — rotating equipment such as generators.  Can be mitigated by harmonic filters or design/operation of capacitors and SVC to avoid resonance at critical harmonic frequencies.

  3. Torsional interaction.  Only an issue when there is a generator nearby whose shaft may have resonant frequencies that can be excited by a mill’s mix of harmonic and non-fundamental currents.  The effect is similar to what may be found when series capacitors excite subsynchronous resonance.  However, mill operations have a cyclical characteristic that may accumulate a relative small level of excitation to a stress failure.  Mitigated by designing and operating the system carefully to avoid matching a shaft resonant frequency.  Torsional relays may provide protection.  These are relatively new.

The whole ball of wax …

Note that solutions to resolve flicker lead to other issues — locating close to a generator for the higher short circuit ratio exposes that generator to harmonic heating and torsional stresses, while increasing the SVC capacity to introduce stiffness can lead to increases in harmonic currents and torsional interactions.  Solutions that may work for the initial operations of the mill may no longer do so as the surrounding system evolves over time.  This is especially a concern for resonance and torsional stresses.

The full solution tends to be an intricately planned, designed and operated marriage of mill and power system that addresses the whole range of quality and reliability issues.  The careful engineering is required because some of the mitigation have side effects that require additional forms of mitigation.


  1. The impact of large steel mill loads on power generating units, Solanics, P. Kozminski, K. Bajpai, M. Esztergalyos, J. Fennell, E. Gardell, J. Mozina, C. Patel, S. Pierce, A. Skendzic, V. Waudby, P. Williams, J., Power Delivery, IEEE Transactions, Jan 2000, Volume: 15, Issue: 1, pages 24-30.
  2. An EMTP study of flicker generation and transmission in power systems due to the operation of an AC electric arc furnace, Ramos, B.N.; Parga, J.L.deC.; Harmonics and Quality of Power, 2000. Proceedings. Ninth International Conference on Volume 3, 1-4 Oct. 2000.
  3. Updating a steel mill power system to improve reliability and decrease energy costs, Albaugh, J.R.; Kornblit, M.J.; Industry Applications, IEEE Transactions on, Volume 30, Issue 5, Sept.-Oct. 1994.
  4. Static Var compensator upgrade in a steel mill, Depommier, B.; Stanley, J.; Power Engineering Society General Meeting, 2003, IEEE, Volume 1, 13-17 July 2003.
  5. Harmonic and transient overvoltage analyses in arc furnace power systems, Mendis, S.R.; Gonzales, D.A.; Industrial and Commercial Power Systems Technical Conference, 1990, Conference Record. Papers Presented at the 1990 Annual Meeting, 30 April-3 May 1990.
  6. Comparison of some active devices for the compensation of DC arc furnaces, Carpinelli, G.; Russo, A.; Power Tech Conference Proceedings, 2003 IEEE Bologna, Volume 2, 23-26 June 2003.
  7. Retaining Ring Cracking at Wisconsin Electric Power Company’s Port Washington Unit 1 — Root Cause Analysis, EPRI Report, June 2002.
  8. Retaining Ring Cracking at Comanche Unit 2 — Root Cause Analysis, EPRI Report, July 1996.

For questions, comments and further discussion, contact us at