Rules and Guidelines
Want to participate in the discussions? Please read our Community Rules and Guidelines. Need some help? Visit Support and Answers.

  • State Not Answered
  • Replies 4 replies
  • Subscribers 18 subscribers
  • Views 95 views
  • Users 0 members are here
Options
You may also be interested in

Voltage Dip Propagation from Shared 220 kV Grid into Heavy Industrial 11 kV Plant Network - Mitigation Strategies and Simulation Methodology

Abhishek Prasad

I am currently undertaking a HV system resilience study for a large heavy industrial facility (steel manufacturing including arc furnace, electric smelting, cold rolling, and hot rolling mill drives) drawing power at 220 kV from a grid substation that is shared with the city distribution network. The study has been commissioned following repeated and well-documented process disruptions caused by voltage dips originating from the 220 kV grid, propagating through the plant's step-down transformers (100 MVA, 220/11 kV) into the 11 kV plant network.

I would like to share the problem in detail and invite the forum's experience on both the simulation methodology and practical mitigation strategies, particularly from those who have worked on similar heavy industrial or grid-connected plant studies.

 

The Problem:

The 220 kV supply is taken from a grid substation shared between the industrial facility and the city distribution network. SCADA and power quality monitoring data has confirmed recurring voltage dips at the 220 kV Point of Common Coupling (PCC) with the following characteristics:

  • Dip magnitudes ranging from 10% to 60% of nominal voltage
  • Durations ranging from 10 ms to 400 ms
  • Both symmetrical and asymmetrical events recorded
  • Events correlating with switching operations, fault clearance, and load changes elsewhere on the shared grid

These dips propagate through four 100 MVA, 220/11 kV step-down transformers into the 11 kV plant busbars, where they cause:

  • Undervoltage trips on large induction motors (arc furnace, electric smelting reactor, cold rolling drives, hot rolling mill)
  • VFD undervoltage faults causing process-critical drives to trip
  • Motor contactor drop-out on auxiliary plant
  • Production interruptions with significant financial and process safety consequences
  • Over 32 documented dip events in a single financial year (FY2025-26) captured across multiple plant areas

The plant has no local generation and is entirely dependent on this 220 kV grid connection.

 

Study Approach:

I am building an ETAP Transient Stability model of the system, scoped from the 220 kV PCC down to the 11 kV plant busbars, to characterise the propagation of dips and assess the vulnerability of each load bus. The modelling approach includes:

  • Calibrating fault impedance at the PCC bus to replicate SCADA-recorded dip magnitudes
  • Dynamic motor models for drives confirmed tripping in SCADA/PLC logs (inertia, load torque characteristic, locked rotor current)
  • Lumped static loads for non-critical buses
  • Simulation time extended beyond 1 second to capture post-dip motor recovery behaviour
  • Sensitivity runs across the recorded dip magnitude and duration envelope (10% to 60%, 10 ms to 400 ms)

The key outstanding data item is the grid fault MVA and X/R ratio at the 220 kV PCC, which I am pursuing through the utility. Everything else including motor nameplates, VFD setpoints, connected load lists, and timestamped trip logs has been compiled from plant records.

 

What I Am Seeking From the Forum:

On Simulation Methodology:

  1. For those who have modelled voltage dip propagation in ETAP or PSS/E for similar heavy industrial plants, what modelling approach gave you the most reliable correlation with measured plant behaviour? Specifically around motor dynamic models and VFD representation.
  2. How have you handled the grid equivalent at the PCC when utility fault level data is unavailable or uncertain? Have you used sensitivity studies, worst/best case bounds, or something else?
  3. Is there value in running a frequency scan or harmonic analysis alongside the transient stability study for an arc furnace plant of this nature, or is that a separate workstream entirely?

 

On Mitigation Strategies:

  1. For plants of this scale (100 MVA transformers, arc furnace and rolling mill loads on 11 kV), what mitigation measures have proven most effective in practice? I am currently evaluating the following options:
  • Dynamic Voltage Restorer (DVR) on the 11 kV bus
  • STATCOM or SVC at the 220 kV PCC
  • BESS for critical load ride-through
  • VFD ride-through parameter optimisation (kinetic buffering, undervoltage trip delay)
  • Motor protection relay setting review, specifically extending undervoltage trip time delay to ride through transient dips
  • Busbar sectionalisation to separate sensitive process loads from dip-exposed buses
  • On-load tap changer response optimisation on the 220/11 kV transformers
  1. Has anyone implemented a DVR or STATCOM on a comparable heavy industrial plant? What were the practical challenges around sizing, response time, and integration with existing protection?
  2. For VFD ride-through optimisation, what is the realistic ride-through capability of modern VFDs during dips of 40% to 60% magnitude, and is kinetic buffering a viable strategy for high-inertia rolling mill drives?
  3. Where the utility grid is the source of dips and the plant has no control over the 220 kV network, what contractual or regulatory mechanisms have others used to engage the utility on Power Quality obligations, and has this led to any grid-side improvements?

 

Broader Question:

This appears to be a systemic issue in facilities that draw power from shared utility substations serving mixed industrial and distribution loads. The plant has no means to control or predict the grid-side events yet bears the full operational and financial consequence.

Is there published IET, CIGRE, or IEEE guidance specifically addressing voltage dip mitigation strategy selection for large HV-connected industrial plants? I am aware of IEC 61000-4-11 and IEC 61000-4-34 for equipment immunity testing but am looking for something more at the system study and mitigation design level.

I would be grateful for any experience, references, or critique of the approach described above.

 

Abhishek Prasad, MIET
Electrical Engineer, Power Systems

 

Parents
  • 0

    EN 50160 might be worth a look - it may not be intended to apply directly to the 220kV side of things but something similar should and the numbers ought to scale.
    (BS EN 50160:2022 outlines the voltage characteristics of electricity supplied by public networks).
    To ride through a drop out, you need to store rather more energy than would 'fill in the gaps' either as L_C storage or as a spinning flywheel., and the numbers you mention are very high for that sort of trick. 
    I;d be looking at delays on the low voltage trips if that is at all possible, and maybe changing some alumnium for steel in the rotating parts to give more inertial storage if that is appropriate. 

    In terms of mitigating the shocks from the arcs themselves, this is a fun read of some Spanish research. In summary, arc furnaces are not good partners to share a supply with but things can be done to help a bit.  

    Mike.

  • 0 in reply to mapj1

    Mike - thanks, both very useful.

    On EN 50160: agreed it won't transfer numerically to 220 kV, but the classification framework (residual voltage × duration bands) is exactly what I needed as a template - I'll bucket our 32 recorded events the same way and use that structure when I take this to BBMB. The standard doesn't assign expected dip frequency or operator responsibility, so it won't hand me a contractual limit outright, but it gives me a clean, recognized way to present the data rather than a raw list of numbers.

    On storage sizing - good catch, and worth clarifying where I'm actually headed: I'm leaning STATCOM at the 220 kV PCC rather than DVR/BESS, so it's a shunt reactive-current injection rather than a real-power/energy device. That sidesteps the "need to store more than fills the gap" issue directly, since it's not storing energy to bridge the event at all.

    But your point still bites in a different form. STATCOM voltage recovery is capped by injected current relative to how weak the grid has gone during the event, not by energy - and our recorded envelope goes down to 10% residual voltage in places, which is a genuinely deep event. A shunt device sized off the average dip could under-perform on the worst 10-20% residual cases specifically. So I'm going to run our full recorded event set through the ETAP model and check STATCOM MVAr sizing against the deepest events individually, not just a typical case, before I commit to a rating.

    On the arc furnace paper - useful confirmation. The results back up what I'd assumed: SVC there did essentially nothing for voltage/current THD (current THD actually roughly doubled with the SVC in, from the TCR's own harmonics) but cut long-term flicker substantially. That lines up with keeping SVC in our study as a furnace-side flicker/PF tool only, separate from the PCC dip-mitigation question.

    Will also look at UV trip delay and the inertia point for the rolling mill drives - cheap to explore relative to any power-electronics option.

    Thanks again - this is exactly the kind of sanity-check I was hoping for.

    Abhishek

  • 0 in reply to Abhishek Prasad

    My understanding of the 'STATCOM' (Static Synchronous Compensator) unlike an SVC (Static Var Compensator )approach is the electronics looks like most of the invertetr guts of a UPS without the battery bits, but with some large capacitors storing the DC voltage (labelled DC-link in the drawing below, shamelessly nicked from google.)

    These DC capacitors still need to be sized to store rather more than the maximum energy you wish to in-fill if the drop out is multi-cycle and affects all phases. During a drop out that affects only one phase-phase voltage,  or that lasts only for one part of a cycle, there is some chance to refill that reservoir from the other  bits of supply waveform that are not missing, so it can be rather smaller. 


    Mike.

  • 0 in reply to mapj1

    Yes it makes sense, thanks. So my ETAP transient stability run gives me the MVAr rating and response time needed at the PCC from the recorded dip events, but the DC-link capacitor design itself sits with the vendor. I'll pass on our worst-case symmetric, multi-cycle dip data specifically in the RFQ so they're not sizing the DC-link off an average case.

    Abhishek

Reply
  • 0 in reply to mapj1

    Yes it makes sense, thanks. So my ETAP transient stability run gives me the MVAr rating and response time needed at the PCC from the recorded dip events, but the DC-link capacitor design itself sits with the vendor. I'll pass on our worst-case symmetric, multi-cycle dip data specifically in the RFQ so they're not sizing the DC-link off an average case.

    Abhishek

Children
No Data

Community Rules & Guidelines