Electrostatic precipitator upgrade: measure first, then fix the real limit

Step 1 — Controls / automatic voltage control (AVC), when spark-limited. If the V-I map shows fields sparking heavily with voltage sagging well below the sparkover ceiling, the limit is control response, not hardware. A modern AVC with fast spark quench, controlled ramp recovery, back-corona detection and intermittent energisation raises the useful corona power the existing TR set can deliver. Indicatively worth a 10–25 % reduction in outlet dust, executed in days on a normal shutdown.

Step 2 — Transformer-rectifier or high-frequency power supply, when voltage-limited. A conventional 50 Hz single-phase TR delivers roughly 35–45 % voltage ripple, so the mean field voltage sits well below the peak that triggers sparkover. A high-frequency switch-mode supply cuts ripple below 1 %, holding the field 2–5 kV closer to sparkover continuously — and since migration velocity rises roughly with the square of field strength, this is indicatively worth 15–30 % lower outlet dust on fields that were genuinely voltage-limited. See electrostatic precipitator design for how field voltage enters the sizing equations.

Step 3 — Internals renewal, when mechanically degraded. Broken or swinging discharge electrodes, plates out of alignment beyond roughly ±5 mm on a 300–400 mm duct (indicative), seized rapper drives and eroded gas distribution screens all show up as electrical instability with a mechanical cause. The remedy is targeted: re-align, replace failed electrodes, rebuild rapping, restore distribution screens — not blanket plate replacement. Renewal returns the field to its design curve; it does not exceed it.

Step 4 — More collecting area, when fundamentally undersized. If the unit is electrically healthy, mechanically true, and still over the limit, only area helps: an added field (casing extension) or an upstream Electrocyclone pre-collector removing 85–95 % of the inlet load before the ESP, in 60–65 % of the plot an equivalent ESP stage would need and at 120–140 kW absorbed power. The design basis below shows the scale of what staged collection buys.

Step 5 — Full retrofit. When casing corrosion, foundation constraints or a step-change in the emission limit make incremental work uneconomic, a new precipitator is built to current duty — usually around the existing foundations, with tie-ins prefabricated to compress the outage.

Upgrade sizing uses the same physics as new design: Deutsch-Anderson (η = 1 − exp(−w·A/Q)) to frame the area–efficiency trade, corrected by Matts-Öhnfeldt with k ≈ 0.5 because real polydisperse ash (design-basis inlet d₅₀ 21 µm) never behaves like a single migration velocity. Electrical work follows IEC practice for HV equipment; mechanical tolerances follow Arrow fabrication standards with alignment records per field.

Performance is proven on one basis only. Before the outage, an isokinetic traverse to EPA Method 5 or ISO 9096 establishes the baseline in mg/Nm³, corrected to the stated reference O₂ — 6 % O₂ dry for biomass — dry basis, within a defined boiler load window (typically 90–100 % MCR) and with soot-blowing state recorded. After the work, the identical protocol runs at the same ports and load window. Numbers quoted on different O₂ corrections or different loads are not comparable, and we do not sign test reports that mix bases. Guaranteed figures are stated per project after the technical assessment, on that stated basis.

How long does an electrostatic precipitator upgrade take?

Indicatively: the diagnostic survey takes 3–5 days on load with no shutdown. A controls or AVC upgrade fits 3–7 outage days; TR or high-frequency power supply replacement 5–10 days; internals renewal 2–4 weeks depending on fields opened; a casing extension or Electrocyclone pre-collector is erected largely pre-outage, with tie-in windows of 2–4 weeks.

The schedule discipline is pre-assembly: for area additions, structural steel, casing modules and the pre-collector are erected beside the running unit, so the outage consumes only duct tie-ins, electrical connection and commissioning. Long-lead items — TR sets, electrodes, insulators — are secured against the outage date, and critical consumables are held as ESP and boiler spare parts stock so a failed insulator found on opening does not extend the shutdown.

Why not just replace the collecting plates?

Because plates are rarely the limit. If rapping is dead, new plates foul in weeks; if gas distribution is skewed, part of any plate area — old or new — sees almost no gas; if the field is voltage-starved, clean plates collect little. Replacing plates without a field map risks spending a full outage for single-digit improvement. Measure first; replace metal only when metal is the problem.