Power Factor Correction Testing and Maintenance

Power factor correction maintenance assesses how capacitor steps, switching devices, protection, any reactors and the controller work together. A favourable cosφ display is only one observation. A step may be unavailable, the measuring chain may be misconfigured, or harmonic stress may have changed. Meaningful testing relates the actual equipment and operating conditions to documented as-found and as-left results.

Review the single-line and control diagrams, capacitor/reactor ratings, controller model, CT details, step arrangement and previous reports. Record when recurrent fuse operations, capacitive running, alarms or heating occur. Preserve the settings and fault history before changing them. Period reactive-energy registers describe longer-term behaviour; a brief step test answers a different question and cannot establish a whole billing period’s performance.

Current logging, harmonic measurements and thermography need defined operating conditions. Capacitance testing, cleaning, connection work or replacement may need the relevant circuit safely isolated. The work plan must consider all sources and possible backfeeds, with access and methods selected by competent personnel. These activities are not one checklist to perform on an energised cabinet.

Capacitors can retain voltage after disconnection. Isolation, prevention of re-energisation and the required absence-of-voltage and residual-voltage verification follow the actual equipment and work procedure. Waiting a stated number of minutes, seeing an indicator extinguish or reaching a controller’s reclosing timer does not itself establish safe access. A fitted discharge resistor is not proof that it still works.

A conventional current-output CT introduces another hazard: opening its secondary while primary current flows can produce dangerous voltage. A suspected polarity or phase error is not permission to move live secondary wiring. Use the actual measuring scheme and an appropriate safe method.

Inspect for swelling, leakage, enclosure damage, operation of protective disconnectors, terminal deterioration and unsuitable temperatures. Bypassing a protective device is not a repair. A blown fuse prompts investigation of the capacitor, switching device, connections and operating stress rather than automatic replacement followed by a pass.

Record phase voltages and currents with the true step state. Low current may reflect capacitance loss, an open fuse, a contactor or connection problem, or the applied voltage. Imbalance also needs the supply and topology considered. Where capacitance testing is appropriate, follow the equipment-specific method. A reading between two terminals of a connected delta group must not be treated as one branch capacitance without understanding the circuit. Other connected components can affect the reading; a general insulation test must not be applied indiscriminately to connected capacitors or electronics.

For a plain balanced three-phase bank, the fundamental-frequency relation Q[kvar] ≈ √3 × Vline[V] × Iline[A] / 1000 provides a useful check. An idealised 50 kvar bank at 400 V corresponds to about 72.2 A. This is an explanatory example, not a universal acceptance current. Detuned-bank net output, capacitor-element rating, actual voltage, harmonics and the instrument method must be reconciled. Capacitance tolerances also belong to the actual model.

Capacitor switching has an inrush-current duty that must suit the installed device. Relevant inspection may include the coil, mechanism, power and auxiliary contacts, connections and any pre-insertion arrangement. A coil operating or an auxiliary indication does not prove that every power pole conducts correctly.

Pre-insertion resistors limit switching inrush; discharge components reduce stored voltage after disconnection. They are different functions. Minimum reclosing time, permitted switching frequency and replacement criteria come from the matched equipment documentation. Operations per hour cannot be converted into a universal safe-work waiting time. Rapidly forcing every step on and off is not a general test method.

Assess the matched reactor/capacitor ratings, ventilation, thermal protection and any sensors. Noise, smell, discolouration or heating may justify investigation, but do not all identify one fault. Ambient, case and winding temperatures are different quantities and must not share an invented common limit.

Retain the measurement point, voltage and current spectra, loading and connected-bank state alongside THD. Process changes can alter the harmonic environment. A fitted reactor does not prove that all resonance risks have been removed or that a required filtering performance has been achieved. Repeated step failures warrant investigation beyond repeated component replacement.

Compare the real CT current path, orientation, phase-voltage association and secondary rating with the controller scheme. Menu definitions matter: one product may request a ratio, another a primary rating, or separate primary and secondary entries. A negative PF display is not universal proof of reversed wiring; generation and the instrument’s quadrant convention also matter.

An authorised functional check relates the command to the actual current/reactive-power response, allowing for concurrent load changes. Resetting an alarm does not establish that its cause has been corrected or that the system can safely return to service.

Thermography records temperature patterns on safely visible surfaces. Compare like components under comparable loading and known step states. Retain ambient conditions, emissivity/surface information, reflections, viewing angle and access limitations. Shiny metal can reflect another heat source and mislead the temperature interpretation.

An unexpected pattern can direct further connection or component checks. It does not alone prove a torque error, internal fault or overload. An unseen point behind a closed compartment remains unseen. Use applicable equipment limits and the actual conditions rather than universal temperature-difference bands; energised observation needs its own risk assessment.

Acceptance criteria come from the applicable installed-equipment documents, approved design, project requirements and competent technical assessment. A blank internet checklist or another site’s tender is not a universal acceptance standard. The report should state the load range observed and tests not performed. A short good-PF result cannot establish an entire period’s compliance.

There is no single interval or identical test package for every bank. Manufacturer requirements, switching duty, contamination, temperature, previous findings, criticality and applicable obligations inform the plan. Revisit it after faults or operating changes.

To define a Pow-Sys assessment, share the equipment list, observed problem, previous reports and available outage window. Operating measurements and work requiring a shutdown are planned separately. Site timing and report scope are confirmed together.