Case Study: Voltage Transients Cost Thousands
The conveyor system at an electronic component supplier, responsible for transporting product around the distribution centre, is entirely reliant on communications between various modules to synchronise the stock movement.
Unfortunately, a number of incidents of failures of the communications systems resulted in the necessity to restart the entire conveyor system; a lengthy process resulting in lost revenue in the order of £50,000 per hour.
It was, not unreasonably, felt that the likely origin of the communications failure was deficiencies in the quality of the power supplied to the sensitive electronics associated with the communications. In order to determine whether or not this was case, a week-long power quality survey was commissioned, the results of which are presented below.
Results of analysis
During the analysis period, a number of voltage transient events were recorded. In addition to the power quality analysis, the site agreed to log the time of any communication failures during the analysis period.
All of the major communication failures had associated with them a voltage transient such as that shown in Figure 1. In addition to the voltage waveforms, the phase current waveforms were captured as these can offer further clues as to the origin of the transient behaviour.
The coincidence of the voltage transients with the communication failures confirmed the suspicion that the origin of the problem lay with the quality of the power (voltage). The next problem was to identify the source of the transient and by doing so, implement an effective solution to minimise / eradicate further instances of the same problem.
From Figure 1, it can be seen that the nature of the voltage and current transient was an oscillation superimposed over the top of the normal mains waveform. The frequency of the oscillation was relatively low (compared to normal transient timescales) and was reminiscent of a capacitor switching transient.
Fortunately, as mentioned above, the current waveforms were captured as well as the voltage waveforms. The post processing software associated with the analyser used in this project allows the user to create plots of power (both active and reactive) on a cycle by cycle basis.
Figure 2 shows the half cycle profile of the phase reactive powers as captured during the transient event shown above. Shown on the plots are the changes in phase reactive power during the transient event. This showed that the sum of the changes in the reactive power across the three phases was approximately 46 kVAr. Upon further investigation, it was found that this site had power factor correction installed which was arranged in a number of automatically controlled 50 kVAr steps.
Figure 1 – Voltage and current waveforms captured during transient event
Figure 2 – Change in reactive power consumption during transient event
Having inspected the power factor correction, it was discovered that each 50 kVAr capacitor step was switched using ‘block contactors’. Block contactors are generally applied to resistive or inductive loads and whilst, if appropriately rated, they are capable of switching capacitive loads, they are far from ideally suited for such an application.
Contactors are available that are specifically designed for switching capacitors. These contactors incorporate pre-connect resistors which switch in series with the capacitor very briefly in order to limit the capacitor inrush current and the consequential voltage transient.
In order to prevent future occurrences of capacitor switching based transient issues, the following was recommended: –
- The capacitor steps were to be broken down into 25 kVAr steps to help limit the magnitude of the capacitor inrush current
- Each capacitor switch was to be switched with a ‘soft switching’ contactor, specifically designed for capacitor switching applications
These recommendations were accepted by the client and implemented, the result of which was a profound reduction in communication failures, with the consequential reduction in down time.
The cost to implement the measures detailed above was approximately £20,000, which when one considers the lost time cost of £50,000 per hour, rendered the expenditure almost insignificant.