Just because nonlinear loads are present doesn't automatically mean they are the source of system problems.
Harmonics are power-related. That is, they will repeat, cycle by cycle, even as their waveshapes are integer multiples of the fundamental (60 Hz). While harmonic currents and voltages may vary with differing loads (or speed cycles in the case of variable speed drives), their presence results in the consuming of space on the power system or the creating of voltage distortion across the entire distribution bus.
In certain instances, however, system operation problems may be intermittent, coming and going without pattern or apparent correlation. This usually indicates a wiring and grounding problem, where multiple pathways for circulating electrical noise exist. As a result, the noise may interfere with a sensitive signal circuit. Because the signal circuit is not protected from such phenomena, it may become disrupted or damaged by the transient energy.
Let's take a look at one example where intermittent problems were experienced and mistakenly attributed to harmonic interaction.
Frozen food processing case history
A medium sized industrial customer in the food processing industry called because it was experiencing operating problems with several pieces of sensitive electronic equipment that controlled part of its manufacturing process. This equipment was a group of three 5 hp adjustable speed drive powered conveyor systems that moved product through the preparation process. The speed of the drive systems depended upon the level of production. The drive systems had recently been purchased and installed. The disturbances to the conveyor systems were so severe that the system controls had to be manually reset in order to continue production.
The facility manager, who was familiar with the problems associated with harmonics, believed there was a harmonic interaction problem on the electrical distribution system, primarily due to the new drives. Further discussion revealed that the building service transformer was rated at 3000kVA and that the three drive systems were the only harmonic producing loads in the processing area.
Doubting seriously that harmonics interaction was the cause of the trouble (because of the small number of nonlinear loads relative to the electrical service size), we recommended that a further survey of the site might provide some additional information. We expected to find some other phenomena correlating with the upsets.
Voltage and current measurements were taken with a true rms digital multimeter at the building service and found to be well within specified operating parameters and capacity. Using a current harmonic analyzer, we observed that the THD (total harmonic distortion) in current was less than 2% and individual harmonic distortion, in current, for the odd harmonics up to the 21st to be less than 3% each.
A study of the utility distribution system operation over the prior three months did not reveal any correlation between its operation and the upsets experienced at this site.
We then moved to the drive locations in the building to examine the installation. Each drive system had its own dedicated circuit from a subpanel and each of these circuits fed a 120V/120V, single-phase, electrostatically shielded, wall-mounted transformer near each drive. The transformer's secondary then fed the drive control system.
From a visual standpoint, it was apparent that the installer had given thought to the installation with the use of dedicated circuits and shielded transformers, which assist in the control of electrical noise while establishing a solid ground reference for the sensitive electronics referenced to ground.
To verify that the transformers were indeed providing noise rejection and a stable ground reference, we again used the true rms multimeter to measure voltage between the case of one drive control system and the power system neutral.
What would you expect to measure? Zero volts, right? Instead we measured a power system neutral-to-case ground voltage of 35V! When the other two transformers were investigated, approximately the same voltages were measured. At least we were on the track of something consistent with the possible explanation of the system's behavior.
Fig. 1 (see page 18) shows the proper installation of these transformers. Notice how the incoming ground, shield, core, secondary neutral and secondary ground are all bonded at one point. This is called a single point ground. This single point ground is then grounded to the building ground system.
When the covers of the transformers were removed to verify the wiring connections, what was found is shown in Fig. 2 (see page 18). The neutral was not bonded to the single point ground and the transformer was not locally grounded to the building ground system. When these connections were made, the neutral-to-case ground voltage returned to zero and the unexplained upsets ceased to occur.
This case history provides an interesting example of the difference between a power related anomaly, or something that repeats with a cyclical pattern, and a disturbance that behaves in a random or transitory fashion. Here, the end user had not given any thought to wiring-related problems, assuming instead that the equipment installer had taken care of the necessary details. Mistakenly, the transitory disturbances were attributed to harmonics.
There are many aspects of power quality. Before you address the more in-depth concerns, make sure that the "housekeeping" is cared for, especially that the wiring and grounding of the installation have been carried out properly so that safety and signal reference protection is provided.