Your questions answered: Power event analysis in mission critical facilities
Ram Kaushik from Schneider Electric replied to questions from “Power event analysis in mission critical facilities,” originally broadcast on July 23, 2020.
Ram Kaushik, U.S. Offer Manager, Digital Power Division, Schneider Electric, replied to questions not answered during “Power event analysis in mission critical facilities,” originally broadcast on July 23, 2020.
Question: Can you please shed some light on the importance of time setting accuracy in smart switchgears?
Ram Kaushik: Time synchronization is absolutely critical for medium- and low-voltage switchgear. When events that span less than one second happen, it is impossible to piece together what happened if clocks are drifting or not set properly. Even if you have data, it can be useless if not time-stamped accurately.
Question: On backup generation, when mechanical loads hit the generator set, what startup loads are considered power events? I’m looking at a six-time running factor for startup.
Ram Kaushik: Good scenario. I think for emergency generators — you absolutely should consider monitoring closely with power quality meters and digital status inputs for generator status input/output.
Question: Sometimes you will find an obvious upstream breaker trip or main breaker trip at the source switchboard. That could be due to utility loss, correct?
Ram Kaushik: Yes, absolutely. In some cases, the evidence is easy to identify like the example you give. If the smoking gun is easy to spot, it makes the facility staff’s analysis easy. Unfortunately, in many cases, the detective’s trail is littered with false evidence.
Question: Are there limitations of how much data is stored before and after and event? For example: Utility power is lost to the mission critical facility. The uninterruptible power supply goes to battery until the generator starts and is able to pick up the load. The UPS is configured to walk the load onto the generator over 30 seconds. Can this metering system capture data for this entire event say five seconds before the utility outage thru five seconds after the UPS walk-in on the generator is complete?
Ram Kaushik: This is a very interesting example. Typically, PQ meters will capture the waveform at loss of utility, and then restoration. The full extended period will be challenging from a waveform perspective. I would also suggest that the full waveform picture for the entire minute or so might not be particularly interesting. How about capturing all the digital I/O status changes in the interim period — UPS start, generator start, transfers, switchover, etc. — with accurate timestamps? Between the two waveforms and the sequence of events, you should have a pretty complete picture.
Question: Concur with you that proper maintenance/records/power conditioning within the campus to avoid false trips, ground and short circuit faults go a long way. All larger campuses are introducing automatic load shedding, power conditioning and power monitoring controls.
Ram Kaushik: Thanks for your perceptive comment. Good to know that you are seeing campuses and facilities adopt best practice as far as power monitoring and record-keeping goes.
Question: What are facilities are considered as mission critical?
Ram Kaushik: Traditionally, large industrials, data centers, hospitals, military campuses, etc. have been considered mission critical. However, that has been steadily changing. Very few larger buildings can withstand even partial outages without serious disruption and business expense. I’d say the definition is more inclusive than ever before.
Question: How do you do the cost benefit analysis for the power event analysis?
Ram Kaushik: Typically, we have seen organizations not factor in risk properly. Partial outage or disruption cost is not factored into return on investment calculations. Opportunity cost of avoided problems because of the presence of analysis tools, prevention of incidents etc. are also not factored in. Once all this is factored in, the business case becomes self-evident — and in some cases — a slam dunk.
Question: How big of a problem are utility power factor correction capacitor switching transients on the operation of variable frequency drives?
Ram Kaushik: It is still an ongoing issue, but varies by region, utility and quality of variable speed drives. VSD technologies have certainly gotten more advanced and more robust but some of the controls are still sensitive to transients.
Question: Can a power quality meter be installed in the switchboard (high- or low-voltage) to monitor the disturbances?
Ram Kaushik: Yes. PQ meters have a variety of installation options. Custom enclosures are also possible.
Question: Which monitoring model device would you recommended for power quality monitoring?
Ram Kaushik: This of course depends on the vendors you are considering. For Schneider Electric, the ION9000, PM8000 and PM5000 series all have solid PQ features and are deployed at different mains and feeder configurations.
Question: Is there a rule or guidance on determining when a subcycle voltage disturbance, dV/dt, is an issue or a nuisance?
Ram Kaushik: I’ll have to research this some more, but based on my recollection, the actual impact of subcycle disturbances varies widely. For example, many drives/motors will not respond to this at all. However, controls on semiconductor equipment (electronics manufacturing) have been known to trip off. So, my initial response is that it varies. I will look into this some more.
Question: Does Schneider Electric offer power quality analysis training courses?
Ram Kaushik: Yes, we have offered several through our Power Management University. Schedules are in disruption because of COVID-19, but a lot of virtual training is also available.
Question: Is there a good source of waveform examples to explain typical power system anomalies, such as capacitive switching, transformer inrush, etc.?
Ram Kaushik: As you can imagine, we’ve collected several examples over the years. There are some websites that have published papers with typical waveforms. The IEEE PQ 1159 standard itself has waveform examples for the various types of disturbances.
Question: Could you please explain-demonstrate how the PQ equipment data get us from the question of why the mixer unit (one of your examples) tripped to the actual culprit in step by step detail using the PQ instruments’ data as evidence?
Ram Kaushik: The evidence trail is the sequence of events report that tracks exactly what tripped first. Once we determined which location saw the disturbance first, we examine the waveforms captured by the PQ meter at that location. Once we see the waveform, we can easily identify it as a capacitor switching transient. We can also look at severity of the event in the CBEMA curve, plus we can look at the disturbance-direction-detection to unambiguously point out that the origination was outside the facility.
Question: Does grounding systems help reduce power quality issues?
Ram Kaushik: Yes, good grounding practice will help to some extent. However, not all PQ issues can be mitigated by improving grounding, transients or sags, for example. Also, grounding will depend heavily on the harmonic content being injected. For example, lots of odd harmonics will mean different sizing requirements. No magic bullet unfortunately, just a lot of smaller things in best practice.
Question: It is proper to sort out the problem of harmonics in the power system during the system design or wait until any appearance of harmonic problems during operations?
Ram Kaushik: We have found proactive PQ management is best. Diagnosing harmonic problems after the fact and then mitigating can be expensive and time consuming. We recommend specifications include PQ monitoring and mitigation right at design phase. For example, drives can be fitted with PQ correction right at source — plus software monitoring can ensure no ongoing issues.
Question: Percentage speaking, what gives more power disturbances, the consumer electrical system, the consumer loads, the utility?
Ram Kaushik: Multiple studies have confirmed that ~80% of PQ events originate on the facility side and ~20% on the utility side.