Voltage Dip 30 October 2013 @ 1418hrs

Earlier today, at about 2 plus in the afternoon, you might have noticed your office lights flickered, the escalators came to a standstill or that the chillers had tripped.
Enough clues to guess what could have been the cause?

Yes, you guessed it right. There was a transmission level fault, hence causing an islandwide voltage dip.

Here, in Singapore. a voltage dip is defined as a drop of more than 10% of the nominal voltage (Line voltage). It typically lasts around 200ms.

In Singapore, the transmission network consists of a 400kV network, overlayed onto 4 x 230kV blocks. The 230kV blocks were split up in the mid 2000s, for controlling of fault level.
It has inherently brought an advantage: Minimize the impact of a voltage dip due to a 230kV transmission fault to just that particular block.

Hence in a 230kV fault, only connected customers in that particular block which the fault occurred will ‘see’ severe dip values (in the ranges of 40-50% dip by magnitude); Customers in the lower voltage levels (66kV, 22kV etc) of that block will see similar (as seen at 230kV) dip severity or less (slightly). The other 3 blocks will see significantly less severe voltage dip or just a slight variation of the nominal voltage.

This is what happened earlier. Below is the 22kV (Line to Line) voltage dip snapshot, captured from one of the monitoring sites in the western part of Singapore.

22kV Voltage Dip Waveform 30-10-2013 1413hrs
22kV Voltage Dip Waveform 30-10-2013 1413hrs

Monitoring sites in other parts of Singapore also registered similar dip patterns, suggesting a transmission level fault.
This brings me to the conclusion that

1) This was a 230kV single phase fault on the L3 (Blue phase) in the western part of Singapore.
UPDATE 31/10: Official Letter from SP PowerAssets: 230kV Cable damage – between Tuas Substation and Jurong Pier Substation at about 2:13 PM, 30/10/2013.

2) An L3 fault occurred. This affected V23 and V31 significantly. Hence you can notice both V23 and V31 ‘dropped’.

Singapore’s 230kv  network is solidly grounded; hence any single phase fault will cause the other two non-faulted phases to ‘drop’ slightly too. This is unlike in a resistive earthed grounded system, (thru the Neutral Ground resistor) where the other two non-faulted phases will swell.

Here it also shows the advantages of having a power quality monitoring system installed. With both voltage and current waveforms recorded during a dip, it will assist the experienced engineer in analyzing whether the drop in voltage was caused by a fault downstream (internal fault) or originated from the grid (external fault); and hence assist him in quicken the restoration time for equipment that could have tripped due to the voltage dip.

Back to Basics – True RMS

There used to be a time whereby the price between a non-True RMS (aka Average Responding) and a True RMS meter/clamp is pretty significant; so much so that if one is just to measure voltage or current for checking if the circuit is ‘Live’ or not, one will go for the cheaper Non-True RMS device.

Recently, price difference have narrowed down quite a bit and in my opinion, one should just get a True RMS meter/clamp straight away.

Loads today are pretty much almost non-linear these days anyway.

For a pure sinusoidal wave;

true rms vs average responding
true rms vs average responding

Hence for an average responding meter, it will scale the rectified average of the ac waveform by 1.11.
This holds true only for pure sinusoidal waveform; which do not exist in the practical world.

The difference in reading can vary from between 5 and 40%, depending on what type of waveform that is being measured.
An example below shows a difference of about 10%. The load being measured was a combination of a couple of CFL and LED bulbs.

It is funny to me to see some contractors out there who uses a non True RMS meter/clamp to verify the readings obtained from their expensive PQ meters.


True RMS vs Average Rectified Clamp
True RMS vs Average Responding Clamp

*Besides True-RMS capability, one should also check for its Safety CAT category. This is an important safety consideration that should not be disregard.


Power Quality monitoring

Power quality monitoring serves two main purposes:-

1) Benchmarking purposes: tracking loading and PQ indices (eg, harmonics, unbalance, flicker) over time, comparing against utility guidelines or known standards from the United Kingdom Engineering Recommendations, IEC and IEEE.

In Singapore, limits introduced in the Energy Market Authority’s Transmission Code took many references from the UK especially. An example is the voltage flicker / fluctuation limits; they are referenced to UK E.R P28: Planning Limits for Voltage Fluctuation Caused By Industrial, Commercial and Domestic Equipment in the United Kingdom. Another example is the Voltage harmonics limits which were referenced to the UK E/R G5/4: Planning levels for harmonic voltage distortion and the connection of non-linear equipment to the transmission networks and distribution networks of the United Kingdom.

In the market, there are two types of PQ monitoring available: permanent or portable. Traditionally, permanent PQ monitoring was exclusive to the utility companies. However over time, as PQ monitoring devices became more affordable and due to increased awareness regarding power quality, many buildings and manufacturing plants started to have PQ monitoring in place, sometimes integrated together with their internal Building Management System (BMS).

One known use of such system here in Singapore is to assist the facility managers whether can they restart / normalize their equipment (eg chillers) after they tripped following a voltage dip in the network. By having both current and voltage waveforms captured during a dip, it becomes a handy tool to inform the manager whether the fault (and hence the dip) originated upstream or an internal plant fault. Knowing that the fault originated upstream in the network will enable the manager to skip some checklists, hence faster restoration time (and less downtime).


2) Investigation purposes: whereby it has been determined that a power quality meter / instrumentation is necessary to aid in solving a particular power quality problem. In this particular case, a portable power quality meter is almost used. Note: there are instances too whereby a permanent type of PQ meter was made portable by installing it in some portable hard-case housing.

ION 7650 made portable
ION 7650 made portable

I have used several portable power quality meters, the first one being the HIOKI 3196 during my university internship days. I then went on to use other PQ meters like the Fluke 435 Series I, Fluke 434, Fluke 1750, Fluke 345, Hioki PW3198, Dranetz-BMI PX-5 Power Xplorer and Elspec G4500 Blackbox. These are the meters that I have used on an almost daily basis and I will be describing their strengths / weaknesses here based purely on my own experiences. (Note: I am not tied to any PQ meters manufacturer)

1) Fluke 435 Series 1
The Fluke 435 is an IEC 61000-4-30 Class A PQ instrumentation. It has several recording modes like Harmonics, Flicker, Dips/Swells, Inrush and a “Logger” mode.

Hardware: There were little differences between the Fluke 434 n 435. The F435 however have the “Power Logger” mode, hence trending of various parameters was a useful function. The major drawback is that when in “Power Logger” mode, the meter would not be able to capture any waveforms. In the event of a voltage dip for instance, a text one-liner will indicate that a dip has detected during the measurement, but no waveforms whatsoever. Another drawback is the maximum recording of 100 parameters only. This is a major drawback for me personally, as that will mean cutting down on the trending of the no. of individual harmonic orders. There is also no high speed transient capture capability.

Software: Powerlog has improved over the years, with the latest version (4.02) having a “clean and nice look” to it. It is adequate for most power quality trending analysis. It has also a useful data distribution histogram for statistical analysis. However my personal preference is to display harmonic current in terms of RMS amperes or in Total Demand Distortion (TDD). To do that, I will have to export out the fundamental current and various individual harmonic current % and re-do the calculation manually in excel. In certain applications, one may need to record the various parameters in 10-min trending (for eg) but requires 30-min values for maximum demand purposes. There is no feature to do that here, other than setting the F435 to record in 30-min intervals in the first place.

Fluke 435
Fluke 435

2) Fluke 434
Hardware: I had used the F434 version without the logging memory function. Hence there is no capability for it to log down trending values (eg aggregated 10-min trending). Hence this was its major handicap.

Software: Having used other software, Flukeview isn’t really good for me, but that’s just my personal opinion.

3) Fluke 345
The Fluke 345 is a single-phase AC/DC power quality clamp that I personally used to measure DC amperes in the presence of AC in Grid-Tied PV applications, to check for the presence of DC injection back to Grid.

Truth to be told, there isn’t any instrumentation available in measuring small amount of DC in the presence of large AC. This is a great tool nevertheless for spot-measurements and verification of loadings when installing a portable power quality meter.

via PowerLog. Same comments as above.

Fluke 345
Fluke 345

4) Hioki PW3198.
This is the latest PQ meter that recently arrived in my hands. It is a Class A PQ meter, able to record trending and capture waveforms too if certain thresholds set are breached.

In my opinion, this is a much better PQ meter than the Fluke 435 Series 1. There is no limitation on no. of parameters to be recorded and it can record trendings and waveforms together (that is how a PQ meter should function, in my opinion). It is also able to capture high speed transient. Two disadvantages that I come across are as follows: 1) Two voltage inputs (+ and -) for every phase, hence the physical connection for a three phase wye will be different from a three phase delta. I have witnessed minor accidents where users were not familiar in a three phase delta setup, causing short-circuits in the PQ meter. 2) For connection of custom Non-Hioki clamps to it, it can only accept clamps with certain mv/A ratio only (hence not fully 100% customizable)

Similar to the PowerLog, it is adequate for most power quality analysis. Also to add, it is able to trend out harmonic current values of the individual orders in terms of RMSamperes, which is good. It cannot however total up these values and give the user the total harmonic current (in RMS amperes). I have to export them out and calculate them manually via Excel. Another plus is that it is able to give 30min interval values for maximum demand, as compared to PowerLog via the F435.

Hioki pw3198
Hioki pw3198

5) Dranetz-BMI PX-5 Power Xplorer
This is a Class A PQ meter from Dranetz-BMI with hi-speed transient capturing capability. It is able to record trending and capture waveforms at the same time.

Hardware: It has similar capability as the Hioki PW3198. Even its voltage input configurations are similar, meaning the physical configuration at the voltage inputs will be different between a 3 phase wye and a 3phase delta connection. IEC 61000-4-30 gives guidelines on how the average aggregation of the trending values must be:

  • By default; data assessed per 10min RMS values
  • 10min interval -> 3000 10-cycle measurements
  • Average: from 3000 x 10-cycle RMS values
  • Minimum of 3000 x 10 cycle RMS values
  • Maximum of 3000 x 10 cycle RMS values

If you are familiar with PQ meters, there is also a maximum / minimum trending plot. On how this is implemented very much depends on the manufacturer, as it is not specified in the Standard. Dranetz-BMI used a single-cycle maximum/minimum value, meaning the maximum/minimum RMS value you will see is from a single cycle, not from a set of 10cycles. This helps greatly when one is doing troubleshooting purposes, and the waveform capturing mode didn’t activate as it was still within the ‘thresholds’.

There are also downsides for the PX-5. The flexible clamps that came with the meter were the LEM Flex RR3035A. Its connection to the PX-5 is via a ‘coaxial cable’ connection. Over time due to wear and tear, the ‘coaxial cable’ became loose easily and resulted in many unusually high current spikes recorded. Another problem I encountered was the overload of the circuit signals due to the wrong setting of the clamps’ measuring range (it has a range of 30/300/3000A). There was one occasion when I had set a low 300A measuring range and left it over night. Due to increase of loads in night time (loads went up to above 1000A, the clamps became saturated and caused many high frequency voltage transients to be recorded. I admit that it was my mistake, but I would have though that voltage and current were separate issues altogether!

Software: Dranview is by far the best power quality analyzing software ever. Nothing comes close to it in my opinion. The option of changing harmonic current to THD%, TDD% and its absolute values in terms of its RMS can be simply be done with a few clicks. Statistical analysis can be done with just adding on some tables to the trending graphs. I could go on and on over here !

Dranetz-BMI PX-5
Dranetz-BMI PX-5

6) Fluke 1750 Power Recorder
The F1750 prides itself as a PQ meter that “records everything” thru its “adaptive” threshold. It was previously an RPM PQ meter, acquired by Fluke. It has also a high speed transient capture capability.

Hardware: Though it has the benefits of “recording everything”, every cycle of it without setting up any thresholds it is not a Class-A compliant meter. You will not see in any of the specs saying it is Class A compliant but you may read that it measures according to Class-A algorithms. This is because its voltage accuracy is +/-0.2% instead of as required +/- 0.1%. And it does not have the ability to be time-synched to a GPS (though this is largely a non-issue to me personally as the practicality of connecting a GPS module to a PQ meter inside a switchroom is almost zero). Despite its shortcomings, the F1750 is useful in troubleshooting purposes and finding faults in a system for a short period of measurement. When a PQ meter captures everything, it will be up to the PQ investigator to be the smart one to decipher what is important or not.

Software: The current version is Power Analyze 2.4. It is still very much a basic analyzing software. I have used it since its early days whereby things like unbalance and flicker were not available options in the software. Things were then added on as time goes by. And if you notice properly, voltage/current unbalance can be found under the ‘THD’ tab. I cannot understand why though. I suspect that it ran out of tabs to display!

It can however gives me the individual order harmonic current in RMSamperes. Similar to the Hioki PW3198, I have to manually export the individual harmonic current (in amperes) out to calculate the total harmonic current in RMSamperes. However the ridiculous thing that I found is that the software only allowed me to export one individual order at each time only. That means I have to export out 49 times for harmonic orders 2nd to 50th!

Fluke 1750
Fluke 1750

7) Elspec G4500 Blackbox
This is a port-over product from its permanent monitoring series. It is another PQ meter that prides itself in measuring every waveform (1024 samples per cycle for voltage) from start to end (I used the model which has 8GB memory). It is also a Class A compliant meter. It however lacks a high speed transient capture capability.

Hardware: It has an in-built router enabling wireless access to it. Physical inputs to the PQ meter are similar to F1750 / F435. Configurations whether it is measuring a three phase wye or three phase delta is configured via software. It has also an optional temperature probe. And it has these nice blue LED lights above every current and voltage input; to indicate if it is connected and measuring. Setting up is via web interface and is quite straightforward.

Software: An SQL database is firstly needed to be installed on the downloading / viewing PC.  It has a basic software, adequate for most power quality analysis. It is able to trend out values in terms of both the IEC 61000-4-30 Class A guidelines or even in cyclic RMS (which will be more accurate). It can also trend out individual and total harmonic current in RMSamperes. Its biggest drawback and main handicap is its downloading and “unzip portion”, whereby it can easily takes half an hour to display the trending / waveform recorded over one week of measurement. In this sense, it is not very practical in my opinion, as many times, a PQ investigator or the Customer will want to view the data recorded almost immediately. A half an hour is simply too long. (used a Lenovo Thinkpad X220 I5-2520 4GB). My opinion is that it was originally designed for a permanent installation, whereby data is being brought back to the PC/server to download and process at fixed short intervals and not in the portable setup scenario whereby one week of data is downloaded and processed at once. It is a neat meter, if only the the downloading/viewing issue can be resolved.

Elspec G4500 Blackbox
Elspec G4500 Blackbox

In my opinion, if I am to purchase new PQ meters it will still be
5) Dranetz-BMI PX5 for its per cycle min/max and also its powerful software.

Trending vs Waveform

Ever since introduced to the area of power quality and its definitions by my first manager, Azmi Rahmat, it has become a pet-peeve for me everytime someone used the words “trending” and “waveform” interchangeably.

To put the record straight, it is not the SAME.

And it has become my personal way in differentiating a PQ engineer from a non-PQ engineer 🙂

waveform vs trending
Waveform (top) vs Trending (bottom)

In relating to this, I had a previous experience whereby the Customer demanded to know why their voltage dip duration is far off from utility’s official voltage dip results at 22kV. Their monitoring was at low voltage side (230V Phase voltage / 400V Line voltage).

While it is known fact that nearer to the loads, the dip recovery may be slightly slower. However this was incredibly long (I was firstly given the trending only).

long recovery trend
Long recovery trend

Only after I compared with their recorded waveforms, then I realised (and chuckled too); their monitoring system had a wrong time scale!

long recovery wave
20 cycles = 2000ms ???
Singapore: 50Hz System. 1 cycle = 20ms

Causes of Power Quality problems

The causes of PQ problems can be divided into two main causes

1) Deficiencies and disturbances in the supply

and / or

2) Caused by the nature and behaviour of the consumer’s load and installation

Here it shows the results of a survey done by Georgia Power Co, from the perspectives of the Customer and Utility

causes of pq problems

However from my experiences, the percentage of blame towards the service provider takes almost 99% of the time here in Singapore.For instance, in a large manufacturing plant setting, the blame on any failure on the “Tools” side will almost exclusively go to the Facilities department (sometimes even without thorough checks on the equipment).
For smaller Customers, tripping issues on their equipment will usually leads to a call / complaint to the Utility almost immediately.

Back in my time working for the utility company, I have experienced countless times whereby my power quality measurements / monitoring showed that the cause of such trips, were actually caused by the Customer themselves (usually a case of high starting in-rush of Customer’s own equipment).

I guess this is the ‘Blame Culture’ here in Singapore. Nothing related to engineering here. Will love to hear experiences from overseas.

As with most things, solving a power quality problem ranges from the very obvious / simple (for instance the above example) to the very complex (especially when it comes to “intermittent” issues).

Personally, I have developed my own checklists (improved over time from own experiences and others) and takes a structured approach at every power quality problem. You should too, if you are new in this area.

Typically, it will begin with a site visit to gather facts/information from relevant parties and conduct baseline site measurements (eg snapshot PQ measurements, IR thermal checks, checking the earthing system etc). It is also important to ask many questions to different people involved. Sometimes you may even get contradicting answers from them! But nevertheless, their answers will give you some clues on what the problem is.

It is from the review of these data (facts and measurement results) that will enable the PQ investigator to form hypothesis what are the likely causes and determine the next course of action necessary to best serve the Customer’s needs.

If the problem happens to be very direct and obvious to identify, the investigation will end here and a report will be formulated on possible ways to solve the problem.

However most PQ investigations will require a further in-depth survey into the electrical system concerned to identify the underlying problem. This will usually require strategic placement of  a number of power quality meters simultaneously over a period of time (typically one business cycle) at various levels of the electrical network for data comparison and correlation. Things like sources of harmonics can usually be determined thru the use of power harmonic flow. Sources of flicker on the other hand can be determined thru comparison of its loadings and the flicker trends recorded.

I was a trained combat medic back in National Service, so I will like to use an analogy from the medical field here.

It is akin to seeing a doctor when one is ill. The doctor will firstly ask questions and conduct simple tests on the patient. If the illness is obvious, the patient will be prescribed specific medication and sent home. Otherwise, the patient will be recommended to undergo further tests at the hospital to narrow down / determine the causes of his illness.

What is a PQ problem?

What is a power quality problem, one may wonder?

There are many definitions to this out there but my personal favourite will be
“Any power problem manifested in voltage, current, or frequency deviations that result in failure or misoperation of customer equipment”.

The keywords here are “Customer Equipment”. If it does not affect the operation of the equipment, it is not a problem (yet).
For instance, almost every load that we have today is a non-linear load, contributing harmonic current and hence distorting the supply voltages.

At the beginning, it will probably not be a an issue, but over time if not keep in check and with more non-linear loads added, the supply voltages will be distorted further and eventually equipment may start to fail/misoperate.
The problem is made worse if design considerations were not made to ‘cater’ for this harmonic loads (eg undersized neutral cables) and no control was made to what type of equipment (whether it meets relevant product emission standards or not, etc) can be connected to the network. In this situation, a combination of equipment control and benchmarking of power quality indices is needed (more on this in my future posts).

Other than harmonic distortion, other power quality problems include voltage transients, frequency variations, flicker, leakage currents, in-rush currents and voltage dip/swell (just to name a few).

Here in Singapore. the utility company, SP PowerGrid, states in their FAQ that the most important power quality concern in Singapore is voltage dips. This is because a dip will have significant economic impacts on the sizeable semiconductor industries primarily located in Marsiling / Woodlands and Pasir Ris / Tampines area. These semiconductor plants have equipment which are sensitive to these milliseconds variation of the supply voltage. A dip may result in their equipment to trip, halting productions which eventually lead to a host of other costly problems.

My experiences showed that while many of such equipment in the “Facilities” side (Eg variable speed drives to the pumps) are quite well-protected (SEMI F47-rated at least), many equipment in the “Tools” side are not. And one has to take into consideration too that the SEMI F47 has its own limitations as well (more on this in my future posts).

This is not to say that other power quality problems do not exist in Singapore. Things like excessive harmonic distortion for instance are “localized” problems, usually more apparent in the Customer’s end (low voltage side). It is only seen in the Grid’s high voltage side (6.6kV and above) if the problem has escalated considerably.

Having worked for the utility, I do know that there are pockets of the transmission / distribution network that have breached the limits set aside for things like harmonic distortion and flicker. No equipment were known to be affected (yet) and at times, it can also be a complex problem for the utility company to pinpoint the chief cause of the problem, as the grid network is very vast. I believed this is a challenge for many other utility companies worldwide.

Linking back to my favourite definition of a power quality problem, it is a misconception that a PQ problem can only be solved by putting expensive PQ meters / monitoring instrument. A PQ meter is just one of the many tools to aid the PQ investigator. Other tools include earth loop measurement device, earth tester, Infrared Thermal Scan, etc. It is also another misconception that once a PQ monitor is installed, it definitely can pinpoint/solve the problem.

Remember the inputs to the PQ meter is just voltage & current. It is not some “magical”  troubleshooting instrumentation. The selection of what PQ meter to use, its settings, the placement of such meters and more importantly the experience of the PQ investigator matters.

p/s: I will share my experiences on the use of various PQ meters in future posts.


About the author

Muhammad Najmi Bin Bohari is a Singapore-registered professional engineer (and an ASEAN Chartered PE) whose passion lies in the area of power quality. He graduated from Nanyang Technological University with a Master of Science Degree in Power Engineering and a Bachelor’s Degree in Electrical and Electronic Engineering.

Najmi is formerly from Quality Power Management Pte Ltd, where he led its Power Quality & Engineering Analysis Unit in conducting works like power system & power quality studies, assessment on buildings’ earthing & lightning protection system and investigations on possible causes of equipment failures or nuisance tripping issues.

Prior to that, Najmi served in SP PowerGrid’s Power Quality & Transient Management section, where he specialized in power quality investigation & analysis, voltage dip sensitivity testing and working with Customers on compliance to power quality requirements. He was also a regular speaker in SP PowerGrid’s biennial public seminar, the “Power Systems Reliability & Quality Workshop”.

In recent years, Najmi has continued to conduct numerous presentations / training in the area of power quality. His most recent presentation was at the 18th Annual PQSynergyInternational Conference and Exhibition 2018 in Bangkok Thailand, where he shared his experiences in Singapore’s PQ scene over the last ten years.

Currently he runs an engineering & consulting company, Potentia Dynamics Pte Ltd. In his free time, he writes on this blog to share his views and experiences in this fascinating world of power quality.

For queries, kindly contact him.

Recent works – selected

  • Third Party Consultancy Study on Power Distribution System Condition Study of a Light Rapid Transit System (Review of Load Flow and Power Quality aspects) – Singapore
  • Analysis of Power Quality Data (Assessment of the quality of supply for the connection of Medical Diagnostic Equipment) – Cambodia, Philippines, Singapore
  • Power Quality Assessment for the Connection of Distributed Generation / Grid-Tied Solar – various projects, Singapore
  • Harmonics Analysis & Study at a specialty chemical plant – Singapore
  • Power Systems Study (Protection Discrimination Review) at a hard-disk manufacturing plant – Singapore
  • Arc-Flash Study at a Test & Manufacturing Facility – Singapore
  • Protection Coordination Study (HT/LT) for a High-rise Telecommunications Building – Singapore

Recent Presentations / Short Courses

  • Harmonics – An Overview
  • Power Quality Training (2-day course for GE Healthcare)
  • Unitest Instruments Seminar on “Know Your Power Quality” – Keynote speaker
  • SS601: COP for Maintenance of Grid-Tied Solar PV Power Supply System (Aspects on Local PQ requirements)
  • Power Quality In Building Services – Case Studies and Examples from Singapore
  • Unitest Instruments Workshop – Power Quality
  • Unitest Instruments Workshop – Electrical Safety

Updated: 21-02-2020

Welcome to powerquality.sg

Welcome to powerquality.sg

Power quality isn’t a new word or term. It has been around for ages. However if one is to search around, there isn’t much information available with regards to this topic here in Singapore.
It is very much still a niche area in the local electrical engineering scene.

Hence, the birth of this blog/site. It is the author’s intention to fill this gap of information here in this website and at the same time a platform to share the author’s views & experiences gained in the subject.

The author is currently the principal consultant for Potentia Dynamics Pte Ltd, an engineering & consulting company.

© powerquality.sg