Sunday, January 31, 2016

Is it possible to save fuel on a four-engined aircraft by using engines of two different sizes

I got this thought when I looked at this photo of an Ilyushin Il-106 plane. Since maximum thrust is required only occasionally, does it make sense to have two large engines that operate all the time, and two smaller engines which are turned on/off only as needed? Will this system result in overall cost savings for the operator? What are the engineering implications of this system? Can the two smaller engines be electric?

Of course, if all four engines are of equal size, then it probably won't be possible to turn off one pair of the engines [think of quad-engined Airbus A340 or Antonov An-124].

Update [14-Jun-18]: This same thought occurred to me again yesterday, and I didn't remember that I had already written it here. Got here after some quick searches. Think of ARM's big.LITTLE architecture. Will something similar prove to be beneficial from an engineering perspective for civil aviation? Imagine a new A340 widebody plane that's smaller than the current one, and uses a pair of large engines and a pair of small engines. At take-off and during climb, all four engines run at maximum or near-maximum power. When the plane starts cruising at its cruising altitude, either the two large engines are turned off [and the smaller ones continue to run], or the two smaller ones are turned off [and the larger ones continue to run]. It's even possible to imagine that at cruising altitude the smaller engines are run at full throttle, while the larger ones are powered to only a small level [or the opposite of this - larger ones are run at full throttle while smaller ones to only a small level]. This is somewhat analogous to the concepts of dynamic frequency/voltage scaling in microprocessors.

None of these exact values can be specified at this point. Only real-world calculations can provide proper answers. However, the important point here is that this idea, if it works, can partially provide the traditional redundancy/safety benefits of quad-engined planes, without the full baggage of higher costs that have unfortunately made beautiful and extremely safe aircraft such as the Airbus A340 out of fashion today. Just imagine that at cruising altitude when only two larger engines were being run, one of the larger engines turns off due to a malfunction. The plane senses this unexpected shutdown and immediately/automatically turns on the corresponding smaller engine at full throttle, to somewhat compensate for the loss of thrust on one side. This scenario looks much better/safer compared to if a single engine shutdown happens on a twinjet like the A350 XWB, with the airplane left with only one working turbine.

If this concept works, then a natural evolution that engineers should evaluate is analogous to the architecture used in MediaTek's Helio X30 SoC - Max.Mid.Min. Planes such as the Antonov An-225 or the Boeing B-52 Stratofortress which use six and eight engines, respectively, might benefit from carrying three different sizes of engines on each wing.

Of course, if architectures used in electronics work in aviation, then it'll be possible to have a new, quad-engined Boeing 747 flying once again in the future, with the same beautifully-shaped body that it has always had.

Friday, January 22, 2016

If Hyundai or Suzuki or Honda or Mitsubishi or Fiat produced a car with specifications and features equivalent to a Mercedes-Benz S-Class, what could the price of this car be?

I've frequently wondered this thing - why isn't there a car from any manufacturer that provides the same [or similar] level of comfort/features/luxury, safety and size/space as a Mercedes-Benz S-Class, but for half - or maybe 2/5th - the price? If someone wants a car that's like the S-Class, but is willing to drop the "Mercedes-Benz" and "S-Class" tags in return for a huge decrease in price, why isn't there a product for him?

As an engineer, I properly understand that the S-Class that starts at INR 1.14 crore in India today costs not more than a few lakhs - maybe twenty or thirty - per car to build. As a MBA, I also understand the interplay between the volume of sales and the contribution margin per unit sold, as far as total profits are concerned. Market segmentation, branding, etc., are other concerns at play here.

However, what cannot be denied are three things:
  1. There exist people out there who would like to own a car that provides S-Class level of comfort/features/luxury, safety and size/space, while paying only as much as the base model of Audi A6 or BMW 5 Series. These people - wherever they are - are willing to forego the Mercedes-Benz S-Class tag in favor of an ultra-luxurious product from a respected manufacturer such as Mitsubishi or Fiat.
  2. Given decent sales, it is possible to profitably build a car like the one outlined above.
  3. It's definitely technically possible for a company such as Honda or Hyundai to build a car equivalent to the S-Class. This case isn't similar to Bombardier trying to enter the large aircraft segment with the CSeries and facing several technical difficulties owing to the larger size of the aircraft. It's already possible for manufacturers such as Suzuki or Ford to build cars like the S-Class.

Two other considerations exist:
  1. How will a potential buyer of an Audi A6 or a BMW 5 Series evaluate this car [since all three are priced similarly but vary vastly in terms of brand and comfort/features/luxury/safety/size/space]? This is a very important question to answer. Will most buyers go for an A6 or a 5 Series [thus choosing brand over car]? Or will very many buyers go for this car too?
  2. Will putting multiple brands on this car compensate - even if somewhat - for the lack of brand power?

Sunday, January 17, 2016

Ideally, dedicated mobile wallets and online payment services shouldn't exist - or are these services acting like branchless, lean, direct banks that operate only online?

These thoughts occured to me when I looked at these three things:

Why do dedicated mobile wallet/online payment services such as MobiKwik, Paytm, FreeCharge, etc., exist at all? Are these services really separate services in themselves, or are these services merely [temporarily] filling in for the features missing in the banking services provided by banks in India? Sometimes it seems to me that ideally, it is the mainstream banks - ICICI, HDFC, Axis, Deutsche, RBS, SBI, etc. - that should be providing all of the services of MobiKwik, etc.

After all, tasks such as bill payments, DTH/mobile recharge, etc., are things that should be doable directly from within the online banking platforms of the banks [and also from the applications]. There is no reason and no extra benefit of having a separate, third-party service where you have to transfer funds first [from your mainstream bank account], only to spend it later. In any case, this extra step of having to add money to your mobile wallet seems like an unnecessary and redundant step that can and should be eliminated. If regular banks add all of the features of these third-party services, coupled with the quick and hassle-free experience of these services, will these services quickly go bust? Seems likely.

On the other hand, it seems like these services are increasingly looking like online-only banks, providing more and more features of regular banks, minus the high costs of having physical branches.

Tuesday, January 12, 2016

It may be possible to reduce upfront as well as running costs of an electrical backup inverter system by including LED illumination bulbs in the viability calculations

Suppose a guy wants a guaranteed X lumens of light for Y seconds during power failures [for a total of Z lumens of light, where Z=X*Y]. Suppose that this amount of light can be obtained by using an electrical inverter of power rating P VA and a battery of capacity Q Ah, when this backup system is operated along with standard fluorescent tubelights.

Now, since the luminous efficacy of LED bulbs is higher than that of fluorescent tubelights, but since LED lamps are significantly costlier than fluorescent tubelights [especially for higher power rating LEDs], we have an opportunity for optimization/trade-off here.

I theorize that it should be possible, at least in some cases, to achieve the same amount of light output [in lumens] for the same desired/required duration with a lower total equipment cost [cost of inverter + battery + bulbs] by using LED lamps instead of fluorescent tubelights [LEDs will raise the equipment cost] coupled with a smaller/cheaper sized inverter system and a smaller/cheaper battery [LEDs consume less electricity per lumen of light delivered], thus lowering overall equipment costs. It is to be noted that usually, power-efficient illumination sources like CFLs are sold with the promise of long-term/overall lower costs, and not lower upfront costs. This is the key difference here. By using expensive LEDs coupled with cheaper smaller-sized inverter and battery, we might also achieve lower upfront cost.

Finally, it should be noted that even if the upfront equipment cost is higher when LEDs are used instead of fluorescent tubelights, it's possible/likely that the lifetime TCO is lower for the LED-based system for the following reasons:
  • LEDs have longer life than fluorescent tubelights.
  • Electricity delivered by an inverter system is significantly more expensive than the electricity we receive from the utility company, because there are lots of energy losses in chemically storing energy in the inverter battery as well as when electrical power is drawn from this battery and also when it's converted from DC to AC, etc. Since LEDs consume much less power than fluorescent tubelights, naturally much less of this more expensive power is consumed, thus lowering running costs.

Thursday, January 7, 2016

Anti-malware computer security software sold online in India can and should be offered without costly and unnecessary packaging, physical media, etc.

Just like there are two types of mutual fund schemes in India - Direct and Regular - there should be two types of ways to deliver anti-malware/anti-virus software from vendors such as Kaspersky, ESET, F-Secure, Trend Micro, etc.. Currently such privacy/security software is sold in India in a nice-looking box, complete with physical media [a CD], etc., and is delivered by a physical courier service such as Blue Dart or Delhivery. Costs can be significantly reduced by eschewing things such as box packaging, physical media, etc., and by delivering the serial key via email, along with a link to download the latest version of the security suite. Just like applying directly to mutual fund schemes gives higher returns than applying through someone, purchasing only the serial key will do away with unnecessary costs and will thus either boost profits of sellers or lower end-user prices [thus increasing sales].

Actual photos of the KIS package that arrived at my home from Snapdeal:

Sunday, January 3, 2016

It's possible - or rather likely - that some markets are of the type winner-takes-all by their very structure

Both existing players and potential new entrants to those markets that are structurally of the type winner-takes-all [a monopoly eventually] should assess their level of dominance and thus their chances of survival from time to time. If it can be confidently determined in advance and if it's confidently determined that they're not going to be the absolute #1 winner [implying that someone else is the winner or is going to be dominant], then it's possibly better to rapidly exit being a player in such markets, since being a #2 or a #3 or a #n isn't going to get these firms anything net positive on a sustained and long-term basis in such markets.

Although I'm not sure at this point, from the way economics seems to work in this market, it appears that online retail/shopping could be a winner-takes-all type of market [dominant player keeps getting stronger, making it near-impossible to beat it after a point]. Amazon is the winner in USA, Alibaba/Taobao/Tmall is the winner in China, and a bloody battle for supremacy is on in India [Amazon will likely be the eventual winner here as well].

Before hastily concluding that Amazon is forever going to be the #1 winner in the US, and Alibaba/Taobao/Tmall similarly in China, pause for a moment. Because there's no worldwide winner in online retail/selling/shopping yet. US, China, India, etc., can be considered full markets by themselves only up to the point in time till which these country-specific markets are closed [or effectively closed] to direct competition from foreign players. Once the online commerce markets of various countries are fully opened-up and thus integrated, it's again going to be a fresh battle for the #1 spot in this new, unified, larger, single global market, where a new winner will be determined.

Other types of winner-takes-all interpretations here, here, here.

Unresolved stuff:
  • Is it possible to accurately identify winner-takes-all markets/industries/products/services using a fixed set of characteristics?
  • Can a market be of winner-takes-all type today but not remain so tomorrow? Or must certain markets always remain either in or out of the winner-takes-all type?