Powered by Invision Power Board

Страницы: (6) « Первая ... 4 5 [6]   ( Перейти к первому непрочитанному сообщению ) Reply to this topicStart new topicStart Poll

> Преимущества MagNet
k.slabenko
Отправлено: Июн 21 2016, 20:14
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



Цитата (rp2004 @ Дек 22 2015, 21:43)
Цитата (V_aryt @ Дек 18 2015, 08:47)
from www.infolytica.com

I got it.. thanks.

You are obviously working for infolytica and therefore your opinion has difficulties being objective.
It would be interesting to hear somebody else who is not involved with the any of the developers and could compare objectively, maxwell and infolytica.

For any objective comparison - evaluate the software, try it for yourself, then tell us your own opinion about the software.

Over the years I found out that maxwell may be number 1 in mechanical analysis, but it is absolutely no good for magnetics, especially with their PC configuration requirements to solve even basic problems. Infolytica seems to be a lot faster... Where Maxwell solves a problem in 19 days, Infolytica solves it in 4 hours... That's how fast it is. But all this is redundant until you actually try the software.

BR,
K.Slabenko.
Infolytica Sales Support Engineer


--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
logout
Отправлено: Июн 22 2016, 11:00
Quote Post


Пользователь
**

Группа: Эксперты
Сообщений: 96
Пользователь №: 175
Регистрация: 23-Августа 11



Посмотрел бы на задачу, где Инфолитика так уделывает Максвелл
PMEmail Poster
Top
rp2004
Отправлено: Ноя 16 2016, 10:24
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



Цитата (logout @ Июн 22 2016, 08:00)
Посмотрел бы на задачу, где Инфолитика так уделывает Максвелл

Exactly! Mr. Slabenko, why don't you share an example where Magnet performs much better than Maxwell.

P.S By the way Maxwell is a pure EM software and has never been good for mechanical analysis (simply because it's not designed for it). Mr. Slabenko most probably meant ANSYS, not Maxwell in his comparison
PMEmail Poster
Top
rp2004
Отправлено: Ноя 16 2016, 14:25
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



[QUOTE=PavelDergachev,Сен 16 2014, 09:07] [QUOTE=k.slabenko,Сен. 15 2014, 03:17 PM] [QUOTE=PavelDergachev,Сен. 15 2014, 12:57 PM]

С коэффициентами не интересно считать тепло, т.к. это можно и в бесплатных прогах прекрасно сделать, это вообще становится тривиальная задача.

Но а вычисление этих коэффициентов и есть суть теплового расчета и как это делать руками большой вопрос, можно только в некоторых простых случаях - не более. [/QUOTE]


Дело в том, что в 1978 и отдельно электромагнитный расчет был уже очень крутым.
Но сейчас то 2014 и теперь нужна мультифизика.
И совсем разные вещи проектировать стандартный трансформатор, двигатель или современный, максимально оптимизированнный. Раньше это было вообще проблематично из-за слабых компьютеров, а теперь это уже реальные вещи и их реально делают при массовом производстве. [/QUOTE]

Totally agree..

Nowadays, using film coefficients for modeling heat transfer (HT) tasks is a joke and the thermal model accuracy is laughable as well. The integration of EM and HT in Maxwell+Ansys, Infolytica modules (don.t know about Comsol) is done on a primitive level (from the point of view of the industry needs) and therefore good for making presentations and university papers only.
PMEmail Poster
Top
rp2004
Отправлено: Ноя 20 2016, 20:49
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



Цитата (k.slabenko @ Июн 21 2016, 17:14)

Over the years I found out that maxwell may be number 1 in mechanical analysis, but it is absolutely no good for magnetics, especially with their PC configuration requirements to solve even basic problems. Infolytica seems to be a lot faster... Where Maxwell solves a problem in 19 days, Infolytica solves it in 4 hours... That's how fast it is. But all this is redundant until you actually try the software.

BR,
K.Slabenko.
Infolytica Sales Support Engineer

I was intrigued by Mr. Slabenko's comments and have played for a week with Infolytica Magnet.
I set and run the same models in Maxwell and Magnet and compared the speed and the results.
I must admit Magnet is a very good product - simple and intuitive interface, reliable convergence, realistic results.
I didn't notice significant difference ion the speed - some models run faster in Maxwell, some in Magnet. It depends on the level of mesh refining the solver goes to (and it's different for Maxwell and magnet), so it is difficult to compare "apple with apple". In short I wouldn't bet on Magnet's speed. Both seem to be faster than JMag however (when running the same models).
The post-processing of Magnet is lagging however - clumsy interface requiring a lot of mouse clicks to get the required result. Also, contrary to Maxwell and Jmag, it doesn't seem to plot the results as the simulation goes (very useful in my opinion for modes running for many hours).

To summarize - Magnet is a very good product and probably much cheaper than Maxwell. Unfortunately our company bought Mawell licenses long time ago, otherwise Magnet would have been very seriously considered.
PMEmail Poster
Top
logout
Отправлено: Ноя 21 2016, 09:36
Quote Post


Пользователь
**

Группа: Эксперты
Сообщений: 96
Пользователь №: 175
Регистрация: 23-Августа 11



При сравнении скорости счёта Jmag, MagNet, Maxwell нужно опираться на точность полученных результатов, количество КЭ. Нужно учитывать, что тот же результат, что и в Jmag, MagNet можно получить в Maxwell на более легкой сеточной модели за счёт элементов второго порядка. В это же время не корректно сравнивать скорость счета решателей на сеточных моделях одинакового порядка.
PMEmail Poster
Top
rp2004
Отправлено: Ноя 22 2016, 12:46
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



Цитата (logout @ Ноя 21 2016, 06:36)
При сравнении скорости счёта Jmag, MagNet, Maxwell нужно опираться на точность полученных результатов, количество КЭ. Нужно учитывать, что тот же результат, что и в Jmag, MagNet можно получить в Maxwell на более легкой сеточной модели за счёт элементов второго порядка. В это же время не корректно сравнивать скорость счета решателей на сеточных моделях одинакового порядка.

I agree, by the way, the polynom's order is optional in Magnet which is a nice feature. The highest order I could get for my models was 3.
What I was trying to say before was that there wasn't any significant difference in the calculation speed of Maxwell and Magnet. Yet, some models would run a bit faster in Magnet while others will run faster in Maxwell. I forgot to mention that I was using the default mesh and solver settings for both progs.
In any case, the difference in calculation speed I observed was nothing like Mr. Slabenko was claiming above.
An annoying drawback of Magnet I forgot to mention is the circuit editor - it is very, very basic. The initial conditions on the caps and inductors cannot be assigned directly and one has to do tricks with external sources to initiate them.
PMEmail Poster
Top
rp2004
Отправлено: Ноя 30 2016, 21:50
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



Цитата (PhX @ Июн 2 2010, 12:32)
Цитата (Админ @ Июн. 2 2010, 12:20 PM)
"Если бы у бабушки были яйца, это был бы уже дедушка" (С)

Theoretically, time-harmonic analysis is only possible when all the materials in the problem are linear. If they are non-linear, sinusoidally-varying sources will not give rise to sinusoidally-varying fields, and time-harmonic analysis is not possible. However, MagNet’s time-harmonic 2D and 3D solvers are actually quasi non-linear solvers, taking into account the approximate material non-linearities by trying to find the operation point on a non-linear B-H curve, using its first few data points.

А если у бабушки только одно??? Кто это?

Кто верит в чудеса! Но все их ждут. (с)

What about just running transient analysis instead of time-harmonic - then all nonliearities will be accounted for (and will probably be more accurate too)
PMEmail Poster
Top
k.slabenko
Отправлено: Дек 11 2016, 01:59
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



rp20014, can you please send me the model you used to test the speeds?

I'm curious to see which settings you used. This was an official Ansys reply, when they reported their solution with 840 000 tetrahedra solved in 15 days. MagNet with 900 000 mesh points solved it in 11 hours.

As far as I can tell, every customer of MagNet says one and the same thing, that MagNet is by a mile faster, than Ansys, so it is very interesting to see a single problem, where Ansys would beat Infolytica in the solving speed.

Please share the models.

Furthermore, can you please let me know what version of MagNet have you used? Is it a full version? Evaluation Licence?


--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
k.slabenko
Отправлено: Дек 11 2016, 02:04
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



Цитата (rp2004 @ Ноя 20 2016, 20:49)
Цитата (k.slabenko @ Июн 21 2016, 17:14)

Over the years I found out that maxwell may be number 1 in mechanical analysis, but it is absolutely no good for magnetics, especially with their PC configuration requirements to solve even basic problems. Infolytica seems to be a lot faster... Where Maxwell solves a problem in 19 days, Infolytica solves it in 4 hours... That's how fast it is. But all this is redundant until you actually try the software.

BR,
K.Slabenko.
Infolytica Sales Support Engineer

I was intrigued by Mr. Slabenko's comments and have played for a week with Infolytica Magnet.
I set and run the same models in Maxwell and Magnet and compared the speed and the results.
I must admit Magnet is a very good product - simple and intuitive interface, reliable convergence, realistic results.
I didn't notice significant difference ion the speed - some models run faster in Maxwell, some in Magnet. It depends on the level of mesh refining the solver goes to (and it's different for Maxwell and magnet), so it is difficult to compare "apple with apple". In short I wouldn't bet on Magnet's speed. Both seem to be faster than JMag however (when running the same models).
The post-processing of Magnet is lagging however - clumsy interface requiring a lot of mouse clicks to get the required result. Also, contrary to Maxwell and Jmag, it doesn't seem to plot the results as the simulation goes (very useful in my opinion for modes running for many hours).

To summarize - Magnet is a very good product and probably much cheaper than Maxwell. Unfortunately our company bought Mawell licenses long time ago, otherwise Magnet would have been very seriously considered.

You can stop the simulation in MagNet at any time and check the results. After you do that, you can then resume the analysis at the point, where you stopped it. So I guess MagNet does include this feature one way or another.

Regarding the delays and lags, please provide more information.

I'm not sure when or how you tried MagNet, but I do know that our support usually responds to that and assist in helping people with their model settings.

If your company would like to try and do a proper comparison vs Ansys, please drop me an email on support@infolytica.co.uk. We can arrange an evaluation and show you exactly how it is.

Regards,
Konstantin


--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
k.slabenko
Отправлено: Дек 11 2016, 02:11
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



Цитата (logout @ Ноя 21 2016, 09:36)
При сравнении скорости счёта Jmag, MagNet, Maxwell нужно опираться на точность полученных результатов, количество КЭ. Нужно учитывать, что тот же результат, что и в Jmag, MagNet можно получить в Maxwell на более легкой сеточной модели за счёт элементов второго порядка. В это же время не корректно сравнивать скорость счета решателей на сеточных моделях одинакового порядка.

Jmag Вообще нельзя сравнивать с MagNet и Maxwell. Философия программ отличается.

Jmag сравнивают обычно с MotorSolve и Maxwell RMxprt.

В таком случае будет более менее честное сравние.

MagNet сравнивается с Maxwell и Comsol.

По поводу точности, наши клиенты её сравнивали сразу в трёх программах (Maxwell, MagNet, Comsol). Везде результат был + - одинаковым. Лидеры по скорости решения были:
1: MagNet
2: COMSOL
3: Maxwell

Как и что проверялось - данных нет, но поэтому мы и предлагаем тестирование, чтобы люди, которые готовы приобрести качественный софт, могли сами убедиться в производительности софта, скорости, лёгкости изучения и каких-то возможных ограничениях. Не советую бросаться выводами до тех пор, пока мы сами не скажем "этого невозможно сделать", ибо часто люди делают поспешные выводы.



--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
rp2004
Отправлено: Июл 6 2017, 05:25
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



I must admit now - Infolytica is faster than Maxwell when running complex models (in some cases considerably faster).

This conclusion comes after a month of running these models with both Maxwell and Infolytica.
Too bad our company is stuck with Maxwell.
PMEmail Poster
Top
k.slabenko
Отправлено: Июл 17 2017, 23:47
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



Цитата (rp2004 @ Июл 6 2017, 05:25)
I must admit now - Infolytica is faster than Maxwell when running complex models (in some cases considerably faster).

This conclusion comes after a month of running these models with both Maxwell and Infolytica.
Too bad our company is stuck with Maxwell.

Well, depending on what you really need. Maxwell is a very nice tool, but not for everything, same as Infolytica and any other software.

It really depends what is your core business. You choose the software by the most applicable projects, if you can't afford everything or you stick to something, which mostly covers your needs.

In Magnetics, Electrostatic and coupled Thermal analysis - Infolytica is very good, but there are also some limits, which may or may not affect you.

Always evaluate an up to date version smile.gif


--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
rp2004
Отправлено: Дек 20 2017, 03:43
Quote Post


Старик
***

Группа: Эксперты
Сообщений: 360
Пользователь №: 113
Регистрация: 5-Февраля 11



Siemens/Mentor bought Infolytica, so Infolytica is part of Siemens now.
Wander if Mr. Slabenko knows more about the changes expected.

P.S. (from July2018): It is sad Mr. Slabenko is not commenting anymore on what is going on with Infolytica and what the plans (of the company) for the future are

Anybody with any comments about Infolytica? Mr. Slabenko is not online anymore, but most of what he said before turned out to be true. It is a very good product indeed. Fast and reliable compared with Maxwell and even more with Jmag
PMEmail Poster
Top
k.slabenko
Отправлено: Май 10 2019, 15:11
Quote Post


Новичок
*

Группа: Пользователи
Сообщений: 48
Пользователь №: 711
Регистрация: 26-Августа 14



Цитата (rp2004 @ Дек 20 2017, 03:43)
Siemens/Mentor bought Infolytica, so Infolytica is part of Siemens now.
Wander if Mr. Slabenko knows more about the changes expected.

P.S. (from July2018): It is sad Mr. Slabenko is not commenting anymore on what is going on with Infolytica and what the plans (of the company) for the future are

Anybody with any comments about Infolytica? Mr. Slabenko is not online anymore, but most of what he said before turned out to be true. It is a very good product indeed. Fast and reliable compared with Maxwell and even more with Jmag

We are still here smile.gif Just too busy to check anything.

Please don't hesitate to check our new website at www.infologicdesign.co.uk

You can always contact me directly by sending an email to: K.Slabenko@gmail.com.

I'm sorry, but I simply cannot spend any time in here.

Best Regards,
Konstantin


--------------------
Konstantin Slabenko
Sales Support Engineer

Infolytica Europe
Tel: +44 (0)1327 810383
Fax: +44 (0)1327 552383
Skype: "k.slabenko"
Email: sales-support@infolytica.co.uk
PMEmail PosterUsers Website
Top
0 Пользователей читают эту тему (0 Гостей и 0 Скрытых Пользователей)
0 Пользователей:

Topic Options Страницы: (6) « Первая ... 4 5 [6]  Reply to this topicStart new topicStart Poll


 


Текстовая версия