back2049 发表于 2018-3-21 15:22:15

marcos 发表于 2018-3-21 11:46 static/image/common/back.gif
课件上的内容,这其实就是一个最简单的,一个油缸推动一个质量块的数学模型
液压油就是一个理想的弹簧
...

我三年前买过一本,大概翻阅了一遍

蜻蜓 发表于 2018-3-21 20:23:42

数字液压 发表于 2018-3-21 11:27
先把最经典的PID公式真正的物理含义研究明白(P、I、D的意义与作用),基本就能应付70%的控制需求了。

谢谢,我先学学,不明白再请教。

蜻蜓 发表于 2018-3-21 20:38:24

marcos 发表于 2018-3-20 14:59
我打算从常老师的《液压控制系统》一书,并结合实际的工作项目,液压拉伸垫(Die Cushion control system) ...

边学边做是对的,学习起来还容易些,只看书理解不深。

蜻蜓 发表于 2018-3-21 21:08:06

PEN 发表于 2018-3-21 01:29
蜻蜓,你太担心编程语言。重要的是公式和算法的知识。
你明白吗:
拉普拉斯变换?


谢谢,我刚开始学习经典(古典)自动控制原理和数学,拉普拉斯变换,伯德图,奈奎斯特图,这些教程是唯一的,没有其他可以选择,我知道需要学习这些。但是编程语言太多,不知道该选择什么,数字液压也推荐ST,应该不会错。

现代控制原理,线性代数,矩阵这些知识,对电液伺服控制有多大帮助,我不知道。

PEN 发表于 2018-3-22 00:25:48

马科斯,
我一直无法提取.rar文件。
------
4.png和6.png中的传递函数不适用于液压缸。 4.png和6.png对于弹簧上的质量是正确的。

有人知道如何纠正?
提示,看看我的符号证明,即PD控制器中的比例增益高于P only控制器中的比例增益。

传递函数如何帮助控制?
必须有更多的信息。没有足够的关于如何使用传输功能的信息。

如何使用传递函数来计算闭环增益?

K,B和M的值是从哪里来的?
如果没有值,这些方程就没有用。

很久以前我有这些问题。有些答案不在书中。

蜻蜓,
奈奎斯特图表从未用过。
他们证明的是该系统是稳定的。
还有更多有用的方法来做同样的事情。

学习ST是一个好的开始。

学习状态空间的矩阵是好的,但学习如何求解微分方程更好,因为液压系统有许多非线性分量,可以用微分方程很容易地建模,但不能用状态空间很好地建模,因为状态空间期望模型是线性的。人们喜欢状态空间的原因是由于Matlab和使矩阵运算变得容易的类似软件。

我自学成才。我于1975年毕业于大学,拥有电子和计算机工程学位。控制理论没有教导。有人教了关于用运算放大器来制作PID控制器。

Marcos,
I have not been able to extract the .rar files.
------
The transfer function in 4.png and 6.png is not correct for hydraulic cylinders.The 4.png and 6.png are correct for a mass on a spring.

Does anybody know how to correct?
Hint, look at my symbolic proof that the proportional gain in a PD controller is higher than the proportional gain in a P only controller.

How does the transfer function help with control?
There must be more information.There isn't enough information on how to use the transfer functon.

How can the transfer function be used to compute closed loop gains?

Where due the values for K, B and M come from?
The equations are not that useful without values.

A long time ago I had these questions.Some answers were not in the books.

Dragonfly,
I never found much use for Nyquist charts.
All they prove is that the system is stable.
There are more useful ways to do the same.

Learning ST is a good start.

Learning matrices for state space is OK but learning how to solve differential equations is better because hydraulic systems have many non-linear components that can be modeled easily with differential equations but can't be modeled well with state space because state space expect the model to be linear.The reason people like state space is due to Matlab and similar software that makes matrix math easy.

I am self taught.I graduated from college in 1975 with an electrical and computer engineering degree.Control theory was not taught.A little was taught about making a PID controller with op-amps.

PEN 发表于 2018-3-22 02:45:48

本帖最后由 PEN 于 2018-3-22 02:52 编辑

台达计算机系统公司每年在美国多次举办基础培训班,每年举办2至3次高级培训课程。
基本班级限制在8人左右。
高级培训班有16人的限制。
课程是一系列需要30-60分钟的课题。
老师讲几分钟,然后学生在培训手册中做练习。每个学生在这个阶段都有一个模拟模式的RMC。
基础培训第二天的后半部分是调整一个真正的系统和通信。

高级培训有许多不同的实验室。这16人被分成4人组,他们在实验室中轮换。实验室很好,因为课室讲座会让学生和老师感到疲倦。

今年夏天,我们将在中国举办第一次基础培训。它将在无锡举办,由我们的经销商无锡永恒赞助。

这些课程并不是技术性的。主要侧重于建立,调整和编程RMC。

http://deltamotion.com/peter/Pictures/Hydraulic%20Training/Basic%20Training%20B.jpg

http://deltamotion.com/peter/Pic ... ferent%20valves.jpg
我不会教这门课。

我们已经培训了大约2000名来自各地但主要是美国和加拿大的人。但是,我们教过来自俄罗斯,中国,韩国,新西兰,澳大利亚,
土耳其,德国,意大利,芬兰和英国。

我们的目标是培养许多人。我们希望我们的客户自立。当客户遇到问题时,他们无法解决,我们可以使用互联网在任我们喜欢有两个频道。一个用于RMCTools,另一个用于视频。

Delta Computer Systems has basic training classes in the US many times a year and advanced training classes 2 to 3 times a year.
The basic classes are limited to about 8 people.
The advanced training class has a limit of 16 people.
The classes are a series of topics that take 30-60 minutes each.
The teacher talks for a few minutes then the students do the exercises in the training manual.   Each student has a RMC in simulate mode for this phase.
The last half of the second day of basic training is tuning a real system and communications.

The advanced training has many different labs.The 16 people are broken up into groups of 4 and they rotate through the labs.The labs are good because class room lectures can get tiring for the student and teacher.

This summer we are going to have our first basic training in China.   It will be at Wuxi and sponsored by our distributer Wuxi Forever.

The classes are not that technical.The focus mainly on setting up, tuning and programming a RMC.

This class will not be taught by me.

We have trainged about 2000 people from all over but mostly the US and Canada.However, we have taught people from Russia, China, Korea, New Zealand, Austraila,
Turkey, Germany, Italy,Finland, and England.

Our goal is to have many people trained.We want our customers to be self supporting. When customers have problems they cannot solve we can support anywhere using the internet.We like to have two channels.One for RMCTools and another for video.




Ysx317 发表于 2018-3-22 09:08:27

蜻蜓是一个上进心很强的年轻人,希望能够在本群学会伺服系统的控制,非常好,我估计这也是许多上网年轻人希望得到的进步,我希望PEN先生和其它的伺服系统控制专家多多贡献知识,让年轻人能够快速掌握这门技术,以便更好的工作。尤其是告诉年轻人怎样能够用数学方法去指导一个系统的设计和调节,包括怎样设定参数、怎样去解大量的微分方程、然后根据这些方程能解决什么问题,能指导那些工作,如果把这些都让年轻人掌握并能够在实际中指导工作,哪真是太好了。希望专家们贡献力量。我先赞一个。

PEN 发表于 2018-3-22 11:11:34

我终于下了长通力的书。 rar压缩不是标准。

长通里的书中有许多其他文献的参考文献。我所知道和尊重的是Taco Viersma。 Taco Viersma开发了许多其他复制的公式。我推荐阅读Taco Viersma的书或pdf。

1965年乔治凯勒做了一些计算,后来被复制为VCCM方程。乔治凯勒的版本更好,因为它考虑到了加速度。长通里的消息并未列出乔治凯勒。乔治凯勒训练波音工程师。他的书更像是一本培训手册。

无关。
Delta RMC用于在波音737s上安装铆钉。 Delta Delta RMC已经这么做了多年。我们负责该项目的工程师多年来培训了多组波音工程师。
回到主题。

没有一个消息来源解释如何计算增益以获得所需的响应。
这是我能做的。我曾多次说过。控制器闭环增益和前馈有公式并可以计算。

除了你阅读的内容之外,还有一个完整的理解水平。在过去,我阅读了一些相同的书籍,并且非常沮丧,因为他们没有告诉我我需要知道什么才是最好的控制。我必须自己学习。

在PTC2017展会上,我尽可能在短时间内向我们展示了安教授。 25年的家庭工作。系统和控制方法的各种组合已经制定出来。 35 MB的例子。我有足够的例子为数十本书。大多数永远不会被发布。

如果没有这种优势,像穆格和博世力士乐这样的公司会像昆虫一样压扁我们。

I finally made the effort to down load Changtongli's book.   rar compression is not standard.

The Changtongli's book makes many references to other documents.   One that I know and respect is Taco Viersma.Taco Viersma developed a lot of the formulas that other copied.   I recommend reading Taco Viersma book or pdf.

In 1965 George Keller did some calculations that was later copied as the VCCM equation.George Keller's version is better because it takes into account acceleration. George Keller is not listed in Changtongli's sources.George Keller trained Boeing engineers.His books is more of a training manual.

Off topic.
Delta RMCs are used to install the rivets in Boeing 737s.The Delta RMCs have done this for years.   Our engineer in charge off that project has trained many sets of Boeing engineers over many years.
Back on topic.

None of the sources explain how to calculate gains to get the desired response.
This is what I can do.I have said many times.Controller closed loop gains and feedforwars have formulas and can be calculated.

There is a whole level of understanding beyond what you read.In the past I read some of the same books and was extremely frustrated because they didn't tell me what I need to know for the best control.I had to learn on my own.

At the PTC2017 show I showed Professor An as much as I could in a short period of time we had.25 years of home work.Every combination of system and control method has been worked out.35 MB of examples.I have enough examples for dozens of books.Most will never be published.

Without this advantage companies like Moog and Bosch-Rexroth would squash us like an insect.


蜻蜓 发表于 2018-3-22 18:14:50

杨老可以写写创新思维方面的书,PEN老师可以出本控制理论与实践方面的书,(PEN老师那本指南文字太少)附上签名,买数字缸或控制器的用户可以得到赠送。

Ysx317 发表于 2018-3-23 17:52:49

蜻蜓 发表于 2018-3-22 18:14
杨老可以写写创新思维方面的书,PEN老师可以出本控制理论与实践方面的书,(PEN老师那本指南文字太少)附上 ...

蜻蜓:现代控制原理,线性代数,矩阵这些知识,对电液伺服控制有多大帮助,我不知道。
你这句话问得很到位,如果我给你举个实例,你可能很伤心:
我的同事,他的学历是职业技术高中,没学过液压,前不久,他帮别人调试了一套4缸同步系统,2天就搞定了,同步精度达到0.05毫米,位置精度达到0.01毫米,他说这是一套高精度4缸同步系统,原来是一个教授负责,调了半年,达不到要求,正好他探亲回家,请他帮忙,于是就帮别人解决了这个难题。我问他你了解这个系统的数学模型吗?你知道他的传递函数吗?你知道系统的频宽吗。。。。。,他一脸茫然,问我干嘛要知道这些?我说我也说不清。还是请有经验的专家解释吧!
页: 1 [2] 3 4 5
查看完整版本: 怎样学习电液伺服控制