发送我们看到比例控制器的最后一个视频。
Send the last video we have seen the proportional controller.

这就是您从西门子 TIAA 获得的帮助。
And this is the go you have got from our Siemens TIAA.

所以我们在这里看到我们有这个稳态误差。
So we have seen here that we have this steady state error.

这是我们的错误，并且始终存在于控制器中。
This is our error and this is always there in the controller.

我们研究过这一点，我们已经看到了液体的实时反应。
This we have studied we have seen the response of liquid live.

这是因为控制器只知道当前的错误。
Now this is because the controller only knows the current error.

因此，即使在这个时间点和控制器的这个时间点，误差始终是相同的。
So even at this point of time and this point of time for the controller the error is always the same.

这一次是因为控制器只关注电流和控制器。
This one time because controllers only focus on the current and the controller.

现在想象一下这样一种情况，我们的控制器实际上此时出现了错误。
Now imagine a situation we have a controller that virtually got the error at this point.

也就是说，这是x。
That said this is x.

过了一段时间后，它再次看到这个错误，就像现在的 X 一样。
And again after a certain time at this time it again sees this error which is X like now.

它还会看到最后一个错误，再次是 X。
It also sees the last error which was X again.

所以他会积累这个，这会让UPS积累这个。
So he will accumulate this it will make it UPS accumulate this.

所以这会减少X但只是X的情况。
So this will make it less X but just to X case.

因此，如果此处的误差增加，尽管电流相同，但由于上次误差，它已累积到最后，并且 Gallanter 变为两周。
So if the error is increased here although the current are the same because of the last error it has accumulated the last and Gallanter becomes two weeks.

所以这已经是过去式了。
So this is the past.

这是在 或 Plaskett 处传递的正确误差 K，并开始添加到 X，现在控制器动作将增加，并且往往会减少误差，或者动作增加，并且往往会减少误差。
This is the correct error K passed at or Plaskett and added begun to X and now the controller action will be increased and it will tend to decrease the error and or action is increased and it will tend to decrease the error.

它可能会尝试达到这一点。
And it may try to reach that point.

所以我们需要定义这种类型的控制器，它也会考虑过去的两个值。
So we need to define this type of controller which will also consider the past two values.

好的。
OK.

我们的控制比例加上积分来了。
There comes our control proportional plus Integra.

好的。
All right.

让我们看看这里有什么。
So let's see what we have here.

比例和积分纵容它所说的积分控制被称为复位控制器。
Proportional and integral to condone it says in integral control is known as a reset controller.

好的。
OK.

它是我们的信号对误差信号的积分。
It is the integration of our signal integration of error signals.

也是误差信号的总和。
Also in summation of error signals.

因此，它是随时间推移的误差的连续求和指示，意味着迄今为止的一些完整的控制器误差和完整的​​控制误差历史记录。
So it says it is a continuous summing indication of error over time means that some of the complete controller error complete control error history up to the present time.

好吧，假设这是这里的曲线，让我改变一下就OK了。
OK so suppose if this is the curve here and let me change the OK.

绿绿绿就OK了。
Green green and it's OK.

现在，如果我将其视为一个阶跃点，那么这就是一个阶跃信号。
Now if I consider this as a step point this is a step signal.

现在这里有什么错误。
Now what is the error here.

这部分是错误的。
This part is the error.

我的意思是选择另一种颜色黑色。
I mean just take another color black.

好的。
OK.

所以这部分是错误的。
So this part is the error.

所以它是这个错误的总和，然后是这个错误的总和，然后是这个错误的自动化。
So it is a summation of this error and summation of this error and then automation of this error.

我用这条曲线这条曲线，这就是错误。
I'm taking this curve this curve and this is the error.

这些是过去和现在的错误，这就是错误。
These are the past and current errors and this is the error.

这就是错误。
This is the error.

所以这是一个错误的总和。
So this is a summation of error.

因此控制器知道过去的错误是什么。
So the controller knows what was the past error.

他知道过去的错误是什么，并且我们继续这样说，它会相应地给出控制信号。
He knows what was the past error and that we keep on saying that and it will give the control signal accordingly.

所以这就是积分动作。
So this is the integral action.

因此，如果您的控制器知道过去和当前的错误，那么您就可以知道您是否看到了稳定状态。
So if you if your controller knows the past and the current error then you can you can you know did you see a steady state.

因此，在控制中，输出与误差成正比，并且也随着过去误差的值而变化。
So in control the output is proportional to error and also changes with the value of past errors.

所以你可以看到这是比例增益。
So you can see this is the proportional gain.

这是比例控制加完整性。
And this is a proportional control plus integrity.

所以你必须将这两者相加才能得到控制输出。
So you have to sum up these two together to get the control output.

好的。
OK.

现在控制器中可能会出现一些严重的问题，但我们有算法来控制它。
Now there would be some serious problems which might come in a controller but we have the algorithm to control that.

现在您可能想知道我们是否继续对控制器输出的误差求和，这是一个 UTI K.P。
Now you might be wondering if we keep on summing the error the controller output which is here this is a UTI K.P.

很好。
is fine.

这与我们看到的误差成正比。
This is proportional to the error we have seen it.

现在这是误差的总和。
Now this is the summation of error.

如果继续添加此错误，它将变成一个很大的值。
If you keep on adding this error it will go to a large value.

现在这个错误在这里太多了。
Now this error is too much here.

那么你就会遇到错误。
Then you have the error.

在这整个过程中，作为一个风险点积累，这是给予，因为这与该值的输出同样重要，可能会增加你知道很多价值损失，范围可能会更大，这可能会导致你将要至饱和。
And in this whole process as a risky point accumulating and this is giving because this is equally integral to the output of this value may increase to you know a lot of value loss a larger range which may which may go it take you're going to to saturation.

所以有两件事。
So there are two things.

一是再次控制。
One is the control again.

这又是控制程度，定义了误差求和的效果。
This is the degree of control again this defines the effect of summation of error.

好的。
All right.

就像我们有这个比例增益，它乘以当前误差，Indigo 增益乘以误差总和。
Just like we have this proportional gain which is multiply the error the current error the Indigo gain is multiplied at the summation of error.

因此，通常该方案非常小，其值定义了误差总和将再次影响整个控制器的程度。
So generally the scheme again is very small and its value defines how much the summation of error will effect the overall controller again.

好吧，让我们再举一个例子。
OK let's take let's take another example.

我们有一个练习，说通过校准他们的文字来读取唐的反馈，所以我们知道使用按比例移动到控制器的 CM 生成算法，并再次使用 Indiaman 和旋钮。
We have an exercise of it says read the feedback of the Tang by calibrating their words and so we know that generate algorithm using a CM moving to that controller proportional and again using Indiaman and knobs.

好吧，那很好。
OK that's that's pretty fine.

现在看看那个。
Now look at that.

它说包括饱和和错误逻辑包括防止饱和的逻辑。
It says include saturation and bumblers logic include logic to prevent saturation.

那么我们来看看那是什么。
So we'll see what is that.

当您从自动切换到模型时，优化代码以实现完全控制。
And when you switch from automatic to model optimize the code to have complete control.

现在我们来谈谈饱和度饱和度的意思。
Now let's talk about saturation saturation means.

假设让我们回到同样的逻辑。
Suppose Let's come back to this same logic.

正如我告诉过你的，这是错误的总和。
As I told you this is the summation of error.

好的。
OK.

因此，如果箭头继续增加，继续增加，这将达到一个很大的值。
So if the arrow keeps on increasing keeps on increasing keeps on increasing this is going to go to a large value.

好的。
OK.

这将使我们的控制输出或控制器达到饱和。
And this is going to make our control an output or controller to saturation.

现在我们知道我们的液位溢​​出的输出是感觉OK，感觉逻辑填充的最大值是10.0。
Now we know that our output in overfilling our liquid level is the feeling OK and the feeling logic the maximum value of filling is 10.0.

或者你可以说 100%。
Or you can say 100 percent.

好的。
OK.

现在 KP 具有某种价值。
Now KP has some some sort of value.

好的，我。
OK I.

我有一些价值。
I have some value.

让我拿一些颜色并拍摄。
Let me take some color and take.

这有一定的价值。
This is having some value.

好的，我们将根据错误继续增加。
OK we'll just keep on increasing based on the error.

好的。
All right.

因此，如果填充空隙的值从 110% 开始，这对金库物理上没有影响，因为金库中超过 100 的值只有 100%。
So if the value of filling voids goes from let's say 110 percent which physically has no effect on the vault because beyond hundred was in the vault only a hundred percent on.

所以这个10%的增长是没有意义的。
So this increase in 10 percent is meaningless.

这对我们的控制系统没有任何意义。
This doesn't make any sense to our control system.

假设该值继续增加并分解增加，直至达到 200%。
And suppose the value keep on increasing disintegrating increasing as it goes to 200 percent.

理论上，在逻辑算法中它会达到二十点零。
Theoretically in the logic algorithm it goes to twenty point zero.

这有可能发生。
It can happen.

好的。
OK.

因此，这意味着当它达到 20 点零并且纳博讷时，它必须达到 5％，比方说，由于要求它必须从 20 点零下降。
So it means when it does 20 point zero and Narbonne it has to come to 5 percent let's say due to the requirement it has to drop from twenty point zero.

到 19 比 18 和 17，然后到 10，因此该控制器从 20 天到 10 天所花费的时间是没有意义的。
To 19 than 18 and 17 and then up to 10 so the time taken by this controller to go from 20 to 10 days is meaningless.

请原谅我的头写这篇文章很卑鄙。
Excuse my head writing this is mean.

哦抱歉没有意义。
Oh sorry meaningless.

所以我们应该保持我们的控制和输出200%到100%。
So we should maintain our control and output 200 percent to only 100 percent.

这是需要预弯的一件事，也是我们向上的原因。
This is the one thing to prebend and the why we up.

其次，你必须确定这种感觉的价值是否会随着价值增加超过百分之百（假设是 101%）而增加。
Second of all you have to make sure if the value of this feeling will as soon as the values increase beyond hundred percent let's say it's 101 percent.

你必须稳定稳定一些黑靛蓝的总和，靛蓝的总和不算什么。
You have to stabilize stabilize the sum of some of it nigrum the sum of Indigo is nothing.

这里只是求和求和。
It's just the summation the summation here.

好的。
OK.

因此，为了稳定这个 Indigo，我们需要制定一个逻辑，所以让我们看看我们在这里做什么。
So to stabilize this Indigo's some will make a logic so let's see what we do here.

那么如何预防这种情况呢。
So how to prevent the situation.

这很简单。
It's very simple.

正确的。
Right.

如果过度填充，我们称 T 的 U 大于 10.0 为填充。
If overfilling let's call filling as U of T is greater than 10.0.

这是 if 逻辑，我们需要什么，我们首先需要这个填充，即我的 U of T，它将等于 10.0 我不会让超过 10.0，所以你会确保感觉等于 10.0。
This is the if logic what do we need we need first of all this filling which is my U of T that will be equal into 10.0 I will not let go beyond 10.0 so you will make sure the feeling is equal to 10.0.

现在，这里的靛蓝与一些与旧的靛蓝相同的靛蓝融合在一起，或者在与一些靛蓝一起进入之前，以便成长。
And now this indigo right here Indigo with some that is equal to old integrate or be up before that before coming to indigo with some in order to grow.

我想在这里提一下，你有 t 等于比例加，这是第二个左右。
I would like to like to mention here that you have t is equal to proportional Plus and this is second or so.

因此，在这种情况下，我们的 Indigo 外溢或整体外建等于误差总和。
So in this case our Indigo outpoured or integral outbuild is equal into summation of errors.

好吧，就这样吧。
Ok so it to be.

输出。
The output.

因此，当您执行此命令并降级输出时，将具有其积分输出加上误差的值。
So when you execute this command and degrade output will have the value of it's integral output plus error.

因此，这将继续增加一定程度的产出。
So this will keep on adding to a degree of output.

所以这就像你的误差总和。
So this is like this is your summation summation of error.

因此 P-A 控制器的输出将为 p，即 K.P。
So the output of P-A controller will be p which is K.P.

陷入错误加上我是关键是你的收获就像 K.P..
into error plus I is key is your gain just like K.P..

这是唯一一个错误。
And this is the one error.

但在积分的情况下要在输出中。
But in the case of integral to be in the output.

就那么简单。
Simple as that.

好的。
OK.

这是 T 的控制器 U 的输出。
This is the output of a controller U of T.

你有D，这是你的感觉。
You have D which is your feeling.

所以现在如果填充的输出超过 10，我们将把这个输出移至 10。
So now if the output of filling goes beyond 10 we will move this output to 10.

所以只是为了避免让我感到困惑。
So just to avoid confusion to me it is.

运算符将 9:59 以度为单位的某个整数 some 等于所有以什么度为单位的事物，我们最终会做什么。
The operator will be 9:59 at in degrees some integral some is equal to all things with what is degree what we do in the end.

我们将把所有积分写成这是最后一行逻辑，它是否等于组间输出。
We will write all integral This is the last line of logic is it equal into inter-group output.

现在我们知道 Nozick 中的 A C 执行第一条语句、第二条 Citrin 和第三条语句以及第四条和第五条语句。
Now we know that the A C in the Nozick execute first statement and second Citrin and third statement and fourth and fifth.

然后，在 10 逻辑的最后一部分中，我将添加度数输出，并将其输入到所有积分中。
Then in the 10 the last part of our logic I would adding in degree output is going into the all integral.

这就是我的变量，所有积分都是变量。
So that's my variable all the integral is a variable.

这是另一个变量。
This is another variable.

因此，我们默认将愤怒输出的值存储在以 K 为单位的度数中。
So we are storing the value of indignant output in the all in degree with a K by default.

那么当输出值超出时发生了什么。
So what happened the moment value of output goes beyond.

然后积分某愤之和相等且做旧又大。
Then the integral some indignant sum is equal and do old and big.

现在，这里的愤怒总和是我的意思是我们只是根据使用积分的评论编写积分总和，我们在这里定义了一些逻辑。
Now the indignant sum here is I mean we just write integral sum based on the reviews that use integrals some logic which we have defined here.

这是我的积分。
This is my integral sum.

这是指示性输出。
This is indicative output.

抱歉，我采用了另一个变量。
Sorry I took another variable.

所以基本上这是输出不可或缺的一部分。
So basically this is integral to output.

让我再解释一次，这可能会令人困惑。
So let me explain one more time this might be confusing.

首先控制输出是B。
First of all the control output is B.

好吧，现在我的积分增益误差累积误差是什么累积误差。
Ok now I is what accumulation of error in integral gain accumulation of error errors.

这是输出中的一个变量，等于积分输出加上误差。
This is one variable in the output equal to integrate output plus error.

因此，我们继续添加负输出，并将其存储在这个精细变量 k 中。
So we keep on adding to a negative output and this will be stored in this fine variable k.

乘以 kedai 就可以了。
Multiply by kedai this is OK.

好的。
OK.

如果你看到这里，这就是我乘以一些积分加上 KPN 来吃掉它的关键。
If you see here this is the key I multiply by some of integral plus KPN to eat it.

所以坚持下去就好了。
So keeping to it is fine.

Q 进入 Indigo 就这罚款这是一个输出。
Q into Indigo on this fine this is a to output.

不，我说过这种靛蓝输出可以超出我们系统不需要的更高值，这将使我们的系统从 100% 增加到 200 人。
No I said that this indigo output can go beyond some higher value which is not required by our system which will take our systems from 100 percent to 200 person.

我之前已经解释过，以避免该值超过我们当时所做的值的 100%。
I have explained it before to to avoid this value go beyond 100 percent what we did the moment value.

戈宾百分百。
Gobion hundred percent.

我们在U-T上值十点。
We the value of ten point on the U-T.

因此，这对于 10.0 秒的水平来说是平凡的，以稳定这愤怒的夕阳。
So this this would be mundane at 10.0 second level to stabilize this indignant sun setting sun.

因此，我们所做的就是我们将 Indigo 输出视为所有带有旧的它降解器的 Indigo，顺便说一下，等于 Indigo 输出。
So what we do is we took Indigo output as it were to all Indigo with an old it degraders by the way is equal to indigo output.

这是之前的一个扫描周期。
This is one scan cycle before.

所以这个会稳定下来。
So this will be stabilized.

这将是稳定的价值。
This would be at stable value right.

这就是 if 语句的执行方式。
This is how you do the if statement.

好的，现在您将为相反的值编写另一个语句。
Okay now you will write another statement for the opposite value.

因此，如果您是史蒂夫，则负值的输出再次小于零点零。
So if youre Steve what is the output is less than zero point zero again for the negative value.

你必须使用等于，直到你得到它和零。
You have to use equal until you get it right there and zero.

积分输出在这里也做了同样的事情。
And again integral output does the same thing here.

好的。
Okay.

然后你关闭你的 if 语句也更关闭或这里的语句。
And then you close your close your if statement also closer or statement here.

这就是如何防止结束和结束逻辑的程度。
So this is how you can prevent the wind up and wind up logic in degree.

现在这是一个大问题，伴随着颠簸。
It is a big problem right now it comes with the bumps.

现在，如果您在从自动切换到手动时看到此处的警告，请优化逻辑以使用泵 Plissken。
Now if you see the caution here when you switch from automatic to manual optimize the logic to have pump Plissken.

现在会发生什么。
Now what will happen.

想象一下，您在 Degrader 中正在进行自动操作，如果可以的话，如果我在这里制作曲线，您可以让我拿另一支笔来使其更具交互性。
Imagine you have automatic operation going on in Degrader and your if I may if I make the curve here you have me take another pen to make it more interactive.

好吧，现在假设如果我有一个步骤输入，然后我采取。
OK now suppose if I have a step input and then I take the.

所以这是我的倾诉，它确实试图稳定下来，一切进展顺利。
So this is my outpoured and it does try to stabilize and this is going fine.

现在这就是这将发现这已经被控制稳定了。
Now this is this is going find this is has been stabilized by being controlled.

现在，如果我转向手动控制。
Now if I move to manual control.

所以手动控制控制器动作中发生的事情没有执行任何操作。
So what happened in manual control controller action is not performing anything.

输出我的输出填充是由手动输入控制的。
Output my output filling is controlled by manual input.

好吧，让我们讨论另一个潜在的问题。
OK let's have another potential matter.

这是控制我男人的感觉。
This is controlling the feeling my man.

在这种情况下，如果您在曼韦尔情况下注意到挖掘机输出指示输出中的这一部分，您的错误会更大，因为现在您在手动情况下没有任何设定点，您只需手动控制轮子。
In that case in that case if you notice this part in digger's output indicative output right in manwell case your error will be more because now you don't have any set point in manual case you just control the wheels manually.

您没有任何设定点。
You don't have any setpoints.

所以在这种情况下会发生什么，拿这支笔在这里，这在某些情况下会增加到更大的值，因为我们的错误确实增加了，因为现在错误是你知道错误是输入减去反馈。
So in this case what will happen take this pen here this this in some will increase to a larger value because our error really increased because now error is you know that error is what input minus feedback.

K.

K.

因为我们的行动没有手册。
because our action is no manual.

反馈将是不同的输入集，因为您不知道手动操作在做什么，所以反馈会有所不同。
The feedback will be a different input set one will be different because you don't know what manual operation is doing.

所以发动机的误差就会增加。
So the error will increase the engine will increase.

这是 Indigo 再次大幅减少树木的一部分。
And that's part of increased again Indigo are reducing trees to a larger extent.

因此，当您在数字输出中拥有很多价值时，然后当您跳转到该内容时，就会说这是马诺利斯操作。
So when you have a lot of value in digital output and then when you jump to that say this is a Manolis operation is here.

所以你从这里有很多错误。
So you have a lot of error from here.

我们应该继续补充。
We should keep on adding.

现在，当您在这个位置跳到自动控制时，您已经在输出程度方面获得了很大的价值，这可能会对您的系统造成像这样的大碰撞或对电线产生负面影响。
Now again when you jump to the auto control at this spot you already have a lot of value in degree of output which may cause a big bump like this to your system or negative bump to a wire.

这个凹凸朝向这个凹凸。
This bump toward this bump.

我们所做的实际上是设置积分输出的值。
What we do is really set the value of integral output.

那么如何防止炸弹然后我们转移到一个人让我在这里取一些颜色。
So how to prevent bombs then we moved to a man let me take some of the color right here.

如果您有手动操作，可以说我是一个旧按钮，因为它等于一个。
If you have a manual operation let's say I'm an old button because it is equal into one.

你做什么作为你的输出。
What do you do as your output.

感觉输出将受到婴儿的保护，您知道这将由手动输入控制。
The feeling output will be defended by baby you know this will be controlled by manual input.

首先，因为这是一种感觉将由手动感觉控制，这是一个手动电位器，它将控制感觉，当然，因为它是手动模式，输出将由手动控制，也是您的数字输出，是您应该使其为 0 的积分和。
First of all because this is the feeling will be controlled by a manual feeling which is this one you have a manual potentiometer which will be controlling the feeling of course because it's a manual mode the output will be controlled by manually and also your digital output which is the integral sum that you should make it 0.

好吧，那么会发生什么，你的学位将为零。
OK then what will happen your degree will be zero.

然后当你转到 auto 时它就会开始生成。
And then when you go to auto It will start generating.

开始添加新的错误。
Start adding the new errors.

它不会计算手动过程的错误以及为什么要这样做。
It will not count the error of manual process and why should it do it.

它永远不应该手动执行。
It should never do it manually.

手册说你不应该关心这个错误。
The manual says you should not care about the error.

这就是为什么你把 Indigo 归零。
That's why you make Indigo out to zero.

我希望这一点是有道理的。
I hope this point makes sense.

否则我也会放一些电子书和笔记，让它更清楚。
Otherwise I will also put some e-books and notes to make it more clear.

但当我们开始做实验的时候你可能就会明白这一点。
But when we start doing the experiment you may understand this point.

现在，如果您在 Manue 负担过重时转到 else 语句，则当管理业务时，您会看到此人在此处运行，而不是在您所拥有的情况下。
Now if you go to the else statement when Manue burden is on you have this man in operation going on here when managed business not on which is in the as you have.

T 的 U 是进入 P 的输出加上 I。
The U of T which is the output going into P plus I.

这与我们在这里所做的类似。
And this is similar to what we did here.

所以这整个方程。
So this whole equation.

整个方程已经到这里了。
This whole equation has come here.

K.

K.

所以这些都是关于如何编写控件代码的。
So those are all about how you vav you can code your control because in.

我们还采取了最后一个错误。
We also took the last error.

以及它的逻辑怎么写。
And how to write its logic.

但别担心，我们会在 Seamans TIAA 中做到这一点，我们会看到结果。
But don't worry we'll do that in Seamans TIAA and we'll see the result.

根据结果​​，我们将在此处粘贴一些响应，以便您可以了解它如何减少偏移，从而减少起始点，并且控制器知道它是过去的错误。
And based on the result we will paste some response of here so that you can understand how it reduces the offset so the onset is reduced and controller knows it's past errors.

好的，如果控制器确实知道当前的错误。
Ok if controller does know it's current errors.

这是对稳态误差的坚持。
It was the stick to the steady state errors.

所以这个人将会非常有趣，所以我们在下一个视频中见。
So this guy this is going to be very interesting so see you in the next video.