The DCS is an extremely important piece of equipment that can be called the brain of a chemical plant.
Operators also tend to specialize in DCS, and the era is truly DCS!
That was a long time ago (lol)
This technology is almost commonplace now.
Even though DCS does not involve a machine shop, it is very important for a factory.
It is also a problem for machine shops who want to understand the equipment comprehensively that they have been avoiding it for a long time.
If you understand just the outline, it may be easier to understand when talking to an instrumentation engineer.
The software aspect may seem difficult, but I think it will be easier to understand if you can imagine manual operation.
It will be easier to understand if you imagine manual operation on site.
- Sequence flowchart
- block diagram
- relay timer counter
- scan time
A sequence describes how to operate the automated parts of equipment.
The more highly automated a factory is, the more information you will have on how to operate the equipment.
The first thing a machine shop does to learn how to operate equipment is
Interview with the manufacturing department
This certainly gives you an overview.
However, the manufacturing department takes up a lot of time.
It’s the same scenario as a manufacturer’s sales representative casually interviewing, visiting, or making phone calls.
The manufacturing department is extremely busy and doesn’t have time to just deal with inquiries from machine shops.
Many machine shops don’t even understand this obvious fact.
Machine shops are so busy that they don’t have time to attend to trivial phone inquiries from manufacturer sales representatives.
That’s the same thing.
- Listen by email if you can listen by email
- You can understand to some extent how the equipment operates by looking at the sequence.
If you don’t study sequences, you won’t know how to operate the equipment, and you won’t be able to use it for verifying equipment capabilities or analyzing failures.
Studying sequences is essential for chemical plant mechanics!
Let’s use the example below.
If this is the case, you can predict the movement without looking at the sequence.
- filling the tank: With valve B closed, check that the liquid level is lower than switch A, then open valve A and fill the tank. When the liquid level reaches switch A, close valve A.
- draining liquid from the tank: With valve A closed, check that the liquid level is higher than B, then open valve B and drain the liquid from the tank. When the liquid level reaches switch B, valve B is closed.
Is this the idea from a batch standpoint?
Here is an example flowchart that corresponds to this flow.
The left side is the flowchart corresponding to the flow.
Although this is sometimes called a sequence, it is actually a flowchart.
The first thing to keep in mind here is what state it is in at all times.
The route on the right is constantly moving.
The condition for operating this sequence is a push button.
When a person presses the start button, driving begins.
It’s exactly the same as the world of games and computers.
If you press the start button, valve A will open.
Although this example is simplified, various conditions are actually checked, such as the condition of the liquid level in the receiving tank and the liquid level in the source tank.
If we rely on humans to do this, we can only save labor by having a machine open and close valve A instead of humans.
Valve A is so large that it cannot be turned by human power. This is a case like this.
When the liquid level reaches switch A, it will automatically try to close valve A.
This is the same behavior I expected.
Rather, this is probably the biggest point of automation.
There is no point in automation if a person monitors the amount of liquid in the tank and closes the valve when it reaches a certain level.
Tables are used to create a general representation for machines in order to translate the flowchart movements into sequences.
The notation is as follows.
The conditions are written on the top, and the operations are written on the bottom.
The step number is written on the horizontal axis.
Step number starts from 1 in order.
If the condition of pressing the start button Y at number 01, that is, pressing the start button, is met,
Operate valve A to Y, that is, open valve A.
You can perform conditional branching by stating which step to proceed to when the condition is met or not.
In this example, the next step number will not be written because the process will proceed to the next step regardless of whether the start button is pressed or not.
Then, let’s look at step number 02.
Also in No. 02, look at switch A’s Y, and then set valve A to N.
Using this representation, it is easy to convert it into a control language as a sequence.
When we say sequence, is it more common to refer to this table?
However, visual flowcharts are preferred in the field because they are not visual for humans.
You can think of PID as a control function.
A function is a mathematical expression that outputs an output through a function in response to an input.
y = f(x)
It is meant to emphasize the idea of
Technically, this is called control action.
It uses a combination of three functions: P, I, and D.
We will briefly explain each of P, I, and D , but first we will explain on-off operation, which is the basis of control operation.
On/off operation is the basic of control !
Although it is often overlooked by people who have studied control engineering,
In actual driving, this is over 90% !
This is because the automatic ball valve is on-off.
The on/off operation is a pattern like the one below.
- ON if the temperature is below a certain value
- OFF if temperature is above a certain value
It is a mechanism that activates if the device deviates from a certain setting range.
An easy-to-understand example of temperature control would be an electric kettle that heats water using electrical resistance.
In batch chemical plants, there are not many examples where temperature, pressure, flow rate, and liquid level are used as judgment conditions for on/off operation.
In overwhelmingly many cases, valves open and close according to a sequence when a predetermined condition is met.
- When you press the start button, there is a waiting period and then the valve opens.
- Press the local button to start the pump, and the valve will open after a certain amount of time.
This is a big difference from a continuous factory.
P is proportional operation
Proportional operation is an operation that returns a value proportional to the deviation.
Using the control variables PV/MV/SV, it can be expressed as follows.
MV = α * (SV-PV)
The general behavior of P operation when SV-PV suddenly steps up at a certain timing is shown below.
The above graph can be qualitatively expressed as follows.
- At the moment when SV-PV rises above the step, SV-PV changes significantly, so α * (SV-PV) also changes significantly.
- Since PV gradually approaches SV, α * (SV-PV) gradually decreases.
- It moves with a certain amplitude and tries to settle down to a certain value.
- Eventually it will settle at a location offset from the initial value. This is related to α * (SV-PV).
a person sits on a chair
A familiar example would be a chair.
The initial value is when no one is on the chair, and when a person sits down on it
After the chair springs are released, the chair will settle down slightly below its initial value.
Same as this.
In the case of a chair, the spring force of the spring is
f = k(x)
This is because it is proportional to the displacement, and it is exactly a proportional operation.
Offset and amplitude
Increasing the sensitivity of the proportional action P is the same idea as increasing the force of the spring, but a little more care is required in terms of control.
- Increasing the sensitivity of proportional action will reduce the offset.
- The amplitude of the fluctuation becomes larger and the convergence becomes slower.
An offset means that there is always a difference between SV and PV.
Even if you want to control the inside of the system at 60℃ and set SV=60℃, it will settle at PV=58℃ with only proportional operation.
Since this deviation varies depending on the system, determining how much SV should be set becomes a world of tricks and tricks.
Automatic control will try to eliminate human judgment, but it will be a system that incorporates human intuition and tricks.
This is not good.
Increasing sensitivity also means applying greater control to even small changes.
If you touch it even a little, it will make a big noise and shake it around.
Trying to understand this using mathematical formulas is a pattern that people who studied control engineering in university fall into.
The people who try their best to understand proportional bands and proportional constants are those who have studied hard for exams.
I is integral action
Integral operation is an operation that returns an amount proportional to the value obtained by integrating SV-PV over time.
Let’s consider the integral operation using a pattern in which SV-PV changes in steps as an example.
Since SV-PV is increasing in steps, it means that the amount of integral operation continues to increase when integrated over time.
In reality, there are no examples of this behavior.
This is because the moment the integral action is applied, the system acts in the direction of alleviating SV-PV.
The same is true for proportional action examples.
Changes in steps can be discussed using mathematical formulas in the field of control engineering, and their characteristics are easy to understand.
However, I think that there are almost no systems in chemical plants that exhibit such behavior.
Integral action removes offset
It is possible with integral action to solve the problem of offset that occurs in proportional action.
This is clear from the graph above.
Even if the deviation SV-PV continues to remain as an offset,
In order to add movement proportional to time,
As time passes, the system tries to calm down.
Then, SV-PV becomes 0.
This is the aim; the integral action cannot be eliminated.
D is differential operation
Integral operation is an operation that returns an amount proportional to the time change value of SV-PV.
Care must be taken because the control is proportional to the amount of change in deviation !
It is not used in batch chemical plants unless there is a special need!
- The system is not as stable as in a continuous factory. Stability time is short
- In chemical reactions, the reaction time and reaction amount are not uniform, so there is a large fluctuation.
- If the control is sensitive to vibration, a failure will result in a major accident.
Risky differential behavior is something to avoid in chemical plants.
A block diagram is used to show the overall picture of process control using a controller.
Let’s simulate the block diagram and show it as below.
The block diagram itself shows the control logic in control engineering, so it is expressed using characters and mathematical formulas.
This was reused as a tool to express functional elements.
Detection → Adjustment → Operation
The block diagram above has many elements, so let’s take out only the main parts.
- Detect process target
- Adjust using detected value
- Manipulate adjustment values for process systems
This detection → adjustment → operation is the main part.
Internal functions of adjustment
Among these detection → adjustment → operation functions, adjustmentThe block diagram above shows a more detailed view of this part.
- Load the detected value PV into the regulator reception
- SV which is the target value setting
- The difference between target value SV and detected value PV comparison
- Using the PV-SV value adjustment
- to the control panel Outgoing
The flow is reception → setting → comparison → adjustment → sending .
Here, the values PV/MV/SV will appear.
Adjust the control part
The part that becomes the control function in the actual block diagram is the part called adjustment in the block diagram above .
Convert SV-MV to PV by doing some conversion
This is the element. Mathematically
PV = f(SV-MV)
It will look like this.
The comparison part is conscious of sequence control.
Control is performed if PV is within a certain range
Because that’s common.
- If the temperature PV is higher than the certain range SV, cooling
- If the temperature PV becomes lower than the certain range SV , heating starts.
Comparison is the part that makes such judgments.
In the world of DCS, several alphabetic abbreviations appear.
If you listen to the words without knowing the meaning, you won’t feel like you’re listening to Japanese.
I will limit my introduction to what even a machine shop should know at least.
PID control is a famous story that appears in classical control at universities.
- P proportional control Proportional
- I integral control Integral
- D Differential control Differrntial
University studies focus on the academic aspect.
PID is used for temperature control, drip control, and vacuum control in batch chemical plants.
In practice, in a batch chemical plant, P is 90%, I is 10%, and D is less than that.
It is recommended to change only P for ease of understanding, especially ease of fine adjustment for fluctuations in operating conditions.
Regardless of I, changing D can instantly cause big trouble.
I will write an article about this on another occasion.
This is a DCS term.
- PV current value Process Value
- MV Manipulated Value
- SV target value Setting Value
This is not limited to control, but is a management value that is necessary for all DCS control, including operation instruction values.
For daily operation management, just watching the PV is enough.
SV is a parameter that is easy to change.
If you adjust the parameters during prototype manufacturing, it will basically be fixed.
PID I and D are fixed reliably, but SV is useful as a parameter to change when changing daily operation.
I would like to control the temperature at 20℃, but since the concentration is rising and the flow rate cannot be changed, the reaction temperature will rise.
Therefore, we changed the temperature control to 19.5℃.
This is the pattern.
I don’t usually watch MVs.
This refers to the actual valve opening during automatic control.
With the adjustment valve, set MV to 0% to force the valve to close, and set MV to 100% to force it to open.
I often see this scene.
You can think of this as indicating the movement of the automatic valve cylinder.
This shows what happens when air is no longer supplied.
- DA normal operation Direct Action Closes when air stops
- RA Reverse Action Opens when air stops
For instrumentation engineers, it is understood as a combination of cylinder operation and valve operation.
From the driver’s perspective, it doesn’t matter.
On the contrary, it’s annoying.
Overall, it is important which direction it acts on.
In an emergency when the air stops, the cylinder is DA and the valve is RA, so RA!
There is no need for information that would confuse your decision.
In a similar story, there is the issue of zero cuts.
In any case, this is a scene where you can get a glimpse of the delusions of an instrument engineer who doesn’t know the actual worksite.
relay timer counter
When I think of sequences, I think that the real thing is almost a black box.
Only control experts will understand.
However, if you limit it to its constituent elements, it is described in various textbooks.
For example, the following components are included in a sequence circuit.
Because these words are difficult to associate with actual control,
In the first place, people don’t want to look at this kind of notation in textbooks.
However, only the relay, timer, and counter in this
This allows considerable control over batch chemical plants.
Let me introduce it to you.
Nitrogen replacement with timer
Nitrogen replacement is always necessary when starting production.
This does not change whether batch or continuous.
Change the air in the system to a nitrogen atmosphere.
The standard is to continue until the oxygen concentration falls below a certain percentage.
In reality, the operational control value is how long nitrogen replacement is performed.
The timer comes in handy here.
Since it’s a job that only happens a few times a year, I didn’t create it as a sequence, and just did it with a stopwatch.
In reality, it’s a matter of deciding what time to start and what time to start replacing the gas with nitrogen.
Relay is a judgment condition
Relays can be used for a wide variety of judgment conditions.
For example, the nitrogen replacement example mentioned earlier.
What should I check before opening the nitrogen valve?
The control valve to be replaced with nitrogen is closed.
This is required.
Otherwise, sufficient nitrogen replacement may not be possible.
If the control valve has an automatic valve.
After clearing the judgment condition that the automatic valve is closed
You can open the nitrogen valve.
You can use a relay here.
Converts the automatic valve open/close signal into an electrical signal and inputs it to the relay.
It can be used as a switch in an electric circuit, or in other words, as a judgment condition.
The person who controls the relay controls the control of the batch chemical plant.
You can say that.
The counter is after the process
This is not very much.
However, if you use a counter, you can transfer the data to the next process.
Is there a pattern where you can manually intervene without using a counter?
For example, let’s say there are five processes performed in a mixing tank in one day.
There was a problem in the middle of the third one.
From there, I proceeded manually until the end of the third step.
After this, how can I automatically proceed from the 4th step again?
The idea is that once you restart the program and set the counter to number 4, you’re good to go.
You can also skip this manually.
If there are many steps, the number of skips will increase, so it’s just a hassle.
If there are few problems in the first place, there is no need to rely on such a counter.
Although it is a bit technical, I would like to introduce the concept of DCS scan time.
Although the information is only updated once per second for more than 1000 instruments, it can be important depending on the situation.
Just knowing the order is enough.
Instrument score is over 1000 points
How many instruments are used in one plant?
There are easily over 1000 points.
There are about 30 input/output signals for one tank.
Each tank has about 10 automatic valves, one thermometer, one liquid level gauge, and one flow meter.
Automatic valves have two input and output points for one valve, inputting the signal for the automatic valve’s open/closed status, and outputting the signal for opening/closing operation.
Thermometers and liquid level gauges only require one point by using the status signal as an input.
Since the flowmeter controls the flow rate, there are two points in total: the flow rate status signal is input, and the flow rate control signal is output.
In addition to this, if you include information such as current and rotation speed, there are about 30 points.
There are over 20 tanks, so 30*20=600.
In addition to tanks, there are other equipment necessary for operation, such as refrigerators and centralized abatement equipment.
There are over 1000 input and output points.
DCS aggregates control
DCS is a collection of control parts to solve this problem with analog instruments.
It is impossible to achieve a score of 1000 points or more using one control device.
So I decided to distribute it to several control devices.
We recognize that it is a distributed control system.
Although it is intended to be distributed, in reality, the information from each instrument is aggregated.
DCS cannot be controlled at once
The DCS aggregates the control of each instrument.
Using the word “aggregated” gives the impression that it is a convenient machine that can process over 1000 instruments at the same time.
Each signal is scanned periodically.
- Signals with more than 1000 points are numbered sequentially starting from 1.
- Scan the first instrument at 0.000 seconds after starting scanning
- Scan second instrument at 0.001 seconds
- Scan the 1000th instrument at 0.999 seconds
- Rescan first instrument at 1.000 seconds
In this way, the scan is repeated every 1.000 seconds.
Scanning is easy to imagine as detection, but
1 second scan for each detection, control, and operation.
No monitoring outside of scanning hours
Scanning every 1.000 seconds means
From 0.001 to 0.999 seconds, no information from that instrument is being monitored.
The idea behind DCS control is that this is a good idea.
From the perspective of mechanical engineering masters, in order to measure physical quantities per unit time.
You may also use data changes of 0.1 seconds.
In terms of mechanical engineering, the temperature changes by 0.1°C in 0.1 seconds, so the idea is 1°C/second.
This cannot be achieved in chemical plants that use DCS control.
In a chemical plant, the temperature changes by 2°C in 2 seconds, so we think of it as 1°C/second.
Just knowing that it’s that accurate
I think you can get a sense of the accuracy of data processing at a chemical plant.
The DCS operation screen has the following structure.
Think about the number of monitors you need
You may have a strong impression of the control room as shown below, with many DCS screens lined up.
There are a lot of monitors lined up.
How many monitors do you actually need?
Maximum of 2 operators
Typically, one DCS operator is required to operate a single plant.
One person operates the DCS, one works on-site, and the remaining one is free.
This movement is basic.
Some companies may rotate upper roles during one shift of the day, while others may have fixed rotations throughout the day.
Or, some companies may have fixed roles for a long time.
In particular, it depends on the number of people who have learned DCS operations.
In any case, in terms of plant operation, I think it would be a good idea to have a maximum of two people operating the DCS.
I forgot to mention this, but it is based on a batch chemical plant.
If it is a continuous factory, there will be fewer people on site, so there will be more people involved in monitoring.
However, I don’t think the number of monitors will change much.
This is a basic tool for checking the situation at the site.
In a batch chemical plant, one plant prepares about 20 graphic screens.
Operators always go through this group of screens at the beginning of every shift.
This is because it is easier to understand the situation than any other handover.
If the operating method has been determined from the beginning, it will be easy to remember which process has been reached at the time of handover.
All you have to do is figure out the deviations from there, so it only takes a few minutes to go around.
At least one of this graphic is always displayed.
2nd graphic or trend or overview
The screen to be used for the second image has not been decided yet.
Even though it’s free, I always use it.
- Graphics when processing multiple processes in parallel
- Trends if you want to focus on important processes
- Overview when taking over
I will use it differently like this.
More and more people are working with two computer screens side by side.
DCS operators have been doing this for a long time.
It doesn’t matter, but if you’re an equipment engineer at a chemical plant, you don’t need to use two monitors.
It would be convenient to have it, but it is not required.
I have various thoughts here as well.
The third one is a spare.
I don’t use it.
Or maybe it’s to monitor and operate when the third free person returns to the control room.
4th piece Engineering station
The fourth picture is the engineering station.
This is essential for DCS equipment.
However, DCS operators will have little opportunity to use it.
It is mainly used when producing new products.
When introducing a new product, the optimal parameters have not yet been determined, so parameters must be changed each time.
Without this station, the settings cannot be changed each time.
We have provided an overview of DCS that mechanical engineers should also know.
Sequence and flow・PID・Block diagram・Relay timer counter・Scan time・HIS
You won’t be able to understand specific driving operations unless you look at flowcharts and tables and visualize manual operation.
This time’s content is about the DCS system. It may not be directly usable in practice.
I think it would be a good idea to use it as reference information to deepen your understanding when you have time.
Please feel free to post any concerns, questions, or concerns you may have regarding the design, maintenance, and operation of chemical plants in the comments section. (The comment section is at the bottom of this article.)
*We will read all comments received and respond seriously.