After 1 full year of working in a food processing plant, I have to admit that there are a lot of interesting design regarding how to make a perfect batch dosing.
Let's take my simple design:---> pump ----> flowmeter ----> valve
My first concept to dosing was a naive design statement:
- start the dosing system
- if totalizer count is enough, stop the dosing system (pump and valve)
That design statement works relatively fine in theory, but, theory is not actual work. With valve having opening and closing time of a second or two, and pump having ramp up and ramp down time of 5 secs or so, the system will always overdose.
To solve this, I have seen in multiple implementation something like:
- if totalizer reach certain point, slow down the pump, and when totalizer reach target, stop pump and valve
of
- if totalizer reach certain point 1, slow down the pump, and when totalizer reach certain point 2 before target, stop pump and valve
That certain point normally determined by a fixed value, normally by trial and error from the 1st statement. Example:
- I need 50L
- After first statement, the system will consistently give me 50.3 ~ 50.6
- I will stop my valve earlier at 49.6, so the final dose will be 49.9 ~ 50.2
This is how normally predictive 1 stage or 2 stage dosing work. This, was due to valve and flow rate behaving consistently (say, valve closing time is 0.3 secs, and flowrate is within 1L/secs)
This method have a major weakness: if system is not well maintained, pump and valve might not behave that consistent overtime (valve close slightly slower, and pump slightly weaker), thus, the dosing system need to be constantly readjusted. Personally, I hate this design, because our plant have bad water supply design, and constantly affect the water dosing system.
I have done some improvise on that methods:
- instead of a fixed cut off point, introduce a dynamic cut off point
- instead of fixing the opening / closing time of valve, I introduced active monitoring from valve feedback, to measure closing time
- using measured flow rate from flowmeter as reading for dynamic cut off
The cut off point will be (valve closing time* flowrate).
In the system I have just finished commissioning, this formula works wonderful. A batch is well controlled within only 1 reading from flowmeter (which is 0.2L), no matter what batch size, no matter what flow rate (which is somewhere in between 600-2300L/h), and does not matter which valves ( I have 7 dosing valves sharing 2 flowmeters and only 1 pump).
This cut off also works better than previous methods, due to not slowing down dosing, thus, save times.
Let's take my simple design:---> pump ----> flowmeter ----> valve
My first concept to dosing was a naive design statement:
- start the dosing system
- if totalizer count is enough, stop the dosing system (pump and valve)
That design statement works relatively fine in theory, but, theory is not actual work. With valve having opening and closing time of a second or two, and pump having ramp up and ramp down time of 5 secs or so, the system will always overdose.
To solve this, I have seen in multiple implementation something like:
- if totalizer reach certain point, slow down the pump, and when totalizer reach target, stop pump and valve
of
- if totalizer reach certain point 1, slow down the pump, and when totalizer reach certain point 2 before target, stop pump and valve
That certain point normally determined by a fixed value, normally by trial and error from the 1st statement. Example:
- I need 50L
- After first statement, the system will consistently give me 50.3 ~ 50.6
- I will stop my valve earlier at 49.6, so the final dose will be 49.9 ~ 50.2
This is how normally predictive 1 stage or 2 stage dosing work. This, was due to valve and flow rate behaving consistently (say, valve closing time is 0.3 secs, and flowrate is within 1L/secs)
This method have a major weakness: if system is not well maintained, pump and valve might not behave that consistent overtime (valve close slightly slower, and pump slightly weaker), thus, the dosing system need to be constantly readjusted. Personally, I hate this design, because our plant have bad water supply design, and constantly affect the water dosing system.
I have done some improvise on that methods:
- instead of a fixed cut off point, introduce a dynamic cut off point
- instead of fixing the opening / closing time of valve, I introduced active monitoring from valve feedback, to measure closing time
- using measured flow rate from flowmeter as reading for dynamic cut off
The cut off point will be (valve closing time* flowrate).
In the system I have just finished commissioning, this formula works wonderful. A batch is well controlled within only 1 reading from flowmeter (which is 0.2L), no matter what batch size, no matter what flow rate (which is somewhere in between 600-2300L/h), and does not matter which valves ( I have 7 dosing valves sharing 2 flowmeters and only 1 pump).
This cut off also works better than previous methods, due to not slowing down dosing, thus, save times.