widelto
Member
Hi guys:
This one goes to PID gurus.
I got a system where i got to cool down edible oil in a tank from 70 degrees centigrade down to 18 degrees following following a degrees per hour pattern. The tank has a jacket on the outside and a coil in the inside, the jacket and the coil have cold water used to cool down oil temperature. Water that runs inside the jacket and coil comes from a pump that recirculates, on the suction side of this pump there is a circuit of cold water 6 to 8 degrees centigrades from a chiller that is used to enter cold water into the pump circuit thru a proportional valve in order to mantain water on the jacket cooler every time in order to cool down oil temperature, when there is an excess of water in the jacket and coil this falls by gravity to another tank then this watter is cool down thu a tower and sent back to the chiller.
The last two years I´ve been controlling this system thru a compact logix plc (PIDE) and used a Delta of temperature between oil and water jacket , oil is always hotter than water jacket temperature so the delta is always positive, this delta is used as a setpoint for the PIDE and the PV is the diference between oil and water jacket temperature measured thru RTD´s, the CV of the PIDE is sent to the proportional valve I mentioned before, with this approach I can not control directly degrees per hour pattern but I can change the Delta and as a result the changes is degrees per hour, the bigger the delta the faster the changes in degrees per hour.
This week I changed my mind and try to use a PIDE cascade approach where the primary loop is the one that has oil temperature as the PV and a Setpoint that follows the degrees per hour pattern from a table, the secondary loop has the ouput from the primary as the SP and the jacket water temperature as the PV and the CV is connected to the proportional valve that controls the cold water entering into the water pump circuit, for some reason the secondary loop is saturated ( i.e. always 100 %) and the proportional valve is 100% opened and my approach do not work properly.
I disconnected the second loop from the primary and use only the first one as the only controller and connect the proportional valve to its output, this way the system works fine but I have no control over water jacket temperature and its RTD is not used.
I guess that my error was to use the water jacket temperature as the PV for the second loop since this temperature is going down together with oil temperature.
This system is called fractionation or winterization.
Any comments are welcome.
William Delatorre
This one goes to PID gurus.
I got a system where i got to cool down edible oil in a tank from 70 degrees centigrade down to 18 degrees following following a degrees per hour pattern. The tank has a jacket on the outside and a coil in the inside, the jacket and the coil have cold water used to cool down oil temperature. Water that runs inside the jacket and coil comes from a pump that recirculates, on the suction side of this pump there is a circuit of cold water 6 to 8 degrees centigrades from a chiller that is used to enter cold water into the pump circuit thru a proportional valve in order to mantain water on the jacket cooler every time in order to cool down oil temperature, when there is an excess of water in the jacket and coil this falls by gravity to another tank then this watter is cool down thu a tower and sent back to the chiller.
The last two years I´ve been controlling this system thru a compact logix plc (PIDE) and used a Delta of temperature between oil and water jacket , oil is always hotter than water jacket temperature so the delta is always positive, this delta is used as a setpoint for the PIDE and the PV is the diference between oil and water jacket temperature measured thru RTD´s, the CV of the PIDE is sent to the proportional valve I mentioned before, with this approach I can not control directly degrees per hour pattern but I can change the Delta and as a result the changes is degrees per hour, the bigger the delta the faster the changes in degrees per hour.
This week I changed my mind and try to use a PIDE cascade approach where the primary loop is the one that has oil temperature as the PV and a Setpoint that follows the degrees per hour pattern from a table, the secondary loop has the ouput from the primary as the SP and the jacket water temperature as the PV and the CV is connected to the proportional valve that controls the cold water entering into the water pump circuit, for some reason the secondary loop is saturated ( i.e. always 100 %) and the proportional valve is 100% opened and my approach do not work properly.
I disconnected the second loop from the primary and use only the first one as the only controller and connect the proportional valve to its output, this way the system works fine but I have no control over water jacket temperature and its RTD is not used.
I guess that my error was to use the water jacket temperature as the PV for the second loop since this temperature is going down together with oil temperature.
This system is called fractionation or winterization.
Any comments are welcome.
William Delatorre
