Emmerson1.docx

ABSTRACT The objective of the Emerson 1 Experiment is to demonstrate an open loop test, which the outcomes from the test can analyse the response rate (RR), dead time (𝑇𝑑 ) and time constant (𝑇𝑐 ). At the beginning in the experiment, the Gain and Reset values is set to 4.00 and 20.00. Then, the controller is set to manual mode and the manipulated value (MV) is set from range 30-40%, the MV value must not exceed than 10%. After that, the experiment is stabilised by using automatic mode and at the end of the experiment a graph is plotted. From the graph, the RR, Td, P, and I value are calculated. The Ziegler-Nichols formula is used in the calculation. The load disturbance test follow up after the calculation. P value stands for Gain while I value for Reset.

INTRODUCTION A process control loops consists of controller, final control element, process and sensor. There are two types of process control loops which are closed loop system and open loop system. The closed loop system occurs when all of the elements in the process control loops are connected, while open loop system occurs when any of the elements is disconnected. The function of the closed loop system is to compares the process measurement signal (PV) from the set point (SP) and makes necessary corrective action to the final control element. In the Emerson 1 Experiment that we conducted, the open loop system is performed. The open loop system is a process which the controller has no control over the final control element. In the other words the final control element are conducted manually by an operator. In industrial practice, the process becomes a closed loop system when the controller is set to automatic mode. However, it becomes open loop system when the controller is set to manual mode. In the experiment, in order to get the desired set point (SP), the controller must be set to automatic mode. Automatic control is performed by a set of mathematical equations which consists of proportional (P), integral (I), derivative (D), set point (SP) and process measurement (PV). This automatic control is also known as controller algorithm. It is required for the computation of controller action (MV). For example, a controller action of “100” may represent a full opening of a control valve which is equivalent to the maximum flowrate. Some techniques to find the optimum P, I, and D values have been introduced. One of the technique is open loop method where it started by performing an open loop test. The result obtained from the test is analyses for response rate (RR), dead time (𝑇𝑑 ) and time constant (𝑇𝑐 ) which will be used in the tuning rule to obtain the calculated optimum P, I and D values. The calculated optimum P, I and D shall be tested for the actual performance in handling a

change in set point and a change in process loading of load variable. The controller must first be set to automatic mode to perform the tests. The fine tuning of the P, I and D is required if the process response to any of the tests does not meet the desired settling area. Furthermore, in open loop process system, it can be self regulating process or non self regulating process. The self regulating process is used for flow, temperature, pressure and pH. While for non self regulating process, it usually involves level control system. To differentiate, the self regulating process has a final steady state, PVf while the non self regulating process does not have final steady state. In the open loop test, the analysis about the RR value (response rate), death time (Td) and time constant (Tc) values are made. There are two ways to calculate this values which are graphical and numerical. Graphical method is divided into two paths, where the first one is using tangent line. Tangent line was drawn at the maximum gradient of the open loop process response curve. The second one is reformulated tangent method. It is a little bit similar with tangent method, but the difference is RR, Td and Tc is in trigonometric function. The angle must be calculated to obtain the data. For numerical method, we have discrete tangent method. In this method, data can be obtain using three ways which are distributed control system (DCS), supervisory control and data acquisition (SCADA) and paperless recorder. In addition, after performing the tests, the P and I value must be calculated by using table tuning rule of Ziegler-Nichols. Besides this table, we also have tuning rule table of CohenCoon and Takahashi. For tuning rules by Chien, Hrones & Reswick, it have three tables which each table is differ from the value of the overshoot and its operation.

The demonstration of an open loop test is made in this experiment, which the outcomes from the test can analyse the response rate (RR), dead time (Td) and time constant (Tc). But before that, a graph of the open loop test must be plotted so that a tangent method can be applied in order to obtain the data for the SP test. After print out the graph, a straight vertical line must first be drawn at MV line. Then, a tangent line is drawn to get the steepest at the maximum loop of the open loop curve. By using protector, the angle is read (appendices 1). From the line, its length is read to calculate its death time (Td). Td= Td (length) x b a and b values is calculated at which a is PV scale and b is Time Scale, where, a= PV/length of PV (mm) b= time/length of the time (mm) Then, from the graph, ∆MV is calculated in order to obtain RR value. The formula for RR value is, tan 𝜃 𝑎 × ∆𝑀𝑉 𝑏 Next calculate Pb and I value using formula from tuning rule of Ziegler-Nichols. MODE

P

I

P

100 RR Td

PI

111.1 RR Td

3.33 Td

PID

83.3 RR Td

2 Td

D

0.5 Td

PROCEDURE 

Performing open loop test

1. The process was stabilized either in manual or automatic mode. 2. The controller was switched to manual mode after the controller in automatic mode when the process has stabilized. 3. The value of Gain and Reset were recorded as 4.00 and 20.00 respectively. 4. Then, the initial value of manipulated variable (MV) was set as 35% (the manipulated variable must be from range 30% to 40%). 5. Then, the controller was switched back to automatic mode in order to stabilize the graph. 6. The graph was printed. 7. The a, b, Td, RR, I, P and 𝑇𝑐 values were calculated. 

Performances Tests (Change in Set Point)

1. The controller was set to automatic mode. 2. The Gain and Reset value were recorded by inserting the P and I values obtained as PGain and I-Reset. 3. The controller was switched to manual mode and the MV value is recorded as 40%. 4. The process was left alone for about 3 seconds to allow the process to work properly. 5. Then, the controller was switched again to automatic mode to stabilize the graph. 6. The graph was printed. 

Change in Process Loading of Load Variable (Load Disturbance Test)

1. The controller was set to manual mode. 2. A change of MV was made of about 5 to 10 %. 3. Then, we waited for about three seconds. 4. The controller was set to automatic mode.

RESULT, DISCUSSION AND SAMPLE CALCULATION Based on the graph in appendices 1, after drawing the tangent line, the angle of the triangle form is calculated. A 24° angle is obtained. To calculate the a value, the reading from 40% to 50% is taken and the length of the two points were measured as 17 mm. The formula to calculate the a value is: a= 10% / 17 mm = 0.59 % / mm Then, the value of b is calculated. The time interval during plotting the graph is one minute. Based on the graph, the scale for time interval is four. It means that each scale represents 15 s. Length for each scale is 7 mm. The formula used to calculate b value is: b= 15 s / 7 mm = 2.14 s /mm Next, the length from tangent line to the straight line on MV line was measured. The length obtain is 1 mm. the formula to calculate the value of death time, Td is: Td= Td (length) x b =1 mm x 2.14 s / mm = 2.14 s After that the response rate (RR) value is calculated by using formula:

tan 𝜃 𝑎 × ∆𝑀𝑉 𝑏 Where ∆MV is the difference between the stable line and the loop on the graph. Based on graph in appendices 1, the ∆MV value obtained is 5 %. The formula to find RR is: RR=

tan 24° ∆𝑀𝑉

×

𝑎 𝑏

= (tan 24°/ 5) x (0.59/ 2.14) = 0.02 𝑠 −1

By using the formula from the tuning rule of Ziegler-Nichols, the value for P and I is calculated as below. P= 111.1 RR Td

= 111.1 (0.02) (2.14) = 4.76 % While for I, I= 3.33 Td = 3.33(2.14) = 7.13 s After performing the SP test, the load disturbance test must be conduct in order to ensure that the P and I values obtained from the experiment is valid in the process control. The load disturbance test can be conducted by following the procedure as in the procedure section. The graph plotted from the test was printed out. Comparing the SP test graph and load disturbance test graph, both of the graphs is almost the same in term of stability. This shows that the P and I value are correct and this values can be used in the process control. The load disturbance test graph can be seen in the appendices 2. However, some error occurs during conducting the experiment where the temperature rose up rapidly more than specified temperature. During the experiment, the specified boiler is 500°C, but it reached up to 800°C. Therefore, the temperature was cool down back by resetting the value of Gain and Reset to 4.00 and 20.00 respectively. This is also a result from parallax error where the observer took a wrong value of response rate (RR) which the outcomes are high values of P and I. After the graph stabilized back by changing the controller mode to automatic, the load disturbance test is performed. The graph during this error

occurs

is

shown

in

appendices

3.

CONCLUSION AND RECOMMENDATION Based from this experiment, we can concluded that even though this Emerson 1 experiment might looked like a simple experiment but it actually can lead to failure even by a slightest mistake. For example, an error likes parallax error which usually a very small value can effects on the graph. All of the values obtain in this experiment are all depends on the graph and the calculation such as response rate (RR), time dead (𝑇𝑑 ) and time constant (𝑇𝑑 ). Furthermore, after print out the graph, we must get the steepest and maximum operating loop tangent line. It is to ensure that we get the closest value and correct value in this experiment. We must not take easy on this aspect because a little difference in this aspect can make all the difference far away from the suitable values. We also must always be alert for any harm and take a proper action to control or avoid the harm. For example, in this experiment at which the temperature rises rapidly, if the observer did not notice the harm, the boiler might be overheated and explodes.