Development of a Multi-Controller Tuning Strategy and Rules for Oily Water Treatment Plant

Abstract

Abstract The Oil and gas industry is one of the biggest industries worldwide. Global oil output has increased dramatically in recent years. This surge in supply and production is the result of multimillion dollars investments on new exploration as well as enhancement of existing field in methods such as artificial lifting using gas or liquid injection to sustain the reservoir pressure. Evidently, these technologies have been successful in withstanding or increasing current oil production rates. However, as the reservoirs age, the water cut increases rapidly in conjunction with the increase in production. This water cut can only go up with the reservoir age but at different rates according to reservoir topology. As a result, there is a need to address the issue of handling this enormous amount of effluent water coming along with the crude. For that reason, Oil companies have established dedicated oily water treatment units for handling incoming flux. The importance of these units arises from economic as well as environmental considerations. From the environmental side, much legislation is in place to limit the oil in water content of the effluent discharge to the environment. Oil companies choose to abide by this legislation to avoid hefty fines. From the economical perspective, increased efficiency in separation means that more crude will be extracted and sent back for processing. Many methods are suggested for separation; however, the gravitational separation method is chosen in this research. Gravity separators are widely used in oil and gas. This approach relies on providing sufficient residence time for the mixture to split into different layers because of the different specific gravities. These separators are designed to provide optimum separation at a specific level. The role of the control system is to maintain the level in the tank as close as possible to the set point despite various fluctuations. In this research, an existing oily water treatment plant was mathematically modeled. This includes mainly the models of the tank, pumps, control valves and the disposable well head pressure. After that, the level control loop on the separator was tuned to provide optimum results. The unit output of separated water is injected to the sea basin via the offshore disposable water well head. Real field data was used in the simulations. The level loop was left to reach steady state and then tuned using the Modified Relay Feedback Test (MRFT). That is done through excitation of the system and measuring the frequency and amplitude of oscillations for each mode of operation to obtain the optimum controller gain. MRFT is considered as one of the non-parametric tuning methodologies which is very beneficial in oil and gas industries since the dynamics of the complex process do not need to be identified or mathematically modeled to tune the controller. The model was developed to illustrate the methodology of loops tuning in the considered oily water plant. The system consists of three pumps which are utilized as per flow requirements. The model was tested for all three modes of operation ranging from 1 pump to 3 pumps running at the same time. A Curve fitting tool algorithm was utilized to approximate the performance of the controller parameters in monotonous behavior. That was done across all modes of operation. The obtained approximation equations were used to control the process in the gain scheduling approach. In each mode, five inlet flow set points were applied to the system as a disturbance to the level process and the behavior of the level loop was observed. For each operating point, a set of controller parameters is obtained. Moreover, the obtained controller performance was compared with another controller obtained using Ziegler-Nichols closed-loop test to demonstrate the advantages of the derived controller. It is evident from simulation results that the gain scheduling controller results in varying disturbance of inlet flow. It is shown that the modified relay feedback test can be effectively used for controller tuning, with astonishing results.