Volume 13, Issue 3 -

The need for linear model, of the nonlinear system, while tuning controllers limits the use of classic controllers. Also, the tuning procedure involves complex computations. This is further complicated when it is necessary to operate the nonlinear system under different operating constraints. Continues Stirred Tank Reactor (CSTR) is one of those non-linear systems which is studied extensively in control and chemical engineering due to its highly non-linear characteristics and its diverse operating range. This paper proposes two different control schemes based on reinforcement learning algorithm to achieve both servo as well as regulatory control. One approach is the direct application of Reinforcement Learning (RL) with ANN approximation and another is tuning of PID controller parameters using reinforcement learning. The main objective of this paper is to handle multiple set point control for the CSTR system using RL. The temperature of the CSTR system is controlled here for multiple setpoint changes. A comparative study is also done between the two proposed algorithm and from the test result, it is seen that direct RL approach with approximation performs better than tuning a PID using RL as oscillations and overshoot are less for direct RL approach. Also, the learning time for the direct RL based controller is lesser than the later.

For the real time monitoring of a Reactive Distillation Column (RDC), a Recurrent Neural Network (RNN) based soft sensor has been proposed to estimate the bottoms product composition of the RDC for the synthesis of n -Butyl Acetate using esterification reaction. This soft sensor acts as a measuring element in a closed loop involving a PI controller for the direct control of the RDC’s product concentration. The RNN acts as a dynamic network, which works on the sequential input data and output data with a recurrent connection. While using the RNN based soft sensor in the open loop, it has been observed that the sensor estimated the composition of butyl acetate in the bottoms with such an accuracy that it can be used for the control purpose. Closed loop results demonstrated that the system has been showing precise controlled results and soft sensor is showing small prediction Mean Square Error (MSE) when disturbances in feed flow rate and set point changes are introduced.

Alcohol based biofuels, such as bio-butanol, have considerable potential to reduce the demand for petrochemical fuels. However, one of the main obstacles to the commercial development of biological based production processes of biofuels is end-product toxicity to the biocatalyst. We investigate the effect of end-product toxicity upon the steady-state production of a biofuel produced through the growth of microorganisms in a continuous flow bioreactor. The novelty of the model formulation is that the product is assumed to be toxic to the biomass. The increase in the per-capita decay rate due to the presence of the product is assumed to be proportional to the the concentration of the product. The steady-state solutions for the model are obtained, and their stability determined as a function of the residence time. These solutions are used to investigate how the maximum yield and the reactor productivity depend upon system parameters. Unlike systems which do not exhibit toxicity there is a value of the feed concentration which maximises the product yield. The maximum reactor productivity is shown to be a sharply decreasing function of both the feed concentration and the toxicity parameter. In conclusion, alternative reactor configurations are required to reduce the effects of highly toxic products.

Recently, we have proposed a new formulation approach for control degree of freedom (CDOF) analysis of process systems. This formula interrelates the CDOF and elements of a process flow diagram (PFD). The formulation is accurate, easily applicable by process engineers, and needs little prior knowledge of the process under study. This paper describes further about this new formulation and its advantages by applying it to batch and cyclic processes. The results demonstrate the correctness of the method in  determining CDOF for all case studies at both steady state and dynamic conditions. The CDOF for “reheat regenerative Rankine” and “vapor absorption refrigeration” cycles have been determined to be 9 and 8, respectively. The steady state CDOF values for the Rankine and refrigeration cycles have been calculated as 8 and 4, respectivley. Also the CDOF for the process of  “heat pump integrated with batch distillation column”  is 4 that verifies the suggested formula of CDOF. The method also gives beneficial insights about the manipulated variables (MVs) in a control system.

Transportation of natural gas from gathering station to consumption centers is done through complex gas pipeline network system. The huge cost involved in transporting natural gas has made pipeline optimization of increased interest in natural gas pipeline industries. In the present work a lesser known application of Ant Colony in pipeline optimization, has been implemented in a real gas pipeline network. The objective chosen is to minimize the fuel consumption in a gas pipeline network consisting of seven compressors. Pressures at forty-five nodes are chosen as the decision variables. Results of Ant Colony Optimization (ACO) have been compared with those of GAMS that utilizes ‘Generalized gradient principles’ for optimization. Our results utilizing ACO show significant improvement in fuel consumption reductions. Similar procedures can be adopted by researchers and pipeline managers to help pipeline operators in fixing up the pressures at different nodes so as the fuel consumption in compressors gets minimized.

A novel method for simulating the torque and temperature curves from a torque rheometer thermal stability test on poly(vinyl chloride) (PVC) was developed. A mathematical model was proposed which combines the chemical kinetics involved in the thermal degradation and stabilisation process of PVC with heat transfer and viscosity relationships within the torque rheometer. The model coefficients were fitted to data obtained from previous experiments with a program written in the Python programming language using the Levenberg-Marquardt Algorithm (LMA) with multiple starts. The mathematical model fits the torque rheometer data successfully depending on the characteristics of the torque curve. The model fit is successful when the torque curve follows the expected behaviour. An unsuccessful model fit occurs when the torque curve deviates from the expected shape in a very specific way. In the degradation phase the torque curve increases to a local maximum, decreases to a local minimum and then increases again. The exact reason for the dip in torque is unknown.