Self-Tuning Microprocessor Universal Controller

Recent advances in microprocessor technology make it possible to implement on-line modern control algorithms at reasonable cost. Self-tuning regulators can, in general, be used to regulate any industrial process and this paper describes one such implementation on a microprocessor in three different software formats, i. e., 8-bits, 32-bit floating-point, and a modified version of the 32-bit floating-point. Studies of the process control specifications and the corresponding modifications required are described.     

Dynamic behavior of mixer-settlers

A mathematical model has been developed to simulate and predict the dynamic behavior of liquid extraction counter current operations in mixer-settler units. Murphree efficiency has been included in the model. The behavior of a multistage glass mixer-settler has been studied under start-up transient conditions. Very good agreement between the simulation results and the experimental findings have been obtained. The simulation results showed that this model can efficiently be used to determine the optimum number of stages, the capacity of each stage, the phase ratio and the time required for each unit to attain steady state conditions. Other controlling parameters in operating such units are investigated.     

On-line adaptive optimal load flow

This paper presents a new on-line approach for solving the optimal power flow problem without sacrificing the accuracy and constraints of the system components. A procedure which merges non-linear programming (NLP) and linear programming (LP) techniques is utilized to make use of the merits of both algorithms. The NLP and LP form a closed loop with a decision maker (DM) which directs the solution algorithm either to NLP or LP according to the analysis of the permissible errors during the calculation process. The input to the closed loop is fed either from the telemetered data or the state estimator.

To adapt the LP algorithm, the objective function and the system constraints are linearized, using Taylor expansion series, around the exact solution obtained as a result of employing the NLP technique. A new linearization algorithm is presented.

The proposed approach has been applied to a 19-bus system and the results are judged satisfactory. Its use as an on-line computational tool for power system operation and control seems very promising.