Advanced teaching-learning-based optimization algorithm for actual power loss reduction

Received Jul 19, 2019 Revised Oct 6, 2019 Accepted Oct 20, 2019 In this work Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) is proposed to solve the optimal reactive power problem. Teaching-Learning-Based Optimization (TLBO) optimization algorithm has been framed on teaching learning methodology happening in classroom. Algorithm consists of “Teacher Phase”, “Learner Phase”. In the proposed Advanced Teaching-Learning-Based Optimization algorithm non-linear inertia weighted factor is introduced into the fundamental TLBO algorithm to manage the memory rate of learners. In order to control the learner’s mutation arbitrarily during the learning procedure a non-linear mutation factor has been applied. Proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) has been tested in standard IEEE 14, 30 bus test systems and simulation results show the proposed algorithm reduced the real power loss effectively.


INTRODUCTION
Reactive power problem plays an important role in secure and economic operations of power system. Numerous types of methods [1][2][3][4][5][6] have been utilized to solve the optimal reactive power problem. However many scientific difficulties are found while solving problem due to an assortment of constraints. Evolutionary techniques [7][8][9][10][11][12][13][14][15][16] are applied to solve the reactive power problem. This paper proposes Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) to solve optimal reactive power problem. Teaching-Learning-Based Optimization (TLBO) optimization algorithm has been framed on teaching learning methodology happening in classroom. Algorithm consists of "Teacher Phase", "Learner Phase". In the proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) non-linear inertia weighted factor is introduced into the fundamental TLBO algorithm to manage the memory rate of learners. In order to control the learner's mutation arbitrarily during the learning procedure a non-linear mutation factor has been applied. Preceding information gathering of learners is determined by the weight factor and through that new-fangled values are calculated. In a learning cycle individuals will try to explore various regions of the exploration space in initial phase. Afterwards individuals progress in a little range to regulate the trial solution to certain extent such that it can investigate reasonably little local space. Proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) has been tested in standard IEEE 14, 30, bus test systems and simulation results show the projected algorithm reduced the real power loss effectively.

PROBLEM FORMULATION
Objective of the problem is to reduce the true power loss: Voltage deviation given as follows: ( Voltage deviation given by: Constraint (equality): Constraints (inequality): (8)

ADVANCED TEACHING-LEARNING-BASED OPTIMIZATION ALGORITHM
Teaching-Learning-Based Optimization (TLBO) optimization algorithm has been framed on teaching learning methodology happening in classroom. Algorithm consists of "Teacher Phase", "Learner Phase" [17].
In ith learner the jth parameter is assigned values capriciously found by For the production "g" parameters of the ith learner are given by,

Teacher Phase
Creation of "g" ; mean parameter E g of each subject learners in the class is defined by, New set of better learners are found by Value of mean to be altered is decided by -teaching factor. Value of can be 1 or 2. ,

Learner Phase
For a specified learner ( ) a different learner ( ) is capriciously chosen( ). In the learner stage the X new is given as: In the proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) non-linear inertia weighted factor is introduced into the fundamental TLBO algorithm to manage the memory rate of learners. In order to control the learner's mutation arbitrarily during the learning procedure a non-linear mutation factor has been applied. Preceding information gathering of learners is determined by the weight factor and through that new-fangled values are calculated. T is number of iteration in single learning cycle. Then the inertia weight factor is described by, In a learning cycle individuals will try to explore various regions of the exploration space in initial phase. Afterwards individuals progress in a little range to regulate the trial solution to certain extent such that it can investigate reasonably little local space. Subsequently replicate the learning cycle over and over again.
The random number "r" is modified by -Dynamic inertia weight. The mean value of the novel random number is amplified from 0.5 to 0.75, and then the stochastic variations are augmented. Mainly difference value added to the current learners. In the meantime, augment from little to big in single learning cycle. Underneath of joint outcome of , the projected algorithm will not engender premature convergence. It will perk up population diversity, shun prematurity in the exploration procedure and augment the capability of the fundamental TLBO to flee from local optima.
In teaching phase new-fangled set of enhanced learners are defined by, In learner stage, the new-fangled set of enhanced learners is defined by, Mutation procedure is very easy, and design variables are initialized arbitrarily in the exploration space: Step a: parameters are initialized Step b: population generated Step c: non-linear inertia weight factor, dynamic inertia weight computed by ) ( ) ;

( )
Step d: individual with the most excellent fitness is chosen and average value is computed Step e: new marks of the learners are computed by ( ) and modernize the old values of the individuals by

SIMULATION RESULTS
At first in standard IEEE 14 bus system [18] the validity of the proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) has been tested, Table 1 shows the constraints of control variables Table 2 shows the limits of reactive power generators and comparison results are presented in Table 3. Then the proposed ATLBO has been tested, in IEEE 30 Bus system. Table 4 shows the constraints of control variables, Table 5 shows the limits of reactive power generators and comparison results are presented in Table 6.

CONCLUSION
In this paper Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) successfully solved the optimal reactive power problem. In order to control the learner's mutation arbitrarily during the learning procedure a non-linear mutation factor has been applied. Preceding information gathering of learners is determined by the weight factor and through that new-fangled values are calculated. In a learning cycle individuals explored various regions of the exploration space in initial phase. Proposed Advanced Teaching-Learning-Based Optimization algorithm (ATLBO) has been tested in standard IEEE 14, 30 bus test systems and simulation results show the projected algorithm reduced the real power loss. Percentage of real power loss reduction has been improved when compared to other standard algorithms.