Minimizing the Active Power Losses and Retaining the Voltage Profile of the Distribution System Using Soft Computing Techniques with DG Source

. Te motivation of this research is to curtail the active power losses in the radial distribution system. Tis paper proposes a novel voltage stability indicator to determine the optimum conductor sizes for various distribution lines and fnd the conductor stability limit. Tis work proposes a new BATalgorithm for fnding the optimal switching confgurations using reconfguration, which is used to minimize the active power losses. Te proposed BATalgorithm gives the optimal locations to place the required amount of DG sources to improve the stability and minimize the power losses and maintain the voltage profle of the systems. Te proposed BAT algorithm has been executed with the MATLAB 2016 software and compared with the diferential evolution technique for IEEE 33 bus, IEEE 69 bus, and Indian real time 62 bus systems. Te test results are applied on diferent conductors, and the real power fow on the diferent conductors are mapped in the proposed method using the novel voltage stability indicator. Te radial distribution system has lower power losses, best voltage profle, cost of saving the DG source, optimal placing of a suitable rating of the DG source, optimal location of the DG source, and cost of saving the DG source which were achieved in this work. In this research, the distributed load fow analysis is implemented to fnd the parameters by the forward-backward sweep algorithm. Te minimum bus voltage 0.95pu is virtual in all the conductors. Te loss reduction in the DE for IEEE 33 bus has been reduced from 42.4% to 45.3% and the loss reduction in IEEE 69 bus from 36.2% to 38.8% and the real time 62 bus from 47.5% to 49.15%. Te loss reduction in the BATalgorithm for IEEE 33 bus has been reduced from 38.46% to 45.3% and the loss reduction in IEEE 69 bus is from 32.2% to 38.8%. Te Indian standard real time 62 bus has been reduced from 38.92% to 49.15%. Te proposed results are compared with the DE and BAT algorithm. Te proposed BAT algorithm is found to be more efectual in reducing voltage deviation (VD) and reducing the power losses in the system.


Introduction
In today's situations, the usage of electrical power has suddenly increased. Te power system operation is very critical to operate. Te power system is a very huge network for generation, transmission, and distribution to provide electricity to tail end consumers. Many components are interlinked from generation to consumer end. Each and every component has high cost, depending upon its size and operation. At present, the increase in electrical energy demand wants more priority given to the power distribution. However, the construction of new substations, new transmission systems, and construction of the new lines are major problems due to environmental issues and increasing costs.
Nearly 30% to 37% of the supply has been lost including the components [1]. Te power losses in the distribution systems are due to improper conductors, aged instruments, and improper operation of the systems.

Novel Voltage Stability Indicator
Te assumptions made on the balanced three-phase line diagram are converted into a single-phase single line diagram in a radial distribution system. Tis single line diagram is used to fnd the real and reactive power fow, the real and reactive power losses, the bus voltages, and the angle of buses using the forward-backward sweep algorithm with a branch exchange method using a novel voltage stability indicator [2]. Normally, this branch exchange method is one of the greater methods to calculate the power fow and power losses of the entire network. Figure 1 shows the expressions of the two-bus system [3]. Te line joins buses "i" and "j" and is demonstrated as a single line diagram. Current fows and voltage drop across the buses are easily found using recursive equations [4] with two assumptions: sending end voltage angle is zero, and the other is receiving end voltage angle δ [5].
Let it be assumed that bus "i" is a slack bus; then, the current I ij is found by Te current fow equation is derived from active and reactive power fows as given below [6].
By using the receiving end powers, the current fow I ij is Te abovementioned equations are current fow equations. Tese equations are used to calculate the apparent [7], active, and reactive power fows and power losses using the forward-backward sweep algorithm [8].
"k" is the line number of the buses between "i" and "j." Te transmission line losses have been calculated from the power equations (9): Te complex equations have been modifed [10].
Te real and imaginary parts are separated and gives Shortening the quadratic equation (11), the real roots found the constancy of the system.
Te positive values are always assumed in the right-side quadratic term and compared with the minimum value of the remaining equation as [12].
Te value of Δ � �������� � (b 2 − 4ac) is diferentiated to zero; the inner term is always zero of V j lies from zero to one, and it yields multiple indicators due to the real roots limitation. If the constraint lies below one and above zero [13], the system stability is determined. Te system stability depends upon the real roots of the equations. Any system, if the power losses are increased, then the system will be unstable [14]. Te system line losses are curtailed if the value of stability lies between zero and one.  Journal of Electrical and Computer Engineering Te novel voltage stability indicator equation has been presented as In this equation, the insufcient [15] of reactive power fow has been given below In this research, the novel VSI has been derived and found the breaking point very clearly. Te VSI value has been found and placed from zero to 1, so the system is pure stable [16]; otherwise, the system has been found unstable. If there is any sudden changes in the lines or sudden increase in the load, the VSI value increased more than one or decreased less than zero. Te novel VSI is the best tool for determining the system conditions.

Soft Computing
Techniques. Now, soft computing techniques are the greatest tool in the intelligence exposed by machines and softwares. Te intelligent techniques are talented to think, reason, invent the meaning, simplify, extricate, learn from past familiarity, and correct their mistakes. Soft computing techniques are the cleverness of a proposed machine or computer which can achieve any rational assignment which a mortal being can achieve. Te conventional techniques [17] are implemented in the area of analysis, design, and control of power systems. Te conventional techniques are difcult and more complicated.
Te soft computing techniques are more critical to understand, intricate, fexible, and large amounts of statistics are used in the calculation for fnding the faults and learning. Te raise in the processing speed and computational time and accurate results are due to wide and vast system data management. Te recent power system operations are near to the boundaries due to the sudden increasing energy consumption [18] and existing electrical transmission systems. Tis condition involves a less predictable power system operation, control, and design, which is probable only by constantly checking the system conditions in a much more detailed manner than required conditions [19]. Now, so many computer tools are available to solve difcult problems like power system planning, design, operation, analysis, and controls. Among these computer tools, soft computing techniques have been growing in the recent years and have been applied to power systems. Recent techniques are differential evolution and the BAT algorithm [20]A, which have been applied in this research to curtail the power losses.

Differential Evolution
Diferential evolution has been planned over 1994-1996 by Storm and Price at Berkeley as a new stochastic direct search optimization method. Several techniques have been used to reconfgure the systems to minimize the power losses. Te soft computing techniques, artifcial intelligence techniques, Genetic algorithms, Artifcial Neural Network, Artifcial Bee Colony, Particle Swarm Optimization, ant colony algorithms, fuzzy logic, Cuckoo Crunch algorithm, Harmony Search algorithm, CAT algorithm, and Grey Wolf algorithm are used for optimizing the power losses. Differential evolution is a profcient investigative process for search and optimization techniques. Te diferential evolution technique is a great tool, which is capable of handling non-diferentiable and difcult optimization on large problems. DE is a real mutation method to ensure diversity and objective function directly. Te standard DE is a group of characteristics for retention of the optimal result and information allocation within the population with a faster convergence rate because of one-to-one competition among the adequate issues with the consequential parent. Te main advantages of diferential evolution include parallel processing in nature, current global optimization profciency, self-referential mutation operations, actual on integer, dissertate, and mixed constraint optimization. Te main advantages of diferential evolutions are the ability to handle nondiferentiable, noisily, time dependable autonomous function; tasks on fat surfaces; capacity to send many solutions in a single route and operative in nonlinear optimization problems with penalty problems; well-organized algorithm without sorting matrix multiplications; and fast and simple for applications and modifcations [21]. Te algorithm is similar to the principles with genetic algorithms, including four basic evolutionary steps of initialization, mutation, crossover, and selection. Standard diferential evolution utilizes a random path to mutate individuals that may point to the capable area during the evolutionary process.

Initialization.
Te frst stage is to develop a random initial population in "D" dimension and suppose "n" dimension with decision space; so, the individual "i" of differential evolution can be exemplifed as follows [22] XU/ij and XL/ij are the upper and lower limits of j th variable in the population. Te above equation is the initial equation of the problem.

Mutation.
Te generation evolutionary target vector is X i,G ,, i � 1, 2, ...., N p ; then, the mutation operations are as follows Te above equation is the mutation of the function. F is the scaling factor, always more than 1 that controls the strait of direction. (X r2,G − X r3,G ) procedures a path that is the starting point of diferential evolution and a reason for direction-based search. r1, r2, r3 ∈ [1, N p ] are the three cooperatively diferent numbers and also diferent from the successively index "i". F is one of the main limits of differential evolution given as [0, 2].

Crossover.
Increasing the diversity of the population, crossover equation is given below: Te above equations are the crossover of the function, where r and b (j) variable denotes an equally distributed random fraction with [0, 1] and also used to state limit governing the efect of crossover, r and j are randomly selected indexes to confrm that at least any one of the variables should be altered, and CR is a user-defned limit supervising the efect of crossover within the limit of b (0, 1). U ij,G+1 is not a duplicate of X ij,G .

Selection.
Te fresh individual is better than the original one. Te new individual has to be an ofspring for the new generation, and the original one is retained as the new generation.
Te above equation is new individual of the function. Te target vector U I, G competes with X i is called target vector survival of the succeeding generation. Mutation, crossover, and selection have been carried out for Np individuals in one generation. f(U i,G ) is the suitability task. If the objective function value of the trail vector is better than the value of the individual vector, the trail vector will be selected as the new individual vector X i,G of the next generation [17]. Te single vector X i,G is kept as the individual, and the vector X i,G+1 is the succeeding generation. Te optimization loop of DE tracks iteratively until the stop conditions are met.

BAT Algorithm
Te BAT algorithm is one of the soft computing techniques for optimizing the power losses and improve the stability of the system. It is a population-based evolutionary optimization problem [23] based on the voice or echolocation actions of natural bats in detecting their prey or food. Normally, bats emit sound called echolocation that they use to fnd the food or prey in and around them and detect their way even in full darkness. Bats are eye-catching animals, which have wings and advanced echolocation capability to detect their prey or food.
(1) Individual bat employs the echolocation practice to sense the distance, and they also recognize the difference between food/prey and background sprints in some mystic way using echolocation. (2) Each bat fy in the direction of xi fies randomly with the velocity Vi producing pulse with wavelength λ, frequency fmn, and loudness A0 to try to catch the prey.
(3) It is an ability to control the emitted pulse and adjust the rate of the emission of "r" range of [0, 1] believing the closeness of its aim. (4) Te loudness suggestions in all the ways decrease from the higher position A0 to lower positionA min. (5) Further generating the initial accidental bat population, the objective function is calculated for all the bats, and the G best bat is stored.

Initialization of Population.
Primarily, the population is the number of major bats for the BAT algorithm created randomly. Normally, the number of bats for this objective function is from 10 to 40. After succeeding to get the initial ftness of the population for an expected function, the values are based on the loudness, movement, and pulse rate.

Movement of Virtual Bats.
In this BAT algorithm, the three rules for updating the positions and velocities of the virtual bats are as follows: where β, ∈[0, 1] is a random vector drawn from an even distribution and x * [24] is the present total best solution among all the entire bats. New solution for all the bats using random march is given below.
where "ε" is the topping factor in the range of [−1, 1]. Tough A t � | (A t i )| is e loudness of all the bats at this time step.

Loudness Values and Amount of Pulse Production.
Te volume and the pulse emission rates of each bat are updated with iteration using the dealings. Te pulse rate is inversely proportional to loudness.
where α is a constant value. Iter is the number of iterations during the optimization processes and generally taken as 0.9. For any rate of 0 < α < 1, c > 0, we have A t i ⟶ 0, r t i ⟶ 0, as t ⟶ ∞ . Te initial value of loudness A 0 can be in the range of [0, 1]. While emission rate r i can be in the range of [0, 1], loudness and pulse rates are selfadjustable. DE and BAT algorithms can be used for designing the physical components of the power systems. It can be used to raise the efciency of the apparatuses used in the power systems. DE and BAT algorithms can be abundantly used to grow a stable, accurate, and ambiguity-free output. Te computer programs have executed the operations of the power system better than manually done [25]. Modifcations are easily possible even after designing the computer programs. Any values coding has been modifed and estimated virtually for improving the system stability. It is eternal and consistent, and it can be easily accepted for paper works. Some refning steps have been taken for a stable and acceptable system. Suppose the optimal conductor of RAC-COON is used for power transmission.

Identify Optimum Location Using BAT Algorithms
A distribution system has many transmission lines for power transmission from one transformer to another transformer and transformer to the consumer extremity end. Te name of the lines is called by animal names (manufacturer name) such as mole, squirrel, saber, weasel, dog, wolf, panther, zebra, beaver, leopard, lion, tiger, bear, goat, sheep, and camel [26]. Table 1 shows the technical specifcations and parameters of the diferent lines. Tere are four conductors taken for this research such as raccoon, beaver, weasel, and rabbit. All the transmission lines are stranded, twisted, and aluminum conductor steel reinforced (ACSR) [27]. Tables 2-6 show the stability point of diferent conductors. Te diferent power factors have been implemented to calculate the stability point in the Raccoon conductor in the distribution systems as shown in Table 3. Te apparent power of the transmission line is gradually increased, and at a particular point, the conductor reaches the instability limit; in this instability limit, the conductor may lose the stranded, or produce sparks or short circuit between the conductors. Using the novel voltage stability indicator equation, applying apparent power LMF (load multiplication factor) gradually increased the load from 1.0 pu. Te value of voltage stability indicator (VSI) is obtained. If the VSI lies between 0 and 1, the system will be in stable condition and the power fow is easy. If the VSI lies between less than 0 and more than 1, the system will lead to instability. So, the transmission line creates problems of either losing the stranded or short circuits happen. Based on the VSI equation, stability limits of all four conductors have been found correctly and also implemented in the electricity board in real time. In any distribution system, the active power fow mainly depends on the load angle and the reactive power fow depends on the voltages of the buses.

. Reactive Power and Voltage Control
Reactive power control can enhance the power system's voltage profle. Reactive power only control and enhance the power system's stability, and the voltage stability is an insufcient reactive power fow or excess reactive power fow of the distribution systems. In this research, the voltage level of the test system decreased below the determined level on the buses node 11, node 34, and node 58 as shown in Table 7. Before reconfguration, the node voltages are below the stability value, and when the reactive power injected Table 8 on the same bus is at the appropriate level, the bus voltages are improved, then the system will go to the stability limit, at the same time, active power losses will also decrease and stability will be maintained. Te two soft computing   techniques DE and BAT have been applied for network reconfguration and placing the reactive power injected in the instability buses are shown in Table 9 Te voltage abnormality and supervisory variables are converted into DE and BAT algorithm symbolizations to construct the afairs between voltage deviation and the governing ability of the

Conclusion
Te radial distribution system has lower power losses, best voltage profle, cost of saving the DG source, optimal placing of a suitable rating of the DG source, and optimal location of the DG source, and cost of saving the DG source have been achieved in this work. Reconfguration of the distribution system and optimal placement of the DG source have been done in fast convergence and saving the minimum cost level. Tere are two works carried out in this research that, one is the reconfguration technique to fnd the electrical parameters of the system and another one is applied soft computing techniques for fnding the optimal location and optimal sizes of DG source. DE  Te loss reduction in the BAT algorithm for IEEE 33 bus has been reduced from 38.46% to 45.3% and the loss reduction in IEEE 69 bus from 32.2% to 38.8% and the real time 62 bus from 38.92% to 49.15%. More numbers of research works have to be performed to achieve the power losses reduction. At last, network reconfguration techniques using the FBS algorithm using reduction techniques has been faster than other load fow analysis. Te BAT algorithm has been utilized to fnd optimal cost saving, optimal location fnder, and suitable DG source injected compared to other soft computing techniques.

Abbreviation
ACSR: Aluminum conductor steel reinforced DE: Diferential evolution DG: Distributed generation LMF: Load multiplication factor VAr: Reactive power unit VSI: Voltage stability indicator. VD: Voltage deviation

Data Availability
Te IEEE 33 bus and IEEE 69 bus system data of this research paper have been taken for executing the program and comparing these results with soft computing techniques, and the Indian standard 62 bus test systems data were measured manually and implemented for this research. Tese data   Journal of Electrical and Computer Engineering 7 have not been copied from any other papers and are not implemented wrongly.

Conflicts of Interest
Te authors declare that they have no conficts of interest.