^{1}

^{*}

^{2}

This article introduces a group search optimization (GSO) based tuning model for modelling and managing Smart Micro-Grids connected system. In existing systems, typically tuned PID controllers are engaged to point out the load frequency control (LFC) problems through different tuning techniques. Though, inappropriately tuned PID controller may reveal pitiable dynamical reply and also incorrect option of integral gain may even undermine the complete system. This research is used to explain about an optimized energy management system through Group Search Optimization (GSO) for building incorporation in smart micro-grids (MGs) with zero grid-impact. The essential for this technique is to develop the MG effectiveness, when the complete PI controller requires to be tuned. Consequently, we proposed that the proposed GSO based algorithm with appropriate explanation or member representation, derivation of fitness function, producer process, scrounger process, and ranger process. An entire and adaptable design of MATLAB/SIMULINK also proposed. The related solutions and practical test verifications are given. This paper verified that the proposed method was effective in Micro-Grid (MG) applications. The comparison results demonstrate the advantage of the proposed technique and confirm its potential to solve the problem.

Many changes have occurred in the electricity sector to improve the energy efficiency and environmental sustainability of generation, transport and final use segments, involving the need to update and improve the electric grids [

At the same time, another concept related to the distribution grid has been proliferating: the MG concept. A MG is a small electricity network, generally in LV level, connected to other MGs and/or to the main grid (generally in medium voltage, using a single point of connection (PC). A MG generally includes DG, electrical loads of different types (i.e., PHEVs) and energy storage devices in a reduced space. In a MG, each actor could take part into: reactive power supply and voltage control; active power supply and frequency control; and fault ride-through capability and power quality control (flicker, harmonics) [

Group Search Optimization (GSO) is a naturally sensitive algorithm which is taken by the animal (lions and wolves) food finding activities. He et al. [

The remainder of this paper is organized as follows: Section 2 reviews the related works on energy management approach in MGs. In Section 3, the system description and the proposed approach is formulated. In Section 4, we present the experimental results. Finally, Section 5 concludes the work.

In related researches, numerous connected tasks are accessible which depend on designing and control of micro grid. Some of them are given here. To optimize the method of the microgrid, a stylish energy management system (SEMS) has been obtainable by Chen et al. [

Conti et al. [

A centralized control system which balances parallel process of different distributed generation (DG) inverters in the micro grid maintains which is taken from Tan et al. [

In the micro grid (MG), the method depends on the cost-benefit review for an optimal aspect of an energy storage system maintains to be proposed by Chen et al. [

Dasgupta et al. [

Mohammadi et al. [

Pablo Arboleya et al. [

Waleed Al-Saedi et al. [

The common formats of the Custom power device (CPD) established at one of the Italian National Agency for New Technologies Energy and a Sustainable Economic Development (ENEA) building is illustrated in

In the next subsections, the different elements that compose the CPD will be described.

A. Li-PO Batteries: The energy storage system is depending on Li-Po batteries. The entire module is repre- sented by 280 cells linked in sequence with charged energy storage of 70 Wh per cell. The entire pack capacity is 16 kWh. The system is capable to release at a greatest rate of 20 kW in 15 min or 10 kW in an hour. The relation between the greatest output power and the charged energy is 4 - 8 h^{−1}. The greatest functioning temperature is 45˚C at a comparative humidity range of 40% - 90%. The approximate life is about 2000 cycles at 80% release. The batteries pack charged voltage is 270 V at complete charge. Though, the pack is divided into numerous parts with voltages not larger than 60 V.

The battery pack is linked to the 600 V dc bus by using of a bidirectional dc?dc Buck-Boost converter. The BMS handle the converter demanding to maintain a stable voltage in the dc bus but preventive the charge and discharge charge to certify the accurate process of the battery pack. A complete description of this manage will be given in the next part. The BMS obtains the cells voltage, current, and temperature so it also calculates the storage system state of charge. The dc bus over-voltage defence is also certified by a 60 VLR of 60 linked by a chopper.

B. Main AC-DC Converter: The power converter has been constructed by using concrete condition static apparatus; it is furnished with IGBTs. The controlling over-voltages must not beat 200 V at complete power. For this reason particular preventive courses are integrated. IGBTs are also furnished with recognition circuit de-saturation. The recognition diffusion has been nearby done with a hardware circuit associated to the anode of every transistor and must guide to the instant detain of its transmission. At the same time, a jamming signal should be sent, through the logic control, to seize of the entire converter. The overheating defence of the device is realized with temperature sensors within the semiconductor module. The inactive converter is offered with suitable input and output sort for restraining conflict, according to the directions 89/336/EEC and 92/31/EEC. The association to the 400 V ac grids is fulfilled with the needs of the Italian Standard CEI 0-21. The 600 V dc bus-bars are sized to attach both a dc load with a greatest power of 20 Kw and a construction system with a highest power of 30 kW. A recognition and defence system offers the information about the apparatus position: power deliver for the electronics out of scope, communication mistake with Digital Signal Processor, bar voltage out of scope, inverter over-temperature, outflow currents to ground, gear time out, inverter over current, defence IGBT modules, thermal security inverter, line voltage out of scope, and low line voltage alarm. The rule of the major converter will be extremely illustrated in the next part.

C. Loads: The loads are located at the ac side and at the dc side of the major converter. In the ac side, the load is the ENEA building utilization, which can swing between 36 and 228 kW. In the dc side, the loads are collected by a few supplementary dc loads being the majority significant the system freshening. The entire amount of dc load is about 5 kW.

D. PV module control: The PV module is connected to the dc bus through a unidirectional dc/dc Buck−Boost converter. If the reference voltage is higher than the measured value,

E. VRL control: As it can be observed in

This section describes about the objective function of the proposed approach. The management of distributed energy with the MG is one of the multi-objective problems in energy management in Custom power device (CPD) based scheme. The primary objective of load frequency control problem is to minimize the error. The PID controllers are extremely popular owing to their simple structure and robustness. According to, efficient tuned PID controller based approach is proposed to solve energy management problem using multi-objective GSO algorithm in this paper. The detail explanation of our optimized management approach is depicted in

The active and reactive power references are calculated as follows:

To determine if the power reference

subtracted from the

Similarly, the PI output is given by

In

The value of the

The required multi-objective function is described in the following (5),

Where,

The solution of the problem can be obtained using optimization procedures for minimization of

Depend on the penetrating activities by the animals like lions, wolves and etc Group search optimization (GSO) [

In general, GSO contains three operatives such as producers, scroungers, rangers who achieve arbitrary walk. The population of the Group Search Optimization is known as a group and every entity in the population is known as a member. Process of iteration, a member is represented by its location and head angle.

In a m-dimensional search space,

On iteration, a group member situated in the generally hopeful area, implements animal scanning to get the optimal supply and the finest fitness value is selected as the producer. A number of group members excluding the producer are arbitrarily chosen as scroungers and then the remaining of members are rangers. Scroungers execute region copying to unite the reserve originate by the producer and do limited searching approximately. Rangers utilize ranging activities by arbitrary walk in the penetrating space to enlarge the possibility of GSO to getaway local optima.

A. Producer operation: Producer by its visualization capability examines the search space for the improved conditions. The producer examines three points approximately its location in definite distances and head angles.

At the k th iteration, the producer behaves as follows:

a) The producer scans at zero degree and tests three points toward its position using Equations (13)-(15).

where

b) The producer will detect the finest summit. If the finest summit has an improved value in contrast with its existing location, producer will fly to that point. If not, it will continue in its present location and turn its head using Equation (16).

c) If the producer cannot determine an improved region after iterations, it will revolve its head back to zero degree as follows:

where,

B. Scrounger operation: Processing on iteration, a few of group elements are chosen as scroungers which will preserve penetrating for occasions to unite the sources establish by the producer. At the

where,

C. Ranger operation: On iteration, a few of group components are chosen as rangers which are distributed from their locations and arbitrarily walk at investigate space. At the

where c is a constant and is a normally distributed random number with mean 0 and standard deviation 1.

The GSO algorithm is presented in Algorithm 1.

Algorithm 1. Computational process for GSO.

Overall our proposed tuning model is comprised into following important steps:

Step 1: In the first step, initialization of all parameters of the algorithm like as input parameter limits and random population N limits.

Step 2: The random solutions are applied in the fitness Equation (1) and find the best solution.

Step 3: Producer operator: The producer operation is performed for selected individual using Equations (13), (14) and (15).

Step 4: Scrounger operator: The scrounger operation is performed for selected individuals using Equation (18).

Step 5: Ranger operator: The ranger operation is performed for remaining individuals by Equation (20).

Step 6: Check the stopping criterion. If satisfied, terminate the search, otherwise go to step 2.

Step 7: Assign the new population to generate new solutions. Go to Step 2.

The proposed technique is implemented in MATLAB/Simulink 7.10.0 (R2015a) platform, 4GB RAM and Intel (R) core(TM) i5.

The initial head angle

where

The most important control parameter of GSO is the percentage of scroungers and rangers. In this paper scroungers are taken as 70% and rangers are taken as 30% of the total population. The parameters required to establish the goal programming model are illustrated in

In this section, we present the comparison results of the existing approach of Pablo Arboleya et al. [

In

In

Parameter | Value |
---|---|

Number of population | |

maximum pursuit angle | |

maximum turning angle | |

maximum pursuit distance | |

initial head angle |

leased. While, the voltage at the ac subsystem is equivalent to its charged value (400 V), the load 1 will soak up 3.2 kW and 3.2 kVA. Initially, element of the power is removed from the grid because the speed of reply of the major converter is restricted to protect the dc bus voltage steadiness. That is why up to 2.5 kW are fascinated from the grid at the initial time immediate, but beforet1is achieved the core converter offers the necessary 3.2 kW, and the dynamic power from the grid becomes zero as shown in Figures 4-6. As the voltage in the ac subsystem is equivalent to its charged value, no spontaneous power will be inserted or fascinated by the major converter because of the droop rule, being the spontaneous power necessary by the ac load (3.2 kVA) acquired from the grid [see Figures 7-9].

In the dc bus, the PV module is operational with permanent irradiation at 1000 W/m and temperature at 298 K. Under these circumstances, the PV module is inserting 6.4 kW to the dc bus. The charged dc load is 6 kW, though, the voltage is about 595 V, and so the dc load will attract 5.9 kW. In the dc bus, there is an additional generation of 0.5 kW and

At

At

At

At

At

At

At

At

At

At

Traditionally tuned PID controllers are employed to address the load frequency control (LFC) issues using various tuning methods in recent times. On the other hand, improperly tuned PID controller may exhibit poor dynamical response; in addition, wrong choice of integral gain may even destabilize the overall system. Conventional tuning is particular suited for simple systems, however fine-tuned heuristics based methods are particularly suited for complex systems e.g. power systems. In this article, we have developed an optimization technique to tune the PI parameters for building integration in smart micro-grids (MGs) with zero grid-impact. The necessity for such method has shown to be addressed when the entire PI controller need to be tuned to improve the MG efficiency. For that, we presented the proposed GSO based algorithm with suitable solution or member representation, derivation of fitness function, producer operation, scrounger operation and ranger operation. A complete and versatile model in MATLAB/SIMULINK was also presented. The simulations results and the experimental test validation were included. This research proved that the proposed system is usefulness in Micro-Grid (MG) applications. The control design through hybrid algorithm for other heuristic algorithms and for other complex systems remains a topic for further future research.

S. Bhagawath,S. Edward Rajan, (2016) Managing of Smart Micro-Grid Connected Scheme Using Group Search Optimization. Circuits and Systems,07,3095-3111. doi: 10.4236/cs.2016.710263