Keywords

1 Introduction

Comparing the conventional energy technology with the renewable energy technology it was found the later having limitless potential. RES are unlimited and cleanest energy sources and have an inconsiderable impact on the environment. Previously, generation capacity expansion would have solved the issue, but it has a negative impact on the environment which in turn increases the operational cost. So to have an eco-friendly solution, some smart grid technology such as renewable energy sources based hybrid microgrids and demand-side management was introduced [1, 2].

Although renewable energy sources have many advantageous features because of its intermittent nature it leads to problems relating to power quality such as frequency and voltage fluctuation. Several authors have found out that the fluctuation that arises can be reduced by energy storage devices particularly battery [2].

The reliability, efficiency, and stability of the grid can be improved by implementing a demand response program as the energy usage pattern can be reshaped. Several kinds of literature were studied and it has been found out that Demand-side management persuades customers to reduce their demand in peak hours by shifting the load along with generation and in return, customers are rewarded in to have low electricity prices. Various types of Demand-side management have been studied and analyzed [7].

The overview of DR in the smart grid, various optimization methods, and strategies based on the target of the optimization procedure has been explained by the authors in [8]. The author in [9] has analyzed the four major aspects, such as programs, issues, approaches, and future extensions of demand response.

Thermostatically controlled loads like such as air conditioner, water heater, and refrigerator, etc. hold a large portion of the residential loads. The control mechanism for the thermostatic load to reduce the demand in peak hours by shifting the loads and to maintain the inside temperature in a controllable range by varying the thermostat set point with the system grid frequency and Real-time pricing (RTP) has been discussed in [10].

With the recent emergence of solar thermal power generation, there has been an acute increase in small and medium solar thermal power demand [2]. One promising approach in meeting this demand is the application of low-temperature Organic Rankine cycle (ORC) solar thermal power system [2].

The management of load and to reduce the peak load demand, various optimization techniques have been used. In this system PSO & SCA optimization techniques are applied and depending upon the convergence time the best one will be selected.

In view of the literature survey, the research work presented here has been outlined as follows:

  • To develop a hybrid energy system which consists of Renewable Energy Resources (RES) such as Solar thermal power plant &Wind turbine generator. Diesel driven generator, controllable and critical loads such as refrigerator, hybrid electric vehicle and Heat pump are used in the model.

  • Demand Response technique is used for better management of energy usage by the consumers.

  • To apply optimization techniques to the PID controller by using particle swarm optimization (PSO) and Sine Cosine algorithm (SCA) and use the best techniques among the two.

Section 2 presents the modeling proposed system components and the SCA is briefly discussed in Sect. 3. The simulated results are discussed in Sect. 4 and conclusions are drawn in Sect. 5.

2 Modeling of Proposed System

The hybrid energy system comprises of an ORC Solar Thermal Power System (1 MW), Wind Turbine generator (500 KW), A diesel-driven generator (3 MW), Refrigerator (0.8 MW) and Hybrid Electric Vehicle (250 KWh). The Critical load is 0.4 MW and the Base load is taken to be 2 MW. The parameters required for the Hybrid energy system is given in Table 1.

Table 1. Constraints of the proposed hybrid system.
Full size table

The transfer function model of the hybrid energy system is shown in Fig. 1:

Fig. 1.
figure 1

Transfer function model of the hybrid energy system

Full size image

2.1 ORC Based STP System

In this type of system power generation is possible even at less capacity and with less temperature of the collector, and so the possibility for low-cost, small scale decentralized CSP units. The ORC system consists of the CPC and ORC working with hydrochlorofluorocarbon (HCFC-123) [2]. The organic working fluid is heat in the collector field and pumped into the heat exchanger where it is vaporized at high pressure [2]. The vapor formed in heat exchanger runs the turbine thus incorporating thermo mechanical energy conversion. The turbine is coupled to a generator producing electricity. Expended vapor from the turbine is condensed and pumped into the reservoir tank ready to be dispersed to the heat exchanger thus repeating the cycle [2]. The transfer function of the ORC system is given as (1):

$$ GORC = \frac{{K_{s} }}{{sT_{S} + 1}} \cdot \frac{{K_{T} }}{{sT_{T} + 1}} $$
(1)

2.2 Wind Turbine Generation (WTG)

With an average growth of 21% of wind energy have been developing rapidly since the last few decades [1]. The capacity of wind by the end of 2013 is 318.105 GW [3]. One of the major setbacks of the wind energy system is its inconsistent nature and also not.

The transfer function of WTG is given by (2):

$$ GWTG = \frac{{K_{WTG} }}{{sT_{WTG} + 1}} $$
(2)

2.3 Diesel Driven Generator (DDG)

Electrical power output is generated from the torque that is produced by DDG which drives the synchronous machine. In certain situations, the demand for load may increase abruptly. To cope up with this condition the dynamic response of DDG should be fast and it should be able to reject the disturbances [1].

The simple first-order transfer function of DDG is given by (3):

$$ G_{DDG} = \frac{{K_{DDG} }}{{sT_{DDG} + 1}} $$
(3)

2.4 Plug-in Hybrid Electric Vehicle (PHEV)

Conventional fuel, battery or super-capacitor is used to power the PHEV [1]. The Energy is stored during the off-peak hours by the PHEV and the energy that is stored, supplied to the grid at the peak hours. The transfer function is expressed as (4):

$$ G_{PHEV} = \frac{{K_{PHEV} }}{{sT_{PHEV} + 1}} $$
(4)

2.5 Refrigerator (REF)

The ability to get disconnected rapidly and its storage capability is one of the most attractive features of the refrigerator. When generation increases the refrigerators are switched on so that the power can be absorbed and when generation decreases the refrigerators are turned off so that the frequency is maintained. The transfer of the refrigerator is given by (5):

$$ GREF = \frac{KREF}{{sT_{REF} + 1}} $$
(5)

2.6 Heat Pump (HP)

With more advancement heat pump can also be used nowadays to reduce energy consumption. A heat pump can provide year-round climate control for the home by supplying heat to it in the winter and cooling it in the summer [6]. Overall, using a heat pump alone may not be economical to fulfill all the heating needs. It can be useful if it performs its function with another supplementary form of heating such as oil, gas or electric furnace. The transfer function of HP is given by:

$$ G_{HP} = \frac{{K_{HP} }}{{sT_{HP} + 1}} $$
(6)

2.7 System Dynamics Model

The net change in power of the proposed system could be estimated by (7) considering the coordinated power deviations of STP, WTG, DDG, HEV, REF, HP and critical loads.

$$ \Delta P_{E} = \Delta P_{STP} + \Delta P _{WTG} + \Delta P _{HEV} \pm \Delta P _{HEV} - \Delta P_{REP} - \Delta P_{HP} - \Delta P_{LOAD} $$
(7)

The linearized transfer function model of the system dynamics could be expressed as (9) considering an equivalent inertia M(0.1667p.u) and load damping D(0.015p.u/Hz) for the net change in system frequency (∆f).

$$ GSYS = \frac{\Delta f}{\Delta PE} = \frac{1}{sM + D} $$

2.8 Objective Function

The function Integral absolute error (IAE) is considered to formulate the objective function (J) of the proposed system to minimize ∆f as (9). Here, Kc represents the PID controller gains (Kp, Ki, Kd) with their upper (Kmax) and lower (Kmin) limits.

$$ J = IAE = \int_{0}^{tsim} {|\Delta f|dt} $$
$$ \text{s}.\text{t}. \, \text{K}_{{\text{min}}} \le \, \text{K}_{\text{c}} \le \text{K}_{{\text{max}}} $$

3 Sine Cosine Algorithm

Sine-Cosine Algorithm (SCA) for solving optimization problems was developed Mirjalili [12]. The SCA is a mathematical optimization technique that obtains the global best solution by simulating the behaviors of sine and cosine functions. A set of random solutions are initialized for SCA and updated based on sine and cosine functions by fluctuating these solutions whether outwards or towards the global best to create a new population [12].

Mirjalili [12] reported the SCA algorithm is proficient enough to reveal an effective performance in comparison with other state-of-the-art metaheuristic algorithms. The detail steps of SCA optimization are briefly illustrated as a flowchart in Fig. 2, considering N populations with specified lower and upper bounds (lb & up) of given dimensions (dim) of a problem for maximum iteration of MaxItr. The PID controllers of the proposed system are tuned with PSO and SCA to study the system responses, whose results are briefly discussed in the next section.

Fig. 2.
figure 2

Flowchart for SCA

Full size image

4 Result and Discussions

The proposed microgrid model as shown in Fig. 1 is studied considering two major cases of Step load perturbation with stepped and varying renewable sources including some sub-scenarios of available based sources. The proposed model was simulated with MATLAB using PSO and SCA considering Scenario 1, to optimize the objective function in (10) and compared the convergence performances in Fig. 3 that reveals the superiority of SCA over PSO to tune the controllers. The optimization parameters for PSO and SCA algorithms are listed in Table 2, considering lb(lower bound) = 0, ub(upper bound) = 100, dim = 12, N = 50, MaxItr(Maximum iteration) = 100, and tsim(simulation time) = 120 s for both.

Fig. 3.
figure 3

Convergence comparison of PSO and SCA

Full size image
Table 2. Parameters for optimisation algorithms
Full size table

The first case is further sub-divided into four conditions (Scenario 1.1 to 1.4) considering different climatic conditions and availability of resources (Table 3).

Table 3. Optimized controller gain
Full size table

Case1: This case is considered using step load and renewable energy sources

Scenario 1.1: Availability of All Forms of Renewable Energy: This is a normal scenario with a sufficient amount of sunlight and wind for power generation for the microgrid. In this scenario, a 20% drop in solar energy is considered at the 40 s and a 5% rise in wind generator at 80 s to give the supply. The contribution of power by the generating units (STPS, WTG & DDG) and the loads (HEV, REF & HP) are shown in Figs. 4 and 5 respectively.

Fig. 4.
figure 4

Power contribution by STPS, WTG, and DDG in scenario 1.1

Full size image
Fig. 5.
figure 5

Power contribution by various loads in scenario 1.1

Full size image

Scenario 1.2: Unavailability of Both the Renewable Sources of Energy: In some unfavorable weather conditions, both the sources may not be available. The power coordination of the generating units and the loads in this scenario are shown in Fig. 6.

Fig. 6.
figure 6

Power contribution by various loads in scenario 1.2

Full size image

Scenario 1.3: Unavailability of Wind Turbine Power: The wind turbine operates within a limit of minimum (cut-in) and maximum (cut-off) speeds, e.g., Speed required for rotating small wind turbine is 8kph. Considering some critical climatic issues, when the wind speed may not be suitable to operate WTG for power generation, whereas the SPV units are smoothly operating. The power coordination of the system during this scenario is shown in Fig. 7.

Fig. 7.
figure 7

Power contribution by various loads and STPS

Full size image

Scenario 1.4: Unavailability of Solar Thermal Power Plant: This case can be considered for bad weather conditions or during the nighttime with appropriate wind generation. In the absence of solar energy, the response of the system is shown in Fig. 8.

Fig. 8.
figure 8

Power contribution by various loads and WTG

Full size image

The frequency response curve of all the four scenarios that have been taken into consideration is shown in Fig. 9. It is seen that when all the sources and the load are available in scenario 1 there is a reduced oscillation and it settles at 0.07662 s.

Fig. 9.
figure 9

Frequency response of all scenarios under consideration

Full size image

The decision parameters for system response comparison like settling time (Ts), overshoot (Osh), undershoot (Ush) and the optimized objective function (Jmin) for all the four cases are listed in Table 4:

Table 4. Decision parameters for different scenarios
Full size table

Case 2: In this scenario stepped load along with varying renewable sources is taken. The response of the system is shown below (Figs. 10, 11 and12):

Fig. 10.
figure 10

Power contribution by STPS, WTG, and DDG in case 2

Full size image
Fig. 11.
figure 11

Power contribution by various loads in case 2

Full size image
Fig. 12.
figure 12

Frequency response of Case 2

Full size image

5 Conclusion

The overall response considering different scenarios discussed in this work shows the robustness of the proposed hybrid system towards stabilizing the system frequency using SCA optimized PID controllers. The controllable loads play a vital role in demand-side management without any storage systems. The optimization technique was chosen for the proposed system for proper economic operation and optimal usage of renewable energy. The work could be further extended for demand-side management of some interconnected hybrid microgrids.