Model of a hybrid renewable energy system: Control, supervision and energy distribution

Model of a Hybrid Renewable Energy System: Control, Supervision and Energy Distribution Dada Delimustafic, Jasmina Islambegovic, Abdulah Aksamovic and Semsudin Masic Faculty of Electrical Engineering University of Sarajevo, 71000 Sarajevo, Bosnia and Herzegovina Email: {dd14509, ji14589, abdulah.aksamovic, semsudin.masic}@etf.unsa.ba

Abstract—This paper presents a concept design of a hybrid renewable energy system (HRES). The proposed HRES design specifies the operation of the following units: a pumped-storage hydro power plant, a wind power plant and a solar power plant. Since the pumped-storage hydro power plant represents the most complex entity of the HRES, special emphasis is placed on its two control loops: the water tank level control loop and the load-depended frequency control loop. Models of these control loops provide the means for efficient controller design and implementation. In order to achieve effective energy distribution and reliable power supply of the consumers, a switch logic architecture has been developed. Since the HRES’s final energy production highly depends on the process of energy conversion, the performances of three types of power converters are analysed. This includes the parametrisation, modelling and simulation of such converters. Finally, a graphical user interface (GUI) for purposes of data monitoring, control and supervision within the designed SCADA system, is presented. The proposed HRES design represents the foundation of the system implementation that aims at serving as a platform for research, education and a broad range of projects on renewable energy.

I. I NTRODUCTION As the world’s energy consumption increases due to population growth and the ever evolving industrialisation process, mankind faces the challenge of preventing resource depletion. Renewable energy sources have proven to be an efficient alternative for conventional energy production based on nonrenewable resources, mainly fossil fuels [1], [2]. Although we will continue to rely on non-renewable energy sources for most of our energy needs, the use of alternative sources of energy is expected to expand even more so over the next few decades. Since the future large-scale usage of renewable energy is a worldwide priority [3], intense research is being conducted on the integration of renewable energy sources into existing power generation systems and the substitution of fossil fuelbased power generation with renewable power supply. Several critical factors have been identified that make such integration and substitution possible, one of them being the obvious need for trained workforce [4]. Because of this, incorporating renewable energy education into the course syllabi of educational institutions remains an important issue. Zahedi [5] developed a multilevel undergraduate course dealing with the principles of renewable energy. The purpose is teaching renewable energy and environmental technology by means of lectures and tutorial sessions using computer simulation packages. Ortiz-Rivera et al. [6], provided the study

of three alternative sources: photovoltaic systems, fuel cells and thermoelectric generators for undergraduate students. To perform circuit analysis and simulation the students were given behavioural models that emulate the typical electrical characteristics of those energy sources. The research culture emerged from the model studies resulted in the growth of the renewable energy development research team, with the number of members increasing from 1 to 12 during the course of three years. The construction of a grid-tied hybrid demonstration system consisting of a 2 kW solar (PV) array and a 2 kW wind turbine is discussed in [7]. The objective was to create an example facility to demonstrate affordable and effective sustainable energy generation options. Data values collected during the first year of the system’s operation indicated that the average total energy generation was 221 kWh/month, enough to supply a small building. A similar system design was proposed by Wies et al. [8], in the form of a standalone distributed generation system integrating existing fossil fuel based energy sources with renewable energy sources and smart grid technologies for off-grid communities. The projects discussed in [7] and [8] were realised within undergraduate and graduate course frameworks encouraging student research on renewable energy through working with an actual system. In this work we propose a HRES design that combines power generation from hydro generators, wind generators and photovoltaic panels.1 Local measurement, control and microcomputer monitoring is realised for each of the system units. The microcomputers form a computer controlled unitary network. A configured network connection allows for monitoring and control of the system from distant locations. The notably multidisciplinary nature of the HRES operation offers possibility for extensive research. The scheduled implementation of the herein presented concepts aims at serving as an educational demo system upon which student research can be carried out. The remainder of the paper is organised as follows. In Section II, the developed system configuration is presented, with special attention directed toward the control design for the pumped-storage hydro power plant. Power converter design with respect to the given requirements is discussed in Section 1 The presented HRES design is aimed to be implemented within a model framework scheduled to be constructed at the Faculty of Electrical Engineering, University of Sarajevo, B&H.

Fig. 2. Water level control loop: The water level is kept at the set value of href = 2 m. The PID controller reacts to level variations detected by the ultrasonic sensor by adjusting the centrifugal pump’s angular velocity.

The upper tank model realises the connection between the current water level h and the water flow Q provided by the centrifugal pump. The dynamical dependence between h and Q can easily be obtained as: ab

Fig. 1. System configuration: MAN represents the Metropiltan Area Network that provides power supply for the feeder network in case the HRES’s energy production is insufficient.

III, while Section IV presents a solution for effective energy distribution using a switch logic architecture. Section V goes into detail about the proposed SCADA system specification. Concluding remarks are given in Section VI. II. SYSTEM DESIGN The designed HRES derives 3 kW of installed capacity from three power generating units: a 1.8 kW pumped-storage hydro power plant, a 600 W wind power plant and a 600 W solar power plant. Each of these units is composed of three substations equal in construction. The produced energy is transmitted from built-in accumulators to feeders via a distribution network composed of indispensable power converters and circuit breakers. The system configuration is shown in Fig. 1. It is important to stress that our choice of the pumpedstorage hydro power plant is based on its interesting nature that can be subject to student research, despite it not being an energy efficient solution. A. Pumped-storage hydro power plant: Level control The pumped-storage hydro power plant operation is defined by the control of two quantities: the water level in the upper tank and the output voltage frequency. Water stored in the upper tank flows through the hydro power plant to generate electricity. Used water flows into the lower tank from where it is being pumped into the upper tank. By controlling the water level in the upper tank we ensure satisfactory energy production. The Simulink model of the water level control loop is shown in Fig. 2.

√ d h(t) + αA 6gh = Q(t), dt

(1)

where a, b and A define the tank’s geometric characteristics.2 The leakage factor α depends upon the construction of the tank orifices through which the water flows out, and is defined as [11]: √ 1 α=µ . (2) 1+ζ The coefficient of contraction µ and the loss coefficient ζ have been computed with respect to the orifice shape. The orifice size has been computed in order to fulfil given requirements regarding the water level and flow. In a stationary state the desired values of these parameters are href = 2 m and Qref = 3 0.03 ms , respectively. The centrifugal pump is modelled as a cascade connection of an electrical and a hydraulic unit. The electrical unit emulates the pump’s motor characteristics and is modelled as a first order system. The hydraulic unit that models the linear dependence of the flow upon the pump’s angular velocity, is given by: Q(t) = ζ ′ ω(t), (3) where ζ ′ is a constant by means of which flow and frictional losses are taken into account [9]. The specific values of the model parameters have been obtained from data sheets of the system components. The tuning of the PID controller with the transfer function G(s) = Kp (1 + T1i s + Td s), was performed using the root locus technique. Fig. 3 shows the system performance for the obtained values of the controller parameters: Kp = 85, Ti = 0, 01 s, Td = 0, 13 s. B. Frequency control Since changes in the connected feeder network effect the frequency of the hydro power plant’s output voltage, a frequency control loop has been designed to ensure the stability √ that the term 6gh results from the fact that the upper tank contains three identical output orifices through which water flows to three water turbines. 2 Note

2.5

105 State of Charge (%)

Set value href System response h

1.5

95 90

0

5

10

15

20 Time t(s)

25

30

35

40

1

1

0.5

0

100

85

Rotor speed ωm (pu)

href (m), h (m)

2

0

2

4

6

8

0.8 0.6 0.4 0.2 0

10

0

2

4

Fig. 3. The water level reaches the set value href = 2 m in less than two seconds.

of the output frequency. Frequency control requires the operation of various mechanical and electrical subsystems. We use a PI controller that produces a certain current signal dependent on the variable output frequency. The current signal is then converted to an equivalent electromagnetic force by a solenoid, as defined by: 1 ∂L(x) F = i2 , (4) 2 ∂x where L is the inductance of the solenoid and x is the position of the plunger that is being moved [12]. The solenoid is connected to a hydraulic valve actuator consisting of a piston and a spring that is placed at the output orifice of the upper water tank. The piston displacement h is actuated by the electromagnetic force of the solenoid through the spring, according to:   if F ≤ Fpr , 0 h = C · (F − Fpr ) · or if Fpr < F < Fmax , (5)   stroke · or if F ≥ Fmax , where C is a constant, Fpr is the spring preload force, Fmax is the spring force at maximum piston displacement, stroke is the piston stroke and or is the actuator orientation with respect to the globally assigned positive direction. As the piston moves over the output orifice of the upper tank, the water flow Q through the orifice changes according to: √ { CD · A ρ2 |p|sgn(p) if Re ≥ Recr , Q= (6) H if Re < Recr , 2CDL · A D ν·ρ p where CDL is a constant, CD is the flow discharge coefficient, ρ is the fluid density, p is the pressure differential, DH is the hydraulic diameter, ν is the fluid kinematic viscosity, A = A(h) is the instantaneous orifice passage area, Re is the Reynolds number and Recr is the critical Reynolds number. It can be seen that the solenoid and the hydraulic valve actuator constitute a subsystem through which the output signal of the PI controller effects the water flow to the turbines and thereby the frequency of the system’s output voltage. Relevant

6

8

10

Time t(s)

Time t(s)

Fig. 4. The output voltage frequency, which is proportional to the rotor speed of the hydro generator reaches the set value after two seconds. At t = 3.7 s the accumulator is fully charged (top) causing the termination of the hydro generator operation and the drop of its frequency to zero (bottom).

responses of the frequency control loop are illustrated in Fig. 4. III. POWER CONVERSION ANALYSIS The primary function of the power converters is to serve as an interface between the generators and the units that either use or store the generated energy. The HRES design requires the use of three types of power converters (see Fig. 1). A. Buck converter Three identical buck converters installed at the outputs of the solar power generators convert the generated 18.3 V DC voltage to the accumulator voltage of 12 V. In order to determine the buck converter parameters, we used the standard buck converter scheme for which (7) and (8) apply [10]. uc = DE

(7)

(1 − D)T 2 uc (8) 8LC The duty cycle value D = 0.655 is defined by the input voltage E = 18.3 V and the output voltage uc = 12 V. The capacity C = 10 µF and the inductance L = 1.2 mH have been computed using (8) for the given period T and output voltage ripple ∆ucmax .3 ∆ucmax =

B. AC–DC converter The three hydro generators of the pumped-storage hydro power plant produce AC voltage of 230 V/50 Hz. Since the generated energy is stored in accumulators rather than being directly transported to the feeders, an AC–DC converter ought to be implemented. As shown in Fig. 5, the desired accumulator DC voltage of 12 V is obtained through three stages of energy conversion. The first conversion is performed by a full-wave bridge rectifier that converts the three-phase input voltage to a rather noised signal with oscillation amplitudes reaching up to 6 V. In order to attenuate such high frequency 3 Note

that the buck converters operate in continuous mode.

Fig. 5.

Powersim model of the AC–DC converter Fig. 7. Powersim model of the DC–AC converter: the IGBT switching frequency is determined by the frequency of the MAN through a frequency synchronisation device.

Fig. 8. The resulting signal of the circuit comprising the boost converter and inverter is the desired 230 V three-phase voltage, suitable for feeder supply.

Fig. 6. Electrical signals specifying the AC–DC conversion: three-phase input voltage (top), rectifier output signal (middle), filter output voltage and AC–DC converter output voltage (bottom)

signal components a passive low-pass filter is applied. The low-pass filter produces a satisfactory DC voltage of 220 V for the following parameters: Rf = 1 kΩ, Cf 1 = 47 µF, Cf 2 = 100 µF. Finally, the use of a Buck converter is required to decrease the value of the filter output voltage to 12 V. The highlights of the signal conversion process are shown in Fig. 6. C. DC–AC circuit The designed distribution grid provides power supply for three-phase loads. Consequently, 12 V DC voltage stored in the accumulators has to be converted to 230 V three-phase voltage suitable for the feeder network. Said conversion has been realised through a circuit that combines a boost converter and an inverter. The circuit diagram is shown in Fig. 7. The role of the boost converter is to increase the accumulator DC voltage to a DC level that after performed DC–AC conversion will produce a three-phase voltage of 230 V. A relatively simple inverter circuit that uses IGBTs as switches performs the DC–AC conversion. Since the inverter produces a rectangular signal, low-pass filters were applied to extract the first harmonic, the desired sinusoid. The cutoff frequency of the low-pass filters was selected with respect to the inverter output signal’s fundamental frequency: π ⇒ fc = 25 Hz, (9) 2πfc = T

where T = 0.02 s represents the period of the inverter output signal. The use of low-pass filters in this case is reasonable considering that the DC component of the inverter output signal is equal to zero. The output signal of the DC–AC circuit is shown in Fig. 8. IV. SWITCH LOGIC ARCHITECTURE The generated energy is used to supply the 1 kW water pump operating in the pumped-storage hydro power plant as well as connected feeders. To optimize the system’s energy distribution we use a network of properly arranged switches shown in Fig. 1. The control of the switch operation is based upon several principles. Firstly, the wind power plant represents the main power source, whereas the solar power plant and the pumped-storage hydro power plant deliver energy if the energy storage of the wind power plant is not sufficient to supply the feeders. Note that the pumped-storage hydro power plant has been accorded lowest priority because it requires the operation of the 1 kW water pump to produce energy. Secondly, each power plant of the HRES contains one main accumulator and two backup accumulators, from which power is derived if the capacity of the main accumulator has dropped beneath 5% of its maximum value. Lastly, the switch operation is determined by the current energy demands. Therefore, the number of accumulators from which power is acquired depends on the total energy consumption of the connected feeder network. Switches SB and SC are closed whenever the HRES’s energy production is higher than the energy consumption of the connected feeders. In case the energy demands do not exceed

200 W, the main accumulator of the wind power plant will supply energy by closing the switches Pw2 and Pw3 . If the discharge of the main accumulator of the wind power plant is detected (meaning that its capacity has dropped beneath 5% of its maximum value), the feeder network is being connected to the first, and if needed, the second backup accumulator. This will result in a change of the switches states, as follows: • Closed switches: Pw1 , Pw3 , Pw5 , Pw6 , Pw8 , Pw9 , SB , SC • Open switches: Pw2 , Pw4 , Pw7 , SA The same logic applies to the switches placed within the solar power plant and the pumped-storage hydro power plant. In a stationary state switch SA is opened, whereas its closing is triggered by energy consumption exceeding the HRES’s stored energy capacity. Should this occur, switches SB and SC are automatically opened, while the feeder network is being redirected to the Metropolitan Area Network through the closed switch SA .

Fig. 9.

GUI main window

V. SCADA SYSTEM SPECIFICATION Monitoring and supervision of the measured parameters and system states requires the implementation of a communication protocol for the microcomputer unitary network. The main computer of the SCADA system collects the measured data and allows for monitoring and supervision through a specifically designed GUI. A. Communication network design Each of the HRES’s units is equipped with a controllerbased system that realises local data acquisition and control while it is being supervised by a microcomputer. All the microcomputers are connected to the main microcomputer through an Ethernet TCP/IP communication network that uses the MODBUS TCP/IP protocol for communication. For the hardware implementation of the MODBUS TCP/IP protocol, we use the 32-bit microcontroller-based EasyARM development system with an attached Ethernet Serial board. The MODBUS TCP/IP software implementation of the server side involves the following steps: • Deserialization of the received data – the input buffer content is copied to a predefined memory structure. • Message decoding • Command execution • Transmission of the response message to the client The MODBUS TCP/IP software implementation of the client side is realised separately. It is based on modelling the MODBUS protocol as a finite state machine. B. SCADA system visualisation The system’s main microcomputer receives current data values of the system parameters from the substations through the Ethernet TCP/IP network. A GUI application on the main microcomputer allows for supervision by creating a visual representation of the collected data. The application starts with the main window that provides access to the specific system units, as shown in Fig. 9. By choosing the pumped-storage hydro power plant on the main

Fig. 10.

Pumped-storage hydro power plant monitoring window

menu, a new window is opened from which the current state of the hydro power plant can be examined (see Fig. 10). Controls on this window provide the operator with information on the current tank level values, hydro generator currents and voltages, accumulator capacities and parameters of the control loops. The Controller Data button enables setting the reference values for the tank level and frequency control loops, as well as the loop’s controller parameters. It can be seen from Fig. 11 and Fig. 12 that the windows for supervision of the wind power plant and solar power plant are organised in a similar manner. The current accumulator capacities of the plants are shown directly on the window, while details on the operating power converters and plant components such as the wind turbine and the photovoltaic panels can be obtained through the Energy Data control. The switch logic architecture is graphically represented on the Feeder Network window (see Fig. 13). It is possible to track the system’s energy distribution and the state of the switches, which varies dependent on the network’s energy demands. The presented GUI was developed within the Microsoft Visual Studio framework and is suitable for Windows platforms.

Fig. 11.

Wind power plant monitoring window

Fig. 13. The energy distribution determined by the energy demands of the feeders is displayed on the Feeder Network window.

Future work prior to the demo system construction will include optimizing the HRES operation by taking into account economical criteria. We also intend to explore different control strategies as well as more sophisticated power converter solutions that have been proposed in recent literature. R EFERENCES

Fig. 12.

Solar power plant monitoring window

VI. C ONCLUSION The herein presented work proposes a complete design of a HRES through three stages: control, providing efficient energy distribution and supervision. We investigated the performance of three different renewable energy subsystems by modelling them and implementing adequate control algorithms. Specific power converters responsible for accurate power supply have been designed. Finally, we created a GUI-based framework within which supervision and control from distant locations can be carried out. The developed models and SCADA system appear to be a promising platform for research on renewable energy. Therefore, the presented concepts aimed to be implemented on a real demo system within a graduate course on renewable energy. By providing a basic HRES design we hope to evoke student interest in renewable energy principles, encourage further research and offer practical experience through working on a versatile design project and its implementation.

[1] U. Bossel, Does a Hydrogen Economy Make Sense?, Proceedings of the IEEE, vol. 24, no. 10, pp 1826 - 1837, nov 2006. [2] The Nuclear Information and Resource Service, False Promises: Debunking the Nuclear Industry Propaganda, www.nirs.org, may 2008. [3] S. Fara, D. Finta and M. Iancu, Efficient Renewable Energy Systems Exploitation Aiming at Reducing Electricity Generation from Conventional Power Plants, Automation, Quality and Testing, Robotics, 2008. AQTR 2008. IEEE International Conference on, pp 91-95, may 2008. [4] P. Meisen, There is no energy crisis, Power Engineering Society Winter Meeting, 2002. IEEE, vol.1, pp 117, aug 2002. [5] A. Zahedi, Computer-based multi-media system on education of ’Renewable energy technology; design and use’, Systems, Man, and Cybernetics, 1998. 1998 IEEE International Conference on, vol.2, pp 1231-1234, oct 1998. [6] E. I. Ortiz-Rivera, J. Gonzalez-Llorente and A. Salazar-Llinas, Bringing renewable energy to the electrical engineering undergraduate education and research at UPRM, Frontiers in Education Conference, 2009. FIE ’09. 39th IEEE, pp 1-6, oct 2009. [7] O. A. Soysal and H. S. Soysal, From wind-solar energy educational demo system (WISE) to sustainable energy research facility (SERF), Power & Energy Society General Meeting, 2009. PES ’09. IEEE, pp 1-6, jul 2009. [8] R. W. Wies, R.A. Johnson and J. Aspnes, Design of an energy-efficient standalone distributed generation system employing renewable energy sources and smart grid technology as a student design project, Power and Energy Society General Meeting, 2010 IEEE, pp 1-8, jul 2010. [9] R. Isermann, Mechatronic Systems: Fundamentals, Springer, Berlin, 2008. [10] M. H. Rashid, Power electronics handbook: devices, circuits, and applications, Elsevier, London, 2007. [11] Z. Andreic, Fluid Mechanics, Faculty of Mining, Geology and Petroleum Engineering, University of Zagreb, Zagreb, 2008. [12] S. E. Lyshevski, Electromechanical systems, electric machines, and applied mechatronics, CRC Press, 2000.