Comparative Study Regarding Control of Wind Energy Conversion Systems Based on the Usage of Classical and Adaptive Neuro Fuzzy Controllers

Transcription

1 Coparative Stuy Regaring Control of Win Energy Conversion Systes Base on the sage of Classical an Aaptive Neuro Fuzzy Controllers Iosif Szeiert Departent of Autoation an Inustrial Inforatics, Faculty of Autoation an Coputer Sciences, Politehnica niversity fro Tiisoara, Av. V. Parvan, No. 2, , Tiisoara, Roania, Phone: ( , Fax: ( , Abstract: The paper presents a coparative stuy regaring the control (using an aaptive neuro-fuzzy controller an a PD controller base on the siulation of win energy conversion systes functioning. There are consiere several siulations base on asynchronous generator usage, by using the eicate MATLAB-PSB (Power Syste Blockset toolbox ipleentations. Keywors: Win energy conversion systes, oeling, siulation, control, asynchronous generator, aaptive neuro-fuzzy controller, MATLAB-Siulink environent 1 Introuction The technical an scientific progresses, cobine with the stiulation politics of ecological energy prouction an consuption conucte to a spectacular evelopent of WECS (Win Energy Conversion Systes. The econoicalpolitical worlwie trens inicate that the win energy inustry looks forwar to a significant expansion. The paper escribes a WECS structure ipleente in the MATLAB-Siulink siulation environent by using the specialize PSB toolbox, esigne for oeling an siulation of energetic an electrical coponents. In Figure 1 there is presente the MATLAB - Siulink iagra of the consiere winill, which presents the following coponents: asynchronous achine, synchronous achine, win turbine, frequency regulator an up loa. The practical probles in the gri integration of winill represent special cases of esign an analysis of energetic power systes. [2], [3], [5]

2 2 Win Energy Conversion Systes Control Structures Issues The asynchronous achine oel is base on the Park q equations (etaile in appenix. The asynchronous achine operates in the generating regie. In this regie, the input into the achine is the echanical energy an the output is the electrical energy. [1] The rotor spee ust be ajuste in concorance to the changing win spee in orer to achieve the axiu aeroynaic efficiency. In this WECS, the win turbine is oele as siple controlle echanical torque source that supplies the asynchronous generator. [7], [6] The frequency regulator input is represente by the voltage s frequency. There is use a three-phase Phase Locke Loop (PLL syste to easure the frequency of the 3-phase voltage of the network. Therefore, the easure frequency is copare to the reference frequency in orer to obtain the frequency error. Afterwars, this error is integrate in orer to obtain the phase. A classic PD type controller erives the error phase. Figure 1 The MATLAB-Siulink WECS Diagra The obtaine signal (analogue signal is afterwars converte to an 8-bit signal that coans the switching eleents fro the up loa. The regulator s output represents the esire power of the up loa. The up loa is use to issipate the excess power prouce by the winill an siultaneously to aintain constant the frequency. The up loa consists of eight three-phase resistors

3 connecte in series with GTO (Gate Turn-Off type base switches. The up loa uses an 8-bit binary coan. [3] In the consiere winill control structure, the synchronous achine presents only the role of a synchronous copensator. In Figure 2 there is etaile the structure of the consiere frequency regulator. This structure has been use in the first stuy cases (SC1. Figure 2 The frequency regulator The secon stuy case (SC2 was consiere by using a oifie control structure. The PLL block an conventional PD controller has been replace with an aaptive neuro-fuzzy controller. The controller consiers the prescribe frequency an phase reference values constant. The propose oifie structure is presente in Figure 3. Figure 3 The frequency regulator base on an aaptive neuro-fuzzy controller

4 3 Win Energy Conversion Systes Stuy Cases Siulation Results 3.1 Stuy Case SC1 The Asynchronous Generator s Noinal Functioning Regie at Rate Power The first stuy case is consiere the noral functioning regie at rate power an rate win spee. The win spee is consiere having the average value of 8 (/s uring the siulation tie perio. The win spee presents a rano variation. The siulation interval was set to 3 secons. There is connecte only a ain consuer of 50 kw. At t 1 =1 [s] there is also connecte a seconary consuer with noinal power of 40 kw. At t 2 =2 [s] the seconary consuer is isconnecte. This fact can be notice in Figure 6, which represents the evolution of the consiere loa power. In Figure 4, is represente the evolution of the electrical energy frequency. It can be observe that in the conition of a continuous win spee variation an of a iffrent loa power the entire WECS control structure succees to control the electrical energy paraeters (voltage an gri frequency. There can be notice that the frequency presents only slightly variations. In Figure 5, is represente the asynchronous achine spee (rp in pu (per unit units. There can be notice that the rotation spee is slightly over the synchronous spee because the achine operates in generating oe. The win turbine s power evolution (kw is represente in pu units (Figure 7. It can be conclue that the consiere WECS s control structure presents quite goo control perforances. [8], [9] Figure 4 Frequency (Hz Figure 5 Asynchronous achine s spee (pu

5 Figure 6 Loa power (kw Figure 7 Win turbine power (kw 3.2 Stuy Case SC2 The Asynchronous Generator s Noinal Functioning Regie at Rate Power (Consiering the Aaptive Neuro-Fuzzy Controller In this stuy case there were consiere the sae siulation conitions as in the previous case (case SC1, but using the oifie control structure base on an aaptive neuro-fuzzy controller. The obtaine results are presente in Figures 8 an 9, the frequency an respectively, the asynchronous achine s spee. As it can be rearke the control structure presents siilar perforance level, which still satisfies the ipose goal of the winill control. The oifie control structure sees to represent a viable alternative control solution for this WECS structure. Figure 8 Frequency (Hz Figure 9 Asynchronous achine s spee (pu

6 Conclusions There can be conclue that both controller structures present siilar control perforances. The usage of the aaptive neuro-fuzzy controller presents the avantage of an easier ipleentation on specialize coputational structures, such as the case of control algoriths ipleente on DSP processors or icrocontroller base control systes. The consiere stuy cases present significance in the oain of the global renewal energies usage, especially in the case of WECS. However, in orer to stuy thoroughly the WECS perforances there are anatory etaile analysis of the behavior of all ain coponents of the WECS line: win turbine, electrical generator, converter an gri. [4] There can be conclue that in the consiere stuy cases (noinal functioning regies the consiere WECS structure presents an overall goo perforance regaring the ipose energy power paraeters in the context of the integration into a istribute power gri (non-autonoous win fars. References [1] Mihet-Popa L., Blaabjerg F., Bolea I., Win turbine Generator oeling an Siulation where rotational spee is the controlle variable, IEEE Transactions on Inustry Applications, Vol. 40, Issue 1, pp. 3-10, Jan.-Feb [2] P. Gipe, Win Power: Renewable Energy for Hoe, Far, & Business, Chelsea Green Publishing Co, ISBN , 2004 [3] N. Buisan, Probles of inuction generator systes at win/iesel/hyro/etc. unconventional electric groups /plants /fars, Eitura Politehnica, Tiisoara, 2003 [4] J. F. Manwell, J. G. McGowan, an A. L. Rogers, Win Energy Explaine Theory, esign an application, John Wiley & Sons Lt, ISBN , 2002 [5] Gasch an J. Twele, Win Power Plants - Funaentals, Design, Construction an Operation, Jaes & Jaes (Science Publishers Lt, ISBN , 2002 [6] I. Szeiert, O. Prostean, N. Buisan, I. Filip, Two Axis Moeling of Inuction Generators for Winills, 2002 Global Winpower Conference, Paris, France, 2002 [7] R. Gagnon, B. Saulnier, G. Sybille, P. Giroux, Moeling of a Generic High-Penetration No-Storage Win-Diesel Syste sing Matlab/Power Syste Blockset, 2002 Global Winpower Conference, Paris, France, 2002 [8] H. Siegfrie, Gri Integration of Win Energy Conversion Systes, Wiley, an Publisher: John Wiley & Sons, ISBN X, 1998 [9] MATLAB Docuentation:

7 Appenix Asynchronous achine oel The atheatical oel of the asynchronous achine presents two subcoponents: the electrical syste an the echanical syste. The equations (1-(9 represent in fact the Park q - asynchronous achine oel: Electrical syste equations: qs = RSiqs + ( Lsiqs + Liqr + ω ( Lsis + Lir (1 s = RSis + ( Lsis + Lir ω ( Lsiqs + Liqr (2 qr = Rriqr + ( Lriqr + Liqs + ( ω ωr ( Lrir + Lis (3 r = Rrir + ( Lrir + Lis ( ω ωr ( Lriqr + Liqs (4 ( L i + L i i ( L i + L i i 3 T e = p s s r qs s qs qr (5 s 2 Where: L + s = Lls L an L r = Llr + L Mechanical syste equations: 1 ω = ( Te T an θ = ω 2J (6-(7 (8-(9 Where: R s, L - stator resistance an leakage inuctance, ls R r, L - rotor resistance lr an leakage inuctance, L - agnetizing inuctance, L s, Lr - total stator an rotor inuctances,, i - q axis stator voltage an current,, i - q axis rotor qs qs qr qr voltage an current, s, is - axis stator voltage an current, r, ir - axis rotor voltage an current, ω - angular velocity of the rotor, θ - rotor angular position, p - nuber of pole pairs, ωr - electrical angular velocity ( ω p, θ r - electrical rotor angular position ( θ p, Te - electroagnetic torque, T - shaft echanical torque, J - cobine rotor an loa inertia constant.