THE STUDY OF WORK OF NEURO FUZZY SYSTEM OF TRACKING THE POINT OF MAXIMUM POWER OF PHOTOVOLTAIC INFERENCE
Ivanets Sergey, Chernihiv National University of Technology (95 Shevchenka Str., 14027 Chernihiv, Ukraine)
Krasnozhon Oleksii , Chernihiv National University of Technology (95 Shevchenka Str., 14027 Chernihiv, Ukraine)
Urgency of the research. Modern electric industry of Ukraine is characterized by high demand in alternative energy sources - photovoltaic - the level of power generation of which is greatly depends on the mode of operation.
Target setting. Operating mode of photovoltaic is determined by flux density of incident solar radiation and its operating temperature. Maximum power generation, under these conditions, is possible only when the load impedance of photovoltaic has a corresponding meaning. Thus, calculating of load resistance value by changing the parameters of the environment is a serious problem.
Actual scientific researches and issues analysis. In recent studies of proposed methods for maximum power point tracking of photovoltaic is proposed to use methods which are based on voltage feedback, current feedback, as well as the observations of the power fluctuations.
Uninvestigated parts of general matters defining. Despite of the fact that a lot of research has been done on this subject, they do not completely solve the problem of calculating the coordinates of the maximum power point of photovoltaic by the change of environmental parameters.
The research objective. Aim of the paper is to offer a structure of the fuzzy maximum power point tracking system of photovoltaic; to justify the choice of system parameters; to perform its initial setup; to form the content of the training set for artificial neural network; to learn the network with corresponding structure; evaluate the efficiency of the system by simulation of its work.
The statement of basic materials. Maximum power point tracking systems are: fuzzy inference systems with Sugeno type, the number of input linguistic variables - 2, output - 1, fuzzy membership functions for input variables - 5 for the output - 25. Fuzzy membership function types are – triangular-trapezoidal or Gauss. Knowledge base of the system contains 25 rules. The training set is based on an analytical description of the surface control, which was obtained before by using hyperbolic approximation.
Conclusions. It is proposed and justified the structure of artificial neuro fuzzy inference system of maximum power point tracking of photovoltaic, completed initial setup, created the content of the training set for neural network, completed its studies for achievement an operation optimum, performed system simulation, investigated the relative error for calculation of the photovoltaic load resistance.
maximum power point control surface, universal approximation, linguistic variable, membership function, artificial neural network, training set, modeling, Mamdani, Sugeno, relative error, fuzzy logic, photovoltaic
1. Ivanets, S.A., Krasnozhon, O.V. (2013). Rozvytok elektroenerhetyky na osnovi kontseptsii “intelektualnykh” elektrychnykh merezh Smart Grid [[The development of power industry on the basis of "Smart Grid" conception of "smart" electrical grids]. Visnyk Chernihivskoho derzhavnoho tekhnolohichnoho universytetu. Seriia «Tekhnichni nauky» – Visnyk of Chernihiv State Technological University. Series “Engineering science”, no. 1 (63), pp. 167–178 (in Ukrainian).
2. Ivanets, S.A., Krasnozhon, O.V. (2014). Vykorystannia nechitkoi lohiky v systemakh vidstezhennia tochky maksymalnoi potuzhnosti fotoelektrychnykh peretvoriuvachiv [The using of fuzzy logic for development of photovoltaic maximum power point tracking systems]. Visnyk Chernihivskoho derzhavnoho tekhnolohichnoho universytetu. Seriia «Tekhnichni nauky» – Visnyk of Chernihiv State Technological University. Series “Engineering science”, no. 2 (73), pp. 180–187 (in Ukrainian).
3. Ivanets, S.A., Krasnozhon, O.V. (2015). Otrymannia analitychnoho opysu poverkhni keruvannia dlia systemy vidstezhennia tochky maksymalnoi potuzhnosti fotoelektrychnoho peretvoriuvacha [Obtaining of an analytical description of the control surface for the system tracking maximum power point of a photovoltaic cell]. Visnyk Chernihivskoho derzhavnoho tekhnolohichnoho universytetu. Seriia «Tekhnichni nauky» – Visnyk of Chernihiv State Technological University. Series “Engineering science”, no. 2 (78), pp. 187–196 (in Ukrainian).
4. Ivanets, S.A., Krasnozhon, O.V. (2016). Poshuk optymalnoi stepenevoi aproksymatsii poverkhni keruvannia dlia systemy vidstezhennia tochky maksymalnoi potuzhnosti fotoelektrychnoho peretvoriuvacha [The search for an optimal degree approximation of the surface control for maximum power point tracking system of photovoltaic converter]. Tekhnichni nauky ta tekhnolohii – Technical sciences and technologies, no. 2 (4), pp. 91–104 (in Ukrainian).
5. Rutkovskaia, D., Pilinskii, M., Rutkovskii, L. (2006). Neironnye seti geneticheskie algoritmy i nechetkie sistemy [Artificial neural networks, genetic algorithms and fuzzy systems] (I.D. Rudinskii, Trans.). Moscow: Goriachaia liniia Telekom (in Russian).
6. Tsai Huan-Liang, Ci-Siang Tu, Yi-Jie Su (2008). Development of Generalized Photovoltaic Model Using MATLAB/SIMULINK. Proceedings of the World Congress on Engineering and Computer Science. October 22–24.
7. Chernykh, I.V. (2007). Modelirovanie elektrotekhnicheskikh ustroistv v MATALB SimPowerSystems i Simulink [Simulation of electrical devices in MATLAB. SimPowerSystems and Simulink]. Moscow: DMK Press (in Russian).