Method and relevant system for converting mechanical energy from a generator actuated by a turbine into electric energy
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
F03D-007/04
F03D-009/00
F03D-007/02
F03D-009/25
출원번호
US-0889308
(2014-05-05)
등록번호
US-9856857
(2018-01-02)
우선권정보
IT-RM2013A0272 (2013-05-08)
국제출원번호
PCT/IT2014/000119
(2014-05-05)
국제공개번호
WO2014/181367
(2014-11-13)
발명자
/ 주소
Vitale, Gianpaolo
Pucci, Marcello
Luna, Massimiliano
출원인 / 주소
CONSIGLIO NAZIONALE DELLE RICERCHE
대리인 / 주소
Harness, Dickey & Pierce, P.L.C.
인용정보
피인용 횟수 :
0인용 특허 :
4
초록▼
A method and system for converting mechanical energy from a generator operated by a turbine into electric energy to be input into an electric network, comprising the following: (A) supplying a power electronic converter connectable to input of the generator and to output of the electric network;(B)
A method and system for converting mechanical energy from a generator operated by a turbine into electric energy to be input into an electric network, comprising the following: (A) supplying a power electronic converter connectable to input of the generator and to output of the electric network;(B) supplying an electrical power outputting the generator to the power electronic converter;(C) adapting the electrical power using the power electronic converter to input the electrical power to the electric network by the following substeps:(C1) calculating rotation speed of the turbine;(C2) calculating wind speed moving the turbine;(C3) calculating reference rotation speed, corresponding to the maximum point of power input to the electric network, using a Maximum Power Point Tracking algorithm;the method comprising substep (C1) implemented by an algorithm carrying out a virtual encoder, and substep (C2) implemented by training a neural network implementing a virtual anemometer.
대표청구항▼
1. A method for converting mechanical energy from a generator operated by a turbine into electric energy to be inputted into an electric network, the method comprising the following steps: (A) supplying a power electronic converter connectable to an output of the generator and to an input of the ele
1. A method for converting mechanical energy from a generator operated by a turbine into electric energy to be inputted into an electric network, the method comprising the following steps: (A) supplying a power electronic converter connectable to an output of the generator and to an input of the electric network;(B) supplying an electrical power from the generator output to the power electronic converter;(C) adapting the electrical power using the power electronic converter to input the electrical power to the electric network by the following substeps: (C1) calculating a rotation speed of the turbine;(C2) calculating an estimated value of a wind speed moving the turbine; and(C3) calculating a reference rotation speed corresponding to a maximum energy of the turbine for the estimated wind speed using a Maximum Power Point Tracking algorithm; wherein substep (C1) is implemented by an algorithm carrying out a virtual encoder, and substep (C2) is implemented by training a neural network implementing a virtual anemometer; and wherein a virtual encoder, implemented by substep (C1), is obtained according to the further following substeps: (C1a) supplying an output voltage of generator to the power electronic converter, the output voltage having a residual ripple voltage component overlapping a DC voltage component;(C1b) increasing the DC voltage component of the output voltage to reach a constant value;(C1c) converting the constant value into a single-phase alternate voltage value for the connection with the electric network;(C1d) supplying a band-pass filter for selecting a fundamental frequency of the residual ripple voltage component;(C1e) evaluating a fundamental frequency of the fundamental frequency of the residual ripple voltage component by a zero-crossing algorithm; and(C1f) calculating the rotation speed of the turbine using the formula: ωT=fripple·π3p, wherein p is the number of polar pairs of the generator. 2. The method according to claim 1, wherein the virtual anemometer, implemented using substep (C2), is obtained according to the further following substeps: (C2a) training the neural network, in an off-line initial calibration mode of the system, giving to the neural network at least a triad of values of rotation speed, wind speed and total power;(C2b) supplying the rotation speed value , calculated in substep (C1f), to the neural network;(C2c) supplying a value of power PT to the neural network, wherein PT is obtained by PT=Pel/η, and η is the efficiency of the electric generator; and(C2d) obtaining from the output of the neural network an estimated value of wind speed necessary for the whole energy conversion process. 3. The method according to claim 2, wherein the algorithm calculates the reference rotation speed value of the turbine, corresponding to the maximum power point, using the formula ωT ref=νstimλopt/R, wherein λopt is the optimal value of the known value λ=ωTR/ν, wherein R is the radius of the turbine and V is an instantaneous wind speed. 4. The method according to claim 3, wherein the reference rotation speed value of the turbine is sent to a control system of the converter for tuning the rotation speed of the turbine and to a control system of an inverter for controlling the input of power into the electric network. 5. The method according to claim 2, wherein in off-line substep C2a, the neural network learns the relationship among three physical quantities V, PT and ωT, and outputs the result of its training consisting of weights in form of numerical coefficients obtained using the characteristic function of the network, and during the conversion of the mechanical energy into electrical energy, in recall substep C2d weights are used by the neural network for estimating the wind speed νstim in real-time using the ability of the neural network of inverting the function νstim=f−1(PT, ωT) starting from instantaneous data of PT and 107T. 6. The method according to claim 5, wherein during substep C2d weighted sums of Euclidean distances between points are executed, comprising the sets of vectors of the input data and corresponding vectors, which constitute weights of neurons, in a tridimensional space stored by a 3×n sized data matrix, wherein n is the number of neurons of the neural network selected for representing the characteristic data of the turbine. 7. The method according to claim 2, wherein the type of neural network is “Growing Neural Gas”.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (4)
Spruce, Christopher; Turner, Judith; Evans, Martin; Bowyer, Robert, Over-rating control of wind turbines and power plants.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.