CN105515029A - Control method and device for flywheel energy storage system - Google Patents

Abstract

The invention provides a control method and a device for a flywheel energy storage system. The method comprises the steps of a, with the angular velocity, the d-axis current and the q-axis current of A, a flywheel as state variables, and the energy storage capacity of the flywheel as a control variable, constructing a model for the energy storage system of the flywheel, and discretizing the model according to the sampling time; B, constructing a performance index based on a second-order Renyi entropy, and determining an optimal input control variable increment with the partial derivative of the input control variable increment to be zero based on the performance index; C, determining a current optimal input control variable based on the optimal input control variable increment. According to the control method and device, based on the simplified energy storage model of the flywheel, a controller for the energy storage system of the flywheel can be designed. Meanwhile, the excess energy of the flywheel is stored, or the deficient energy of the flywheel is supplemented. In this way, a wind turbine generator set is ensured to transmit a smoothing power to the power grid.

Description

The control method of flywheel energy storage system and device

Technical field

The present invention relates to field of renewable energy technology, particularly relate to wind turbine control system field.

Background technology

Wind power generation alleviates the energy demand pressure of China to a certain extent, improves the energy resource structure of China, promotes sustainable economic and social development.But the operation characteristic of the uncertainty of wind-resources and Wind turbines itself determines Wind turbines has very strong intermittence, fluctuation and anti-peak-shaving capability, bring very large impact to the operation of electric power system and planning.Wherein be apparent that most, can not expect and the characteristic such as stochastic volatility because wind speed has, the active power of output of Wind turbines below rated wind speed is fluctuated with wind speed change.The fluctuation of Wind turbines active power of output may produce larger impact to the electrical network quality of power supply, reduces the stability of electrical network.Therefore, require that Wind turbines exports comparatively level and smooth active power while realizing maximal wind-energy capture.

The level and smooth control of active power of output is the key technology of Wind turbines.The present invention is directed to Wind turbines active power and smoothly control existing problem, utilize that the flywheel energy storage system life-span is long, energy storage density is large, energy conversion rate is high, by discharge and recharge number of times restriction, convenient for installation and maintenance and to advantages such as environmental hazard are little, on the basis that maximal wind-energy capture controls, flywheel energy storage system control program based on wind generator system is proposed, by the control to flywheel energy storage system, realize the level and smooth control of Wind turbines active power of output, thus stabilize active power fluctuation while raising wind energy utilization efficiency.

Flywheel energy storage system complex structure, has a lot of uncertain parameter, is a non linear system, needs to carry out Special controlling for its feature.But current existing techniques in realizing is complicated, the flywheel energy storage control system that the randomness being difficult to solve wind speed causes is with the problem of randomness.Consider unsteadiness and the randomness of this system, the present invention devises Probability density functions control device.Probability density function shape controlling (PDF) method is proposed by professor Wang Hong, mainly for the industrial process with random signal, the distribution of such as, fiber length distribution in paper-making process, the grain in grain processing and boiler flame temperature distribution etc.The direct CONTROLLER DESIGN of these class methods exports PDF distribution shape tracing preset PDF distribution shape to make system.

Summary of the invention

Given this, the object of the invention is to overcome control system in prior art and realize complicated, the flywheel energy storage control system that the randomness being difficult to solve wind speed causes, with the problem of randomness, provides a kind of control method and device of flywheel energy storage system.

In order to realize this object, the technical scheme that the present invention takes is as follows.

A control method for flywheel energy storage system, described method comprises step:

A, with the electric current of the angular speed of flywheel, d axle and q axle be state variable, variable resistor is for input control amount, the model that the model that flywheel energy storage is output variable structure flywheel energy storage system, flywheel energy storage are output variable structure flywheel energy storage system, according to the sampling time by model discretization;

B, building performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

C, optimum input control amount increment is utilized to determine current optimum input control amount.

Wherein determine that described performance index are:

The tracking error in system k moment is e _k=P _ref-y _k, for the probability density function of k moment error sample, P _reffor reference power value, y _kfor output variable;

Second order Renyi entropy is: $H_k=-log\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\γ}}_{e_k}{\left(\mathrm{\τ}\right)}^{2}d\mathrm{\τ}=-\log V_k;$

Error mean square value E _kfor: $E_k=\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\τ}}^2{\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)d\mathrm{\τ};$

Performance index are:

J(u _k)＝-R ₁V _k+R ₂E _k+0.5R ₃u _k ²；

Wherein R ₁, R ₂, R ₃for every weight, u _kfor input control amount, V _kfor information potential.

Determine e _kprobability density function for as follows,

Assuming that k moment tracking error sample is S _k={ e ₁, e ₂... e _n, so k moment probability density function is:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=\frac{1}{N}\underset{i=1}{\overset{k}{\Σ}}\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2);$

Wherein N is sample error number, and ψ is gaussian kernel function, and δ is the parameter of gaussian kernel function, and its expression formula is: $\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2)=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\δ}}{e}^{-\frac{{(\mathrm{\τ}-e_i)}^2}{2{\mathrm{\δ}}^2}};$

Pass between the probability density function of adjacent moment is:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=(1-\mathrm{\ξ}){\mathrm{\γ}}_{e_{k-1}}\left(\mathrm{\τ}\right)+\mathrm{\ξ}\mathrm{\ψ}(\mathrm{\τ}-e_k,{\mathrm{\δ}}^2).$

Wherein comformed information gesture V _kfor:

$V_k=(1-\mathrm{\ξ})V_{k-1}+\frac{\mathrm{\ξ}}{L}\underset{i=k-L+1}{\overset{k}{\Σ}}\mathrm{\ψ}(e_i-e_k,2{\mathrm{\δ}}^2);$

Wherein L is window width, and ξ is forgetting factor, be less than 1 coefficient.

Especially, be zero determine that optimum input control amount increment comprises according to the local derviation of performance index to input control amount increment:

Performance index are:

J(u _k)≈Q ₀+Q ₁Δu _k+0.5Q ₂Δu _k ²+0.5R ₃(u _k-1+Δu _k) ²；

Wherein $Q_0=Q_{k|u_k=u_{k-1}},$

$Q_1=\frac{\∂Q_k}{\∂u_k}|u_k=u_{k-1},$

$Q_2=\frac{{\∂}^2{Q}_k}{\∂{u_k}^2}|u_k=u_{k-1},$

Q _k＝-R ₁V _k+R ₂E _k，

Wherein V _kfor information potential, wherein R ₁, R ₂for every weight.

Order determine optimum input control amount increment Delta u _k.

A control device for flywheel energy storage system, described device comprises the system model construction unit, controlled quentity controlled variable increment optimization unit and the input control amount that are linked in sequence and optimizes unit, wherein, comprises step:

System model construction unit be used for the electric current of the angular speed of flywheel, d axle and q axle be state variable, variable resistor for input control amount, flywheel energy storage is that output variable builds the model of flywheel energy storage system, according to the sampling time by model discretization;

Controlled quentity controlled variable increment is optimized unit and is built performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

Input control amount is optimized unit and is utilized optimum input control amount increment to determine current optimum input control amount.

By adopting control method and the device of energy-storing and power-generating system of the present invention, following beneficial effect can be realized.

1, the present invention is based on the flywheel energy storage model of simplification, the controller of design flywheel energy storage system, utilize flywheel to store excess energy or fill up not enough energy to ensure that wind turbine generator carries smooth power to electrical network.

2, consider that flywheel energy storage control system that the randomness of wind speed causes is with randomness, therefore estimate by design density function controller solves the stochastic problems of system.

Therefore the present invention has enriched the method for designing of flywheel energy storage system controller, improves the control performance of wind turbine generator, and further illustrates the control problem that Probability density functions control device can solve stochastic system preferably.

Accompanying drawing explanation

Fig. 1 is wind-driven power generation control system structural representation.

Fig. 2 is the control method schematic diagram of the flywheel energy storage system according to the specific embodiment of the invention.

Fig. 3 is the flywheel drive motors system of equivalence.

Fig. 4 is the schematic diagram that wind speed change at random is described.

Fig. 5 illustrates the information potential of entropy and the variation diagram of error mean square value.

Fig. 6 is the performance index variation diagram according to the specific embodiment of the invention.

Fig. 7 is q shaft current and tracking error variation diagram.

Fig. 8 is the probability density function variation diagram of tracking error.

Fig. 9 is dq shaft current variation diagram.

Embodiment

Below in conjunction with accompanying drawing, the present invention is elaborated.

The example embodiment that following discloses are detailed.But concrete structure disclosed herein and function detail are only the objects for describing example embodiment.

But should be appreciated that, the present invention is not limited to disclosed concrete example embodiment, but covers all modifications, equivalent and the alternative that fall within the scope of the disclosure.In the description to whole accompanying drawing, identical Reference numeral represents identical element.

Should be appreciated that, term "and/or" as used in this comprises one or morely relevant lists any of item and all combinations simultaneously.Should be appreciated that in addition, when parts or unit are called as " connection " or " coupling " to another parts or unit, it can be directly connected or coupled to miscellaneous part or unit, or also can there is intermediate member or unit.In addition, other words being used for describing relation between parts or unit should be understood according to identical mode (such as, " between " to " directly ", " adjacent " to " direct neighbor " etc.).

As shown in Figure 1, wind-driven power generation control system is very complicated, and comprise generator, rectifier, inverter etc., the controller mainly for flywheel energy storage system in the specific embodiment of the invention adjusts.

Before introducing the specific embodiment of the present invention, be first described in conjunction with general principle of the present invention.

Generally speaking, control method of the present invention comprises following flow process.

Invention adopts permagnetic synchronous motor as flywheel drive motors, and according to the knowledge of motor, and consider that the dynamic process of power electronic power device is more a lot of soon than the electromechanical dynamic process of electricity generation system, AC-DC converter and loaded portion equivalence become inductance L _swith variable resistor R _stwo-part parallel connection, as the equivalent load of permagnetic synchronous motor.As shown in Figure 3, equivalent variable resistance R _sresistance change along with handing over the change of the control impuls duty ratio of orthogonal variator, the flywheel energy storage model form therefore built is as follows:

The state variable of system is made to be $x=\left[\begin{array}{c}{i}_d\\ i_q\\ {\mathrm{\ω}}_f\end{array}\right],$ Input control amount u=R _s, output variable is energy and the y=P of flywheel storage, then system model can be written as following form:

$\left\{\begin{array}{c}\left[\begin{array}{c}\stackrel{.}{x_1}\\ \stackrel{.}{x_2}\\ \stackrel{.}{x_3}\end{array}\right]=\left[\begin{array}{c}\frac{-{\mathrm{Rx}}_1+p(L_q-L_s)x_2{x}_3}{(L_d+L_s)}\\ \frac{-{\mathrm{Rx}}_2-p(L_d+L_s)x_1{x}_3+{\mathrm{p\ηx}}_2}{(L_q+L_s)}\\ \frac{-{\mathrm{p\ηx}}_2}{J_f}\end{array}\right]+\left[\begin{array}{c}-\frac{x_1}{(L_d+L_s)}\\ -\frac{x_2}{(L_q+L_s)}\\ 0\end{array}\right]u\\ y={\mathrm{p\ηx}}_2{x}_3\end{array}\right.,$

If the sampling time is T _s, by system discretization, we can obtain:

$\left\{\begin{array}{c}\left[\begin{array}{c}{x}_{1,k+1}\\ x_{2,k+1}\\ x_{3,k+1}\end{array}\right]=\left[\begin{array}{c}{x}_{1,k}\\ x_{2,k}\\ x_{3,k}\end{array}\right]+T_s\left[\begin{array}{c}\frac{-{\mathrm{Rx}}_{1,k}+p(L_q-L_s)x_{2,k}{x}_{3,k}}{(L_d+L_s)}\\ \frac{-{\mathrm{Rx}}_{2,k}-p(L_d+L_s)x_{1,k}{x}_{3,k}+{\mathrm{p\ηx}}_{2,k}}{(L_q+L_s)}\\ \frac{-{\mathrm{p\ηx}}_{2,k}}{J_f}\end{array}\right]+T_s\left[\begin{array}{c}-\frac{x_{1,k}}{(L_d+L_s)}\\ -\frac{x_{2,k}}{(L_q+L_s)}\\ 0\end{array}\right]u_k\\ y_k={\mathrm{p\ηx}}_{2,k}{x}_{3,k}\end{array}\right.,$

On the basis setting up system model, utilize Probability density functions control device, the power smooth controlling Wind turbines exports.

The present invention is according to tracking error estimate by design density function controller, because this system exists randomness, consider that entropy can characterize the uncertainty of any stochastic variable, the output probability density function dispersiveness meaning tracking error that minimizes of this control system tracking error entropy reduces, show as not only point but also narrow distribution in shape, therefore second order Renyi entropy is incorporated in system performance index by the present invention, also need the mean-square value considering tracking error in addition, be introduced in performance index and make mean square of error value little as far as possible, even close to zero.Show that the probability density function of error is in shape for peak value appears at about zero.

The present invention chooses performance index: J (u _k)=-R ₁v _k+ R ₂e _k+ 0.5R ₃u _k ²;

Wherein u _kfor the input control amount in k moment, E _kfor the mean-square value of tracking error, R ₁, R ₂, R ₃for every weight, V _kbe the information potential of second order Renyi entropy, last is the bound term of controlled quentity controlled variable.

Definition Q _k=-R ₁v _k+ R ₂e _k, according to optimal control theory, order optimal control increment can be obtained:

${\mathrm{\Δu}}_k=-\frac{Q_1+R_3{u}_{k-1}}{R_3+Q_2};$

Wherein $Q_0=Q_{k|u_k=u_{k-1}},Q_1=\frac{\∂Q_k}{\∂u_k}|u_k=u_{k-1},Q_2=\frac{{\∂}^2{Q}_k}{\∂{u_k}^2}|u_k=u_{k-1}.$

By u _k=Δ u _k+ u _k-1, the controlled quentity controlled variable u in k moment _k.

Therefore, in the specific embodiment of the present invention, be attached to the characteristic of flywheel energy storage system, flywheel energy storage system control method of the present invention comprises the following steps.

Step 1: build the model based on flywheel energy storage system;

Due to the randomness of wind speed, during by Wind turbines maximum power curve controlled, generator active power of output can fluctuate with the change of wind speed, thus the active power that net side converter is exported to electrical network fluctuates thereupon.In order to realize the level and smooth control of active power, the present invention introduces energy accumulation device for fly wheel in current transformer DC side, utilizes the energy conversion effect of energy accumulation device for fly wheel to realize power smooth and controls.When needs storage power, convert electrical energy into mechanical energy storage in flywheel wheel disc; When needs release energy, be electric energy by the changes mechanical energy in flywheel wheel disc.

At the DC side access flywheel energy storage system of current transformer, we expect that current transformer surveyed by net is P to the level and smooth active power that electrical network exports _g, the active power that outputs to flywheel energy storage system is P, generator exports active power is P _s.By energy relationship, the active power that generator exports equals to net the power needed for surveying and adds the active power flowing into energy accumulation device for fly wheel, namely meets following relation:

P _s＝P _g+P，

The then reference value P of flywheel energy storage control system power output _ref=P _s-P _gas long as the power P of the actual output of flywheel energy storage system exports along reference value, then Wind turbines just can to electrical network output smoothing power.

The present invention only considers the Controller gain variations of flywheel energy storage system, and therefore ignore Wind turbines power stage and control, below rated wind speed, wind turbine generator presses maximal wind-energy capture principle power output, makes wind energy conversion system can catch maximal wind-energy P _max, by aerodynamics its expression formula known be:

P _max＝0.5ρAC _pamxV ³，

Wherein ρ is atmospheric density, and A is blade fan sweeping area; C _pamxfor maximal wind-energy usage factor.

All wind energies flowed through can not be converted to electric power by wind-driven generator, most high conversion efficiency is about 59% in theory, in fact most blade converting wind energy efficiency, about between 30% ~ 50%, converts the total delivery efficiency after electric energy to through electromechanical equipment and is about 20% ~ 45%.Therefore, we set generator system energy conversion efficiency as α, then the active power that generator exports is:

P _s＝αP _max＝0.5αρAC _pamxV ³，

Wherein α is between 20% and 45%.

In order to make grid side power output P _glevel and smooth output, the key realizing the level and smooth control strategy of active power is that the computational methods of smooth power and the active-power P of flywheel energy storage system control.

Adopt mean wind speed to replace actual wind speed in present embodiment and calculate smooth power P _g, that is:

$P_g=0.5{\mathrm{\α\ρAC}}_{pamx}{\stackrel{\‾}{V}}^3$

Wherein for mean wind speed.

Flywheel energy storage system generally comprises flywheel rotating disk, bearing, drive motors, power converter etc.During flywheel energy storage system work, according to the needs of energy flow, inverter control flywheel drive motors flywheel driven rotating disk carries out adjustment of rotational speed, realizes energy buffer effect by the conversion of electric energy and mechanical energy.Permagnetic synchronous motor (PMSM) because its structure is simple, speed governing is convenient, be easy to control, functional density is large, easy to maintenance, again without excitation loss, power speed control wide ranges and be easy to realize the features such as to and fro flow of power, adopt permagnetic synchronous motor as flywheel drive motors in native system.

Fly wheel system is a kind of device utilizing physical method to realize mechanical energy and electric energy to change.When system power is had a surplus, electric energy by convertor controls electric machine rotation, thus makes flywheel follow the rotation of motor and accelerate to rotate, and converting electric energy is that the mechanical energy of rotary flyweights is able to storage in the process; If during noenergy conversion, motor keeps constant revolution, and fly wheel system does not carry out energy exchange; When the supply of system required power, motor obtains reduce-speed sign, and flywheel deceleration rotates, and the mechanical energy of flywheel is converted to electric energy by motor, then is transferred to electrical network through converter.

Permagnetic synchronous motor has multivariable, close coupling and the feature such as non-linear as flywheel drive motors, and adopt rotor field-oriented vector control mode, the Mathematical Modeling under d, q synchronous rotating frame can be expressed as:

$\left\{\begin{array}{c}{u}_d={\mathrm{Ri}}_d+L_d\frac{{\mathrm{di}}_d}{dt}-{\mathrm{pL}}_q{i}_q{\mathrm{\ω}}_f\\ u_q={\mathrm{Ri}}_q+L_q\frac{{\mathrm{di}}_q}{dt}+{\mathrm{pL}}_d{i}_d{\mathrm{\ω}}_f+{\mathrm{p\η\ω}}_f\end{array}\right.;$

Wherein i _dfor stator current d axle component, i _qfor stator current q axle component, p is flywheel drive motors number of pole-pairs, and η is rotor permanent magnet magnetic linkage, L _d, L _qbe respectively the inductance of d axle and q axle, u _dand u _qbe respectively the voltage of d axle and q axle.R is stator resistance, ω _ffor the angular speed of flywheel drive motors.

Consider that the dynamic process of power electronic power device is more a lot of soon than the electromechanical dynamic process of electricity generation system, therefore ignores, AC-DC converter and loaded portion equivalence become inductance L _swith variable resistor R _stwo-part parallel connection, as the equivalent load of permagnetic synchronous motor.As shown in Figure 3, equivalent variable resistance R _sresistance change along with handing over the change of the control impuls duty ratio of orthogonal variator; Therefore

$\left\{\begin{array}{c}(L_d+L_s)\stackrel{.}{i_d}=-(R+R_s)i_d+p(L_q-L_s)i_q{\mathrm{\ω}}_f\\ (L_q+L_s)\stackrel{.}{i_q}=-(R+R_s)i_q-p(L_d+L_s)i_d{\mathrm{\ω}}_f+{\mathrm{p\ηi}}_q\end{array}\right.$

For face dress formula permagnetic synchronous motor, d axle inductance is equal with q axle inductance, i.e. L _d=L _q, so the electromagnetic torque of flywheel drive motors can be expressed as:

T _f＝pηi _q；

Ignore the dynamic process of drive motors and converters thereof in flywheel energy storage system, also fly wheel system can be reduced to single mass.The torque acted on this single mass flywheel energy storage system only has drive motors electromagnetic torque, and its equation of motion is:

$T_f=-J_f\frac{{\mathrm{d\ω}}_f}{dt};$

Wherein J _ffor the moment of inertia of fly wheel system.

If ignore mechanical loss, the active-power P of flywheel energy storage system will equal the electromagnetic power of drive motors output, then the electromagnetic torque T of active-power P and flywheel drive motors _fwith the angular velocity omega of flywheel drive motors _fpass be:

P＝T _fω _f；

According to analyzing above, regulating the resistance of equivalent resistance just can change the dq shaft current of flywheel drive motors, and then controlling the electromagnetic torque of drive motors, make the output of fly wheel system along the power stage expected.

The state variable of system is made to be $x=\left[\begin{array}{c}{i}_d\\ i_q\\ {\mathrm{\ω}}_f\end{array}\right],$ Input control amount u=R _s, output variable is energy and the y=P of flywheel storage, then system model can be written as following form:

$\left\{\begin{array}{c}\left[\begin{array}{c}\stackrel{.}{x_1}\\ \stackrel{.}{x_2}\\ \stackrel{.}{x_3}\end{array}\right]=\left[\begin{array}{c}\frac{-{\mathrm{Rx}}_1+p(L_q-L_s)x_2{x}_3}{(L_d+L_s)}\\ \frac{-{\mathrm{Rx}}_2-p(L_d+L_s)x_1{x}_3+{\mathrm{p\ηx}}_2}{(L_q+L_s)}\\ \frac{-{\mathrm{p\ηx}}_2}{J_f}\end{array}\right]+\left[\begin{array}{c}-\frac{x_1}{(L_d+L_s)}\\ -\frac{x_2}{(L_q+L_s)}\\ 0\end{array}\right]u\\ y={\mathrm{p\ηx}}_2{x}_3\end{array}\right.,$

If the sampling time is T _s, by system discretization, we can obtain:

$\left\{\begin{array}{c}\left[\begin{array}{c}{x}_{1,k+1}\\ x_{2,k+1}\\ x_{3,k+1}\end{array}\right]=\left[\begin{array}{c}{x}_{1,k}\\ x_{2,k}\\ x_{3,k}\end{array}\right]+T_s\left[\begin{array}{c}\frac{-{\mathrm{Rx}}_{1,k}+p(L_q-L_s)x_{2,k}{x}_{3,k}}{(L_d+L_s)}\\ \frac{-{\mathrm{Rx}}_{2,k}-p(L_d+L_s)x_{1,k}{x}_{3,k}+{\mathrm{p\ηx}}_{2,k}}{(L_q+L_s)}\\ \frac{-{\mathrm{p\ηx}}_{2,k}}{J_f}\end{array}\right]+T_s\left[\begin{array}{c}-\frac{x_{1,k}}{(L_d+L_s)}\\ -\frac{x_{2,k}}{(L_q+L_s)}\\ 0\end{array}\right]u_k\\ y_k={\mathrm{p\ηx}}_{2,k}{x}_{3,k}\end{array}\right.,$

Step 2: on the basis of step 1, design output probability density function controller, the power smooth controlling Wind turbines exports.

Control objectives stores excess energy by flywheel or fills up not enough energy with the power ensureing wind turbine generator and carry to electrical network according to smooth power P _gexport.Namely active power flywheel being exported by control q shaft current is according to reference power P _refexport, the tracking error in initialization system k moment is e _k=P _ref-y _k.

The present invention is mainly according to tracking error e _kprobability density function cONTROLLER DESIGN, by controlling the probability density function of tracking error distribution be that sharp narrow reduces gradually to ensure tracking error and is stabilized near zero, also just meet system output tracking expect output.

Step s1: choosing of performance index.

Consider that entropy can characterize the uncertainty of any stochastic variable, the output probability density function dispersiveness meaning tracking error that minimizes of this control system tracking error entropy reduces, show as not only point but also narrow distribution in shape, therefore second order Renyi entropy is incorporated in system performance index by the present invention, makes the probability density function of tracking error be the distribution of sharp narrow.According to definition second order Renyi entropy H _kexpression formula is:

$H_k=-log\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\γ}}_{e_k}{\left(\mathrm{\τ}\right)}^{2}d\mathrm{\τ}=-\log V_k,$

Wherein information potential obvious H _kand V _kbe all monotonic function, and the monotonicity of the two is contrary.The maximization realizing information potential can realize minimizing of entropy.

Also need the mean-square value considering tracking error in addition, be introduced in performance index and make mean square of error value little as far as possible, even close to zero.Show that the probability density function of error is in shape for peak value appears at about zero.According to definition error mean square value E _kexpression formula be:

$E_k=\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\τ}}^2{\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)d\mathrm{\τ},$

Therefore the present invention chooses performance index and is:

J(u _k)＝-R ₁V _k+R ₂E _k+0.5R ₃u _k ²

Wherein R ₁, R ₂, R ₃for every weight, last is the bound term of controlled quentity controlled variable.

Realize minimizing of performance index, first calculate the probability density function of tracking error and information potential;

Because system comprises random quantity, we adopt non-parametric estmation method to estimate the probability density function of tracking error.Assuming that k moment tracking error sample is S _k={ e ₁, e ₂... e _n, so k moment probability density function is: ${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=\frac{1}{N}\underset{i=1}{\overset{k}{\Σ}}\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2),$

Wherein N is sample error number, and ψ is gaussian kernel function, and δ is the parameter of gaussian kernel function, and its expression formula is: $\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2)=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\δ}}{e}^{-\frac{{(\mathrm{\τ}-e_i)}^2}{2{\mathrm{\δ}}^2}};$

Meanwhile, the probability density function of adjacent moment has following recurrence relation:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=(1-\mathrm{\ξ}){\mathrm{\γ}}_{e_{k-1}}\left(\mathrm{\τ}\right)+\mathrm{\ξ}\mathrm{\ψ}(\mathrm{\τ}-e_k,{\mathrm{\δ}}^2),$

According to sliding window technology, the computational methods of information potential are as follows:

$V_k=(1-\mathrm{\ξ})V_{k-1}+\frac{\mathrm{\ξ}}{L}\underset{i=k-L+1}{\overset{k}{\Σ}}\mathrm{\ψ}(e_i-e_k,2{\mathrm{\δ}}^2),$

Wherein L is window width, and in each sampling instant, L sample error is made up of current sample error and a front L-1 sample error.

Step s2: asking for of controlled quentity controlled variable.

Controller design target makes performance index minimum, and available gradient descent method first obtains optimal control increment Delta u _k, make u _k=Δ u _k+ u _k-1, wherein u _kfor k moment control inputs, u _k-1for k-1 moment control inputs.Definition Q _k=-R ₁v _k+ R ₂e _k, by Q _ktaylor expansion bring J (u into _k):

J(u _k)≈Q ₀+Q ₁Δu _k+0.5Q ₂Δu _k ²+0.5R ₃(u _k-1+Δu _k) ²，

Wherein $Q_0=Q_{k|u_k=u_{k-1}},$

$Q_1=\frac{\∂Q_k}{\∂u_k}|u_k=u_{k-1},$

$Q_2=\frac{{\∂}^2{Q}_k}{\∂{u_k}^2}|u_k=u_{k-1},$

According to optimal control theory, order optimal control increment can be obtained:

${\mathrm{\Δu}}_k=-\frac{Q_1+R_3{u}_{k-1}}{R_3+Q_2}$

By u _k=Δ u _k+ u _k-1, the controlled quentity controlled variable u in k moment _k.

Step s3: iteration upgrades.

In each sampling instant, try to achieve input control amount according to the method described above, under the effect of this input control amount, upgrade the parameter of etching system during k, then carry out asking for of next sampling instant input control amount, loop iteration goes down step by step, until terminate to run.

Therefore, the specific embodiment of the invention comprises a kind of control method of flywheel energy storage system, and described method comprises step:

A, with the electric current of the angular speed of flywheel, d axle and q axle be state variable, equivalent resistance for input control amount, flywheel energy storage is that output variable builds the model of flywheel energy storage system, according to the sampling time by model discretization;

B, building performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

C, optimum input control amount increment is utilized to determine current optimum input control amount.

Wherein, determine that the method for described performance index is:

The tracking error in system k moment is e _k=P _ref-y _k, for the probability density function of k moment error sample, P _reffor reference power value, y _kfor output variable;

Second order Renyi entropy is: $H_k=-log\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\γ}}_{e_k}{\left(\mathrm{\τ}\right)}^{2}d\mathrm{\τ}=-\log V_k;$

Error mean square value E _kfor: $E_k=\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\τ}}^2{\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)d\mathrm{\τ},$

Performance index are: J (u _k)=-R ₁v _k+ R ₂e _k+ 0.5R ₃u _k ²; Wherein R ₁, R ₂, R ₃for every weight, u _kfor input control amount, V _kfor information potential.

In addition, e is determined _kprobability density function method be as follows,

Assuming that k moment tracking error sample is S _k={ e ₁, e ₂... e _n, so k moment probability density function is:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=\frac{1}{N}\underset{i=1}{\overset{k}{\Σ}}\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2),$

Wherein N is sample error number, and ψ is gaussian kernel function, and δ is the parameter of gaussian kernel function, and its expression formula is:

$\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2)=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\δ}}{e}^{-\frac{{(\mathrm{\τ}-e_i)}^2}{2{\mathrm{\δ}}^2}},$

Pass between the probability density function of adjacent moment is:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=(1-\mathrm{\ξ}){\mathrm{\γ}}_{e_{k-1}}\left(\mathrm{\τ}\right)+\mathrm{\ξ}\mathrm{\ψ}(\mathrm{\τ}-e_k,{\mathrm{\δ}}^2),$

And comformed information gesture V _kfor:

$V_k=(1-\mathrm{\ξ})V_{k-1}+\frac{\mathrm{\ξ}}{L}\underset{i=k-L+1}{\overset{k}{\Σ}}\mathrm{\ψ}(e_i-e_k,2{\mathrm{\δ}}^2);$

Wherein L is window width, and ξ is forgetting factor, be less than 1 coefficient.

Be zero determine that optimum input control amount increment comprises according to the local derviation of performance index to input control amount increment:

Performance index are:

J(u _k)≈Q ₀+Q ₁Δu _k+0.5Q ₂Δu _k ²+0.5R ₃(u _k-1+Δu _k) ²；

Wherein

$Q_0=Q_{k|u_k=u_{k-1}},$

$Q_1=\frac{\∂Q_k}{\∂u_k}|u_k=u_{k-1},$

$Q_2=\frac{{\∂}^2{Q}_k}{\∂{u_k}^2}|u_k=u_{k-1},$

Wherein V _kfor information potential, R ₁, R ₂, R ₃for every weight.

Order determine optimum input control amount increment Delta u _k.

Match with the control method of flywheel energy storage system of the present invention, a kind of control device of flywheel energy storage system is also disclosed in the specific embodiment of the invention, described device comprises the system model construction unit, controlled quentity controlled variable increment optimization unit and the input control amount that are linked in sequence and optimizes unit, wherein, step is comprised:

System model construction unit be used for the electric current of the angular speed of flywheel, d axle and q axle be state variable, equivalent resistance for input control amount, flywheel energy storage is that output variable builds the model of flywheel energy storage system, according to the sampling time by model discretization;

Controlled quentity controlled variable increment is optimized unit and is built performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

Input control amount is optimized unit and is utilized optimum input control amount increment to determine current optimum input control amount.

Therefore, utilize control method and the device of flywheel energy storage system of the present invention, can based on the flywheel energy storage model simplified, the controller of design flywheel energy storage system, utilizes flywheel to store excess energy or fills up not enough energy to ensure that wind turbine generator carries smooth power to electrical network.

Further by an embody rule example, Advantageous Effects of the present invention is described below.

The wind turbine generator parameter that this embody rule example adopts is: wind energy conversion system sweeps wind radius r ₀=30m, rated power 1500kw, rated wind speed 12m/s, the moment of inertia 2.136 × 10 of wind turbine generator ⁵kg.m ², maximal wind-energy usage factor C _pmax=0.4382, generator energy conversion efficiency is α is 0.45, and atmospheric density is set to 1.2kg/m ³.Flywheel parameter: flywheel radius 2m, moment of inertia 1.2kg.m ², fly-wheel motor number of pole-pairs is 2, flywheel rated angular velocity 2.9rad/s, and permanent magnet flux linkage is 0.2366wb, and rated power is 150kw.Systematic sampling time T _s=0.5s, forgetting factor ξ=0.1, system adjustable parameter δ=0.5, R ₁=0.5, R ₂=0.5, R ₃=0, N=80, L=40.

When wind speed change at random, as shown in Figure 4, the information potential of the entropy of tracking error increases gradually, and mean-square value is reduced to zero gradually, as shown in Figure 5.Performance index are also successively decreased gradually and are stabilized near 0, as shown in Figure 6, the control algolithm that illustrative system is chosen can ensure the power output that the tracking of system stability is expected, tracking error is also stabilized to 0 gradually, as shown in Figure 7, the PDF change of tracking error as shown in Figure 8 in concrete change.As shown in Figure 9, the d shaft current of flywheel drive motors controls near 0, q shaft current along with power change driving on-position between change.When fly-wheel motor runs with driving condition and positive, system accelerates energy storage, absorbs the surplus power that generator exports; Otherwise when drive motors runs with on-position negative power, flywheel deceleration releases energy, export supplemental capacity to grid side converter.Can find out, flywheel energy storage can ensure that wind turbine generator is level and smooth to the power that electrical network exports, instead of along with random wind speed random fluctuation.

It should be noted that; above-mentioned execution mode is only the present invention's preferably embodiment; can not limiting the scope of the invention be understood as, not depart under concept thereof of the present invention, all protection scope of the present invention is belonged to modification to any minor variations that the present invention does.

Claims (6)

1. a control method for flywheel energy storage system, described method comprises step:

A, with the electric current of the angular speed of flywheel, d axle and q axle be state variable, variable resistor for input control amount, flywheel stores or the power of release is that output variable builds the model of flywheel energy storage system, according to the sampling time by model discretization;

B, building performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

C, optimum input control amount increment is utilized to determine current optimum input control amount.

2. according to the control method of a kind of flywheel energy storage system described in claim 1, it is characterized in that, determine that described performance index are:

The tracking error in system k moment is e _k=P _ref-y _k, for the probability density function of k moment error sample, P _reffor reference power value, y _kfor output variable;

Second order Renyi entropy is: $H_k=-log\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\γ}}_{e_k}{\left(\mathrm{\τ}\right)}^{2}d\mathrm{\τ}=-{\mathrm{logV}}_k;$

Error mean square value E _kfor: $E_k=\underset{-\mathrm{\∞}}{\overset{+\mathrm{\∞}}{\∫}}{\mathrm{\τ}}^2{\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)d\mathrm{\τ};$

Performance index are:

J(u _k)＝-R ₁V _k+R ₂E _k+0.5R ₃u _k ²；

Wherein R ₁, R ₂, R ₃for every weight, u _kfor input control amount, V _kfor information potential.

3., according to the control method of a kind of flywheel energy storage system described in claim 2, it is characterized in that, determine e _kprobability density function for as follows,

Assuming that k moment tracking error sample is S _k={ e ₁, e ₂... e _n, so k moment probability density function is: ${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=\frac{1}{N}\underset{i=1}{\overset{k}{\Σ}}\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2);$

Wherein N is sample error number, and ψ is gaussian kernel function, and δ is the parameter of gaussian kernel function, and its expression formula is: $\mathrm{\ψ}(\mathrm{\τ}-e_i,{\mathrm{\δ}}^2)=\frac{1}{\sqrt{2\mathrm{\π}}\mathrm{\δ}}{e}^{-\frac{{(\mathrm{\τ}-e_i)}^2}{2{\mathrm{\δ}}^2}};$

Pass between the probability density function of adjacent moment is:

${\mathrm{\γ}}_{e_k}\left(\mathrm{\τ}\right)=(1-\mathrm{\ξ}){\mathrm{\γ}}_{e_{k-1}}\left(\mathrm{\τ}\right)+\mathrm{\ξ}\mathrm{\ψ}(\mathrm{\τ}-e_k,{\mathrm{\δ}}^2).$

4. according to the control method of a kind of flywheel energy storage system described in claim 3, it is characterized in that, comformed information gesture V _kfor:

$V_k=(1-\mathrm{\ξ})V_{k-1}+\frac{\mathrm{\ξ}}{L}\underset{i=k-L+1}{\overset{k}{\Σ}}\mathrm{\ψ}(e_i-e_k,2{\mathrm{\δ}}^2);$

Wherein L is window width, and ξ is forgetting factor, be less than 1 coefficient.

5. according to the control method of a kind of flywheel energy storage system described in claim 4, it is characterized in that, be zero according to the local derviation of performance index to input control amount increment, determine the increment of optimum input control amount, comprising:

Performance index are:

J(u _k)≈Q ₀+Q ₁Δu _k+0.5Q ₂Δu _k ²+0.5R ₃(u _k-1+Δu _k) ²；

Wherein $Q_0=Q_{k|u_k=u_{k-1}},$

$Q_1=\frac{\∂Q_k}{\∂u_k}|u_k=u_{k-1},$

$Q_2=\frac{{\∂}^2{Q}_k}{\∂{u_k}^2}|u_k=u_{k-1},$

Q _k＝-R ₁V _k+R ₂E _k，

Wherein V _kfor information potential, wherein R ₁, R ₂for every weight;

Order determine optimum input control amount increment Delta u _k.

6. a control device for flywheel energy storage system, described device comprises the system model construction unit, controlled quentity controlled variable increment optimization unit and the input control amount that are linked in sequence and optimizes unit, wherein, comprises step:

System model construction unit be used for the electric current of the angular speed of flywheel, d axle and q axle be state variable, variable resistor for input control amount, flywheel energy storage is that output variable builds the model of flywheel energy storage system, according to the sampling time by model discretization;

Controlled quentity controlled variable increment is optimized unit and is built performance index with second order Renyi entropy, is zero determine optimum input control amount increment according to the local derviation of performance index to input control amount increment;

Input control amount is optimized unit and is utilized optimum input control amount increment to determine current optimum input control amount.