CN104091209A - Wind driven generator set power characteristic assessment method based on BP neural network - Google Patents

Abstract

The invention provides a wind driven generator set power characteristic assessment method based on a BP neural network. The method comprises the steps of only testing a reference unit according to the IEC61400-12-1 standard and then performing power assessment on all of units in a whole field according to testing data and SCADA data, and a large amount of money and time can be saved. Based on the BP neural network, the nonlinear relation of free wind speed and cabin wind speed can be achieved, and an assessment result is accurate and reliable.

Description

Wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network

Technical field:

The present invention relates to a kind of wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network particularly, belongs to technical field of wind power generation.

Background technology: power characteristic is the important base attribute of wind-powered electricity generation unit, and it is directly connected to the economic and technique level of wind-powered electricity generation unit.General IEC61400-12-1 or the IEC61400-12-2 of adopting carries out the test of wind-powered electricity generation unit power characteristic at present.Though IEC61400-12-1 standard can test out the power stage situation of unit accurately, but to the strict requirement of being provided with of landform and anemometer tower, the wind energy turbine set with a varied topography to some, probably cannot or comparatively difficulty is carried out the test of this mode, allow to test, for have tens even the wind energy turbine set of hundreds of platform also can not carry out one by one IEC standard wind power characteristic test, if carry out, will expend a large amount of manpower, material resources and financial resources.Though IEC61400-12-2 standard can be comparatively fast and cost is lower that unit is carried out to power assessments, but cabin transport function (the Nacelle Transfer Functions setting up in assessment, NTF) applicable elements is very strict, be difficult to all meet, this just causes the powertrace evaluating to have larger error, the judgement of impact to operating states of the units.

For IEC standard wind-powered electricity generation unit power characteristic, test is difficult to be widely used, and in running of wind generating set process, SCADA system can be carried out dynamic sampling to nacelle wind speed and corresponding power, automatically draws the powertrace of this unit.Yet, the measured nacelle wind speed of survey wind devices that is arranged on wind-powered electricity generation unit afterbody is the wind speed of receiving wake of rotor impact, because the little deviation of wind speed can cause the deviation that power is very large, so exist larger error with the drafting that the wind speed that is subject to wake effect carries out powertrace.

Yet the simulation to wind-powered electricity generation unit wake effect is difficult, be just can carry out by professional computational tool; The distinguished and admirable impeller through rotating, fluctuations in wind speed is more irregular, and the mapping relations between nacelle wind speed and free wind speed are the non-linear of complexity, are to be difficult to the description carried out with linear function.。

Summary of the invention

Goal of the invention: the invention provides a kind of wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network, its objective is the existing deficiency of solution mode in the past, it utilizes BP neural network to revise nacelle wind speed, and all or part of wind-powered electricity generation unit of wind energy turbine set is carried out to the assessment of power characteristic.

Technical scheme: the present invention is achieved through the following technical solutions:

A wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network, is characterized in that: the method comprises the steps:

Step 1: according to IEC61400-12-1 standard, each unit in wind energy turbine set is carried out to place assessment, for the wind energy turbine set of general landform, select at random a unit as with reference to unit; For the wind energy turbine set of complex-terrain, calculate the rugged index of each unit (RIX), the unit of selecting to approach all RIX mean values is as with reference to unit;

Step 2: carry out power characteristic test with reference to unit according to IEC61400-12-1 standard, meanwhile, in residue set box low pressure side, voltage transformer (VT) summation current transformer is installed, to obtain the net power output valve of each unit;

Step 3: synchronous recording anemometer tower data, power characteristic test anemometer mast data, each unit SCADA data and each power of the assembling unit output data, and set up assessment data storehouse.Nacelle wind speed in assessment data storehouse is collected in the SCADA system of unit, free wind speed is collected in anemometer tower and power characteristic test anemometer mast in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit;

Step 4: in assessment data storehouse, the data time such as nacelle wind speed, free wind speed, power is synchronous, to the data in database, screening with check unified frequency is 10min, free wind speed is anemometer tower and wind speed that anemometer mast is surveyed;

Step 5: data regression, according to anemometer tower data, determine the atmospheric density of testing wind energy turbine set, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform;

Step 6: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, carry out the assessment of wind-powered electricity generation unit power characteristic with this.This method adopts three layers of BP neural network, and wherein input layer is anemometer tower wind speed, nacelle wind speed, and output layer is free wind speed, chooses reference unit data as training data, sets up the mapping relations of free wind speed and nacelle wind speed;

Step 7: using the input value of the nacelle wind speed of each unit in wind energy turbine set, anemometer tower data BP neural network after training, obtain the free wind speed of each unit;

Step 8: through the correction of above-mentioned wind speed, require to carry out the power characteristic test of wind-powered electricity generation unit according to IEC 61400-12-1 wind-powered electricity generation unit power characteristic testing standard, draw out powertrace and power coefficient, extrapolated annual electricity generating capacity, finally provides assessment report.

During applicable this method, in wind energy turbine set, have at least a unit to carry out power characteristic test (this machine consists of reference unit) according to IEC61400-12-1 standard, and place's anemometer tower is at least installed in this wind energy turbine set.

The method step is specially:

The first step: each unit in wind energy turbine set is carried out to place assessment, in evaluation test place, the plane gradient, through the plane on tower frame for wind generating set basis with take angle and landform ground level on the variation of wind-powered electricity generation unit position between the covering of the fan that summit was become and must meet the requirement of IEC61400-12-1 standard to test site landform, table 1 has provided the specific requirement that test site surrounding terrain changes, wherein, L is the distance between wind power generating set and meteorological anemometer mast, and D is the rotor diameter of wind power generating set;

Table 1 test site requires: landform changes

(1), for the wind energy turbine set of general landform, select at random a unit as with reference to unit;

(2) for the wind energy turbine set of complex-terrain, calculate the rugged index of each unit (RIX), the unit of selecting to approach all RIX mean values is as with reference to unit;

Second step: carry out power characteristic test with reference to unit according to IEC61400-12-1 standard, meanwhile, in residue set box low pressure side, voltage transformer (VT) summation current transformer is installed, to obtain the net power output valve of each unit;

The 3rd step: synchronous recording anemometer tower data, power characteristic test anemometer mast data, each unit SCADA data and each power of the assembling unit output data, and set up assessment data storehouse, nacelle wind speed in assessment data storehouse is collected in the SCADA system of unit, free wind speed is collected in anemometer tower and power characteristic test anemometer mast in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit;

The 4th step: the check of data and screening, according to GB/T 18710-2002 and in conjunction with the actual conditions of wind energy turbine set, carry out range check, trend test, in Table 2,3, then carry out data rejecting and correction, guarantee that data can reflect that it is to reject wind-powered electricity generation unit not work or the data that test macro breaks down that the ruuning situation of wind-powered electricity generation unit, data are rejected objective;

The zone of reasonableness reference value of table 2 major parameter

The reasonable change trend reference value of table 3 major parameter

The data time such as nacelle wind speed, free wind speed (anemometer tower and wind speed that anemometer mast is surveyed), power in assessment data storehouse is synchronous, and unified frequency is 10min;

The 5th step: data regression, according to anemometer tower data, determine the atmospheric density of testing wind energy turbine set, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform;

Atmospheric density can have temperature and air pressure measured value to draw according to formula (1):

${\mathrm{\ρ}}_{10\min }=\frac{B_{10\min }}{{\mathrm{RT}}_{10\min }}---\left(1\right)$

Wherein: ρ _10minaverage air density for 10min; T _10minaverage absolute air temperature for the 10min that records; B _10minaverage gas pressure for the 10min that records; R is gas law constant 287.05J/ (kgK);

Automatically the wind-powered electricity generation unit through type (2) of controlling for power just can be converted normal atmosphere wind speed and depress:

$V_n=V_{10\min }{\left(\frac{{\mathrm{\ρ}}_{10\min }}{{\mathrm{\ρ}}_0}\right)}^{1/3}---\left(2\right)$

Wherein, V _nfor the air speed value after conversion; V _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³;

For adopting stall to control, there is the wind-powered electricity generation unit of constant slurry distance and rotating speed, measured power stage data can utilize formula (3) to convert:

$P_n=P_{10\min }\·\frac{{\mathrm{\ρ}}_0}{{\mathrm{\ρ}}_{10\min }}---\left(3\right)$

Wherein: P _nfor the air speed value after conversion; P _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³;

The 6th step: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, carry out the assessment of wind-powered electricity generation unit power characteristic with this.This method adopts three layers of BP neural network as shown in Figure 1, and wherein input layer is anemometer tower wind speed, nacelle wind speed, and output layer is free wind speed.Choose reference unit data as training data, set up the mapping relations of free wind speed and nacelle wind speed;

The 7th step: using the input value of the nacelle wind speed of each unit in wind energy turbine set, anemometer tower data BP neural network after training, obtain the free wind speed of each unit;

The 8th step: the drafting of powertrace, after completing data correction, selected test data will sort according to bin method, and selected data group should cover from lower than incision wind speed 1m/s during to the 85% rated power output of wind-powered electricity generation unit in the wind speed range of 1.5 times of wind speed.Wind speed range should be divided into 0.5m/sbin continuously, and central value is the integral multiple of 0.5m/s.Utilize the corresponding performance number of each wind speed bin after normalization to carry out the drafting of powertrace:

$V_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{V}_{n,i,j}---\left(4\right)$

$P_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{P}_{n,i,j}---\left(5\right)$

Wherein: V _imean wind speed value for i bin after conversion; V _{n, i, j}air speed value for the j data group of i bin recording; N _iit is the data bulk of the 10min data group of i bin; P _iaverage power content for i bin after conversion; P _{n, i, j}performance number for the j data group of i bin recording;

The 9th step: the drafting of power coefficient curve, power coefficient can be calculated and be obtained by formula (6) according to measured powertrace:

$C_{p,i}=\frac{P_i}{\frac{1}{2}{\mathrm{\ρ}}_0{\mathrm{AV}}_i^3}---\left(6\right)$

Wherein: C _p,ifor the power coefficient in bini; V _ifor converting resulting mean wind speed in bini; P _ifor converting resulting power stage in bin; A is the swept area of Wind turbine wind wheel; ρ ₀for standard air density;

The tenth step: the calculating of year generating, annual electricity generating capacity is to utilize to measure the estimated value that resulting powertrace distributes and calculates with reference to wind speed frequency for difference, and distribute with reference to wind speed frequency, can adopt rayleigh distributed to carry out, this distribution and shape coefficient are that the Weibull distribution of 2 o'clock is equal to.4,5,6,7,8,9,10, the l lm/s of annual electricity generating capacity while being to(for) annual mean wind speed (AEP) can be calculated and obtain according to formula (7):

$\mathrm{AEP}=N_h\underset{i=1}{\overset{N}{\mathrm{\Σ}}}[F\left(V_i\right)-F\left(V_{i-1}\right)]\left(\frac{P_{i-1}+P_i}{2}\right)---\left(7\right)$

Wherein: AEP is annual electricity generating capacity; N _hbe the hourage ≈ 8760 in a year; N is bin number; V _ifor the mean wind speed value at i bin after conversion; P _ifor the average power content at i bin after conversion;

The function of rayleigh distributed is:

$F\left(V\right)=1-\exp [-\frac{\mathrm{\π}}{4}{\left(\frac{V}{V_{\mathrm{ave}}}\right)}^2]---\left(8\right)$

Wherein: the Rayleigh Distribution Function that F (V) is wind speed; V _avefor the annual mean wind speed value at wind energy conversion system wheel hub centre-height place; V is air speed value, sets V _i-1=V _i-0.5m/s and P _i-1during=0.0kW, start stack;

The 11 step: report generation.

In the tenth step, annual electricity generating capacity must be calculated two aspects, be on the one hand " measurement of annual electricity generating capacity ", be the extrapolation of annual electricity generating capacity on the other hand, if measured, be not included in cut-out wind speed value, need to obtain by extrapolation method the annual electricity generating capacity that is extrapolated to cut-out wind speed from measured maximum wind velocity value, the performance number that annual electricity generating capacity extrapolation part obtains all wind speed that are all minimum wind speed of powertrace lower than test of hypothesis is 0, and suppose that all is steady state value higher than high wind speed in measured powertrace to the power in wind speed range between cut-out wind speed, it for the constant power level of extrapolation method, should be the performance number of the high wind speed bin of measured powertrace.

Advantage and effect:

The invention provides a kind of wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network, the application only need test according to IEC61400-12-1 standard reference unit, then according to test data and SCADA data, all units of the whole audience are carried out to power assessments, can save a large amount of money and time.It can realize free wind speed to the nonlinear relationship of nacelle wind speed based on BP neural network, and assessment result accurately and reliably;

Concrete feature is as follows:

Utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, with this, carry out the assessment of wind-powered electricity generation unit power characteristic.This method is adapted to the power characteristic assessment of grid-connected or not grid-connected separate unit horizontal axis wind-driven generator group, utilizes the mapping relations of the free wind speed of BP neural network and nacelle wind speed.During applicable this method, in wind energy turbine set, have at least a unit to carry out power characteristic test (this machine consists of reference unit) according to IEC61400-12-1 standard, and place's anemometer tower is at least installed in this wind energy turbine set.

This method is applicable to many units in wind energy turbine set to carry out power characteristic assessment.Assess the SCADA system that nacelle wind speed in database used is collected in unit, free wind speed is collected in anemometer tower in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit.The net power of wind-powered electricity generation unit can be installed voltage transformer (VT) summation current transformer in set box low pressure side and be obtained.This method adopts three layers of BP neural network, and wherein input layer is anemometer tower wind speed and nacelle wind speed, and output layer is free wind speed.Choose reference unit data as training data.In assessment data storehouse, the data time such as nacelle wind speed, free wind speed (wind speed that anemometer tower is surveyed), power is synchronous, and unified frequency is 10min.During applicable this method, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform.This method, according to measured data, provides the estimation of powertrace, power coefficient curve and the annual electricity generating capacity of test unit.

Accompanying drawing explanation:

Figure 1B P neural network structure figure

The graph of a relation of the free wind speed of Fig. 2 and nacelle wind speed

Fig. 3 wind-powered electricity generation unit power characteristic estimation flow figure

Fig. 4 assesses in place the schematic diagram that requires of deformation over the ground

The assessment of Fig. 5 place allows the maximum fluctuating schematic diagram of landform

The assessment of Fig. 6 place allows the ruling grade schematic diagram of landform

Embodiment: the present invention will be further described below in conjunction with accompanying drawing:

Below in conjunction with drawings and Examples, the present invention will be further described, and embodiment described herein only, for description and interpretation the present invention, is not intended to limit the present invention.

As shown in Figure 1, Figure 2 and Figure 3, the wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network that the present invention proposes is based on: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, with this, carry out the assessment of wind-powered electricity generation unit power characteristic.

Basic ideas of the present invention are: power characteristic is the important base attribute of wind-powered electricity generation unit, and it is directly connected to the economic and technique level of wind-powered electricity generation unit.General IEC61400-12-1 or the IEC61400-12-2 of adopting carries out the test of wind-powered electricity generation unit power characteristic at present.Though IEC61400-12-1 standard can test out the power stage situation of unit accurately, but to the strict requirement of being provided with of landform and anemometer tower, the wind energy turbine set with a varied topography to some, probably cannot or comparatively difficulty is carried out the test of this mode, allow to test, for have tens even the wind energy turbine set of hundreds of platform also can not carry out one by one IEC standard wind power characteristic test, if carry out, will expend a large amount of manpower, material resources and financial resources.Though IEC61400-12-2 standard can be comparatively fast and cost is lower that unit is carried out to power assessments, but cabin transport function (the Nacelle Transfer Functions setting up in assessment, NTF) applicable elements is strict, be difficult to all meet, this just causes the powertrace evaluating to have larger error, the judgement of impact to operating states of the units.

The distinguished and admirable impeller through rotating, fluctuations in wind speed is more irregular, and the mapping relations between nacelle wind speed and free wind speed are the non-linear of complexity, are to be difficult to the description carried out with linear function.Because can realizing, BP neural network is input to the non-linear arbitrarily of output, therefore adopt the mapping relations of the free wind speed of BP neural network and nacelle wind speed.Process is trained, thereby draws the free wind speed of each unit.

A wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network, its concrete implementation step is as follows:

The first step: each unit in wind energy turbine set is carried out to place assessment, in evaluation test place, the plane gradient (through the plane on tower frame for wind generating set basis with take the angle of wind-powered electricity generation unit position between the covering of the fan that summit was become) and ground level on landform variation must meet the requirement of IEC61400-12-1 standard to test site landform.Table 1 has provided the specific requirement that test site surrounding terrain changes, and wherein, L is the distance between wind power generating set and meteorological anemometer mast, and D is the rotor diameter of wind power generating set.Fig. 1 has provided wind-powered electricity generation unit power characteristic test site landform and has changed schematic diagram, and Fig. 2 has provided and in 2L, allowed the maximum of landform to rise and fall, and Fig. 3 has reflected in different regions, and place assessment permission is apart from the physical features drop of blower fan pylon base water plane maximum.

Table 1 test site requires: landform changes

(1), for the wind energy turbine set of general landform, select at random a unit as with reference to unit;

(2) for the wind energy turbine set of complex-terrain, calculate the rugged index of each unit (RIX), the unit of selecting to approach all RIX mean values is as with reference to unit.

Second step: carry out power characteristic test according to IEC61400-12-1 standard with reference to unit.Meanwhile, in residue set box low pressure side, voltage transformer (VT) summation current transformer is installed, to obtain the net power output valve of each unit.

The 3rd step: synchronous recording anemometer tower data, power characteristic test anemometer mast data, each unit SCADA data and each power of the assembling unit output data, and set up assessment data storehouse.Nacelle wind speed in assessment data storehouse is collected in the SCADA system of unit, free wind speed is collected in anemometer tower and power characteristic test anemometer mast in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit.

The 4th step: the check of data and screening, according to GB/T 18710-2002 and in conjunction with the actual conditions of wind energy turbine set, carry out range check, trend test (in Table 2,3).Then carry out data rejecting and correction, guarantee that data can reflect the ruuning situation (rejecting wind-powered electricity generation unit does not work or the data that test macro breaks down) of wind-powered electricity generation unit objective.

The zone of reasonableness reference value of table 2 major parameter

The reasonable change trend reference value of table 3 major parameter

The data time such as nacelle wind speed, free wind speed (anemometer tower and wind speed that anemometer mast is surveyed), power in assessment data storehouse is synchronous, and unified frequency is 10min.

The 5th step: data regression, according to anemometer tower data, determine the atmospheric density of testing wind energy turbine set, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform.

Atmospheric density can have temperature and air pressure measured value to draw according to formula (1):

${\mathrm{\ρ}}_{10\min }=\frac{B_{10\min }}{{\mathrm{RT}}_{10\min }}---\left(1\right)$

Wherein: ρ _10minaverage air density for 10min; T _10minaverage absolute air temperature for the 10min that records; B _10minaverage gas pressure for the 10min that records; R is gas law constant 287.05J/ (kgK).

Automatically the wind-powered electricity generation unit through type (2) of controlling for power just can be converted normal atmosphere wind speed and depress:

$V_n=V_{10\min }{\left(\frac{{\mathrm{\ρ}}_{10\min }}{{\mathrm{\ρ}}_0}\right)}^{1/3}---\left(2\right)$

Wherein, V _nfor the air speed value after conversion; V _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³.

For adopting stall to control, there is the wind-powered electricity generation unit of constant slurry distance and rotating speed, measured power stage data can utilize formula (3) to convert:

$P_n=P_{10\min }\·\frac{{\mathrm{\ρ}}_0}{{\mathrm{\ρ}}_{10\min }}---\left(3\right)$

Wherein: P _nfor the air speed value after conversion; P _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³.

The 6th step: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, carry out the assessment of wind-powered electricity generation unit power characteristic with this.This method adopts three layers of BP neural network as shown in Figure 1, and wherein input layer is anemometer tower wind speed, nacelle wind speed, and output layer is free wind speed.Choose reference unit data as training data, set up the mapping relations (as Fig. 2) of free wind speed and nacelle wind speed.

The 7th step: using the input value of the nacelle wind speed of each unit in wind energy turbine set, anemometer tower data BP neural network after training, obtain the free wind speed of each unit.

The 8th step: the drafting of powertrace, after completing data correction, selected test data will sort according to bin method, and selected data group should cover from lower than incision wind speed 1m/s during to the 85% rated power output of wind-powered electricity generation unit in the wind speed range of 1.5 times of wind speed.Wind speed range should be divided into 0.5m/sbin continuously, and central value is the integral multiple of 0.5m/s.Utilize the corresponding performance number of each wind speed bin after normalization to carry out the drafting of powertrace:

$V_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{V}_{n,i,j}---\left(4\right)$

$P_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{P}_{n,i,j}---\left(5\right)$

Wherein: V _imean wind speed value for i bin after conversion; V _{n, i, j}air speed value for the j data group of i bin recording; N _iit is the data bulk of the 10min data group of i bin; P _iaverage power content for i bin after conversion; P _{n, i, j}performance number for the j data group of i bin recording.

The 9th step: the drafting of power coefficient curve, power coefficient can be calculated and be obtained by formula (6) according to measured powertrace:

$C_{p,i}=\frac{P_i}{\frac{1}{2}{\mathrm{\ρ}}_0{\mathrm{AV}}_i^3}---\left(6\right)$

Wherein: C _p,ifor the power coefficient in bin i; V _ifor converting resulting mean wind speed in bin i; P _ifor converting resulting power stage in bin; A is the swept area of Wind turbine wind wheel; ρ ₀for standard air density.

The tenth step: the calculating of year generating, annual electricity generating capacity is to utilize to measure the estimated value that resulting powertrace distributes and calculates with reference to wind speed frequency for difference.And distribute with reference to wind speed frequency, can adopt rayleigh distributed to carry out, this distribution and shape coefficient are that the Weibull distribution of 2 o'clock is equal to.4,5,6,7,8,9,10, the llm/s of annual electricity generating capacity while being to(for) annual mean wind speed (AEP) can be calculated and obtain according to formula (7):

$\mathrm{AEP}=N_h\underset{i=1}{\overset{N}{\mathrm{\Σ}}}[F\left(V_i\right)-F\left(V_{i-1}\right)]\left(\frac{P_{i-1}+P_i}{2}\right)---\left(7\right)$

Wherein: AEP is annual electricity generating capacity; N _hbe the hourage ≈ 8760 in a year; N is bin number; V _ifor the mean wind speed value at i bin after conversion; P _ifor the average power content at i bin after conversion.

The function of rayleigh distributed is:

$F\left(V\right)=1-\exp [-\frac{\mathrm{\π}}{4}{\left(\frac{V}{V_{\mathrm{ave}}}\right)}^2]---\left(8\right)$

Wherein: the Rayleigh Distribution Function that F (V) is wind speed; V _avefor the annual mean wind speed value at wind energy conversion system wheel hub centre-height place; V is air speed value.Set V _i-1=V _i-0.5m/s and P _i-1during=0.0kW, start stack.

Annual electricity generating capacity must be calculated two aspects, be on the one hand " measurement of annual electricity generating capacity ", be the extrapolation of annual electricity generating capacity on the other hand.If measure and be not included in cut-out wind speed value, need to obtain by extrapolation method the annual electricity generating capacity that is extrapolated to cut-out wind speed from measured maximum wind velocity value.The performance number that annual electricity generating capacity extrapolation part obtains all wind speed that are all minimum wind speed of powertrace lower than test of hypothesis is 0, and supposes that all is steady state value higher than high wind speed in measured powertrace to the power in wind speed range between cut-out wind speed.It for the constant power level of extrapolation method, should be the performance number of the high wind speed bin of measured powertrace.

The 11 step: report generation.

The above be only better embodiment of the present invention, but protection scope of the present invention is not limited to this.Anyly be familiar with those skilled in the art in the technical scope that the present invention discloses, the variation that can expect easily or replacement, within all should being encompassed in protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.

Claims (4)

1. the wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network, is characterized in that: the method comprises the steps:

Step 1: according to IEC61400-12-1 standard, each unit in wind energy turbine set is carried out to place assessment, for the wind energy turbine set of general landform, select at random a unit as with reference to unit; For the wind energy turbine set of complex-terrain, calculate the rugged index of each unit (RIX), the unit of selecting to approach all RIX mean values is as with reference to unit;

Step 2: carry out power characteristic test with reference to unit according to IEC61400-12-1 standard, meanwhile, in residue set box low pressure side, voltage transformer (VT) summation current transformer is installed, to obtain the net power output valve of each unit;

Step 3: synchronous recording anemometer tower data, power characteristic test anemometer mast data, each unit SCADA data and each power of the assembling unit output data, and set up assessment data storehouse; Nacelle wind speed in assessment data storehouse is collected in the SCADA system of unit, free wind speed is collected in anemometer tower and power characteristic test anemometer mast in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit;

Step 4: in assessment data storehouse, the data time such as nacelle wind speed, free wind speed, power is synchronous, to the data in database, screening with check unified frequency is 10min, free wind speed is anemometer tower and wind speed that anemometer mast is surveyed;

Step 5: data regression, according to anemometer tower data, determine the atmospheric density of testing wind energy turbine set, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform;

Step 6: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, carry out the assessment of wind-powered electricity generation unit power characteristic with this; This method adopts three layers of BP neural network, and wherein input layer is anemometer tower wind speed, nacelle wind speed, and output layer is free wind speed, chooses reference unit data as training data, sets up the mapping relations of free wind speed and nacelle wind speed;

Step 7: using the input value of the nacelle wind speed of each unit in wind energy turbine set, anemometer tower data BP neural network after training, obtain the free wind speed of each unit;

Step 8: through the correction of above-mentioned wind speed, require to carry out the power characteristic test of wind-powered electricity generation unit according to IEC 61400-12-1 wind-powered electricity generation unit power characteristic testing standard, draw out powertrace and power coefficient, extrapolated annual electricity generating capacity, finally provides assessment report.

2. the wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network according to claim 1, it is characterized in that: during applicable this method, in wind energy turbine set, have at least a unit to carry out power characteristic test according to IEC61400-12-1 standard, this machine consists of reference unit, and place's anemometer tower is at least installed in this wind energy turbine set.

3. the wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network according to claim 2, is characterized in that: the method step is specially:

The first step: each unit in wind energy turbine set is carried out to place assessment, in evaluation test place, the plane gradient, through the plane on tower frame for wind generating set basis with take angle and landform ground level on the variation of wind-powered electricity generation unit position between the covering of the fan that summit was become and must meet the requirement of IEC61400-12-1 standard to test site landform, table 1 has provided the specific requirement that test site surrounding terrain changes, wherein, L is the distance between wind power generating set and meteorological anemometer mast, and D is the rotor diameter of wind power generating set;

Table 1 test site requires: landform changes

(1), for the wind energy turbine set of general landform, select at random a unit as with reference to unit;

(2) for the wind energy turbine set of complex-terrain, calculate the rugged index of each unit (RIX), the unit of selecting to approach all RIX mean values is as with reference to unit;

Second step: carry out power characteristic test with reference to unit according to IEC61400-12-1 standard, meanwhile, in residue set box low pressure side, voltage transformer (VT) summation current transformer is installed, to obtain the net power output valve of each unit;

The 3rd step: synchronous recording anemometer tower data, power characteristic test anemometer mast data, each unit SCADA data and each power of the assembling unit output data, and set up assessment data storehouse, nacelle wind speed in assessment data storehouse is collected in the SCADA system of unit, free wind speed is collected in anemometer tower and power characteristic test anemometer mast in wind energy turbine set, and this database storage is not less than the data of 180h continuous measuring hours, and data acquiring frequency, within 7s～10min, can cover wind speed range and the wind regime condition of certain limit;

The 4th step: the check of data and screening, according to GB/T 18710-2002 and in conjunction with the actual conditions of wind energy turbine set, carry out range check, trend test, in Table 2,3, then carry out data rejecting and correction, guarantee that data can reflect that it is to reject wind-powered electricity generation unit not work or the data that test macro breaks down that the ruuning situation of wind-powered electricity generation unit, data are rejected objective;

The zone of reasonableness reference value of table 2 major parameter

The reasonable change trend reference value of table 3 major parameter

The data time such as nacelle wind speed, free wind speed, power in assessment data storehouse is synchronous, and unified frequency is 10min; Free wind speed is surveyed wind speed by anemometer tower and anemometer mast;

The 5th step: data regression, according to anemometer tower data, determine the atmospheric density of testing wind energy turbine set, the average air density of measuring during test site valid data gather is at 1.225 ± 0.05kg/m ³during scope, without atmospheric density being standardized as to actual average atmospheric density; If not, according to IEC61400-12-1 related content, transform;

Atmospheric density can have temperature and air pressure measured value to draw according to formula (1):

${\mathrm{\ρ}}_{10\min }=\frac{B_{10\min }}{{\mathrm{RT}}_{10\min }}---\left(1\right)$

Wherein: ρ _10minaverage air density for 10min; T _10minaverage absolute air temperature for the 10min that records; B _10minaverage gas pressure for the 10min that records; R is gas law constant 287.05J/ (kgK);

Automatically the wind-powered electricity generation unit through type (2) of controlling for power just can be converted normal atmosphere wind speed and depress:

$V_n=V_{10\min }{\left(\frac{{\mathrm{\ρ}}_{10\min }}{{\mathrm{\ρ}}_0}\right)}^{1/3}---\left(2\right)$

Wherein, V _nfor the air speed value after conversion; V _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³;

For adopting stall to control, there is the wind-powered electricity generation unit of constant slurry distance and rotating speed, measured power stage data can utilize formula (3) to convert:

$P_n=P_{10\min }\·\frac{{\mathrm{\ρ}}_0}{{\mathrm{\ρ}}_{10\min }}---\left(3\right)$

Wherein: P _nfor the air speed value after conversion; P _10minmean wind speed value for the 10min that records; ρ _10minaverage air density for the 10min that obtains; ρ ₀for standard air density 1.225kg/m ³;

The 6th step: utilize the nonlinear fitting ability of BP neural network to remove to revise the measured nacelle wind speed of wind turbine cabin afterbody survey wind devices, thereby obtain free wind speed, carry out the assessment of wind-powered electricity generation unit power characteristic with this; This method adopts three layers of BP neural network as shown in Figure 1, and wherein input layer is anemometer tower wind speed, nacelle wind speed, and output layer is free wind speed; Choose reference unit data as training data, set up the mapping relations of free wind speed and nacelle wind speed;

The 7th step: using the input value of the nacelle wind speed of each unit in wind energy turbine set, anemometer tower data BP neural network after training, obtain the free wind speed of each unit;

The 8th step: the drafting of powertrace, after completing data correction, selected test data will sort according to bin method, and selected data group should cover from lower than incision wind speed 1m/s during to the 85% rated power output of wind-powered electricity generation unit in the wind speed range of 1.5 times of wind speed; Wind speed range should be divided into 0.5m/sbin continuously, and central value is the integral multiple of 0.5m/s; Utilize the corresponding performance number of each wind speed bin after normalization to carry out the drafting of powertrace:

$V_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{V}_{n,i,j}---\left(4\right)$

$P_i=\frac{1}{N_i}\underset{j=1}{\overset{N_i}{\mathrm{\Σ}}}{P}_{n,i,j}---\left(5\right)$

Wherein: V _imean wind speed value for i bin after conversion; V _{n, i, j}air speed value for the j data group of i bin recording; N _iit is the data bulk of the 10min data group of i bin; P _iaverage power content for i bin after conversion; P _{n, i, j}performance number for the j data group of i bin recording;

The 9th step: the drafting of power coefficient curve, power coefficient can be calculated and be obtained by formula (6) according to measured powertrace:

$C_{p,i}=\frac{P_i}{\frac{1}{2}{\mathrm{\ρ}}_0{\mathrm{AV}}_i^3}---\left(6\right)$

Wherein: C _p,ifor the power coefficient in bin i; V _ifor converting resulting mean wind speed in bin i; P _ifor converting resulting power stage in bin; A is the swept area of Wind turbine wind wheel; ρ ₀for standard air density;

The tenth step: the calculating of year generating, annual electricity generating capacity is to utilize to measure the estimated value that resulting powertrace distributes and calculates with reference to wind speed frequency for difference, and distribute with reference to wind speed frequency, can adopt rayleigh distributed to carry out, this distribution and shape coefficient are that the Weibull distribution of 2 o'clock is equal to; 4,5,6,7,8,9,10, the l lm/s of annual electricity generating capacity while being to(for) annual mean wind speed (AEP) can be calculated and obtain according to formula (7):

$\mathrm{AEP}=N_h\underset{i=1}{\overset{N}{\mathrm{\Σ}}}[F\left(V_i\right)-F\left(V_{i-1}\right)]\left(\frac{P_{i-1}+P_i}{2}\right)---\left(7\right)$

Wherein: AEP is annual electricity generating capacity; N _hbe the hourage ≈ 8760 in a year; N is bin number; V _ifor the mean wind speed value at i bin after conversion; P _ifor the average power content at i bin after conversion;

The function of rayleigh distributed is:

$F\left(V\right)=1-\exp [-\frac{\mathrm{\π}}{4}{\left(\frac{V}{V_{\mathrm{ave}}}\right)}^2]---\left(8\right)$

Wherein: the Rayleigh Distribution Function that F (V) is wind speed; V _avefor the annual mean wind speed value at wind energy conversion system wheel hub centre-height place; V is air speed value, sets V _i-1=V _i-0.5m/s and P _i-1during=0.0kW, start stack;

The 11 step: report generation.

4. the wind-powered electricity generation unit power characteristic appraisal procedure based on BP neural network according to claim 3, is characterized in that:

In the tenth step, annual electricity generating capacity must be calculated two aspects, be on the one hand " measurement of annual electricity generating capacity ", be the extrapolation of annual electricity generating capacity on the other hand, if measured, be not included in cut-out wind speed value, need to obtain by extrapolation method the annual electricity generating capacity that is extrapolated to cut-out wind speed from measured maximum wind velocity value, the performance number that annual electricity generating capacity extrapolation part obtains all wind speed that are all minimum wind speed of powertrace lower than test of hypothesis is 0, and suppose that all is steady state value higher than high wind speed in measured powertrace to the power in wind speed range between cut-out wind speed, it for the constant power level of extrapolation method, should be the performance number of the high wind speed bin of measured powertrace.