Predicting the Wind: Data Science in Wind Resource Assessment

Hands-on tutorial about data science in wind resource assessment. You can explore this tutorial interactively via binder at github.com/flrs/predicting_the_wind

The tutorial is about the fictitious scenario of building a wind farm on the hills around the AI incubator The Sandbox San Diego. Using Python code, it explains how the wind can be measured and how the measurement can be used together with climate models and ground station data to generate a long-term estimate of the wind. Subsequently, the tutorial explores how to use that estimate to predict wind turbine power output. Finally, the tutorial puts the output of the fictitious wind farm into the broader context of the California power grid.

From a data science perspective, the tutorial touches on data exploration, modeling and validation, and clarifies where domain knowledge of wind and fluid dynamics can help improve the wind estimate.

The tutorial uses the following tool stack:
- Python for performing calculations
- Jupyter Notebook as an IDE
- RISE Jupyter extension for presenting the notebook
- hide_code extension for hiding some (long-ish) code to build maps
- Folium to display maps
- SciPy's orthogonal distance regression
- scikit-learn's RandomForestRegressor
- brightwind to support wind resource assessment tasks

Transcript

1. P edic i g he Wi d: Da a Scie ce i Wi d Re ce A e me Se f- d e i , Da a Scie i 2020-03-11, So rce Code: I e ded a die ce: Da a anal s s and da a scien is s ho are in eres ed in learning abo ho da a science echniq es are applied in he ind energ domain. F ia R chec P Da a San Diego Mee p gi h b.com/ rs/predic ing_ he_ ind
2. A e da Wha i i d e ce a e me ? H mea e he i d P edic i g l g- e m i d eed P edic i g i d bi e e
3. Wha is Wind Reso rce Assessmen ?
5. S nd g d, b ... H m ch e i in he ind? Will be able ell he gene a ed elec ici a a ? A e he ne 25 ears?
6. Wind resource assessment to the rescue! P edic l g- e beha i f he i d P edic e f i d bi e K if ill ake a !
7. Wind Resource Assessment =  B i di g de f he h ica d i e ci i g U ce ai i j a f e M de da a cie ce i e a i e e i he e d Red ce e i i a d g ba a i g!
8. P edic i g he Wi d: A Da a Scie ce P b e ! Ge i g i d da a C ea i g i d da a A a i g i d da a B i di g a de f he i d P edic i g he i d a d f he i d fa Thi i a ica da a cie ce !
9. G W d Da a: M Ma Me ma l k like hi : Mea e he i d a diffe e heigh Ha e e f i d eed, i d di ec i , em e a e, h midi , a d eci i a i
10. A a i g Wi d Da a Le ' ad da a f a e a a d chec ! I [1]: O e a da a h d eed / a 30, 45, a d 58 he gh , e e a e C, a d d d ec , 10- e e a . (The e da a a e a a ca a d I ge e a ed he .) h eb O [1]: d_30 d_45 d_58 di i e 1999-12-31 16:00:00 3.39 3.73 3.84 12.37 243.68 1999-12-31 16:10:00 3.27 3.65 3.81 12.22 250.02 1999-12-31 16:20:00 3.31 3.63 3.80 12.12 252.29 1999-12-31 16:30:00 3.78 4.26 4.38 12.04 249.54 1999-12-31 16:40:00 3.96 4.38 4.52 11.99 254.50 impor a da as d da a = d. ad_ a ('./da a/ _ a . a ') da a. d(2). ad()
11. Le ' ge a feel f he i d da a b l i g he ! In [3]: ... , ha l k e e ! O t[3]: # Pl i d eed ime e ie impor bright ind as b anemometers = ['spd_30','spd_45', 'spd_58'] b .plot_timeseries(data[anemometers])
12. Le ' a i g e da ee e de ai ! In [4]: Ob e a i : Wi d eed a ie a h gh he da . Highe heigh ea highe i d eed. O [4]: b .plo _ imeseries(da a.loc['2019-03-11',anemome ers])
13. Which di ec i n d e he ind c me f m? Le ' l a f e enc e ! I [5]: O [5]: . _ a ( a a[' _30'], a a[' '])
14. Wind Data Analysis Takeaways Wi d da a = g i e e ie Wi d da a e The e a e d ai - eci c c e a d e e i d da a (e.g. f e e c e )
15. F om Mea emen o P edic ion: B ilding a Model We ant to predict ho much energ a ind farm ill likel produce in 25 ears of operation. P ble : Onl 2 ears of met mast ind measurements to predict 25 ears of ind With that little data e reall don't kno enough about ho the ind ill beha e! S l i : Get more, longer-term data from other sources, co ering as much time as possible Ra i ale: The more e kno about the past, the better e can predict the future.
16. Sec ion O e ie H a d he e ge e l g- e da a B ild a i le del edic i d eed U e a e ad a ced del f ciki -lea edic i d eed I e del i h i d e e g d ai k ledge I e iga e h c e a d c a e i d eed del
17. Ge ing More, Long-Term Da a P a da a ce : G ba c a e de , ea e e b d a e C e e Sa db (b e): 4 c a e de d ( a e), 3 a ea e e a ( e)
18. Le ' ad c i a e de ( ) a d ai a i da a! ERA5 I [6]: C i a e de ica d i d eed a g d e e a d ha e 1-h e i . ( h h I d aded ERA5 da a a d h h I d aded ai a i da a.) Thi eb hi eb O [6]: pd di mp ime 1999-12-31 16:00:00 4.548827 249.010856 12.370013 1999-12-31 17:00:00 3.981960 251.738388 11.915547 1999-12-31 18:00:00 2.607753 249.791620 11.245783 1999-12-31 19:00:00 1.559933 233.880179 9.782310 1999-12-31 20:00:00 1.359054 231.334928 10.454369 f om a b impo Pa impo a da a d _da a = ' a5_0': ' a5_0. a ', ' a5_1': ' a5_1. a ', 'MYF': 'MYF_200001010000_202003070000. a ', 'NKX': 'NKX_200001010000_202003070000. a ', 'SAN': 'SAN_200001010000_202003070000. a ' fo a , in _da a. (): _da a[ a ] = d. ad_ a (Pa ('./da a/'). a ( )) _da a[' a5_0']. ad()
19. From Short-Term to Long-Term Data Our challenges: Airport measurements are taken and climate models are calculated far from our site (the Sandbo ). The have measurements in greater intervals than our met mast. How much can we trust them to tell us about the wind characteristics at the Sandbo ?
20. Le ' plo me ma ind peed da a again he da a from he ERA5 clima e model! I [7]: Good ne : De pi e being ph icall far a a , here eem o be grea imilari ie be een he clima e model and he me ma da a. (Thi i no al a he ca e, b i i here o make hi orial f n and ea .) O [7]: l _da a = d.c ca ([da a[' d_58']. e am le('1W').mea (), l _da a['e a5_0'][' d']. e am le('1W').mea ()], a i =1) l _da a.c l m = ['Mea eme ', 'LT Refe e ce'] b . l _ ime e ie ( l _da a, da e_f m='2017')
21. P blem: Ho do e e ploit the similarities bet een references and met mast data to get an idea of hat our met mast data ould ha e looked like if e had measured for 25 ears? S l i n: 1. Build model describing relationship bet een measurement and references 2. Let model predict hat long-term measurement ould look like Let's build a model!
22. A Sim le M del: O h g nal Lea S a e O hogonal lea a e : D a a be - line be een all ime amp-poin of efe ence and me ma ind peed ha minimi e he o hogonal di ance be een line and ime amp-poin I [8]: I look like e ha e a good amo n of da a (7k+ poin ) and a e pec able of 0.89. Le ' plo o line of be ! 'N da a ': 761, ' ff e ': 0.48915546304725566, ' 2': 0.8682745042710436, ' e': 0.7728460999482368 from b d.a a e.c e a import O a Lea S a e # Re a le dail da a_1D = da a. e a e('1D'). ea () _da a_1D = _da a['e a5_0']. e a e('1D'). ea () _ de = O a Lea S a e ( ef_ d= _da a_1D[' d'], a e _ d=da a_1D[' d_58'], a e a _ d='1D') _ de . ()
23. In [9]: There is some sca er b model s he da a q i e ell. O [9]: l _m del. l ()
24. P blem: To make en e of he model in e m of ho ell i can p edic ind peed , e an o e i o p edic he ind peed fo he ime pe iod hen e ha e me ma mea emen and hen compa e he e mea emen o he model' p edic ion . Fo hi p po e, a e o me ic i inapp op ia e - i ell no hing abo ind peed ! (In e p e a ion of i al o a he ick fo o hogonal lea a e eg e ion in gene al) S l i n: U e RMSE ( oo mean a e e o ) of p edic ed ind peed . ac al mea ed ind peed a e o me ic! I [10]: # Define co ing me ic: RMSE import as def ( d c , ac a ): return . ((( d c -ac a )**2). a ()) a _ d c = a _ c =
25. Le ' c e he im le h g nal lea a e m del ing RMSE! In [11]: The RMSE i 11% f he ind eed. Thi i n eall a g d n mbe . If e ld b ild he jec a ming 11% fa e ind han e ld ac all ha e, e ld ha e made a e e en i e mi ake. S : H can e im e he m del? Le ' lea n ab Ph ic ! RMSE of im le model: 0.315 RMSE a % of ind eed mean: 11% edic ion = (ol _model. a am [' lo e']*l _da a_1D[' d']+ol _model. a am ['off e ']) all_ edic ion [' im le'] = edic ion all_ co e [' im le'] = m e( edic ion, da a_1D[' d_58']) in ('RMSE of im le model: :.3f '.fo ma (all_ co e [' im le'])) in ('RMSE a % of ind eed mean: :.0f %'.fo ma (all_ co e [' im le']/da a_1D[' d_58'].mean()*100))
26. Topography vs. Wind I ag e a d g e e c e ef a a e ca e. A g e e a h f ef gh , acce e a e a d he a d dece e a e a d he b ( e ca e bec e d ag a ). P f h e a e e a e ce e ec ed d eed .
27. Ele a ion Map of A ea A o nd he Sandbo We mmi ed he cale, b i l k a if he e a e me ele a i diffe e ce a d he Sa db .
28. Ho can e ake ad an age of he opog aphic info ma ion? Le ' b d g de f d ffe e d d ec ! Ra a e: We a d ha he e e a d eed be ee a a d efe e ce e d ec ha he .
29. A Be e M del: Binned O h g nal Lea S a e O r simple model j st binned in 12 direction sectors. Q ick remark: We se for b ilding o r model here instead of , since I ant to sho o that o can e en anal e ind data ith more common data science tools. SciP Bright ind In [12]: # B da a b d ec dir_bin = pd.In er alInde .from_break (np.lin pace(0,360,13)) da a_1D = da a.re ample('1D').mean() l _da a_1D = l _da a['era5_0'].re ample('1D').mean() l _da a_1D['dir_bin'] = pd.c (l _da a_1D['dir'], dir_bin ) da a_1D['dir_bin'] = pd.c (da a_1D['dir'], dir_bin )
30. I [13]: RMSE f b ed de : 0.289 # B ild binned or hogonal lea q are model from c . d import ODR, M de , Da a def e _da a_ _d _b (d _b ): _da a_ _b = _da a_1D[' d'][ _da a_1D['d _b '] == d _b ] da a_ _b = da a_1D[' d_58'][da a_1D['d _b '] == d _b ] ret rn _da a_ _b , da a_ _b def de _fc (B, ): ret rn B[0]* +B[1] b _ a = for b _ , d _b in e e a e(d _b ): b _ a [b _ ] = ' _ a e ': None, 'be a ': None, ' e': . a , ' ed c ': None _da a_ _b , da a_ _b = e _da a_ _d _b (d _b ) c c e _ = ( e ( _da a_ _b . de ). e ec ( e (da a_ _b . de ))) b _ a [b _ ][' _ a e '] = e (c c e _ ) if not c c e _ : contin e e = ODR(Da a( _da a_ _b [c c e _ ]. a e , da a_ _b [c c e _ ]. a e ), M de ( de _fc ), be a0=[1., 0.5]). () b _ a [b _ ][' ed c '] = de _fc ( e .be a, _da a_ _b ) b _ a [b _ ]['be a '] = e .be a b _ a [b _ ][' e'] = e(b _ a [b _ ][' ed c '], da a_ _b ) a _ ed c ['b ed'] = d.c ca ([b _ a [' ed c '] for b _ a in b _ a . a e ()\ if b _ a [' ed c '] is not None]). _ de () a _ c e ['b ed'] = . a ea ([b _ a [' e'] for b _ a in b _ a . a e ()]) ('RMSE f b ed de : :.3f '.f a (a _ c e ['b ed']))
31. Score o far RMSE of simple model: 0.315 RMSE of binned model: 0.289 Our binned model performs better already!
32. A More Ad anced Model: RandomFore Regre or E pec a ion: Random forest regressor captures nuances in relationship between reference and measurement better than other models Backgro nd: A generates an estimate from multiple decision trees. Each tree only gets trained on a subset of samples and features. This means that each tree has some "specialized" knowledge about the data and can see a particular aspect of the data better than other trees. By combining the predictions of all trees, a random forest can provide a relatively robust prediction. Approach: Engineer features that model can pick up on Run the model random forest
33. Fea e E gi ee i g Let's come up ith some (simple) features that the random forest can feed on. In [14]: O [14]: d di d_ i g_3 d_ i g_5 d i e 2018-02-01 1.455543 173.896001 15.397280 3.040377 3.040787 1.0 2.0 2018-02-02 1.528880 168.001625 16.170596 3.040377 3.040787 2.0 2.0 2018-02-03 2.025680 251.970139 16.567941 1.670035 3.040787 3.0 2.0 = da a_1D[' d_58'] # Ge c c e ime e conc_inde = o ed(li ( e ( .inde ).in e ec ion(l _da a_1D.inde ))) X = l _da a_1D.loc[conc_inde ,[' d', 'di ', ' m ']]. o _inde () # R lli g mea def make_ olling(da a, indo _ id h): olling = da a[' d']. olling( indo _ id h).mean() olling = olling.fillna( olling.mean()) olling.name = ' d_ olling_ '.fo ma ( indo _ id h) return d.conca ([da a, olling], a i =1) X = make_ olling(X, 3) X = make_ olling(X, 5) # Da e-ba ed fea e ca e em al a e X.loc[conc_inde ,'d'] = X.inde .da X.loc[conc_inde ,'m'] = X.inde .mon h X.head(3)
34. No that e ha e data ith features, let's build 2 models: One model for hich e ill ithhold some alidation data and, for comparison ith the pre ious models, one model that uses all data. I [15]: The RMSE is not too high and de nitel ithin the range of the other models. Let's compare the models in more detail. RMSE (RF de , ai i g e ): 0.303 RMSE (RF de , a ida i e ): 0.330 RMSE (RF de , a da a): 0.300 f om k ea .e e b e impo Ra d F e Reg e f om k ea . de _ e ec i impo ai _ e _ i X_ ai , X_ a , _ ai , _ a = ai _ e _ i (X, , e _ i e=0.2, a d _ a e=42) f_ de = Ra d F e Reg e ( _e i a =100, b_ c e=T e, a d _ a e=100, i _ a e _ eaf=10) f_ de .fi (X_ ai , _ ai ) i ('RMSE (RF de , ai i g e ): :.3f '.f a ( e( f_ de . edic (X_ ai ), _ ai ))) i ('RMSE (RF de , a ida i e ): :.3f '.f a ( e( f_ de . edic (X_ a ), _ a ))) de = Ra d F e Reg e ( _e i a =100, b_ c e=T e, a d _ a e=100, i _ a e _ eaf=10) de .fi (X, ) edic i = de . edic (X) a _ edic i ['f e '] = d.Se ie ( edic i ,i de =c c_i de ). _i de () a _ c e ['f e '] = e( edic i , ) i ('RMSE (RF de , a da a): :.3f '.f a (a _ c e ['f e ']))
35. Com a ing he 3 Wind S eed Model I [16]: All 3 models follo he arge mas nicel . The fores model some imes cap res peaks be er han o her models (J n 18, Ma 6), b also occasionall has bigger misses (Apr 12, Ma 30). from a ib impor a d.Da aF a e(a _ edic i ). c['2018-04':'2018-06',:]. (fig i e=(25,5)) da a_1D[' d_58']. c['2018-04':'2018-06']. (c=' ', =2, abe ='Ta ge Ma ') . ege d() . h ()
36. Time for a direct score comparison! Remember: The lo er the RMSE, the better the model. I [17]: Despite all the fanciness of the random forest model, it does not reach the score of our binned model. This being said an RMSE of 0.3 is still relativel high hen measured in terms of ind speed. .S (a _ c ). .ba ();
37. An In-De h L k a he Binned M del W e e b e b ed de , e d d e d a e ec . Le ' d a . We a de a d b de . F , e ' d a e be f a e de ed e b . In [18]: 6 b a e de 25 a e . F ed c e e b , e de bab e e ab e. bin_ ise_params = pd.Series([stat['n_samples'] f stat in bin_stats.values()], inde =dir_bins) bin_ ise_params.plot.bar(figsi e=(15,5)) plt.grid()
38. I ld be be e e im le m del f he e ca e , ince e kn ha i ha been ained n a l f da a and e f m ea nabl ell. I [19]: (0.0, 30.0]: Sim le m del (30.0, 60.0]: Sim le m del (60.0, 90.0]: Sim le m del (90.0, 120.0]: Sim le m del (120.0, 150.0]: Sim le m del (150.0, 180.0]: Bi - i e m del (180.0, 210.0]: Bi - i e m del (210.0, 240.0]: Bi - i e m del (240.0, 270.0]: Bi - i e m del (270.0, 300.0]: Bi - i e m del (300.0, 330.0]: Sim le m del (330.0, 360.0]: Sim le m del be a = [] fo be a, alid_bi ed, bi _ in i ([ a ['be a '] fo a in bi _ a . al e ()], [ a [' _ am le ']>=25 fo a in bi _ a . al e ()], di _bi ): if no alid_bi ed: be a .a e d( .a a a ([ l _m del. a am [' l e'], l _m del. a am [' ff e ']])) i (' : Sim le m del'.f ma (bi _)) el e: be a .a e d(be a) i (' : Bi - i e m del'.f ma (bi _))
39. No that e ha e a more rob st model, let's re-predict o r time series and check o r RMSE metric! I [20]: I [21]: It looks as if lling the gaps in the binned model had a er bad impact on o r RMSE, making the ne model the orst-performing one. What is going on here? RMSE b ed - e de : 0.403 O [21]: b ed 0.289 e 0.300 e 0.315 b ed_ _ e 0.403 d e: a 64 b _ e = [] b _ ed c = [] f d _b , be a in (d _b , be a ): _da a_ _b = _da a_1D[' d'][ _da a_1D['d _b '] == d _b ] da a_ _b = da a_1D[' d_58'][ _da a_1D['d _b '] == d _b ] b _ ed c = de _ c (be a, _da a_ _b ) b _ ed c .a e d(b _ ed c ) b _ e = e(b _ ed c , da a_ _b ) b _ e .a e d(b _ e) a _ ed c ['b ed_ _ e'] = d.c ca (b _ ed c ). _ de () a _ c e ['b ed_ _ e'] = . a ea (b _ e ) ('RMSE b ed - e de : :.3 '. a (a _ c e ['b ed_ _ e'])) d.Se e (a _ c e ). _ a e (). d(3)
40. We sho ld look at the RMSE per bin to get a better pict re of ho hich binned model is to blame for the increase in error. In [22]: We are getting the highest errors in the bins here e inserted the simple model. B t, if there is no ind in those bins (= directions), the errors in those bins are not important! What e reall need is a bin- eighted scoring metric! pd.DataFrame( 'rmse': bin_rmses, 'n_samples': [stat['n_samples'] for stat in bin_stats. al es()] , inde =dir_bins).plot.bar(s bplots=Tr e,figsi e=(15,5));
41. I ed Sc i g Me ic: Bi -Weigh ed RMSE Fi , e i ca c a e he eigh e gi e each bi . Thi i i a he be f a e i each bi . In [23]: N e ca b i d c i g f c i ! In [24]: bin_n_samples = [s a ['n_samples'] f s a i bin_s a s. al es()] bin_ eigh s = pd.Series(bin_n_samples, inde =dir_bins)/np.s m(bin_n_samples) def rmse_binned(predic ion_spd, reference_spd, reference_dir): sqr_errors = (predic ion_spd-reference_spd)**2 eigh s = bin_ eigh s[reference_dir] error = np.sqr (np.nanmean(sqr_errors. al es* eigh s. al es)) e error
42. Wi h he c i g f c i de be , e ' ca c a e he bi - eigh ed RMSE f a f de ' edic i . I [25]: The bi - eigh ed RMSE h c ea e diffe e ce be ee he de ha he eigh ed c e: O bi ed de ha i i g he i e de ' i f a i i ga c e be O i e de e f e ha he bi ed de O f e de e f S , af e a , e d e b b ed de ed c g- e d eed a a ! O [25]: U e g ed B -We g ed b ed_ _ e 0.4026 0.1328 b ed 0.2888 0.1347 e 0.3152 0.1416 f e 0.2996 0.1807 a _ c e _b ed = f de _ a e, ed c _ d in a _ ed c . e (): da a_1D[' d_58'][da a_1D['d _b '] == d _b ] a _ c e _b ed[ de _ a e] = e_b ed( ed c _ d, _da a_1D[' d'], _da a_1D['d ']) d.Da aF a e([a _ c e , a _ c e _b ed]).T. e a e(c = 0: 'U e ed', 1: 'B -We ed' )\ . _ a e ('B -We ed'). e.ba ( =0, a =0.5, c =' b e').f a (' :.4f ')
43. P edic ing he Long-Te m Wind S eed No e can predict the long-term ind speed at our site. This helps us to get a good idea of ho much ind energ e can potentiall harvest if e build a ind turbine there. Remember: We assume the ind in the future ill behave like the ind in the past. To get to the long-term ind speed, e take all predictions from our bin_fill_ im le model. To ma imi e accurac , e substitute it ith actual measurements from our mast herever possible. I [26]: Let's summari e our long-term ind speed time series! I [27]: R a 761 - a a a a (10.3%). S a : 1999-12-31, L : 20 a , A . 2.86 / _ _a _ a = a _ ['b _ _ '] _ _a _ a [ a a_1D[' _58']. ] = a a_1D[' _58'] ('R a - a a a a ( :.1% ).'\ . a ( a a_1D. a [0], a a_1D. a [0]/ _ _a _ a . a [0])) ('S a : :%Y-% -% , L : :.0 a , A . :.2 / '\ . a ( _ _a _ a . [0], _ _a _ a . a [0]/(365.25), _ _a _ a . a ()))
44. A fe o d of ca ion abo he long- e m ind eed ime e ie W d eed a a a e he da . The ef e, d b e d ce e e g h a f ac f ha 24 h e d. We d ced a dail e e e ha d e e ec he e d eed cha ge d g he da , beca e e a e aged h d eed e ha 24 h e d. A a e , e ge ea c e e g d c be h h e e e . Ac a d e ce a e e ch e c ca ed ha h he e.
45. F m Wind S eed Mea emen P edic i n: Takea a I i ea b i d a i e i d eed edic i de M e ad a ced de e f a a be e D ai k edge ca he a i h di g he igh de ...a d i h a e i g de e f a ce!
46. From Mas Heigh o T rbine Heigh We have: Long-term wind speed time series at 58 m height (height of the wind speed sensor on the mast) We want: Long-term wind speed time series at height of turbine We need: Information about turbine height Information about how wind speed behaves with height In this section: Ver short and simpli ed version.
47. T bi e Heigh We arbi raril choose as rbine model. H b heigh : 119 m (h b: "nose" of he rbine aro nd hich blades ro a e) T rbine heigh : H b heigh + ele a ion of land ha rbine s ands on (ass me 100 m for all rbines) Mas heigh : Mas heigh + ele a ion of land ha mas s ands on (ass me 80 m) Ves as V112
48. Shea : Beha io of Wind Speed i h Heigh We e he o cale ma ind eed o bine heigh : Shea e onen : ind o le o e la
49. Le ' nd he hear e ponen b ing S ea .A e age() me hod. brigh ind' In [28]: Shea e nen : 0.19 anem me e _heigh = [30, 45, 58] a e age_ hea = b .Shea .A e age(da a[anem me e ], anem me e _heigh ) in ('Shea e nen : :.3 '.f ma (a e age_ hea .al ha))
50. Appl Shea La o Long-Te m Time Se ie N ha e kn he hea e nen , e can calc la e he l ng- e m ime e ie a bine heigh . In [29]: On a e age, he ind i 9.1% fa e a he bine heigh han a he mea ed heigh . h_ bine 100.0 + 119.0 h_ma 80.0 + 58.0 l _ eed_a _ bine l _ eed_a _ma *(h_ bine/h_ma )**a e age_ hea .al ha in ('On a e age, he ind i :.1% fa e a he bine heigh han a he mea ed heigh .' \ .fo ma (l _ eed_a _ bine.mean()/l _ eed_a _ma .mean()-1))
51. Predicting Wind T rbine Po er O tp t We ha e c me a l g a . N ha e ha e a l g- e m i d eed ime e ie a bi e heigh , e a e ead edic bi e e .
52. Po er Curve Po er C r e: T rbine po er o tp t as f nction of ind speed. Let's plot the V112 po er c r e! I [30]: Obser ations: T rbine onl starts to prod ce po er at abo t 2 m/s ind speed, po er o tp t is stead bet een ca. 12 and 25 m/s. e _c e = d. ead_c ('./da a/ e a _ 112_ e _c e.c ', i de _c =0).i c[:,0] e _c e.i de . a e = 'Wi d S eed [ / ]' e _c e. ( i e='Ve a V112 P e C e (P e O i kW)', fig i e=(15,5), i =(0,3500));
53. Po er C r e s. Predicted Wind Speed No that e kno ho much po er the V112 turbine produces b ind speed, let's see ho our predicted long-term ind speed ts into the picture. I [53]: Obser ations: Long-term ind speed is er small in comparison to hat the turbine can handle, the V112 turbine is completel o ersi ed for this site! e _c e. ( e='Ve a V112 P e C e (P e O W) . L -Te W d S eed') _ eed_a _ b e. . ( ec da _ =Tr e, a a=0.5, b =20 , f e=(15,5), abe =' - e ');
54. Calculating Po er Output De i e he V112 being e i ed, le ' la a nd i h he ene g d c i n n mbe e ld ge if e e e b ild hi bine. We an ge a "feel" f he e and i in he c n e f he c mm ni a nd he Sandb , i e. I [78]: Alm 25 MWh! I ha a l ? I ha a li le? Le ' e e hi n mbe in he e m : I [79]: Tha i a l f a and a g d am n f elec ic ca cha ge ! Mea b e e ea W : 24,925 T a ab e a e da f 1 ea : 975 F Te a M de S c a e e ea : 249 _ e _ . e ( _ eed_a _ b e, e _c e. de , e _c e. a e , ef 0, 0) _ e _ d.Se e ( _ e _ , de _ eed_a _ b e. de ) e_ e e _d a _ ea _ e _ . a e[0]/(365.25) _ e _ ea _ e _ . ()/ e_ e e _d a _ ea ('Mea b e e ea W : :,.0f '.f a ( _ e _ ea )) (' T a ab e a e da f 1 ea : :.0f '.f a ( _ e _ ea /(3.5/60*1.2)/365.25)) (' F Te a M de S c a e e ea : :.0f '.f a ( _ e _ ea /100))
55. Ho Man Ho seholds Co ld We Po er? 2017: T e ea Sa D e e d c ed 5600 W f e ec c ( ce: ). T e Sa db ZIP c de (92121) ad 1677 e d 2010 ( ce: ). SDGE a E P ec -c de .c I [97]: We , a e a d eadf ce a . B , ad , d ea e e a e a c a . T ea : We d ' ea a b e d a e e Sa D e e d . C ea , e e a c a da a, b d a d b e c e e Sa db d e a e e e. W e bad - aced b e, e c d e 4.5 Sa D e e d (0.3% a a d e Sa db ). W 377 bad - aced b e , e c d e 1678 Sa D e e d (100.1% a a d e Sa db ). e d _ e _ b e _ e _ ea /5600 c _ _92121_ e _ b e e d _ e _ b e/1677 ('W e bad - aced b e, e c d e :.1 Sa D e e d ( :.1% a a d e Sa db ).'\ . a ( e d _ e _ b e, c _ _92121_ e _ b e)) ('W 377 bad - aced b e , e c d e :.0 Sa D e e d ( :.1% a a d e Sa db ).'\ . a ( e d _ e _ b e*377, c _ _92121_ e _ b e*377))
56. N Ca ac Fac (NCF) E e mea e h ell a bine a ind eed di ib i n and elec ici g id en i nmen in e m f ne ca aci fac (NCF). Thi me ic de c ibe h m ch elec ici he bine ill gene a e f m he ac al ind en i nmen , in c m a i n h m ch i c ld he e icall gene a e, if he ind ble en gh make he bine gene a e i ma im m e all he ime. In [100]: Thi ne ca aci fac i eall , eall l ! (T ical NCF : 30% - 50%) We c ld lace hi bine in a be e ! N b d ld b ild a bine cl e he Sandb (gi en a i cal da a)! The net capacit factor is 2.2%. ncf output_per_ ear/(365.25*po er_cur e.ma ()) print('The net capacit factor is :.1% .'.format(ncf))
57. How "Valuable" Would our Power be? Challenge: Rene able energ is not (al a s) produced hen needed Selling energ in high-demand hours can be more pro table s. in lo -demand hours Blue: Demand / Orange: Demand minus solar and ind (Sell energ hen this alue is high at slightl cheaper prices than fossil fuel po er plants to make good pro t.)
58. Let's plot our diurnal pro le to see if we would produce a good amount of electricty during these pro table hours. I [108]: Unfortunately, it looks as if we produce power right when a lot of solar power is in the grid, pushing electricity prices down. Not every wind project is like this – sometimes wind speeds are high just as energy demand peaks. mea ed_da a = da a[' d_58']. e am le('1h').mea () mea ed_da a.g b (mea ed_da a.i de .h ).mea ()\ . l (fig i e=(15,5), i le='Di al P e P d c i P file', lim=(0,23));
59. ...b ha o ld be a good po for a ind rbine? Al h gh Sa Dieg bei g e f he , he e a e e g d f i d bi e i Calif ia.
60. Refe e ce ( le ed ab e) Analy ing Wind Data M : Getting Wind Data: Met Masts L P Getting More, Long-Term Data ASOS : Topography vs. Wind F : E : Power Curve V V112 P C : . "W , M , S D " CC B -SA 2.0 " 3 " CC B -SA 2.0 . :// . . / ? =-GIT N - 4M :// . . / / . . - - . / /7- - 112-
61. R c (c ) Ca c a P O Toas er po er cons mp ion: , ass med 1200 W for 3.5 min es Tesla Model S 100 kWh ba er : H "Va ab " W P b ? D ck c r e image from ...b a b a a b ? California ind map from energ secalc la or.com Wikipedia Wikimedia Commons inde change.energ .go