Key Features

• Open Source with source code on GitHub
• Interactive map with production forecasts for wind power plants of a country
• Diagram with production forecast for individual wind power plants
• Customisable wind power plant attributes
• Crowdsourcing functionality to improve the model: New crowdsourced data are regularly checked for plausibility before being used to enrich the dataset and to retrain the model on this enriched dataset

Technical Model Details

• One deterministic Multi-Layer Perceptron (MLP) per lead time for arbitrary wind power plants (unified model)
• Trained on European wind power plant production data from ENTSO-E Transparency Platform
• ReLU activation function, three hidden and one output layer
• Input features: Turbine Type, Hub Height, Age, Wind Speed
• Model predicts normalised wind powers, which are then scaled with the capacity values. Opposite test results refer to these normalised values, during testing of the model on 20 % of the data.

Usage

• Extreme values in features and long forecasting times reduce accuracy. Only the latter is accounted for by the standard deviation.
• The wind power plants are displayed from zoom level 11 onwards.
• Missing Hub Height (26 % of displayed wind power plants): Mean value from database is used as default
• Missing Commissioning Date (6 %): June 2010 is used as default
• Unknown Turbine Type (16 %): Equal mixture of all known turbine types

Motivation for Contribution

• Improves the forecasting model by allowing it to learn new value ranges and turbine types
• Allows this web application to remain free to use
• Supports sustainability as a Corporate Social Responsibility (CSR) initiative

Contribution Guidelines

• Set correct wind power plant configuration and time zone
• Check the displayed production diagram before contributing
• Submit production values in kW (Download Example File)

Data Sources

• Wind Power Plant Metadata: The Wind Power
• Wind Power Plant Production Data: ENTSO-E Transparency Platform
• Wind Data: ECMWF

Data Security & Privacy

• All uploads are encrypted (HTTPS)
• Data are stored securely on a Microsoft Azure server in Germany until downloaded locally and not shared with third parties
• Data are stored permanently in the aim to build a valuable database with European wind power production data
• If you want to retract contributed data, please send an email with the configuration of the wind power plant and the approximate time of contribution to the email address below
• Contributions remain anonymous

Legal Notice

This web application was developed by Alexander Peters at Imperial College London and RWTH Aachen University.

The content and forecasts displayed are for informational purposes only. No responsibility for losses or damages arising from its use is assumed.

For questions, concerns, or feedback, please contact a.peters24@imperial.ac.uk.

Model Test Results

Lead Time (h)	Huber	MAE	MSE	RMSE	MBE	STD
0	0.014116	0.112665	0.028232	0.168023	0.005455	0.167919
3	0.014713	0.115218	0.029436	0.171568	0.000899	0.171562
6	0.013966	0.114196	0.027933	0.167130	0.013213	0.166600
9	0.013523	0.112362	0.027047	0.164458	-0.006482	0.164318
12	0.014666	0.115518	0.029333	0.171268	0.002889	0.171252
15	0.015538	0.120240	0.031081	0.176297	-0.012355	0.175857
18	0.015005	0.120570	0.030010	0.173235	0.008307	0.173038
21	0.014797	0.119310	0.029594	0.172029	0.000753	0.172029
24	0.015680	0.119206	0.031359	0.177085	-0.009626	0.176785
27	0.016400	0.125365	0.032810	0.181136	-0.001759	0.181139
30	0.015848	0.122604	0.031696	0.178035	-0.006956	0.177900
33	0.015782	0.124985	0.031564	0.177664	-0.000418	0.177658
36	0.017216	0.128337	0.034432	0.185558	0.009351	0.185323
39	0.018059	0.134220	0.036126	0.190067	0.004525	0.190014
42	0.017772	0.132578	0.035544	0.188531	0.016043	0.187844
45	0.018103	0.134621	0.036206	0.190278	0.026019	0.188488
48	0.018413	0.134545	0.036827	0.191902	-0.007000	0.191787
51	0.019527	0.142666	0.039067	0.197654	0.003965	0.197634
54	0.019217	0.141065	0.038434	0.196046	-0.000155	0.196048
57	0.019533	0.142086	0.039066	0.197652	0.007355	0.197514
60	0.020603	0.146156	0.041205	0.202991	0.002537	0.202970
63	0.021654	0.154504	0.043316	0.208125	-0.001461	0.208113
66	0.021484	0.152684	0.042969	0.207289	-0.006130	0.207192
69	0.022200	0.156112	0.044399	0.210712	0.000978	0.210702
72	0.023348	0.157520	0.046697	0.216095	-0.000592	0.216105
75	0.024224	0.165452	0.048468	0.220154	0.000162	0.220159
78	0.023988	0.164340	0.047977	0.219036	0.000586	0.219031
81	0.025103	0.166808	0.050207	0.224069	0.007997	0.223926
84	0.026101	0.171211	0.052201	0.228476	-0.009758	0.228268
87	0.026650	0.176331	0.053314	0.230897	-0.004291	0.230831
90	0.026188	0.174046	0.052376	0.228858	-0.001362	0.228854
93	0.027359	0.177099	0.054719	0.233920	0.004801	0.233881
96	0.028197	0.181478	0.056394	0.237475	-0.000012	0.237480
99	0.027985	0.181607	0.055991	0.236624	-0.003765	0.236565
102	0.028282	0.181937	0.056564	0.237831	-0.006911	0.237729
105	0.029170	0.184500	0.058339	0.241535	0.000559	0.241530
108	0.029906	0.189859	0.059811	0.244563	0.000641	0.244575
111	0.030771	0.192369	0.061560	0.248112	-0.004749	0.248078
114	0.030912	0.192836	0.061824	0.248644	-0.000739	0.248640
117	0.032094	0.194327	0.064188	0.253353	0.008098	0.253219
120	0.034210	0.206702	0.068419	0.261571	-0.004043	0.261534
123	0.035327	0.212137	0.070665	0.265829	-0.007391	0.265711
126	0.035226	0.210736	0.070452	0.265427	0.001915	0.265420
129	0.035710	0.210658	0.071420	0.267246	0.011399	0.267013
132	0.036475	0.216365	0.072949	0.270091	-0.002151	0.270075
135	0.036955	0.218529	0.073924	0.271890	-0.003511	0.271876
138	0.036900	0.216156	0.073801	0.271663	0.006966	0.271572
141	0.037613	0.218698	0.075226	0.274273	0.013024	0.273973
144	0.038517	0.225397	0.077033	0.277549	-0.003806	0.277522