Come meet the poster presenters to ask them questions and discuss their work
Check the programme for our poster viewing moments. For more details on each poster, click on the poster titles to read the abstract.
PO021: Evaluating the effective energy gain due to power performance upgrades by processing historical data with python
Serena Lamberti, Wind Technology Specialist, Eni Plenitude
This work aims at proposing a method for evaluating the effective energy gain after the installation of a power-up solution through python processing of historical data. The method is applied to single WTGs by analyzing wind and power production 10min SCADA data in the time period before and after the power-up implementation (dataset A and dataset B). For each dataset, the power output is described by a scattered plot that in most cases will not clearly fit the theoretical power curve due to the presence of bad quality or abnormal data that should be removed to avoid biases or inaccuracies. The procedure of data cleaning and validation is based on the combination of three filtering criteria: 1. By applying DBSCAN algorithm, points are marked as outliers - and therefore removed from the dataset - when they lie alone in low density regions of the scatterplot. 2. Other customized filters are applied in relation to physical limits (cut-in/cut-off) and grid curtailments. 3. Final filter is applied using quartile algorithms. Any data point outside the range [Q1 – 1.5 * IQR, Q3 + 1.5 * IQR] is considered an outlier (Q3 and Q1 are 75th and 25th percentiles, IQR = Q3 – Q1). Cleaned dataset A and B are organized in a matrix of wind speed bins and direction sectors. All cells not reaching a defined minimum threshold of samples size are discarded. For each remaining cell, the mean power output (MW) is calculated.