The world is racing towards global electrification, poised to provide electricity to the 20 percent of the global population currently without this utility. Leading the race is wind energy—specifically offshore. The transformation of the energy industry from “black gold,” oil, to “green gold,” renewable energy sources, will have varied far-reaching impacts. Fortunately, TIBCO has developed a comprehensive solution that will take current and future wind energy investors and operators from the land to the sea—moving operations offshore.
In fact, over $35 billion was invested in offshore wind projects during the second quarter of 2020. The longevity and significance of these projects make this a hot market with plenty of room for expansion. To tackle the complex variables that impact the outcomes of wind projects, TIBCO Data Scientist Catalina Herrera and her team have worked tirelessly to produce the most comprehensive solution available to date.
Investors and developers can utilize this solution to not only forecast potential for wind farms, but to predict the production power of wind projects and determine the feasibility and potential limitations of a location. Users can also site and plan for the great new frontier—offshore wind farms. Established wind farms will also benefit from operations optimization such as the ability to incorporate real-time data from sensors to prevent downtime and eliminate overproduction concerns.
From planning to operations, siting for wind farms to placing steel in the ground, TIBCO’s Wind Forecasting Analysis is opening doors for energy players. In this blog, you’ll learn the role of analytics and data science in bringing the promise of one form of renewable energy, wind, to fruition.
3 Best Practices for Analytics Professionals to Bring Wind Power to Fruition
Source Meaningful Data
The variables needed to be considered for accurate wind forecasting are vast. In this solution, wind characteristics such as wind speed, wind temperature, and wind pressure comprise just one component of analysis. Wind characteristics are proven variables that are required when predicting power outcomes. Additional variables such as weather seasonality, product power of various wind turbines, and geographical relation to power stations (and more!) also play a role in developing this comprehensive analysis.
Due in part to the multiple players involved with collecting these data sets, many data silos exist. Pulling in this data from weather stations, equipment manufacturers, and current wind farms, produces large quantities of data but not necessarily quality data. Therefore, data cleansing is imperative to take this raw data and produce quality models for prediction. TIBCO solutions make the data wrangling process quick and seamless.
Create Location-Specific Time Series Models
Using the cleansed data, this Wind Forecasting Analysis Solution creates time series models for stakeholders in the wind industry.
- Wind farm developers can use the models for siting optimal locations.
- Energy suppliers can avoid overproduction by coordinating the collaborative production of traditional power plants with weather-dependent energy sources.
To obtain more accurate predictions, extrapolation techniques are required. Obtaining weather data or wind farm data isn’t enough, as this analysis requires weather data from exact longitudes and latitudes. This extrapolation can be done using Voronoi Polygons, a function of TIBCO Spotfire. To create accurate models, weather conditions data are indexed in time order to demonstrate simple models such as wind speed variations during a single day. These smaller models are captured in an ARIMA (Auto-Regressive Integrated Moving Average) model to predict wind power generation. ARIMA is a class of predictive model that captures a suite of different standard temporal structures in time series data.
The user of this Wind Forecasting Analysis benefits from the clean visualizations that TIBCO Spotfire provides, but behind the dashboard, is a complex data pipeline that uses large quantities of data from many services, processed by advanced statistical methods. TIBCO Data Science is responsible for processing all of the data transformation from the multiple data sources, imputing missing values, and aggregating and joining data tables—all in preparation for the ARIMA model to run.
Looking ahead, the solution can incorporate streaming data, effectively offering the analysis in real time. This hyperconverged analytics approach enables users to not only visualize and perform expansive analysis into wind data but also to generate actionable insights, rapidly.
Make Informed Decisions: Onshore and Offshore
As wind energy plays an increasingly important role in the supply of energy globally, implementing a solution for predicting the output of wind farms becomes more crucial. Not just for planning new wind projects, but also for optimizing existing operations, and creating energy ecosystems that are collaborative and efficient.
Here’s what we know about offshore wind:
- 80 percent of offshore wind resources are in waters greater than 60 meters (197 feet).
- Floating wind turbines enable sites further from shore, where they are out of sight, but there’s also better wind!
- Floating wind technology is expected to be deployed at utility-scale by 2024.
The analytics needed behind the scenes will continue to increase in complexity as the number of variables to consider increases with the transition to offshore wind projects. With the push for offshore wind projects to provide a greater amount of energy to the grid within this decade, the capabilities that this analysis provides are all the more relevant. We are proud to be spearheading a data science solution that will empower the renewable energy industry. We are actively working to employ this same science in other types of renewable energy, such as solar. You can learn more about our Wind Forecasting Analytics solution in this webinar.