PosterPresentation
- 1. Data from Global SunFarm Network User Interface (ui.R) Introduction Analytical Tools The Solar Durability and Lifetime Extension (SDLE) collects weather and power data from solar photovoltaic (PV) plants worldwide. Nearly one terabyte of raw data has been generated on-site and is stored on SDLE lab servers. This creates the need for methods by which an end-user can access, visualize, and interact with collected data. This project identified analytical techniques for solar PV power and weather data and presented an interactive “Kiosk Mode” created using the R programming language. The Global SunFarm Network Data Sources and Collection The SDLE network of data sources is worldwide: • 40kW SDLE SunFarm - Cleveland, OH • 1MW MCCo utility-scale plant - Cleveland, OH • 20kW IIT-GN rooftop arrays - Gandhinagar, India • Three rooftop arrays in Cleveland, OH - 35kW Adelbert Gymnasium - 35kW Tinkham Veale Center - 70kW Seidman Cancer Center • Two solar farms in Taiwan - Taitung - Luhju • AEP Dolan Technology Center • Replex Plastics Wind Gust Analysis Rooftop Power Production Analysis R statistical programming language • Open source, powerful, and flexible environment • Large external repository of analytical packages (CRAN) Shiny and ShinyDashboard (R packages) • Creates a web-accessible, end-user facing tool • Allows a user to select, plot, analyze, and download data • Easy to expand capabilities over time Data characteristics • Minute-by-minute resolution • Average and maximum speed (m/s) and direction (degree) Method of analysis • Gust speed = (Max wind speed – Avg. wind speed) • Dual-axis plots for characterizing events (like a storm) Example: Exploratory Data Analysis (EDA) • Plot, visualize data to find meaningful relationships • Performed on 4 years of data at the 35kW Adelbert plant Observations • Progressive, substantial drop in power in June • 7000kW @ 9.6¢ per kWh1 = $680 loss • Cause of loss – human interferenceMetrology - Vaisala WXT520 • Air temperature & pressure • Wind speed (max and avg) • Wind direction • Relative humidity • Rain amount Insolation & Irradiance • Kipp & Zonen equipment • CMP6 (top) – diffuse energy - instant (kW) and total (kWh) energy • CHP1 (bottom) – direct energy - instant (kW) and total (kWh) energy server.R Takes the output of the function & Generates output ID for each element to be presented in Main Panel to user server.R Processes inputs & Calls on following functions accordingly Main Panel Displays results to the user - Histogram, Plot & Summary Sidebar Panel Takes input from user & Generates unique input ID Acknowledgements Kiosk Mode and Data Collection Pathways Conclusions and Future Work We would also like to thank the SDLE Center, the Medical Center Company, Nielsen Holdings N.V., and the Indian Institute of Technology – Gandhinagar for their support of these endeavors. We also want to thank Mohammad Hossain and Pei Zhao for their help. 1 – eia.gov. “Average retail consumer price of electricity in 2013” • Scalable – Easy to replicate the functions and data folders to add new SunFarms as they come online • Accessible – Any user can easily plot data of their interest • Visual - Plots helps to identify any anomalies in the data over a course of time in very simple way. • Online - Can be hosted using an IP address which can be easily accessed by an end-user. • Time – The analysis process involves merging all data files into one data frame, which takes 5-6 minutes. • V: vs CRADLE – Compiling shared data from V: is slow. Our online database, CRADLE, will allow for much faster rates of download and analysis Pros Cons Monthly average total power production at Adelbert Gym: Data Collection Mechanism Kiosk Mode Architecture Data Ingestion into online database CRADLE Data Ingestion into Shared Drive V: Kiosk Mode Data Folders : sdledata, iitgndata, adelbertdata, tinkhamdata Each folder contains the unique SunFarm data Functions – sdle.R, iitgn.R, adelbert.R, tinkham.R Each function grabs the data from the respective folder Processes the data & returns the data to be plotted to the server.R This project identified analytical methods for use with solar PV weather and power data. Two examples of analytical capabilities were shown and an interactive mode for viewing and working with the data, dubbed the “Kiosk Mode” was also presented. Our future work includes: • CRADLE – Start ingesting data from CRADLE. • Speed – CRADLE ingestion will reduce compile time significantly • New capabilities – Integration of lab-developed analytical techniques and R scripts into the Kiosk Mode • Layout – The Kiosk Mode should convey information to the user more effectively. Output plots and layout can be improved. • Connectivity - Hosting the Kiosk Mode using a public IP address. Instrumentation, Analysis, and Outreach at 1MW and 40kW Solar PV Plants Matthew J. Randall 1 , Raj Shekhar 2 , Timothy J. Peshek 1 , Naran M. Pindoriya 2, Roger H. French 1 1Case Western Reserve University - SDLE Center l 10900 Euclid Avenue, Cleveland, OH 44106 2Indian Institute of Technology - Gandhinagar l Palaj Gandhinagar, Gujarat, India 382355