High Performance Monitoring Install and forget? Not quite. Best practices for PV performance management tools Solar Asset Management Asia Conference Arup Barat arup.barat@drakerenergy.com June 24, 2015
Draker Overview Leaders in PV performance, control and asset management solutions Proven experience in solar PV monitoring (15 years). Over 2GW of solar PV under management. Global experience: Japan, USA, Canada, Spain, Italy & Saudi Arabia. Market leaders in North America for third party monitoring in the utility and large commercial segment. Maintain one of the most comprehensive sets of continuous field performance data from very diverse PV systems (500+ module types, 150 inverters types, 60 met stations). Independent, integrated solution provider covering engineering, design, manufacturing, installation & maintenance. Innovation & quality applied with flexibility to comply with market and customer requirements. Key local partnerships to enable deep Customer Commitment and local support. Partner in Japan:
Large-scale PV plant facts 250MW PV Plant Size > 6,300 soccer fields. 7 miles of internal roads. More than 200 inverters. 750,000+ panels. 1,600+ combiner boxes. 35,000 strings. What are the points and modes of failure? What is the impact on the grid? How is the lifetime value of the asset impacted?
What are the sources of underperformance? Commonly known & modeled* Component derate factors Default Low High PV module nameplate DC 0.95 0.80 1.05 Inverter and transformer 0.92 0.88 0.98 Panel mismatch 0.98 0.97 0.995 Diodes and connections 0.995 0.99 0.997 DC wiring 0.98 0.97 0.99 AC wiring 0.99 0.98 0.993 Soiling 0.95 0.30 0.995 System availability 0.98 0.00 0.995 Shading 1.00 0.00 1.00 Sun-tracking 1.00 0.95 1.00 Age 1.00 0.70 1.00 Overall DC-AC derate 77.0% 0% 99.3% *Default values from PVWatts, NREL Uncommon or harder to model Real world examples: DC disconnects accidently left open after annual maintenance (4% energy loss) String fuses not replaced after annual maintenance (-1%) Blown string fuses from incorrect sizing at commission Vandalism leading to weak string production (-2%) Unmanaged vegetation (-8%) Industry Studies Missed cleaning for 3 months (-30-45%) Dust Storm (-60%) Static O&M plans are insufficient. Need to adapt to meet project / portfolio needs.
Underperformance happens Cleaning crew arriving on site Unmanaged vegetation
What is the impact of underperformance? It Depends! Each project and each portfolio is different. Value of energy lost Cost of grid electricity Cost of business loss PPA rates Contractual penalties Required rate of return Lifetime revenue Grid support commits O&M costs Some Real World Examples: (Pay back period for loss mitigation systems) 1. 1 MW-dc array in LA, USA; 1602 kwh/kwp; Payback in 2.89y Value of 1% increase in performance (@$0.12/kWH x 20 Years straight line) 2. 2 MW-dc array in Japan; FiT of $0.50/kWh; Payback in 1.59y 3. 5 MW-dc ground mount, retrofit in Spain. Payback in 1.56y *From SMA published study
Should I actively manage underperformance? That also depends! It s an economics-driven decision. Value of Incremental Energy Output Value of O&M Cost Savings Cost of Advanced Management System PPA/FiT pricing Net metering savings Typically $0.05 - $0.35 kwh, depending on jurisdiction Fixed price O&M contract Cost of cleaning Cost of corrective maintenance Typically $0.015 - $0.025/Wdc Site configuration String or zone-level granularity Cost to install Typically $0.01 - $0.02/Wdc
How can I mitigate the risks of underperformance? Be: data driven. If you can t measure it, you can t improve it. Ask for: the right information at the right time in the right way. Both remote and local monitoring Real time + batch processed Portfolio analytics: comparative performance Track: relevant performance indicators per project and across portfolio Production: Planned, Expected, Actual, Variance Efficiency: PR, Yield, Availability Financials: fixed, variable costs Strive for: data actionable information automation Closed loop on-site control; Open loop remote control Transparent and customizable automation
Types of monitoring systems Enabling data driven decisions Measure Actual Output Inverter Direct Only Insufficient for large plants AC Meter Monitor long term trends With Met sensors, basic performance assessment No reference point for comparison DC Subcombiner Account for inverter losses Comparative performance DC String Soiling, shading, loss analysis Variable Basic degradation Compare Actual to Model Against Predicted Fixed model inputs + variable historical Met data Predicted output Against Expected Fixed model inputs + variable measured Met data Expected output Compare Actual to Optimal Against Optimal Highly situational Based on comparative performance data Means different things to different audiences. Optimize: o Pure Performance o Operating Cost o ROI o Contractual obligations o Resource loading Requires real time and batched analytics and data mining. Advanced
Performance Loss analysis by type of monitoring Configuration Inverter Direct AC only submetering DC String Combiner DC String level DC substring/pa nel level Inverter Failures/System Underperformance R R R R R Dead String Detection O? R R R Comparative string performance (typical accuracy) O O R +/- 5% R +/- 0.5% R +/- 0.5% Dead Panel Detection O O O R R Soiling Analysis (typical accuracy) O O O R +/- 3% Zonal Soiling Analysis O O O R R Mismatch Loss analysis O O O R R Panel Diode Engaged Detection O O O O R Invert MPPT Mistracking O O O O R Array Wiring Losses O O O O R R +/- 1%
Performance Loss Analysis visualization Samples & Screenshots
Solar Asset Management Platform Cloud Data Delivery Intelligent Array TM Platform Intelligent Array Management Suite! Moving beyond visualization! Architectural hierarchy, spatial topology and logical structure fully comprehended! PV impairment models embedded! Fault diagnosis in real time! Intelligent O&M support! Generate information, not data! Accurate root cause differentiation! Business rules based decision making! ROI and threshold based alerts! Prioritized maintenance plans! Differentiated user classes! Self clearing action plans! Data Analysis Network Operations Center (NOC) Page 13 Company Confidential Intelligent Array TM User Base Station Inverters Meters Met Sensors Utility or ISO SCADA System Data Collection Intelligent Maintenance Active Monitoring Preventive Maintenance Directed Dispatch TM Informed Operations Management Reporting Asset Optimization / Document Mgmt Solar Asset Management Platform O&M Provider System Owner Stakeholders
Panel-to-Grid TM Visibility and Control Turn mountains of data into actionable information through: Careful preprocessing Systems learning Insightful analytics Meaningful metrics Surgical response This is only possible if you have: Granular data acquisition Last mile expertise DC/AC/Grid/Portfolio-level visibility Highly robust, extensible, modular, scalable platform Big Data analytics engine
What s next? Portfolio Management 2.0 Even portfolio-level asset management is not sufficient, what the market needs is intelligent asset optimization. This optimization may mean different things to different stakeholders: Optimal operations and reduced service cost for O&M providers Component reliability and optimal usage for developers/epcs Peak performance for financiers Guaranteed returns for owners Big data, systems learning and integration with best of breed vertical applications will be needed.
Approaches to Optimization Dynamic Optimization Reactive Optimization Dynamic Optimization Focus on maximizing energy production Real-time compensation for shading and other sources of power loss Improved energy harvest by keeping DC strings at inverter maximum power point Particularly useful when obstructions can t be designed around neighboring buildings, trees, building orientation More often used for unattended arrays without access to O&M personnel Reactive Optimization Focus on optimizing lifetime ROI No real-time compensation for power losses Improved energy harvest by informing O&M tasks and decisions Detailed power information capture and analysis Informs manual intervention as required i.e., repair wiring, replace modules, clean array
Thank you! Arup Barat arup.barat@drakerenergy.com Sales Contacts Japan: sakamoto@eko.co.jp United States: tom.walsh@drakerenergy.com