summary presentation on the analytical monitoring of grid

IEA INTERNATIONAL ENERGY AGENCY

PHOTOVOLTAIC POWER SYSTEMS PROGRAMME

From “Stamps” to User Guidelines: Case Studies on How to Assess PV Performance

Achim Woyte 3E sa, Brussels, Belgium

www.3E.eu



Outline

• User stories for PV monitoring • Monitoring data • Systematic approach for interpretation

=> Periodic linear regression • Examples • Conclusions



PV Monitoring – Why?

I like to see how much power my installation produces.

If anything goes wrong, I want to be alerted by SMS within 5 minutes.

I want to compare the performances of my plants to see which equipment works best.

For our investors, I want to compare the performance of their portfolio to the initial business plan.



Monitoring Data: What to Measure?

PR

IC Array capture

losses

IS System losses

PV array Inverter Grid GI

VDC IDC PDC

VAC IAC PAC

yr Reference yield

yA Array yield

yf System yield

Tamb , T mod , SW

©



Definitions: Yields & Losses

Yields Reference yield Yr = in-plane irradiation / 1 kW/m2

Array yield YA = DC energy / PV peak power Final yield Yf = AC energy / PV peak power

Losses Capture losses LC = Yr -YA System losses LS = YA -Yf

Performance Ratio PR PRA = YA / Yr PR = Yf / Yr

Capitals: energy [h]; small letters: power [p.u.]



Courtesy of Christian Reise, Fraunhofer ISE

Data Plots – A Visual Tool



Systematic Approach

• Use data for identifying the system and component parameters

• Identify simplified physical relationships • Derive linear model parameters from regression • Update periodically and compare the results • Deviations indicate irregular operation



Energy Flow in a PV System

PR

IC Array capture

losses

IS System losses

PV array Inverter Grid GI

VDC IDC PDC

VAC IAC PAC

yr Reference yield

yA Array yield

yf System yield

Tamb , T mod , SW

©



Physical Relationships of Quantities

Temp. – Secondary Effects Influence of wind speed on module heat transfer kth vs SW

System performance yf vs yr

Influence of module temperature PR vs TMod

System Level Performance

Array performance yA vs yr

Influence of module temperature PRA vs TMod

Array Level Performance Module temp. TMod – TAmb vs yr

Module Temperature Resilience of grid voltage on active power vAC vs yf

Grid Connection

Array voltage VA vs TMod

Array Level – Specific Effects



Final vs. Reference Yield

No shadow, June Growing vegetation, May

8 weeks in summer 2012

Installation in Belgium, monitored by 3E



Irradiance, PR and Module Temp. March 2012: smooth operation

Final vs reference yield Temperature influence

Installation in Sweden, monitored by ABB



Irradiance, PR and Module Temp.

Installation in Sweden, monitored by ABB

Final vs reference yield Temperature influence

May 2012: inverter failure (1 out of 3)




Module temp. TMod – TAmb vs yr

Module Temperature



Influence of module temperature PR vs TMod



Influence of module temperature PRA vs TMod

Array Level Performance Resilience of grid voltage on active power VAC vs yf

Grid Connection

Array voltage VA vs TMod




Thermal Behavior

Normal operation, August Sensor detached, July

8 weeks in summer 2012: detached temperature sensor

Installation in Norway, monitored by Teknova



Thermal Behavior with Wind Speed

∆T vs reference yield (yr) kth vs wind speed (SW)

Irradiance, equivalent thermal resistance and wind speed

Installation in Northern Italy, monitored by EURAC




Module temp. ϑMod – ϑAmb vs yr

Module Temperature



Influence of module temperature PR vs ϑMod



Influence of module temperature PRA vs ϑMod

Array Level Performance Resilience of grid voltage on active power VAC vs yf

Grid Connection

Array voltage VA vs ϑMod




Array Voltage vs Module Temp.

Near shadow, from end of April Near shadow continued (May)

Array voltage as main indicator for the operating point (April/May 2010)

Installation in Southern Spain, data from GL Garrad Hassan



Periodic Linear Regression • Simple standard method

for data analysis • Automate and apply

online or offline • Interpret visually or

analytically • Reconstruct PV system models

The method can confirm smooth operations and

reveal distortions.



Conclusions • Different monitoring users have different

expectations • Detailed analytical monitoring allows to identify

disturbances quickly and draw conclusions on the cause

• Periodic linear regression modeling is a powerful yet simple tool for interpreting your masses of data



Acknowledgements • Contributors from 9 countries:

Mauricio Richter (3E, Belgium), Bengt Stridh (ABB, Sweden), David Moser (EURAC, Italy), Nils Reich (FhG-ISE, Germany), Stefan Mau (GHGL, Spain), Hervé Collin (INES, France), Mike Van Iseghem (EdF R&D, France) Anton Driesse (PV Performance Labs, Canada), Anne Imenes (Teknova, Norway), Sulaiman bin Shaari & Ahmad Maliki bin Omar (Universiti Teknologi MARA, Malaysia)

• Regions of Belgium for funding of our IEA PVPS work – Brussels Capital Region – Région Wallonne – Vlaams Gewest

summary presentation on the analytical monitoring of grid

Documents