Intro

The goal of this project was to provide different tools for the photovoltaic plant (PV) park owned by the Enel group. In particular, we provided optimization in the following areas:

  1. Digitization and digitalization of plant information
  2. Performance Ratio validation: solar sensors checking, satellite-based data provision and reference solar power curve creation
  3. Predictive maintenance services

Executive summary

– Company: Enel Green Power

– Industry: solar power plants management

– Challenge: exploitation of big data analysis to optimize the power plants asset management, with efficiency improvement and Operation and Maintenance (O&M) costs reduction.

Background

Owner needed to improve and homogenize the current digitization and digitalization procedures and information, giving the very different situations thought all the subgroups in different world states with different languages, file formats, archived information, and work-habits.

PV Power Plants (Worldwide)

Challenge

The challenge was to improve the initial assets management thanks to a three-steps procedure.

Solution

In the first step, we provided guidance and operatively created a homogeneous taxonomy and database structure to be used for all the PV-plants.

In the second step, we checked all the plants performance ensuring a quality check both on electrical data (power, current…) both on environmental data from the installed weather-stations. When sensors data were not available or were deemed not accurate enough, we provided satellite-based irradiance data as a reliable and alternative data source.

In the final step, we created a predictive maintenance algorithm specialized for the customer needs, able to predict incoming failures.

Benefits

The predictive maintenance allowed to avoid costs timely detecting failures (details on the table); the sensor check activity was performed on about 600 sensors and allowed to optimize recalibration and purchasing activities. The digitization and digitalization activities boost the systematic activity performances.

Through the activity, data from operating plants are used to optimize their management and quickly identify potential malfunctions. Innovation goals are the asset management optimization, generation efficiency improvement, and O&M costs reduction.

New interfaces were built to allow access to plant monitoring also from a tablet or any mobile device to facilitate the work of plant O&M operators.

Benefit Assumption

Time to Repair

744h (1 month)

U

Time to Detection

504h (3 weeks)

Predictive Model Sensitivity

85%

Failure Probability

8%

O&M Savings

20 ÷ 30 %

Benefit Evaluation

Revenue Gain

from 90 k€/y up to 145 k€/y *

O&M Cost Savings

from 280 k€/y up to 420 k€/y **

Benefit impact on net revenue

from 2.7% up to 4.2% ***

Table 1: evaluation of benefits given by i-EM solution. Benefit values are reported as absolute and as percentage values of the gross revenue of a PV portfolio. *Revenue Gain min=744h (time to repair)*0.17 (utilization factor)*100MW*100 €/MWh*0.08(inverter fault probability)* 0.85 (model sensitivity)

R

For the curious costumer

At i-EM S.r.l. we have what you need (we hope!)… Try us!

Author

Fabrizio Ruffini, PhD

Senior Data Scientist at i-EM