Slide

ABOUT

PVET’ is an advanced performance evaluation service for PV plants thataccomplishes a comprehensive, real- time and low-uncertainty audit of the installation. PVET’ is based on a system that combines Big Data, machine learning and parametric algorithms that allows optimizing the long-term production and O&M costs of the PV plant, leading to a maximization of its economic return.

The main advantage of PVET® is its low-uncertainty procedures for intelligent detection, quantification, diagnosis and prediction of performance failures, which allows adapting and optimizing O&M strategies, as well as accurately forecasting power generation, in arder to meet the requirements of the energy markets.

Slide

Solar operation performance monitoring

The assessments are not limited to individual plants, but they consider cross-analyses through entire portfolios, achieving higher levels of optimization. This approach allows our clients to have a greater strategic vision of the assets and to be able to manage their operation in a more precise way.

PVET® is delivered through an intuitive, simple and convenient front-end that incorporates additional functionalities to the system, giving it a great flexibility.

Cross-platform integrated at web, app and control centr levels, which allows supervising large PV plant portfolios from any place and at any time.

Slide

DATA ACQUISTION & FILTRATION

PVET® works with the all the data collected by the monitoring system of the PV plant. In addition, it can include data coming from other external sources, as reference modules or soiling sensors, or even from non-continuous data sources, as thermographic inspections or I-V curve testing campaigns. Once connected, PVET® synchronizes all this data in dedicated databases in the cloud, as presented in Figure 1.

FUNCTIONALITIES: DATA ADQUISITION

Figure 1. Data collection and synchronozation of PVET®

Before accomplishing the analyses, PVET® applies an advanced filtering process to guarantee the data quality, which is a key condition to assure the quality of the results themselves. More in detail, PVET® applies two types of filters:

> Monovariable filters: they assure that every registered variable is active, well connected, well recorded and behaves according with the expectations. These filters include the requirements established in the IEC- 61724 standard.

> Advanced multivariable filters: they guarantee coherence among variables of different types coming from different sources.

Slide

Big data for clear results

PVET® integrates the SCADA data and synchronizes them real time on cloud together with other additional data sources:

  • Reference modules
  • Soiling sensors
  • Thermographic campaigns
  • I-V curve campaigns with the E-1000 tracer.

DATA COHERENCE VERIFICATION

  • Automatic real time access to the client’s databases.
  • Detection of communication failures and lack of data.
  • Management of the gaps in the data series and separation between communication, registration and operation failures
  • Filtering of anomalous values. Management of data with values outside the usual operation ranges together with validation of the data source by means of value coherence algorithms.
  • Digital recalibration of the measurement sensors.
Slide

DATA ANALYSIS

PVET® calculation engine is divided in 4 main modules that give different information about the PV plant.


> Main performance indices of the plant. For every single element of the PV plant (string, stringbox, tracker, inverter … ), it obtains yields, efficiencies, performance ratios, availabilities, contractual values …

>  Comparison with the initia! production estimates. For every element and for the PV plant as a whole, PVET® compares the performance results with historical records and the initial yield assessment estimates.

>  Fault detection and diagnosis. lt detects, quantifies, diagnoses, evaluates and predicts performance anomalies in the plant. lt separates between:

  • Data acquisition:faults in communication systems, sensors, or invalid measurements of the devices.
  • Generation failures: operating failures of any of the elements of the PV plant, related to energy losses or equipment behaviour different from those expected
  • Predictive failures: anomalous behaviours that, although they do not yet entail an energy loss, may do so in the future, or lead to premature ageing of the equipment.

> Digital twin and energy prediction: lt simulates the behaviour of every element of the PV plant according to the operating conditions, estimating the producible energy. This module also validates the simulation models of the PV plant what, once optimized, allows the prediction of the expected production of the PV plant at different time horizons to ease the market integration of the PV plant.

QPV® ANALYSIS MONITORING OVERVIEW

Figure 2. Data analysis of PVET®

Slide

In search of efficiency ...

1. KEY PERFORMANCE INDICATORS

  • Calculation of the main PV plant KPIs: productivity, PR, PRSTC, PI, uptime contractual and energy-based availabilities, expected yield, and losses breakdown.
  • Calculation of the main equipment KPIs: efficiency, operating voltage response, temperature derating, and MPP tracking accuracy.
  • Calculation of the main O&M KPIs: acknowledge, intervention, response and resolution times, and spare parts levels.
  • Monitoring of the economic evolution of the project and comparison with initial expectations.
  • Comparison with historical evolution according to weather conditions.
  • Calculation of long-term component degradation, which is one of the main reliability and life-time indices.
  • Study of the quality of the AC signal: reactive component, power factor, current and voltage imbalances, response to the grid constrictions.
  • Comparison of operating conditions measurement sensors.
  • Calculation of soiling losses.

2. FAILURES ANALYSIS

  • Fault detection. Low uncertainty procedures for the detection of anomalies. Our advanced algorithms allow detecting failures a ecting a string in only 2% of its power generation capacity (For example, due to a shorted diode).
  • Advanced fault diagnosis based on ‘big data’ and ‘machine learning’ techniques. For example: shorted diodes, low STC power, tracking errors, MPPtracking defects, imbalances between phases, anomalous temperatures, PID, LID, open series, inverter stops, grid limitations, low performance, soiling, hot spots…
  • Inclusion and management of native alarms as a reinforcement to the diagnosis of anomalies.
  • Development of spatial and temporal analysis methods for the improvement of false alarms in detection, especially on days with meteorological variability.
  • Expert system with cross-plant analyses, which allows managing O&M portfolios in a more precise way, anticipating degradation failures and identifying the best-performing equipment.

3. PRODUCTION ESTIMATION

  • Calculation of the losses breakdown: temperature, low irradiance, inverters, transformation, soiling, shading, dispersion, wiring, power factor, and grid restrictions.
  • Estimation of production, with a precision better than 2% from the operating conditions records.
  • Algorithms for production prediction at 24 hours horizon, to satisfy the requirements of the generation markets, based on satellite meteorological predictions. This way, installations will be able to minimize production deviations when attending to the energy sales markets, minimizing the risk and achieving the maximum economic return.
  • Calculation of losses due to grid curtailments.

4. PREVENTIVE AND PREDICTIVE ANALYSIS

  • Analysis of degradation and estimated life time of equipment that allows anticipating the maintenance work.
  • Planning of cleaning campaigns to maximize production and reduce cleaning costs.
  • Forecast of failures due to fatigue or wear and solution planning.
  • Cross-analysis between PV plants and devices.

5. ASSET PORTFOLIO ANALYSIS

  • Comparison of plant operation.
  • Analysis of technical and economic e ciency of O&M tasks.
  • Analysis of technical and economic e ciency of equipment manufacturers and distributors.
Slide

DIAGNOSTICS

DATA

  • Missing data
  • Filtered data monovariable: IEC filter of Range, Dead value, and Abrupt value.
  • Filtered data multivariable: coherence filter

PRODUCTION

  • String: defective
  • String: open String
  • String: defective diodes
  • String: shadows dueto backtracking
  • String: soiling affection
  • Stringbox: open stringbox
  • Stringbox: open string
  • Stringbox: defective string
  • Stringbox: shadows due to backtracking
  • Inverter: stop
  • lnverter: MPPT deviation
  • lnverter: clipping
  • lnverter: late start
  • lnverter: voltage limitation
  • lnverter: temperature derating
  • Tra cker: safety position
  • Tracker: stop
  • Tracker: deviated position
  • Grid: power limitation
  • Grid: voltage limitation
  • Grid: power factor limitation

PREDICTIVE

  • lnverter: anomalous temperature
  • lnverter: voltage unbalance
  • lnverter: current unbalance
  • lnverter: air filter change
Slide

FUNCTIONALITY

PVET®service adds a continuous audit and advice by our expert engineer team:

  • Personal attention to the client during his daily operation.
  • Analysis and performance reports of specific failures.
  • Customized development to meet specific customer needs.
  • Advice on improvement actions in the plant monitoring system.
  • Recommendation of in-field actions for analyzing and solving specific problems.
Slide

Front - end: Endless possibilities

The PVET® is delivered through an on-line platform that allows the continuous monitoring of the PV assets by the client, showing all the functionalities of the PVET® service:

1. MULTI-PLATFORM VISUALIZATION:

  • Web platform as the main service tool.
  • Multi-platform App.
  • Control Centre mode for a continuous visualization of the whole operating assets.

2. USER MANAGEMENT:

  • Different types of users.
  • Control of user’s permissions and information access. • User management by group of plants.

3. GENERATION AND MANAGEMENT OF REPORTS:

  • Daily, weekly, monthly, quarterly and annual reports with the most relevant information. • Reports adapted to types of users.
  • Periodical report libraries and other reports associated to the plants.
  • Automatic mail reporting.

4. NOTIFICATION SYSTEM BY EMAIL AND APP.

5. INCIDENCES AND EVENTS SYSTEM:

  • Calendars to manage incidences and events.
  • Incident management by text and picture tickets.
  • Planning of preventive actions and measurement campaigns.
  • Download tool. Possibility of downloading any PVET system information in a simple and intuitive way.

Co-financed project

The PVET® developments contain the results of the project “Supervision and optimization system based on Artificial Intelligence and Big Data to maximize the performance of photovoltaic plants”, which consists of the development and implementation of a new family of algorithms that allow to carry out market an innovative solution in the field of supervision and management of photovoltaic plants. This project has been financed by the European Regional Development Fund “A way of making Europe” under the Red.es program.

This site is registered on wpml.org as a development site.