Trends and innovations in photovoltaic operations and maintenance

According to the most recent Wood Mackenzie's report on PV O&M economics [1], the annual global capacity additions will average 193 GW_DC over the next ten years. In 2022 alone, an addition of 22 GW_DC is expected, reaching a total global capacity of 175 GW_DC, being the utility-scale the sector that will continue to dominate the global installed capacity (see figures 1 and 2).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The O&M market is expected to grow at a steady pace of 14%. The Asia-Pacific market is expected to reach nearly US$5.7 billion by 2030, followed by Europe, Middle East, and Africa with US$5.2 billion and North, Central and South America with US$4.1 billion. US will be the most attractive country for solar PV O&M, reaching a total opportunity of US$3.5 billion by 2030 (see figure 3).

Zoom In Zoom Out Reset image size

System density and economies of scale continue to be the main drivers for cost reductions across the globe, when considering full scope service agreements. O&M providers, therefore, servicing a significant volume of assets in a certain region, are likely able to offer a more competitive '$ kW⁻¹' (see figure 4).

Zoom In Zoom Out Reset image size

Competitive auction schemes continue to drive increasingly lower capital expenditure (CAPEX) and operational expenditure (OPEX). With O&M costs not expected to decline any time soon, service providers are being challenged to remain profitable. With single-digit margins being the industry norm, economies of scale and advanced digital solutions will be the main enablers to increase margins and overall profitability (see figure 5).

Zoom In Zoom Out Reset image size

Low contract prices are still a misleading reference figure in the industry as critical O&M activities are excluded from the scope to keep contract prices low. Certain low-labour-cost markets enjoy cheap O&M average costs—sometimes at least half as much as equivalent projects in developed markets—but their cost base becomes heavily skewed towards corrective maintenance and cost of spare parts. Projects in such markets are especially vulnerable to manufacturing supply chain disruptions.

Furthermore, Wood Mackenzie's report emphasises that all projects, regardless of size, experience average annual O&M costs risk escalation on various cost factors. For utility-scale projects, the two activities at most risk of cost escalation are site operations and scheduled maintenance costs. In markets with especially high labour costs, their cost escalation has profound impact on the ability of service providers to remain competitive and must do so through operational improvements.

Spare parts escalation has a less profound impact on overall cost escalation over a ten year period to 2030. However, supply chain factors such as tariff introductions and the impact of Covid on manufacturing have already shown a significant impact and will cause substantial near-term cost escalation. Equipment prices are expected to rise as high as 30% in 2022 depending on manufacturing origin and type, which may force some asset owners to delay corrective replacements till the supply chain issues are mitigated.

Commercial and industrial (C&I) projects generally have higher per kW per year O&M costs in many markets. Economies of scale is a key factor driving up price differences between utility-scale and C&I service contracts. Corrective repairs, including inverter and modules replacements, account for nearly half of the global O&M spend over the next ten years, a clear driver for potential repowering activities and advanced analytics solutions. This dynamic also emphasizes the benefits of full-scope contracts where the service providers have warranties in place.

The state-of-the-art regarding O&M best practices is led and coordinated by SolarPower Europe ¹ (SPE), one of the most important member-led industrial associations, which represents over 260 companies across the entire solar value chain. SPE publishes periodically several market outlooks and thematic reports on relevant topics such as lifecycle quality, sustainability, grids, emerging markets, digitalisation and storage, just to name a few. Within the Lifecycle Quality Workstream, updates and launches every two years (free of charge) a new version of the report entitled O&M Best Practice Guidelines. The current version, the fifth edition [2], builds on 2019's fourth edition and is the result of a year of intensive work by 29 leading solar experts, from 20 companies. The contributors work across the solar PV industry and include O&M service providers, Asset Managers, Asset Owners, renewable energy consultants, legal experts, digital solutions providers and technical advisors.

It is highly recommended to the reader to take a deep dive especially into chapter 12, where the most important trends and innovations shaping today's O&M market are summarized, including the latest aerial monitoring techniques, AI approaches for automated plant performance diagnosis and data-driven decision making.

Last but not least, SPE promotes the use of the Solar Best Practices Mark ² , which is a suite of self-certification-based labels based on the guidelines and checklists. Its purpose is to create more transparency in solar services and incentivise excellence by providing visibility to complying companies by featuring their profiles in the Companies Directory and the possibility to display the Mark on their company websites and other publicity materials (see the current existing Marks on figure 6).

Zoom In Zoom Out Reset image size

In addition to the innovation described in the O&M Best Practice Guidelines, the adaptation and application of BIM is one of the topics that is expected to become of relevance in O&M sector, as a result of research conducted at different level by some stakeholders of this sector. As explained in a recently published report by the TRUST-PV project ³ [3], BIM is already a proven concept helping the construction industry to render processes more efficient, reduce costs and risks. The adoption of the BIM methodology in the PV industry requires a collective effort because its benefits are maximized when it is implemented throughout the entire PV plant lifecycle.

Already in the 1970s it was understood that the efficiency of various construction industry processes could be significantly improved by a centralized data management system. During the planning, construction and operation of a built asset, data passes through the hands of multiple teams. It was observed that the information flow between delivery teams is often hindered: loss of information and miscommunication lead to reduced efficiency and delays, hence increased costs and risk. It became clear then, that all stakeholders of a built asset would greatly benefit from a framework supporting the management and production of information during the asset lifecycle, hence the formal concept of BIM started to develop:

BIM is an information management framework that is about getting benefit through better specification and delivery of just the right amount of information concerning the design, construction, operation and maintenance of buildings and infrastructure, using appropriate technologies [4]

As the PV industry grows and matures, there is an increasing demand towards implementing BIM for PV power plants. Current efforts are aimed at translating the concept for PV assets, identifying the main data components, the processes that access and generate information during each lifecycle stage and to describe the mechanics of information flow. BIM relies on strong support from digital technologies, and it is commonly implemented as a collaboration of multiple software tools. Interoperability during the entire lifecycle (figure 7) is best supported when the collaborating software use open standards for data exchange.

Zoom In Zoom Out Reset image size

A common challenge is that these phases are difficult to align, and information loss occurs when handing over the project from one task team to another. BIM can bridge this information loss by storing all data centrally for every party to benefit from. Furthermore, BIM allows to track individual components in a dynamic way, keeping up to date with maintenance tickets, configuration changes, sensor measurements, and control signals, which greatly benefits the monitoring and optimization of plant performance. Additionally, BIM can serve as an ideal database to derive insights regarding, among others, component failures which aids in future PV plant designs to reduce uncertainty and risk, improve predictability and hence grid-friendliness.

BIM is present in each stage of the PV lifecycle, which we have categorized chronologically in figure 8.

Zoom In Zoom Out Reset image size

2.1.1. Digital twin and 3D modelling

The second step in the PV lifecycle should be the creation of a digital twin, which in the scope of this document is understood as follows:

A parameterized digital description of a PV system that exists throughout its life-cycle, and contains all physical information needed to simulate the behaviour and performance of the real PV plant it represents. [3]

Digital twins then, viewed as digital representations of real PV assets, provide the baseline upon which the entire BIM concept is built. In other worlds, digital twins are enablers for the efficient application of more sophisticated data management systems.

As the digital transformation of the PV industry is gaining momentum, a considerable fraction of existing PV installations lacks a digital history. However, the fact that such plants were not provided a digital twin at the beginning of their lifecycle should not hinder the application of BIM in their O&M processes. It is therefore essential to develop a pathway for instantiating digital twins of older PV plants and for enabling the application of BIM.

While the digital twin of a new PV plant can be created directly by the design engineer using PV-specific CAD design software, it is common that the one for an existing plant is built up only from the available as-built documentation. Different methods exist for digitizing various aspects of existing PV installations, but keeping process scalability in mind, let us talk about a process that can be automatized to a certain degree: the creation of the 3D geometry of an existing PV installation using airborne photogrammetry.

The aerial 3D modelling process allows a very accurate spatial representation of solar assets. The process includes collection of visual imagery using very accurate ground control points that are processed using photogrammetry to create 3D models of individual components. The 3D model covers the ground topography, tree shading and building geometry within the site. After further image processing of the 3D model, PV module and table geometry are generated. All 3D shapes can then be included into a combined 3D geometry file using one of the standard file formats [5]. State-of-the-art services by industry leaders such as Above ⁴ , allow the aerial collection of 3D data (see figure 9) in a cost-effective way, helping to overcome the following challenges:

• (a)  
  As-built CADs (location, pitch, azimuth, tilt) are incorrect, inaccurate or absent
• (b)  
  Shading scenes are inaccurate or out-of-date (e.g. trees have grown)
• (c)  
  Accurate 3D model of the site required prior to maintenance work
• (d)  
  Electrical simulation (yield, performance ratio (PR)), e.g. in PVsyst, is based on incorrect or poor assumptions about site layout, topography and shading
• (e)  
  Settlement or subsidence that requires regular monitoring

Zoom In Zoom Out Reset image size

Once the digital twin of an existing plant is created, the information that it contains can be exported for its use at successive stages. For construction monitoring for example, 3D drone mapping could be repeated at different intervals to make sure construction or repowering works occurs according to specifications, a process that currently can lead to errors due to information loss between the designing and constructing party and therefore inconsistencies between expected and actual performance. The digital twin needs to be maintained through any modification the asset may undergo, to always reflect the current asset state. This enables the PV plant digital twin to act as a single source of 'truth' whenever any of the following information is requested: PV plant 3D model, metadata, Bill of Materials, Single Line Diagram, electrical layout, component datasheets and coordinates, exports to yield modelling or O&M software (see figure 10).

Zoom In Zoom Out Reset image size

2.1.1.1. BIM for O&M

During the O&M phase, the BIM methodology can help in identifying individual components, tracking their performance history, and possibly helping with predictive maintenance. This is in stark contrast with current practices where individual components are simply replaced without fully understanding the root cause of their failure, simply because information on its individual properties, history and performance have been lost during the PV lifecycle.

However, some important, performance-determining attributes of a PV plant cannot be controlled from visual data. Attributes such as component properties and electrical connections can only be obtained using existing documentation, with the inevitable involvement of manual processing. Solar design software such as PVcase ⁵ provide semi-automatized ways of adding an electrical design to a 3D model, once the stringing pattern and the locations of inverters, transformers and cable trenches have been identified.

The correct performance of O&M activities highly depends on the inputs (data) that the O&M contractor will gather from internal and external sources. PV plant metadata, such as the plant as-built documentation, quantity and specs from its components (modules, inverters, transformers, etc.), tilt and azimuth of installed modules and the geographical location as well as the information obtained from the monitoring system (Power, Current, Voltage, Irradiance, temperature, wind speed and direction), the maintenance tickets and electrical signatures are considered information from internal sources. On the other hand, 2D/3D plant layouts & yield simulations, satellite data (irradiance, ambient temperature, wind speed, etc.), IR thermal images & EL/PL/UV images are considered information from external sources.

With all the information gathered the O&M contractor is able to produce the desired outputs (services), for example a well-structured performance report from the PV plant including contractual key performance indicators (KPIs) (PR, Availability) calculations, and in necessary cases include the calculation of advanced KPI's (Temperature Corrected Performance Ratio, Performance Index, etc.). This enables to create a 'dynamic' maintenance plan that will include corrective and preventive maintenance processes as well as revamping & repowering projects. Field inspection and measurements such as aerial thermography, IV curve measurements are part of regular O&M activities. In order to plan these activities efficiently, the metadata of the site and plant electrical characteristics, such as string layout, electrical connectivity are necessary from BIM. The results of the inspection and measurement can be processed and analysed to determine the estimated power generation by means of an energy yield model. Comparisons can be made between multiple inspections and measurements to identify any changes in performance or degradation in the system performance. All the results, outputs and analysis can then be accessed and used for reporting of the system performance. These include anomalies (thermal or visual) detected through the aerial inspection, estimated power loss/generation, historical modifications of components and/or performance analysis [3].

2.1.1.2. BIM for end-of-life management

The last step of the PV lifecycle is its End-of-Life and recycling where PV components are dismantled and disposed of, or they can be used in another PV plant or for research purposes. BIM allows to retain all information of the individual component in its new environment and can also greatly help in better interpreting failure analysis or destructive testing results for research purposes.

Current PV monitoring systems typically rely on wired networks that establish connections between string boxes, inverters, irradiance sensors and weather stations towards modems or gateways, the most common technologies in use being (local area network, based on Ethernet cables), RS-485 serial interfaces (with Modbus protocol) and power lines communications. Internet connection is then achieved by modems that use one or several of the following technologies: ADSL, Optical fibre, 3G or 4G cellular technologies or Satellite Communication.

As explained in another recently published report by the TRUST-PV project [6], cellular technologies such as NB-IoT/5G, can offer many advantages for the application of IoT to existing and new PV monitoring systems, such as:

• (a)  
  Wide coverage area
• (b)  
  High number of connected devices
• (c)  
  Low power consumption and long battery life: devices can remain disconnected from the network as long as there is no data to be transmitted: this feature allows to save power and to guarantee a long battery life
• (d)  
  Tight coordination and standardization
• (e)  
  High quality of service
• (f)  
  Low-cost devices
• (g)  
  High stability of connections
• (h)  
  No need of any gateway to send data to the Web

NB-IoT offers also high Maximum Coupling Loss values and good signal penetration both indoors and in locations having solid obstacles (see figure 11).

Zoom In Zoom Out Reset image size

2.2.1. Revisiting the current PV Supervisory Control and Data Acquisition (SCADA) architecture

PV monitoring systems can be based on Wireless Sensor Networks that use cellular technology (NB-IoT/5G) or non-cellular (SigFox, LoRaWAN, ZigBee) protocols. Preliminary analysis by [6] already show that cellular NB-IoT technologies are suitable for complementing or even replacing the current SCADA architecture (see figure 12).

Zoom In Zoom Out Reset image size

NB-IoT provides an efficient transmission of small data packets and guarantees nevertheless a sufficiently high throughput, which makes possible over-the-air (FoTa) updates of device Firmware. The IoT SCADA architecture proposed is structured as follows:

• (a)  
  All meteorological and electrical sensors, string boxes, inverters and production meters will transmit data on NB-IoT network through the Web, towards a monitoring portal. NB communication modules can be directly integrated with the sensors.
• (b)  
  Separate IoT modems can be used, connected to sensors via RS-485 serial ports, for instance. In this way the modems work as converters from Modbus protocol to Narrow Band IoT.

The proposed IoT architecture has several technical, practical and operational advantages:

• (a)  
  No need of any central data-logger for data collecting: with current systems if a data logger fails, all data transmissions are interrupted or even lost
• (b)  
  High reliability due to a substantial connection redundancy
• (c)  
  No more wires nor cables are used, with positive consequences: easier installation, better maintenance and reduced electromagnetic interferences generally collected by the wires

PV plants produce a considerable amount of data that creates a complex analysis and performance calculation process. From the individual module to the inverters and meters, every system component in a PV plant generates data regarding power production, temperature, and other parameters, which are supplied into the monitoring systems in real-time. Without the proper means for processing and evaluating this data in a timely manner, both asset owners and O&M contractors may find themselves drowning in data, unable to capitalize on all of the insights hidden within it.

This is an ideal terrain for Big Data, which uses computational methods to improve PV plant performance estimates and data analysis. The Big Data concept has become increasingly important in the development of new analysis tools to improve the reliability of data-driven solutions.

When it comes to supporting the PV industry, Big Data analytics has a huge advantage. O&M contractors can use it to categorize data and provide relevant insights by researching data patterns. Utilities will be able to control solar wind and radiation fluctuations while accurately estimating the amount of energy that can be redirected into the power grid with these. The information is subsequently processed and applied to improve the efficiency of solar assets. This technology aids in deciphering the massive amounts of data being made available in order to obtain better results [7].

It is worth noting that the massive PV plant facilities have a highly sophisticated infrastructure that comprises, among other things, panels, sensitive equipment, sensors, and wiring. Ground-level asset maintenance is extremely complex and time-consuming in this instance; therefore, O&M Contractors can reduce downtime by doing preventive and predictive maintenance, and finally, by evaluating historical data, Big Data algorithms can be used to predict maintenance needs and streamline maintenance operations.

2.3.1. Failure diagnosis based on field data

Regarding Performance and Reliability of PV plants, the International Energy Agency PV Power System Programme ⁶ through its Task 13 has done a remarkable analysis and compilation effort to put together the state-of-art research-based technical developments on relevant topics, such as:

• (a)  
  Analysis of the effectiveness of predictive monitoring in avoiding failures in real case studies and assess the possibility to integrate the algorithms in monitoring platforms.
• (b)  
  Different approaches and methods to determine soiling losses.
• (c)  
  Methodologies for the calculation of the degradation and performance loss rates
• (d)  
  Methodologies to assess technical risks and mitigation measures in terms of economic impact and effectiveness during operation.
• (e)  
  Best practice on methods and devices to qualify PV power plants in the field.
• (f)  
  Compile guidelines for O&M procedures in different climates and to evaluate how effective O&M concepts will affect the quality in the field.

Furthermore, in the most updated report on the Quantification of Technical Risks [8], an exhaustive compilation of the most important PV failures is provided, not just to expand the existing understanding of them but also to link them to technical risks, degradation rates, power loss and to suggestions of how to intervene or prevent them, providing therefore a strong technical baseline for PV planners, installers, investors, independent experts and insurance companies.

The state-or-the-art for quality assurance features well known field inspections techniques such as aerial thermography and I–V curve tracing. Underperformance issues are detected by trained personnel in control rooms assisted by sophisticated monitoring software and finally contractual KPIs, mainly PR and Availability, are calculated and reported periodically with semiautomated procedures.

Most market ready software solutions that provide automatic failure analysis and diagnosis can be placed at the insight level as depicted in figure 13, and the adoption of predictive maintenance approaches (foresight level) that make use of artificial intelligence and machine learning techniques is experiencing a slower pace than other sectors, such as on-shore and off-shore Wind. Nowadays, one of the main challenges even for very large and consolidated PV O&M contractors is the cost-effective deployment of such advanced tools. Such solutions must be, usually, built on top of existing monitoring systems and are not easily deployable for large portfolios without expensive manual work. On the other hand, all-in solutions require abandoning current software platforms and therefore might not be an option for many players.

Zoom In Zoom Out Reset image size

The main drivers for cost reduction are innovations focused on the seamless integration of maintenance tickets, monitoring data and expert knowledge, being also investigated by the above-mentioned TRUST-PV project (see section 2.1).

A growing amount of waste from PV panels is projected for the coming years, and in Europe it is foreseen a generation of 11 million tons by 2050 [9]. As clearly discussed in [10, 11], PV waste (from modules, cables, inverters, oil, and other materials) is connected to the exponentially expanding PV installations on a global scale, and it is currently posing an emerging environmental issue while simultaneously presenting unprecedented and multifold value generation potential.

It is estimated that up to 80% of the PV waste stream can consist of product defects upon production, transportation or infant failures over the first four operational years, instead of products that actually reach the end of their designed technical life. Experts estimate that about 45%–65% of these PV modules can be repaired or refurbished. Therefore, up to nearly 50% of the PV waste can be diverted from the recycling path [10]. In this context, significant PV business and research and development efforts are shifting towards more sustainable, environmentally friendly, and economically viable end of life (EoL) management for PV modules, such as recycling, raw material recovery, repair/refurbishment, and even reuse of decommissioned or repaired PV modules [12] (see figure 14).

Zoom In Zoom Out Reset image size

As part of the O&M activities, the management of decommissioned PV components usually takes place when the following activities occur:

• (a)  
  Component replacement for yield maximization in a revamping or repowering project
• (b)  
  Decommissioning of components before EoL during ordinary or maintenance activities
• (c)  
  End of designed technical life of a PV plant (EoL).

When the modules need to be disposed, the O&M contractor often carries out the process of dismantling and disposal, on behalf of their clients, contacting a consortium authorized by the government that has partnerships with recycling facilities. This activity is conducted following the national and international legislation regarding the disposal of waste of electrical and electronic equipment.

Considering a scenario where the O&M contractor could be more active, expanding its performance and services and contributing with circular economy initiatives that go beyond the recycling path, the O&M contractor could consider buying the modules for offering services such as:

• (a)  
  Repair & Refurbishment
• (b)  
  Re-certification & Labelling
• (c)  
  Reuse (Sell to a second-hand market)
• (d)  
  Characterization & reliability testing
• (e)  
  Collection & shipment for research purpose

To perform the services mentioned above, the national legislation and market should be carefully assessed in order to check what is possible and allowed to be done. Another important point to consider is the traceability and the type and gravity of the defect of the decommissioned PV modules, in order to study the possible destination those modules can have, be it recycling, second-hand use, or disposal.

The identification of the type and gravity of the defect of the components can be obtained from a basic report on failure analysis usually done after visual inspection. The information required to accomplish the historical data is mainly found in the monitoring data to build up an operating history, as well as to be able to compare this information with the maintenance tickets. This is undoubtedly a time consuming and costly job, due to the time that must be invested to collect and analyse this historical data. However, this work can be reduced and automated with the application of BIM within a PV plant, where all the historical information of the various elements that have undergone changes or modifications can be obtained and analysed really quick. Therefore, the use of BIM would reduce time and costs, being the starting point for focusing the efforts into a 2nd life market for PV assets instead of disposing all the assets after dismantling. Once the future of the modules is decided, a tracking process would be initiated in order to identify the quantity and serial number of each dismantled element.

Furthermore, as stated in [2], to ensure the technical-economical bankability of solar PV re-use and second life solar PV, within the O&M framework and the overall solar PV value chain, it is important to:

• (a)  
  Identify the addressable 'target volume', i.e. the failed solar PV modules (or strings), the repair of which is technically feasible, and the occurrence or distribution of such failures.
• (b)  
  Determine the post-repair efficiency and/or post-revamping reliability of these modules.
• (c)  
  Integrate optimal sorting-repair-reuse and logistics procedures in the current solar PV O&M value chain, embracing circular economy business models.

2.4.1. Reuse of PV modules

2.4.2. Circular economy initiatives—status, drivers and opportunities

In general, circular economy activities (reuse, recycling, eco-design) in the PV sector in Europe exist, but they are in early stages and there are still many obstacles to overcome towards a more complete transition to a circular economy. In general, investments in research projects and experiments are occurring through the Horizon Europe funding program that help to involve many actors through the solar value chain, to establish networks and advocacy coalitions, and to increase the knowledge base on circularity in the PV sector. However, although legislation is partially in place (with some further policies under discussion), only a few organizations are exclusively dedicated to circular PV. Also, the level of interactions and collaborations between them is relatively limited. Furthermore, the market for reused and eco-designed PV modules and recycled materials is not very well-formed, with some uncertainties and informalities and low demand regarding recycled materials, as well as eco-designed and reused PV modules [10]. Nevertheless, we begin to see some innovative solutions in the industry for recycling and revalorization of raw materials, that allow to recover ultra-pure silicon and other metals (copper and silver) currently lost during the production of cells and at the end-of-life of modules (see figure 15).

Zoom In Zoom Out Reset image size

In order to further enable a circular transition in the PV sector, different policy approaches that could be taken into consideration are financial incentives (subsidies, green public procurement), standards, ecolabels, and specific targets on legislation for reuse and for recycling of valuable materials, as well as awareness-raising initiatives.

In that sense, there is a guideline for preparing modules for reuse being developed by the European research project called CIRCUSOL [17]. The guidelines contain the steps for the application of several testing and inspection methods and relabelling of modules, with the aim of ensuring sufficient PV module quality while still being economic. The guidelines will be used by the International Electrotechnical Commission Technical Committee 82 to write a technical report at first and then a standardisation on the preparation for reuse and reuse of PV modules in the near future.

Finally, the drivers, opportunities and advantages of having a more circular PV sector are related to job creation, reduction of risk of pollution, the possibility of recovering or extending the use of valuable materials, reducing costs, and finally to the consequent likelihood of becoming less dependent of the external supply chain.

2.4.3. Progressive repowering

The concept of 'progressive repowering' will allow to intervene several times during the lifetime of a plant, splitting the repowering investment in more than one shot. It is aimed at increasing energy production without requiring additional land. This concept also follows a path of circular economy, that promotes material recovery for further use in the industry.

The concept is built upon three solid foundations: advanced monitoring, smart field interventions and a decision support system, as depicted in figure 16.

Zoom In Zoom Out Reset image size

The main innovative aspects of this concept, being explored within the TRUST-PV project are the following:

• (a)  
  Maximum plant efficiency: constant increase of energy yield instead of accepting the natural component degradation
• (b)  
  Optimization of use of land: this will occur without increasing the land needed, by using the most up-to-date components, with higher efficiency than those that will be substituted.
• (c)  
  Circular economy: material recovery from disposed components for further use in the industry.