Category: Solar

There are now more than 5 million solar installations across the United States alone, and that figure may triple by 2034. Globally, solar PV capacity surpassed 1,200 gigawatts in 2024, with another 655 GW of new installations expected by the end of 2025. That is a staggering amount of infrastructure to track, inspect, and maintain. 

For much of the industry’s history, operations and maintenance (O&M) teams have relied on scheduled walkthroughs, paper-based records, and gut instinct. Technicians would physically walk row after row of panels, clipboard in hand, hoping to catch problems before they snowballed. It worked barely when solar farms were smaller. But it simply doesn’t scale. 

Artificial intelligence is changing that. From drone-based serial number scanning to machine learning models that predict inverter failure weeks in advance, AI is quietly becoming the backbone of modern solar asset management. To help industry professionals understand its practical application, this development is not a distant future scenario. It is happening today on utility-scale farms across Europe, Australia, and North America, and the economics are compelling enough that smaller operators are paying close attention. 

The Core Problem: Scale, Complexity, and Hidden Losses 

Before exploring what, AI can do, it helps to understand what solar operators are actually up against. 

A utility-scale solar farm might house 200,000 individual panels spread across hundreds of acres. Each of those panels can develop its own unique problems of microcracks from hail, soiling from bird droppings, hotspots from cell degradation, or shading from a branch that grew into a sight line. Any single underperforming panel might lose only a fraction of a percent of total output, but multiply that across thousands of modules, and the financial impact becomes significant quickly. 

Traditional inspection methods are slow and expensive. Unplanned downtime across industrial operations costs an estimated $50 billion annually, and solar is not immune to it. Manual inspection crews cannot feasibly survey an entire large-scale installation more than once or twice per year, which means problems often go undetected for months. 

Reactive maintenance, fixing things after they break, and even scheduled preventive maintenance, both leave money on the table. The emerging paradigm, driven by AI, is predictive maintenance: identifying anomalies before they become failures, dispatching crews only when and where needed, and keeping panels producing at peak capacity throughout their service life. 

AI Solar Panel Identification: Knowing What You Have 

One of the most overlooked challenges in solar asset management is surprisingly basic: operators often don’t have accurate records of the installed panels, specifically by type and location. Addressing this can give asset managers a sense of control and confidence in their operations. 

This scenario matters more than it might seem. If a panel needs a warranty service, you need its serial number. If a string of panels is underperforming, you need to know exactly which units are involved. If a module exhibits early degradation, you want to cross-reference its manufacturing batch against similar failures elsewhere in your portfolio. 

The Manual Tracking Problem 

Historically, installers manually logged solar panel serial numbers during installation. Workers would scan barcodes or transcribe alphanumeric strings one by one, then match them to position maps in a process prone to transcription errors, sequence mix-ups, and gaps in documentation. One drone-based inspection company found a 30% error rate in customer documentation created through manual tracking methods. That’s not an edge case; it’s an industry-wide problem. 

Drone-Based AI Serial Number Scanning 

The solution that’s gaining ground fast is drone-mounted computer vision. Drones equipped with high-resolution cameras fly in a systematic grid pattern over solar arrays, capturing images of every module. AI-powered image recognition software then automatically detects and extracts serial numbers from the panel labels, linking each identifier to its GPS-mapped position within the asset management platform. However, challenges such as weather conditions, label degradation, and initial setup costs can impact effectiveness, which industry professionals should consider when planning AI integration. 

Current drone scanning systems achieve read rates of 85–95% even under real-world conditions where labels may be affected by weather, soiling, or partial shading. The remaining unreadable panels are flagged for targeted manual verification, dramatically reducing the total labor required. What once took weeks of ground-level work can now be completed in a single day, even for a large installation. 

Technology is moving quickly. Several large independent power producers (IPPs) and asset managers already use drone scanning for annual portfolio inventory audits. RFID tags, either attached to or embedded in panel frames, are becoming increasingly common among major manufacturers, adding another layer of machine-readable identification alongside traditional solar barcodes. 

The Regulatory Tailwind 

The European Commission’s Digital Product Passport initiative is adding urgency to this shift. Forthcoming regulations will mandate that every solar panel carries machine-readable identifiers linked directly to lifecycle documentation covering everything from manufacturing provenance to end-of-life recycling records. Industry players who adapt now will feel proactive and ahead of compliance requirements. 

For asset managers, getting your serial number tracking right now is not just an operational nicety; it’s increasingly a compliance requirement. 

Computer Vision and AI-Powered Fault Detection 

Beyond identification, AI is reshaping how operators find and diagnose panel defects. The traditional approach, walking the rows, eyeballing panels, hoping to spot discoloration or cracks, is giving way to a far more sophisticated system. 

Thermal Imaging + Machine Learning 

The most powerful tool in the AI inspection arsenal is thermal imaging combined with machine learning analysis. Solar panels generate heat unevenly when they malfunction. A cracked cell, a failed bypass diode, or a hotspot caused by soiling will all produce characteristic thermal signatures that differ from healthy panels. Infrared cameras mounted on drones or fixed monitoring systems capture these heat patterns across entire arrays. 

Machine learning algorithms trained on large defect datasets can then classify what they see with remarkable precision. They distinguish between hotspots caused by soiling (which need cleaning) versus those caused by cell degradation (which may warrant replacement). They can flag panels exhibiting early warning signs before performance actually drops. 

In one documented case study, drone-based thermal inspections helped a solar operator avoid an estimated $296,000 in annual revenue loss by catching degradation early enough to intervene. 

Soiling Detection and Cleaning Optimization 

Dust and biological soiling particularly bird droppings are among the most persistent performance killers in solar O&M. A heavily soiled panel can lose 30% or more of its energy output, and bird droppings create especially damaging hotspot conditions because of their resistance to natural cleaning by rain. 

AI-assisted soiling detection using drone-captured RGB imagery is maturing into a commercial solution. Researchers have developed custom architectures, such as SDS-YOLO (Soiling Detection System, based on the YOLOv5 framework), specifically trained to identify and localize soiling patterns at the module level in aerial images. The system distinguishes between dust which may be manageable without immediate intervention and bird droppings, which warrant urgent cleaning. 

By knowing exactly which panels are soiled and how severely, operators can move away from scheduled blanket cleaning toward targeted, data-driven cleaning schedules. The energy savings and reduced water usage are significant at scale. 

End-to-End AI Inspection Platforms 

Companies like Aispect.ai, Raptor Maps, and SmartHelio are building end-to-end platforms that ingest drone imagery, apply deep learning defect detection, and surface actionable work orders for maintenance crews all within a single interface. Rather than having a technician manually interpret thermal images, the AI performs classification and priority ranking, directing human attention to the issues that matter most. 

These platforms are increasingly integrating robotics as well. AI-controlled autonomous ground robots are being developed for cleaning and targeted maintenance, guided by the same AI systems that detect the problems in the first place. Research from 2025 demonstrated robotic cleaning systems that achieved 91.3% cleaning efficiency, reducing dust density dramatically and restoring up to 31% of energy output on heavily soiled panels. 

Predictive Maintenance: From Reactive to Proactive Solar O&M 

The third major frontier of AI in solar asset management is predictive maintenance using machine learning to anticipate equipment failures before they happen. 

Inverter Diagnostics 

Inverters are the workhorses of any solar installation, and they’re among the most failure-prone components. A failed inverter can take an entire string of panels offline, and the losses accumulate with every hour of downtime. Machine learning algorithms are being trained on inverter error logs, operational telemetry, and environmental data to predict fault conditions days or even weeks before failure occurs. 

This approach matters because it changes how O&M teams operate. Instead of dispatching a technician in response to an alarm, operators can schedule preventive service during planned maintenance windows reducing emergency call-outs, extending equipment life, and keeping energy production consistent. 

Degradation Tracking and Module Performance Analysis 

All solar panels degrade over time. Most manufacturers warrant panels against dropping below 80% of rated output over their 25-year lifespan. But degradation is not uniform. Some panels degrade faster than others due to manufacturing variation, installation conditions, or site-specific environmental stressors. 

AI applied to time-series performance data can track individual panel degradation rates, identify outliers that are declining faster than expected, and help asset managers make informed replacement decisions. More sophisticated analysis can even identify why a panel is degrading distinguishing between shading from a nearby obstruction, cell degradation, or connection issues providing context that a simple performance alert cannot. 

Research projects aggregating data from thousands of solar plants across multiple companies are building the training datasets needed to make these models genuinely predictive rather than merely descriptive. The result is a shift from looking at historical averages to getting forward-looking insight into which panels, strings, or inverters are likely to need attention next. 

AI-Powered Weather and Output Forecasting 

Predictive maintenance extends beyond hardware failure. AI is also transforming how operators forecast energy output a capability that matters for grid management, energy trading, and financial planning. 

Hybrid models combining physics-based simulations with machine learning trained on satellite imagery, sky-camera data, and historical plant telemetry can now predict solar irradiance and energy generation up to 48 hours ahead with significantly better accuracy than traditional weather forecasting methods. Studies have found that AI forecasting models reduce forecast error by more than 27% compared to conventional numerical weather prediction approaches. 

For asset managers running large portfolios, that forecasting accuracy translates directly into better contract performance, reduced balancing costs, and stronger relationships with offtake partners. 

The Business Case: What the Numbers Say 

The ROI argument for AI in solar asset management is becoming hard to ignore: 

• Predictive maintenance can increase productivity by 25%, reduce equipment breakdowns by 70%, and cut maintenance costs by 25%, according to industry analysis. 

• AI-powered inspection and fault detection can reduce maintenance costs by up to 40% at scale. 

• Automated drone-based serial number audits correct documentation error rates that manual processes leave as high as 30% errors that directly affect warranty claim processing and portfolio valuations. 

• Early fault detection visibility to prevent hundreds of thousands of dollars in annual revenue loss at individual sites. 

The global investment signal is clear, too. The energy sector has seen more than $13 billion invested in AI technologies, with over 50 identified applications across the solar value chain. The technology is no longer experimental it’s becoming a competitive differentiator. 

Who’s Building the Future of Solar AI 

The ecosystem developing AI tools for solar asset management includes a mix of established energy software companies and well-funded newcomers: 

SmartHelio (an EPFL spin-off) has built a physics-informed AI platform for predictive analytics and automated fault detection across utility, commercial, and residential solar installations without requiring additional hardware. 

FairFleet has made a name for itself specifically in drone-based PV serial number scanning and asset documentation, operating across more than 70 countries and integrating results directly into asset management platforms. 

Proximal Energy has partnered with major developers, including Excelsior Energy Capital, to deliver AI-powered asset management for utility-scale solar, including an AI “agent” approach to performance optimization. 

Aispect.ai, launched in early 2025, offers computer vision inspection tools that identify cracks, soiling, and misalignment from drone imagery, with plans to expand the technology into agriculture, security, and manufacturing. 

Raptor Maps has built one of the most established drone-inspection platforms in the industry, with documented partnerships with major drone manufacturers and a track record on large installations. 

What This Means for Solar Asset Managers Today 

If you manage a solar portfolio whether a single commercial rooftop installation or a multi-site utility portfolio the practical takeaways from the AI revolution in this space are fairly concrete: 

Start with data integrity. AI is only as good as the data it learns from. If your panel-level serial number records are incomplete or inaccurate, addressing that foundation potentially with AI-assisted drone scanning unlocks every downstream capability. 

Evaluate predictive maintenance platforms. The gap between reactive and predictive O&M is wide, and it’s widening financially every year. Platforms that integrate real-time SCADA telemetry with AI anomaly detection are no longer cost-prohibitive for mid-sized operators. 

Think about the regulatory horizon. Digital Product Passport requirements and other traceability mandates are on the way, particularly in Europe. Getting ahead of these requirements now is cheaper and less disruptive than scrambling to comply after the fact. 

Don’t underestimate thermal inspection ROI. A single drone thermal survey can pay for itself multiple times over by identifying revenue-eroding defects that manual inspections would miss for months. 

Looking Ahead 

The solar industry is on an extraordinary growth trajectory, and the pressure on asset managers to do more with less will only intensify. AI doesn’t replace experienced O&M professionals it makes them dramatically more effective, directing their expertise toward decisions that genuinely require human judgment while automating the time-consuming work of monitoring, identification, and preliminary diagnosis. 

The operators who embrace AI-powered solar panel identification and asset management now are building a data advantage that will compound over time. As models train on more plant data, as drone technology improves, and as regulatory requirements for traceability tighten, the gap between AI-enabled and traditional O&M approaches will grow wider. 

The question is no longer whether AI will transform solar operations. It already is. The question is how quickly the rest of the industry will catch up. 

If you walk into a massive solar panel factory or a place where they’re putting solar panels on the utility grid, you’ll see technicians working on solar panel installations doing the same thing over and over again: Looking at each solar panel and using their hands to hold the solar panel up in front of them until they get to the correct position in the installation site; squinting at the small type on the solar panel; typing in the long code into a tablet or paper at the installation site. This is dreadfully slow, painfully inaccurate (mostly due to human error), and completely contrary to the speed with which the solar industry operatestoday. 

The global growth of the solar market is expected to be $300 billion by 2030, which means the demand for solar panel installations, tracking, and servicing will continue to grow rapidly. Therefore, the days when automation for solar panel serial number scanning was a luxury are gone. As a result, manufacturers, EPCs (engineering, procurement, and construction), and O&M (operation and maintenance) companies need to implement solar panel serial number scanning in an automated fashion to maintain their competitive advantage.

The guide is focused on serial number tracking of solar panels; the following information should provide you with a good understanding of what you need to know: why solar panel serial number tracking is necessary, where manual processes will fail, what types of technology work best in solar panel environments, and how you can create an end-to-end, automated, field-tested process for scanning and tracking solar panel serial numbers. 

Managing a utility-scale project comprising tens of thousands of solar panels from various manufacturers makes it a daunting task to manage accurate and up-to-date data.  

The following describes the consequences of mistakes made while tracking serial numbers:  

• Denial of warranty claims. Because there is no verification of the installation record against the serial number, manufacturers will frequently deny warranty claims. Repairs of one defective solar panel that is disputed warranty data will cost the repair company more in administrative time than the original repair. 
• Product recalls are exceedingly challenging to execute. When manufacturers issue recalls for defective products, it is difficult for solar sites with no records of the serial number to identify affected solar panels among many sites across the country. 
• Increased compliance risk. The EU, US, and AU all have regulations in place that require compliance with recycling and end of life obligations on a solar panels and require traceable records to the solar panels that you own. Commonly, records that are kept manually will not hold up to an audit. 
• Reduced operational efficiency for owners and operators. Many Solar providers provide their fleet operators access to view historical records (including the installation date, inverter association, cleaning schedule, and past performance of solar panels), that allows their field technicians to have everything that they need to work efficiently in just a few clicks on the computer. 
• Increased insurance claim costs. Insurers of large portfolios of solar energy are beginning to take into consideration how well their assets are being managed by comparing quality of records of how their investments are being tracked through the use of serial number traceability and whether or not there are any issues that could cause an insurance claim. 

Where Manual Serial Number Scans Are Not Successful

The pace of manual serial number input, or even one-by-one scanning via a handheld unit, is slow. This can result in failure on multiple levels as time passes. 

Errors in Data Entry  

Studies on manual data entry, primarily in manufacturing, show an error rate from 1–4%. For instance, assuming the installation of 50,000 panels, this would result in an estimated 500–2,000 serial numbers being recorded incorrectly creating potential problems with warranties, compliance and/or maintenance.   

Degradation of Barcodes in the Field  

Solar panels are built to last 25–30 years and are subjected to UV, thermal cycling, moisture and physical wear and tear this all shows up as degradation of printed barcodes or QR codes on the labels of the panel. Laser etched codes on the glass or aluminum frames will last much longer than the adhesive labels that many of the older panels use and the adhesive labels are typically no longer readable past 5-7 years after being installed.   

Bottleneck in Workflow  

Larger utility scale installations generally have the scanning group working independently of the installation. If the scanning group has scheduling overlaps, the slower-running scannings can cause delays to the entire project, if they are too far behind, there is great temptation for the teams to skip records or perform batch recording based on the delivery manifests thus compromising the panel level accuracy completely.  

Data Silos Between Manufacturing and Field  

Manufacturers record serial numbers with their internal ERP systems. 

Contractors generally log their work into their construction management software. Owners and operators often utilize different systems than the contractors they hired. When these systems are disconnected, the same serial number for a solar panel is entered multiple times by various people and creates multiple opportunities for error and inconsistency. 

Solar Serial Number Scanning Solutions: What Is Effective  

One solar panel serial number scanning solution may not be as ideal as others because the solar environment is challenging for different types of barcode scanning. Below is a comparison of the different serial number scanning solutions: 

1. Fixed Industrial Barcode Scanners (Manufacturing Plants)  

Fixed-mounted laser or imager scanners mounted on conveyor lines or end-of-line test stations at a manufacturing plant. They can read 1D barcodes, DataMatrix codes, QR codes, and laser-etched marks at production speeds without any manual intervention.  

Recommended For: 

High-volume solar panel production operations using a fixed point. They typically have a read rate greater than 99.9% when they are lined up properly and maintained correctly. 

2. Handheld Bluetooth Barcode Scanners  

Handheld Bluetooth barcode scanners connected to tablets or cellular phones are the primary option for installing and commissioning solar panels in the field. When used properly, the new generation Bluetooth handheld barcode scanners can read 1D barcodes, QR codes, and DataMatrix codes with limited restrictions in direct sunlight. The most important difference in handheld Bluetooth barcode scanners is the software.  

Commodity hardware is used for the scanner; the critical component of scanned asset data flowing into your asset management system in real-time. 

 Best suited for: 

Panel-level audits or fault finding for O&M teams in small to medium-sized installations should be considered; low upfront cost and flexibility for operational workflow integration.

3. Drone Aerial Scanning

Using drone-mounted camera systems with computer vision capabilities enables off-site scanning of panel serial numbers or asset tags across an entire array. Using drone aerial scanning is most efficient for large ground-mounted installations because ground walking through every row would be very time-consuming. The current scanning systems right now can achieve a read rate of 85%-95% against a good label condition. Therefore, scanned codes that are degraded due to soil or shading require manual verification. 

  Emerging Technology: Rapidly maturing, drone scanning is currently in use by numerous large Independent Power Producers (IPP) and asset managers for their annual portfolio inventory audits. By years end, you can expect scanning systems to demonstrate significant improvements with read rates and cost performance.

4. RFID Tracking of Panels

Several large panel manufacturers and large solar developers are beginning to utilize RFID tags either attached or embedded into the frames of the panels. RFID technology does not require direct line of sight, can be scanned even under moderate soiled panels, and many panels can be scanned simultaneously with batch scanning techniques. Cons of RFID technology are the higher price you pay for each panel and the expense associated with building the required reader infrastructure to use RFID tag technology. 

5. Next Generation Computer Vision + AI 

The development of computer vision systems that can be utilized in the field to recognize solar panels using images has matured from a research study to a commercial deployment activity. These computer vision systems will be installed in various locations and will use a camera technology already available (e.g., standard cameras mounted to tracker rows, drones, or maintenance vehicles) to provide asset management platforms with recognised identifiers associated to the associated panel image. 

Technology	Best Environment	Typical Read Rate
Fixed industrial scanner	Manufacturing / factory	>99.9%
Handheld Bluetooth scanner	Field installation / O&M	97–99%
Drone + computer vision	Large ground-mount arrays	85–95%
RFID reader	Any (no line-of-sight)	95–99%
AI computer vision (OCR)	Arrays, trackers, drones	88–97% (maturing)

 Automation of the Solar Panel Scanning Process 

While scanning hardware is an important piece of this puzzle, automation requires the development of a connected workflow to allow scanned data to be transferred directly from the field to systems that require the data without any human intervention. 

Step 1: Establish a Canonical Serial Number Standard 

Before implementing a new scanning solution, it is critical to establish a definitive standard for the serial number used for each panel. Although this may seem like an obvious step, many organisations either have multiple incomplete databases that contradict each other or have no established standard. Therefore, it is important to clearly define the following items: 

• Master Record from whence the data originates (usually the asset management system, or CMMS)
  •    How manufacturer serial numbers correspond to the company’s internal asset number if it exists
  •    Any additional data captured at scan time, including GPS coordinates, installer ID, string assignment, image of the panel at the time it was installed, and time of image captured. 

Step 2: Choose Your Integration Architecture 

Your scanning system should connect to the databases of your backend platforms through APIs that are using Basic Authentication, most newer asset management systems (Salesforce, SAP PM, IBM Maximo, Infor EAM, SolarEdge or O&M systems) will have Web API’s (REST) available that can receive serialized asset record data The questions you need to ask to get started are:  

– Will scanned data be synced in real time (i.e., as it is scanned) or in batch (i.e., after a technician has finished their shift)? 

– What happens if the technician is scanning from a remote location and there is no internet connectivity? (The answer is to build an offline-first mobile application that queues scans and syncs them later) 

– How will your system validate serial numbers against the manufacturer’s product database (i.e., to catch errors right away) when a technician scans a serial number?  

Step 3: Use Real-time Data Validation 

The biggest benefit of an automated scanning database versus a manual database is being able to validate the data that is scanned as soon as it is scanned. When a technician scans the serial number of a panel, its validation will be:   

– Is this serial number in the manufacturer’s product database? (If you have an API integration or have uploaded the manufacturer’s product manifest) 

– Has this serial number already been scanned and logged at another site? (Working with criminals and catching them in the act) 

– Does the spec of this panel match the expectation for this position in the string (example: wattage, voltage class, cell technology)?   

Step 4: Record Commissioning Data 

Initiate the capture of a complete commissioning record, rather than just the serial number, using the scanning event you just completed. A mobile scanning program should prompt the installer to verify that the following items were performed:  

• GPS location of the installation or manual row/string location of the installation (for structured installations) 
• Inverter/string assignments 
• Completion of initial visual inspection 
• Photographic documentation of the front of the label and location of installation 
• Installer certification and signature.

Step 5: Close the Loop with the Manufacturer’s and Warranty System 

The most sustainable use of automation is when the serial number documentation can be automatically sent back to the manufacturer/insurer. Work with your panel suppliers to develop the following:  

• Automatic warranty registration via API at the time of commissioning scan 
• Batch exports of serial number confirmation per manufacturer’s required format 
• Recall notification triggers to immediately identify impacted assets within your portfolio.

Panel Label Durability-Field Perspective  

A poor scanning process prevents the code on the panel from being scanned. Label durability is often overlooked when discussing solar asset management; therefore, it should be considered when selecting the panels and maintaining existing fleets of panels.   

What Deteriorates Labels  

• Adhesive labels are very vulnerable to UV rays; laminated polyester labels can endure longer periods of UV rays.
• Temperature changes cause the edges of labels to curl, and the labels themselves to crack due to thermal movement.
• The chemical attacks on labels occur from a variety of sources such as cleaning products, bird droppings, and pollutants.
• Physical wear and tear causes damage to labels through mechanical means such as the use of cleaning tools, hailball impact or when panels move on tracking systems.  

Better Options:  

• Laser etched codes on aluminium frames are very durable and will not fail as they do not use any adhesive material, thus lasting the entire life of the solar module.
• Ceramic Ink printed code on the glass back of a solar module is resistant to ultraviolet light and chemical attacks, and is less susceptible to abrasion than the frame mounted labels.
• RFID tags in the junction box are not exposed to environmental elements and can be scanned without a line of site.  

Pro Tip:  

If you are performing an audit on an existing fleet of solar modules with degraded labels, plan to relabel when you next perform scheduled maintenance rather than accepting gaps in read results. A lost serial number during warranty disputes between manufacturers can cost significantly more than a relabeling program.   

The ROI Case for the Automation of Solar Serial Number Scanning 

Developing the business case for automating serial number scanning becomes evident when calculating the real cost of not automating it.  Here’s an outline of how to calculate your return on investment.  

Decrease in Labour Cost  

Approximately 45–90 minutes of the labour required to scan and enter the data for 100 panels is used up to complete an installation that has been performed to industry standards. 

The amount of time that is required for automated scanning with real-time integration and reporting has been reduced from previously up to 90 minutes per 100 panels to only 10-15 minutes. This corresponds to an approximate saving of 150 hours of person time on a 10 MW project consisting of approximately 25,000 panels, when calculated using the average commercial electrician or specialty commissioning rates. This also means that there will be significant dollar savings before accounting for the costs of correcting errors.  

Verified Warranty Recovery 

Solar operators that utilize automated scanning and have the ability to track serial numbers and provide prior to installation commissioning documentation as part of the warranty recovery process have reported significantly improved warranty claim acceptance rates. Some have reported increases in acceptance from 60-70% to 90%+ when validation with time and GPS stamps is able to be provided as part of the warranty recovery process. The difference in warranty recovery between utilization of verification and not will be in the hundreds of thousands of dollars over the panel warranty period for large portfolios. 

Starting the Implementation Process – Getting Started   

Below are some relevant steps that you can take to start the automation of your solar panel scanning program of serial numbers:   

• Audit the quality of your current serial number data (what your current quality is will help guide you in the deployment of new systems) 
• Mapping of the serial number data flow from manufacturing, through delivery to installation, commissioning, and continuous data capture
• Choose suitable scanning devices based on how you will be primarily using the devices (i.e. assembly line, installation in the field, and/or periodic inspecting). 
• Test mobile scanning applications that can function while offline with an API connection to your asset management software. 
• Obtain access to manufacture’s API for product validation or to obtain structured data export files for validation use. 
• Run a pilot project or use a pilot location, the results will be used as a benchmark for your overall project success, including; read/scan rates, error rates, and man-hour rates. 
• Establish label quality criteria to be used for new panel procurement. 
• Plan a re-labeling program for all your existing fleet that have degraded bar coding. 
• Teach your field installers on how to properly perform scanning activities based on how they are trained to use that equipment. 
• Define key performance indicators, target read/scan rate of +99%; target error rate of <0.1%, target commissioning complete at time of new inventory. 

Common Pitfalls: 

1. Treating scanning as an IT project instead of an operations project. If the installer or technician does not perform work correctly, then the system will not work. The scanning process must be considered in conjunction with installers and technicians so that it can be optimised for their operations. 
2. Underestimating the need for offline scanning capabilities. Remote solar sites often do not have good or consistent mobile network connectivity. Therefore, when selecting a mobile scanning solution, it is important that it has the capability of providing offline scanning capabilities seamlessly (i.e. queuing the records to be sync’d once the system is back “online”). 
3. Disassociating the panel manufacturers’ relationship to the data. The effectiveness of your scanning system is only as strong as your supplier’s ability to provide you with accuracy. Therefore, it is very important to work to develop those relationships early. 

Maximizing for scan speeds instead of scanning quality will result in a slower scan at a rate to capture the GPS position, photograph, and string assignment, for this data will provide a much larger total value over an asset’s expected usable life of 25 years. As compared to the value provided through a single scan of only a serial number, the slower scan will produce far more value. 

A failure to have a plan that addresses degradation on existing labels in the fleet will become necessary for asset owners to have a remediation plan to keep the fleet operational using automated scanning methods. All operators will be needing the work needed to remediate existing assets that use degraded labels in order to move forward with the automated scanning technologies.   

As the way forward becomes clear, the expansion of the solar energy sector and the increased complexity of the product offering through a combination of increasing the number of solar technology installations with the expansion of portfolios will be dependent on the development of a digital identity for every solar panel. Digital identities will enable:   

• analytics related to performance,
• predictive maintenance,
• warranty management,
• recycling compliance, and
• secondary market transactions.  

The movement toward solar panel Digital Identities is being rapidly accelerated by:   

• The introduction of EU Battery and Solar Regulations; and
• The Digital Product Passport initiative of the European Commission. Both the regulations and the initiative will mandate that every solar panel is issued with machine-readable identifiers and a direct link to lifecycle documentation.

Provenance tracking based on blockchain technology: An increasing number of manufacturers and industry consortia are now implementing pilot programs utilizing blockchain technology to generate immutable, shared records of a panel’s identity and historical data that can be accessed by all participants in the value chain.    

 First, using Artificial Intelligence (AI)-powered anomaly detection, asset management platforms are now applying panel-level serial number data in conjunction with performance monitoring to identifyindividual panels that are beginning to exhibit early signs of degradation, allowing operators to proactively implement targeted maintenance prior to equipment failures.    

In summary:

Automating the scanning of serial numbers associated with solar panels creates significant enhancements throughout their life cycles by enabling the growth of faster and more accuratecommissioning, streamlining warranty management, improving operational and maintenance services (O&M), and ultimately meeting recycling compliance standards. In addition, the technology necessary to improve upon solar asset management processes is both mature and accessible. The major issue that exists is developing an effective integration strategy for incorporating the technology into cohesive workflows, while also maintaining the data integrity required for long-term use.  

The solar industry is experiencing an inflection point. Projects continue to increase in size with respect to the number of solar panels required (which in turn contributes to overall cost), but also continue to provide less profit margin than they have in the past. Therefore, those operators or manufacturers who establish an automated, robust system of data collection for solar asset management will have a competitive edge as the solar market continues to mature; not only with respect to increased operational efficiency but also with respect to providing greater financial returns than ever before to both their investors and customers. 

It’s important to make an immediate transition from using manual scans and spreadsheets for record keeping to something more advanced like an automated solution, otherwise you may experience disputes on your first warranty claim once the project is completed.