In the first article in this series, we discussed the many factors that impact energy output from solar arrays and why it’s important to improve the insight into array performance if lifetime energy yield is to match expectations.
We challenged the industry paradigm that a Performance Ratio of 0.80 is best-in-class and pointed to the emergence of advanced panel-level monitoring and optimization technologies that can help make arrays “intelligent,” enabling enhanced Operations and Maintenance (O&M) practices that can help drive heightened expectations.
The second article outlined how these new technologies can be used to diagnose array impairments with an unprecedented level of accuracy and specificity. We introduced the concept of cloud-based intelligence, removing the need for human vigilance in the monitoring process and enabling the concept of a Virtual Network Operations Center (VNOC). These advanced systems perform real-time diagnosis of faults, specify site maintenance requirements, dispatch resources when issues trigger predefined financial thresholds and confirm when faults have been corrected.
In this third and final article in this series, we will discuss the deployment options for these advanced performance optimization technologies and more specifically how they can scale to bring “intelligence” to the performance and financial management process of the world’s largest PV arrays.
As PV panel prices have fallen, balance of system (BoS) costs have been gaining more attention. Conventional thinking has been focused on a cost-per-watt-installed metric. Using this lens, any expenditure on BoS items is a capital budget evil that must be minimized. This is a short-sighted view that assumes site design and O&M practices remain static and that the general “rules-of-thumb” for asset overbuild and Performance Ratio (PR) expectation are somehow inviolable facts of life.
A far more useful lens for scrutinizing project viability is Levelized Cost of Energy (LCOE). This takes into account the lifetime energy harvest, ongoing operational costs and the cost of financing, in addition to the initial capital cost. To build a successful project and maximize its potential for return, all four of these terms should be worked, not just the initial capital cost. This alters the dynamics of considering BoS costs, with more focus being brought onto how smart BoS decisions can increase the numerator in the LCOE equation (increase energy harvest, reduce O&M costs, reduce costs of financing and insurance), far more than they impact the denominator (increased initial capital cost).
In the previous article, we explained how improved insight enabled by the precision and granularity of panel-level monitoring, coupled with enhanced O&M practices enabled by machine-based diagnostics, can raise the expectations for Performance Ratio and hence lifetime energy harvest. It can also reduce O&M costs, by automatically diagnosing impairments, pin-pointing their location, calculating the economic impact of intervention and then recommending prioritized action. Furthermore, there is growing evidence of its acceptance by the finance and insurance industry as a key element in risk mitigation, which can be a gate to getting a project financed and may reduce the costs of borrowing and insurance coverage. In Europe, large projects are now hard to finance without string-level monitoring as a minimum. There is also increasing attention from the insurance industry on technologies that can monitor panel degradation rates against warranties, increase physical asset security and increase safety by reducing the risk of fire and electric shock. From 2014, the National Electrical Code in the United States is set to outlaw the installation of any rooftop system that isn’t able to be rendered safe, down to the panel level, within seconds.
Advanced panel-level monitoring and control, such as that provided by the Clarity system from Solar Power Technologies Inc., is designed to address all of these ‘numerator enhancing actions’ in the LCOE equation.
Dynamic Optimization Technology, Design Flexibility and Smart Deployment Options
Project ROI can be enhanced further by the deployment of dynamic optimization technology, such as DC-DC converters, otherwise known as optimizers. These distributed panel-level electronic devices increase energy harvest by dynamically addressing the mismatches that arise within an array due to panel mismatch, damage, degradation, differential soiling, shade, alignment differences, variable clouding or temperature variations. According to NREL, such mismatch terms have the biggest impact on the Performance Ratio of PV arrays, amounting to a 10 percent to 50 percent downside risk to expected energy output. By using DC optimizers, which maintain every panel at maximum power point irrespective of array mismatches, typical recovery can be as much as 50 percent of the energy that would otherwise have been lost to these mismatch impairments.
In a large, well built, well maintained array, that doesn’t suffer any effects of shade or misalignment, the typical energy harvest gain from using optimizers will be in the 6 to 8 percent range over the life of the array. This is on top of the benefits that are gained by leveraging the advanced monitoring capabilities that are provided with many optimizer solutions.
Despite these benefits, large array owners have been reluctant to specify the use of optimizer technology. The potential energy harvest gain has not been viewed as sufficiently large or certain enough to overcome concerns regarding the initial capital cost for the technology. In essence, concern that the LCOE equation isn’t improved enough. However, next generation optimizers provide additional benefits in two key areas, which make this technology a compelling option in large-scale arrays.
The first area is flexibility in deployment. Fully autonomous optimizers, which can work independent of central control, or proprietary inverter technology, can be selectively deployed in areas of a large array that are known or suspected to have mismatch vulnerability. For instance, the part of an array that is subject to daily or seasonal shade; an array edge that gets excessive soiling from adjacent agricultural activity, or roads; the array center that gets less convection cooling and so exhibits higher cell temperatures that cause string drop-outs or even inverter trips in hot weather. These areas can be determined at the time of array design, or they can be highlighted by advanced monitoring technology once the array is operational. Optimizer technology can then be specified or retrofitted into areas with the maximum identified payback. With this intelligent, selective deployment, a site can enjoy the majority of the benefits of optimization technology but at a reduced cost, since optimizers are only committed to areas of maximum impact.
The second area of impact for the new generation of optimizers is in site design flexibility and the ability to actually decrease capital and labor costs. Some next generation optimizers are able to manage inverter bus voltage and operate with large-scale inverters. This enables new array design techniques, such as string stacking, delivering significant deployment cost savings to offset the cost of the optimization technology. For instance, by building longer strings (stacking) and using the optimizers to manage the bus voltage to a narrow range at the top of the specification for the central inverter, the number of strings are reduced, return wiring runs eliminated, combiner box count and installation labor reduced and the inverter runs closer to its maximum throughput at all times, in effect increasing its realized capacity by 50 percent or more. The effective cost of deploying optimization technology can now be reduced significantly below the guaranteed benefit, even in large well managed arrays.
In summary, leveraging the capabilities of next generation optimizers through selective deployment or changing site design methodology, now brings optimization to large-scale arrays as an effective Performance Ratio and ROI enhancement tool.
The three articles in this series have covered PV array performance improvement and discussed the benefits of panel-level monitoring plus next generation dynamic optimization technology as tools to be used to achieve this. However, there has been a reluctance to consider these technologies in larger scale arrays because early generation solutions didn’t scale well, and the perceived benefits were offset by the challenge of deploying and using them. However, next generation solutions now coming to market resolve these scaling challenges and open the way for these advanced technologies to be used in the largest utility-scale projects.
To accomplish monitoring and optimization in large-scale arrays, it is necessary to consider three key attributes – flexibility, scalability and intelligence. Systems must accomplish all three if they are to be a viable consideration in large-scale arrays.
Flexibility: Every large PV site is unique. The combination of its physical location, environmental conditions, asset choices, construction methodology, financing strategy, O&M philosophy and contracted agreements make it unique. As such, it requires a performance enhancement strategy that is tailored to these unique characteristics. Panel-level monitoring technology gives maximum insight into array impairments and the best possible tools for aggressive O&M strategies. DC optimizers provide automatic compensation for dynamic impairments that cannot be addressed by O&M methods. A system that allows these two performance enhancements tools to be used in combination, at varying levels of density and granularity, gives maximum flexibility to site designers to tailor performance enhancement to the unique characteristics of the site. Examples of this are monitoring solutions that can be deployed at the panel or string-level, as well as optimizers that can be selectively deployed only where the energy enhancement benefits are known to be compelling.
Scalability: The ability to physically scale the monitoring and optimization solution across the site, cost effectively and reliably. Physical scale can be enabled by a self-organizing, self-healing, secure wireless mesh network that can scale to tens of thousands of monitored or optimized devices in an array. Large physical areas can be covered, with mixed device densities, independent of array topology or hierarchy.
Intelligence: The ability to handle scale from a data perspective. Gaining unprecedented insight into array performance brings a risk of data overload. It is inconceivable that this level of data from a large-scale site will be monitored through human vigilance. What is required is machine-based intelligence that converts the wealth of data into valuable, actionable O&M information. Systems now coming to market use the precise granular data coming from the array, coupled with PV impairment models that are continually evolving, an understanding of the array hierarchy, physical location and time, plus business rules set by the site owner and O&M team to continually assess array performance, diagnose faults and present the site team with a prioritized action plan. Systems like this make an array “intelligent” — able to inform the owner of exactly what, where and when human intervention is necessary or advisable to keep the array performing at optimal levels.
The solar market is still in its infancy and models of best practices are still evolving. Many companies have made tremendous strides in material cost reduction, but this path to reducing LCOE eventually produces diminishing returns. The next frontier is to make significant progress in raising the performance entitlement and expectation from installed solar assets. With the emerging generation of combined monitoring and optimization systems, site designers and owners now can deploy dynamic optimization techniques, precision panel-level monitoring, machine-based intelligence and advanced O&M methodologies to raise PR entitlement and enhance the actual performance in large scale arrays.