In almost every corner of the solar industry, key players are struggling to standardize design, construction, operations and maintenance, regulation and permitting. As an industry grows and matures, it must go through multiple stages of process standardization.
In the auto industry, this was exemplified by the production line of the Model T, then Toyota (followed later by Ford, Chrysler, GM and others) adopting W. Edward Deming’s total quality framework, and in modern times by the industry standardizing on OBD-2 computer systems for diagnostics.
Solar manufacturing, installation, operation and maintenance comprise a huge number of processes, many of which don’t follow modern manufacturing and process engineering standards. The solar industry has used many of the lessons of mass-manufacturing science in order to produce cells and panels more effectively, but it has not yet ported these lessons to the more effective installation of solar.
In fact, in solar installation, operations and maintenance, the industry deploys some of the same metrics to determine the outcome of installation capacity that were common in a 1920s-1940s automotive factory. These methods revolve around reworking failed components and assemblies, or rolling a truck to re-work a site in the solar example. Often, this inspection occurs at the intake for solar asset management, or system activation in residential systems.
Diagram of a Rework-on-Failure Based Quality Assurance Framework
Engaging with the industry
Over the past few months, the team at Sustainabilist has been conducting interviews with experts who engage with quality assurance and quality control in the solar energy industry. These thought leaders include Jigar Shah (Generate Capital), Cody Oram (Vivant), Rudy Saporite & Richard Lawrence (IBTS), Marvin Harmon (Harmon Engineering), K.C. Radford (Radian Generation), Daniel Roesler (UtilityAPI) and Sandra Kwak (10Power), among others. We’ve heard from those who develop residential, utility-scale and international solar projects on what quality control looks like today and where it could be tomorrow.
From these interviews, several initial trends have become apparent:
We are at the very early stages of focusing on process efficiency, with outcomes currently viewed as a measurement point.
Larger residential players are just in the beginning phases of viewing QA/QC as a driver of savings, instead of a cost center.
Consistent technological advancements in panels and inverters lead to high retraining costs and an inability to standardize processes among technology with different capabilities and compatibilities.
In the commercial and utility scale of solar, regular engineering rework is the norm.
Engineering, procurement and construction (EPC) companies have the responsibility of meeting project specifications, but each EPC uses a different process instead of an industrywide standard.
Based on our interviews and experience in this industry, we believe there are significant opportunities to improve QA/QC processes by implementing approaches based on quality science, even among the companies that have begun to put effort into them. QA/QC could be dramatically improved by leveraging the data that is already being collected. This data collection and analysis could initiate a feedback loop, lowering costs and raising the quality of solar installations. This feedback loop has been implemented in almost every mass-production operation since the 1980s, in a similar and scientific fashion.
When a solar company sells a project to its customer and then to its financier, there is currently little incentive to alert these stakeholders of areas that could benefit from improvement. Instead, each person along the solar system’s life has an idea of what a well-functioning system looks like, and as long as those who have paid for systems are satisfied, installers and engineering firms seldom share technical data on the efficiency of a system after install. This lack of process orientation causes every player along the value chain to lose a big chunk of the pie.
Lighting the way forward
To bring solar into its own process improvement revolution, the industry must begin by embracing process-based quality control. By centering our efforts around processes instead of outcomes, we will be able to provide more feedback to manufacturers and installers on common failures, incompatible technologies and system performance, and we can offer higher-functioning systems with better returns to financiers, while cutting rework and O&M costs.
Closing the cycle won’t just benefit customers. Data insights including failure mode analysis, installation process measurements and contractor assessments from the installation and performance of solar systems very rarely return to the manufacturing firm. This data can be used to better diagnose and prevent future electrical or mechanical failures.
If the industry is to succeed in becoming a mass-manufacturing operation for the installation and operation of solar, we all need to switch to a manufacturing process-oriented mindset. This methodology has lowered costs, decreased build times, and increased quality for every mass-production industry that has adopted it, bar none. Aligning folks from all stages of the space on the definition of solar as a manufacturing process needs to be the first priority. The Solar Energy Industries Association has begun to address this subject by hosting speakers at their events, including at the upcoming Codes and Standards Symposium where I’ll be speaking more about this subject on March 29.
Standardizing quality assurance and control can cut costs and build trust across the many stakeholders from design to inspection. Most critically, a robust quality control program can cut high labor costs for O&M costs. QA/QC and process improvements can substantially reduce the time and subjectiveness of inspections by using pre-existing data to identify common errors as they are specific to certain technologies, scales and geographies. These improvements will lead to overall high consumer satisfaction and more valuable assets as systems have less variability in their function and more efficient performance.
In thinking about specific examples of how this could be used in the solar industry, we refer to the diagram of process-centered production below. This is standard in high-tech manufacturing, and avoids failure-driven rework illustrated above. In solar, we could implement this in the many ways, including the following:
Contractors should receive data-driven education in order to improve their process capability. Across the industry, information that is being gathered on contractor performance should be used to design tools and methods that make it easier for contractors to succeed, instead of just grading them by their work. One company that we spoke to claimed that just the decrease in roof damage on customer sites more than paid for their QA/QC program based on giving contractors feedback.
In commercial and utility scale deployments, EPCs should be gathering a statistical sampling of quality of installation before thousands of panels are installed. If, at 5 percent of installed capacity, measurements were taken to ensure that specifications are being followed and processes were adjusted, that would be much closer to a manufacturing framework.
Manufacturers could use observational data to start to streamline repairs, and this information could be factored into future equipment design, as well as instructional materials for training maintenance technicians. Several of the larger companies we spoke to have a communication line with manufacturers, but as yet, there is no systematic method of gathering this data and using it for improvement.
A Modern, Process-Oriented Manufacturing Cycle
Given the state of solar, the transition from a one-off construction mindset to a mass-production methodology is inevitable -- and imminent. Across innumerable industries, the story is always the same: Early adopters reap the greatest benefits, and the latecomers lose market share. By adapting to the rigorous statistical standardization practices every other modern industry uses, we can accelerate the deployment of solar, increase the profits of industry players, reduce their downside risk, and pull solar into a future of mass-production and ubiquitous deployment.
Jason S. Trager, Ph.D. is the managing partner at Sustainabilist, a firm specializing in using data science, process engineering, machine learning, and artificial intelligence tools to increase the profitability and effectiveness of sustainability oriented businesses. He earned his Ph.D. from UC Berkeley with a focus on the application of mass-production methodologies to energy efficiency in buildings. Jason can be reached at firstname.lastname@example.org.