Making the Case for Quality
Wind Power Company Gets to the
Root of an Icy Issue
• A root cause analysis project saved Clipper Windpower $1 million in lost revenue.
• By identifying the root causes of turbine failure during inclement weather, Clipper increased customer satisfaction through improved turbine availability.
• This project also supported a key supplier’s quality process, as Clipper’s team helped redesign and test an improved anemometer.
• Team members mastered quality tools and strategies, preparing them for future improvement projects.
At a Glance . . . With a steady breeze, a wind turbine—stretching 262 feet high—majestically turns its three power- ful blades, generating enough clean, renewable electricity to power 750 homes for a 24-hour period. When the breeze turns into a driving wind combined with ice, freezing rain, snow, and even freezing fog, the turbine’s anemometer, which measures wind speed and force, can freeze up and result in costly downtime for wind power compa- nies such as Clipper Windpower.
About Clipper Windpower
Headquartered in Carpinteria, CA, Clipper Windpower is a rapidly growing company engaged in wind energy technology, wind turbine manu- facturing, and wind project development. Clipper employs more than 850 people in the United States, Denmark, and the United Kingdom. At the heart of its manufacturing operations is an ISO 9001-certified manufacturing and assembly facil- ity that began operations in Cedar Rapids, IA, in March 2006.
Turning to Quality to Improve Turbine Availability
As Clipper’s first wind turbines came online in northwest Iowa, western Illinois, and western New York near Buffalo, the winter of 2007-08 hit hard and fast with freezing rain and fog causing anemometer units to fail. While the towers continued to run, without the anemometers there was no guidance on which direction to move the 153-foot blades to harness the wind most effectively. Clipper initially tried to address the problem throu gh software upgrades, but soon additional anemometers began to freeze, compounding the problem and impacting turbine availability. Without a quick solution available and with growing numbers of anemometers impacted each day, Clipper initiated a root cause analysis (RCA), an integral part of the Six Sigma define, measure, analyze, improve, and control (DMAIC) problem-solving process (as shown in Figure 1). The rigorous DMAIC approach defines the steps a team follows, starting with identifying the problem and ending
by Janet Jacobsen
ASQ www.asq.org Page 1 of 4
with implementing a long-lasting solution. To evaluate potential RCA projects, Mike Trueg, manager of field quality assurance/ continuous improvement at Clipper’s Cedar Rapids plant, uses a matrix that measures the impact of safety, quality, and turbine availability. “For this project the scoring met the criteria because of the big impact on turbine availability,” explained Trueg, an ASQ Senior member.
An RCA project was chartered to address the weather-related anemometer issues. The project objective was to identify the root cause of the anemometer failures that was leading to downtime and decreasing turbine availability. A project team was tasked with creating an action plan and implementing corrective actions by the start of the next winter season.
Following the DMAIC Approach
Selecting team members for this RCA project was somewhat challenging, recalls Ellen Sennett, who served as the project’s co-leader. “We started with people who had experience with electrical issues since that seemed to be the problem,” said Sennett, an employee of Clipper for two years. In all, seven stakeholder areas were represented on the improvement team, as shown in the table in Figure 2. Not all team members partici- pated during every stage of the project; for example, the vendor representatives came onboard once the root cause was identified.
The team worked through the steps as outlined in Figure 3.
Defining the Problem
Soon after the initial weather-related failures, the company began collecting data each time inclement weather took a turbine offline. This early data collection led to the charter of the RCA project.
ASQ www.asq.org Page 2 of 4
Figure 2— RCA project team members Stakeholder Area Represented Number of Team Members Quality assurance 2 Fleet services 3 Electrical engineering 2 Anemometer supplier 2 Operations 1 Vendor recovery 1 Procurement 1
Figure 1— Clipper Windpower’s Six Sigma DMAIC problem-solving approach
Analyze the process
Find the root causes of the problem and
understand/quantify their effect on process performance (finding
the critical Xs).
Measure the process
Understand and baseline the current performance of the
process through a set of relevant and robust
Define the problem
Develop clear project based on a real problem
that is relevant to the customer and that will
provide significant benefits to the business.
Improve the process
Develop, select, and implement the best
solutions with controlled risks.
Control the process
Ensure the solutions are embedded, the process
has robust controls, and the project has
a clear closure.
Closeout Tollgate Review: Successful projects need clear and visible closure, the key elements of which should include documenting
lessons learned, transfer of the process back to business as usual, and effective controls. A project review is required to gain a consensus that the project is
ready to close.
Remediation Tollgate Review: Projects that require potential fleet remediation require a
project review to gain a consensus that a fleet
remediation is required and a budget has been approved.
</= 5 weeks
>/= 7 weeks
Key Metric: RCA kickoff to root cause(s) identified and approved.
Goal is five weeks or less.
Root Cause Tollgate Approval: The end of the analyze phase is a review point. A project review is required to assess the root cause identification, to gain a consensus that the
root cause has been identified, and to commit any additional
resources required for its success.
Figure 3— Clipper’s DMAIC-based RCA project steps Define the problem in concrete, measurable terms.
• Define and document problem or gap. • Collect data to understand magnitude of failure.
Measure: Quantify the problem and perceived aspects of the root cause.
• Gather data on current situation. • Develop SIPOC and fishbone diagram. • Filter fishbone items through cause and effects matrix. • Analyze top items through FMEA.
Analyze data to determine the root cause of the defect.
• Develop data collection plan for top priorities from FMEA. • Develop test plan. • Confirm root cause with test data.
Improve: Identify and implement the proposed solution.
• Evaluate design improvements through validation testing. • Identify corrective action. • Develop implementation plan.
Control: Confirm improvement gains through monitoring.
• Ensure 100% inspection at vendor. • Mistake proof (poka yoke) wiring. • Update installation instructions and training.
Measuring to Quantify the Problem
The data collected indicated that, although the winter weather conditions were severe, both precipitation and temperatures fell within the supplier’s specifications for the anemometer. The RCA team developed a supplier-inputs-process-outputs-customer (SIPOC) matrix to quantify the problem and any perceived aspects of the root cause. To pinpoint possible root causes of equipment failures, they also completed a fishbone diagram, which generated 45 items for further study. Next, RCA team members entered the potential causes into a cause and effects matrix to focus on the most likely culprits. The matrix tool enabled the team to pare down the potential causes to nine items for a failure mode and effects analysis (FMEA).
Analyzing Data to Determine the Root Cause
The next step for Sennett and her team was to develop a data col- lection plan covering the potential causes with the highest risk priority numbers from the FMEA. In all, data were collected from tests performed on 13 FMEA potential causes—ranging from improper training on work instructions for wiring the heating cir- cuits to issues with heating the transducer cap on the anemometer.
After data collection and testing of the anemometer, the RCA team concluded that the supplier’s design of the heating circuit did not meet the advertised specification. This failure led to an insufficient heating circuit for Clipper’s application and thus caused weather-related failures of the company’s wind turbines.
Sennett remembers that getting the supplier of the anemometers to acknowledge that its product did not work in the field as promised was a real challenge. Eventually, data from the field and the RCA project convinced the supplier. In hindsight, Sennett feels that perhaps her team could have involved the sup- plier in the project a little sooner. “It would have been beneficial to have the supplier go through the DMAIC steps with us and discover the root cause, instead of us finding it and telling them they had a problem,” she said.
Identifying and Implementing a Solution
With the root cause in hand, the team began to evaluate improve- ments to the anemometer’s heating circuits through a series of winter weather-simulated validation tests. Trueg reports that, having 405 units to replace, data analysis was vital: “That’s why we created our own winter weather environment with a wind machine and a misting device to verify our solution. We didn’t want to remediate all these sites and then have to do it again.”
Following military standard 810F section 521.2 for icing/freez- ing rain, the Clipper team directed three rounds of laboratory testing to analyze the performance of three prototypes for an improved anemometer. The first new prototype was immediately rejected because the simulated winter conditions created an ice build up, which quickly caused the anemometer to fail. A second prototype also failed before a third version finally withstood the extreme weather conditions of the lab.
Once the testing was completed, the team created an action plan. The plan goal was to have all anemometers on each of the 405 turbines throughout the country replaced with the newly designed version by March 30, 2010.
Controlling to Confirm Improvement
In addition to the heating circuit improvements based on the lab testing, several other controls were introduced:
• The vendor conducts 100-percent inspection of the product through a three-day burn test of the unit’s heating system. This eliminates the shipment of any defective products.
• All anemometers are tested with a turbine control unit in the manufacturing facility to validate functionality.
• The new anemometer design also incorporates a connector that can only attach one way to the junction box, thus eliminating improper wiring in the field.
• The wiring is color-coded for the operators who install the connectors.
New Design Stands Up to Mother Nature
While field testing began late in the winter of 2008-09, Clipper realized the importance of carrying over the testing into the winter of 2009-10 to confirm the effectiveness of its improve- ment plan. Once again, Mother Nature cooperated by throwing her full bag of winter tricks as 40-50 mph winds, one half inch of ice, four to eight inches of snow, and temperatures of minus 15 degrees and below were reported at various wind farms. Despite these conditions, Clipper recorded only two weather- related anemometer issues for a 1.6-percent failure rate. Clipper soon discovered that the two failures were caused by a sup- plier assembly team issue and were not directly related to the improvements generated by the RCA project. With the improve- ments and control verified, the RCA project was officially closed.
The RCA team kept turbine customers informed throughout the DMAIC process with presentations about remediation steps to reduce the weather-related failures. Team members walked through the entire DMAIC process with key customers and explained how the root cause was determined, as well as plans to implement corrective action. Sennett added that many of Clipper’s customers are familiar with Six Sigma tools, so the RCA process is the type of problem solving they like to see. “This process helps with customer satisfaction as [customers] know we are taking the time to find the root cause and using trained people to do [corrective action] the right way the first time,” Sennett said.
External customers weren’t the only ones who benefited from this RCA project. Employees at Clipper’s remote monitoring dis- patch center, which controls the turbines from the Cedar Rapids facility, saw a decreased workload as fewer turbines required attention during inclement weather.
Sennett believes that this RCA project and others that followed help Clipper’s employees think more proactively and address issues before they become fleet-wide issues. “Our goal is to
ASQ www.asq.org Page 3 of 4
ASQ www.asq.org Page 4 of 4
become more preventive and look at things before they start to fail, and with the Six Sigma processes you can do a better job of designing out the defects in the beginning before implementation,” noted Sennett.
Building a Culture of Quality
Both Trueg and Sennett credit this RCA project for opening their eyes to key issues such as internal testing and expanding the company’s supplier base. As a result of this improvement project, Clipper created a plan for introducing new suppliers to avoid potential problems caused by single sourcing. “We’ve also developed testing here at the manufacturing site so if we have quality issues, we can test before sending something out into the field that potentially causes failures or creates the need for a replacement part,” explained Trueg.
Sennett said that while some team members were initially skeptical about the DMAIC process, they quickly learned the importance of taking the time for each step, recognizing that without the structured process, people tend to collect unnecessary data unrelated to the issue. For several team members, working on this project sparked an interest in learning more about process
improvement and prompted them to request further training and the opportunity to earn Six Sigma Green Belt certification. Trueg is amazed at the change in Clipper’s staff once they serve on an RCA team: “The attitudes and focus on problem solving with data are a strong part of the Clipper culture.”
For more information:
• Sennett and Trueg recommend the following books to guide your process improvement activities: The Lean Six Sigma Pocket Toolbook by Michael L. George, David Rowlands, Mark Price, and John Maxey, and Statistics for the Utterly Confused by Lloyd Jaisingh.
• Visit the Knowledge Center at www.asq.org/knowledge- center to find additional resources on root cause analysis and Six Sigma.
About the Author
Janet Jacobsen is a freelance writer specializing in quality and compliance topics. A graduate of Drake University, she resides in Cedar Rapids, IA.