dcsimg
 

Reducing Construction Costs with Statistics: You Don't Need a Weatherman

An energy company reduced construction costs by factoring weather data into its contractsMark Twain said, “Everybody talks about the weather, but nobody does anything about it.”  But when weather-related delays caused construction costs to soar, people I spoke with at one energy company did do something about it, by using weather data and making it part of their agreements with contractors.

The company needed to nearly double capacity at one of its facilities to better serve customers in several states. Keeping the project on time and on budget was critical. But the facility’s location was prone to extreme weather, including hurricanes. 
 
Every time poor weather halted construction, the entire project was delayed, and the impact of these delays had to be negotiated with the contractor on a case-by-case basis. These negotiations cost time and money. The company knew the weather could have a serious impact on the four-year facility expansion project’s schedule and budget. 
 
The company wanted to minimize construction schedules, reduce expenditures, and minimize time spent in weather-related negotiations by including some protection against variability in the weather in its contract.  Quality specialists at the company used Six Sigma techniques, Minitab Statistical Software, and historical weather data from the site to do it.

Analyzing Weather Data with a Chi-Square Test

The team needed to develop realistic limits for change order claims due to weather. Using Minitab, they analyzed data from five weather stations within 35 miles of the facility. Their first task was to determine whether they could use data from one location, or whether it would be best to use an average of the data from across all the stations. 
 
They used Minitab’s chi-square test to answer this question. This test assessed whether the proportion of rainfall from each of the five local weather stations was equal. 
 
If the data from the five stations were similar enough that it would be safe to use only the data from one of them, the p-value from the chi-square test needed to be at least 0.05. It was 0.01. In other words, Minitab’s analysis showed that using an average of the data across the stations would be more reliable than using just one station’s data. 
 
Because weather patterns change over time, they also needed to select a date range that accurately reflected the current pattern around the site. The team used Minitab to look at the most recent five-year slice of data available. However, this time period included a major hurricane, and Minitab’s analysis showed that the hurricane precipitation created extraordinary variation within the data.
 
So the team expanded their analysis to cover 10 years. Although this longer period actually included two hurricanes, it included sufficient data to absorb the variation they caused while still representing the current weather pattern around the site.

​Using Data Analysis to Specify Weather Limits in Contracts

The team was able to confidently proceed with using the 10-year time frame as the basis for the weather limits in their contracts.  They used Minitab’s analysis of historical weather data to set an upper specification limit for each month that represented 95% of the projected precipitation for the site. These specifications were then written into the construction contract.  
 
With the insight acquired from using Minitab, the team reinvented the negotiation process for the expansion project. The new contracts specified that precipitation could not be the basis for a delay, unless it exceeded the defined limits for any given calendar month. Now when a contractor claimed a change order due to weather, the company was able to review the claim against the limitations set out in the contract. If the weather in question exceeded the limitations, the delay was granted. If not, the delay could not be granted under the contract. 
 
During 45 months of construction, weather-related negotiations were necessary for only three months. Out of 27 days in which more than an inch of precipitation fell, eight days were granted as valid reasons for change delays under the contract. A total of 19 schedule days were saved. 
 
To prove that these savings resulted from the new process, and not from unusual weather during the project, The company used Minitab’s 2 proportions test to compare the weather data from the construction period with the historical data. The test verified that the weather pattern during the construction project was statistically the same as the historical weather pattern.
 
Because of the success of this project, the concepts the team developed have been fully embraced by the company, and now are viewed as a critical tool in managing the company’s large, multi-year construction projects.
  
So maybe you can't do anything about the weather—but by analyzing the right data, at least you can avoid negotiating about it!

Comments

blog comments powered by Disqus