For four decades, the state Department of Transportation has been building bridges and highways while ignoring some of its own rules for ensuring high-quality concrete.

A review of concrete test results from one major project, the new Route 403 highway in North Kingstown, specifications and interviews with officials show that the DOT has selectively thrown out low test results and doesn’t use a standard industry method for checking on suspect concrete after it has been poured.

The files for the $130-million Route 403 project, the bridges and highway the DOT is building to the state’s Quonset Point industrial park, show dozens of cases where the DOT allowed concrete whose strength was below specification to remain in the project’s structures. It also failed in some cases to impose on contractors the monetary penalties called for in the agency’s regulations.

The department also virtually never uses the most powerful sanction against suppliers who deliver substandard concrete: the demolition, removal and replacement of below-specification concrete at the contractor’s cost.

In one case, DOT files show that a state engineer demanded in 2002 that a substandard concrete beam supporting a bridge in the 403 project be junked. He was overruled, the beam was used, and he recanted his objections when the DOT was asked about them.

The DOT and its major concrete supplier said the practices used during the 403 project have also been used on other projects, such as the ongoing Route 195 relocation in Providence.

While acknowledging the practices, the DOT stands by the safety of its projects. However, the practice of accepting below-specification concrete raises questions about how the department oversees millions of dollars in highway projects each year. After The Journal started inquiring into the DOT practices, officials said it had recently changed its practices to conform to its own regulations.

DOT officials said that accepting below-specification concrete was harmless because the specifications assigned to the structures in question were higher than necessary, and because the test results weren’t so low as to raise questions about the bridges’ structural integrity.

“There’s always a margin of safety” in bridge-engineering design, said Frank Corrao III, the DOT’s deputy director for construction management. When below-specification concrete remains in a structure, he said, the DOT made sure “that it has not compromised the integrity of the structure.”

Industry experts bear out that safety claim, saying that bridges are normally designed with generous safety margins.

Stephen A. Cardi, treasurer of Cardi Corp., general contractor on all but one of the 10 Route 403 contracts and the supplier of the concrete poured for Route 403, said the poor test results are so insignificant as to be not worthy of notice. He also said that below-specification test results could have resulted from testing errors, something the DOT disputes.

Cardi said his company applies strict standards to all of the numerous projects to which it supplies concrete, both for its own contracts and to other contractors. He said the reason other contractors, including his company’s competitors, buy concrete from Cardi is its high quality. The company’s computer-controlled production, he said, imposes rigid quality control on the concrete. Even if you wanted to, he said, “it doesn’t allow you to cheat.”

Concrete durability has been a major problem in Rhode Island. The $610-million Route 195 relocation was necessitated partly by the extreme deterioration of the existing bridges in Providence and East Providence, several of which are held up by wood and steel emergency supports.

Do the below-specification test results suggest less durability for the Route 403 bridges?

“No,” said DOT spokeswoman Heidi Cote. Concrete continues to gain strength as it cures, she said, but additives and other measures are more important to achieving durability than strength.

FOR A HOME handyman, concrete is a useful and uncomplicated tool. To sand and gravel, add cement and water, with the proportions depending on the use. The cement reacts with the water in a process called hydration, binding the other materials together and hardening to form a strong and durable construction material.

To meet the complex engineering demands of structures such as bridges, however, the standards for ingredients and their mixing and the placement and curing of the concrete become exacting, even down to the amount of air in the mix.

To get all that right, agencies such as the DOT publish detailed specifications, often based on national standards, and require regular inspections and testing both at the job site and in the laboratory.

Besides the cement itself, DOT rules demand tests of the size and quality of the sand and gravel, the strength of steel reinforcing rods and the amount of water in the mix (too much weakens the concrete).

Asked how concrete could end up below the specification, the DOT said it monitors both the concrete plant and the job site, but that “additional water could be added along the way.”

Because concrete cures gradually, compressive strength — its most important quality — can’t be determined when it is poured. Instead, 6-inch-by-12-inch cylinders are cast from samples. The standard practice is to let the cylinders cure for 28 days and then crush them, measuring their resistance in pounds per square inch, or psi, and comparing that with the specification.

For most of the important parts of bridges — and the ones you can see, like the abutments that hold up the ends, the piers that support the superstructure in between and the beams that support the deck — the usual DOT specification is 5,000 psi.

For a 150-cubic-yard “lot” of about 15 truckloads, the DOT takes samples from the first truckload and from one of each five truckloads after that, according to Mark Felag, the DOT engineer who heads the agency’s Materials Section. One sample represents about 50 cubic yards — equal to a volume 3 feet by 3 feet by 150 feet.

In the vast majority of more than 500 test results from seven Route 403 contracts, the compressive strength met the specification.

However, there were at least 29 cases with flaws. In at least 21 cases, the average of the test cylinders’ strength was below specification, although in several cases the margin was slim. In all 29 cases, half or more of the cylinders tested below specification. Of the project’s 16 bridges, 9 had one or more structural parts where the average concrete strength was below specification.

One bridge, which carries Route 403 over the Davisville Access Road, has at least five parts, including abutment footings, pier footings, pier columns and at least one beam, with tests averaging below specification.

In eight cases in the project, the DOT accepted concrete where more than one test cylinder was below specification by 500 psi or more. That’s a number that two major national organizations, the American Concrete Institute and the American Association of State Highway and Transportation Officials, use to flag suspect concrete. However, it’s a flag the DOT does not watch for.

“That is not one of the criteria we use,” Corrao said.

The combination of practices at the DOT meant that, one way or another, in 20 cases, below-spec concrete remained in the structures, usually in part of a bridge.

The DOT, meanwhile, said that in response to The Journal’s questions, it reviewed records on all 34,851 cubic yards of concrete it poured statewide during the first eight months of this year and found no serious problems. That’s a bit less than the 46,000 cubic yards that have gone into Route 403 since 2001, according to DOT figures.

THE CONCRETE section of the DOT’s “Standard Specifications for Road and Bridge Construction,” known as the “Blue Book,” which DOT officials describe as their bible, is exacting, even specifying how many times truck mixers are supposed to revolve. For concrete strength testing, it calls for selecting three test cylinders at random, crushing them and averaging the results.

The records, however, show that the DOT has routinely tested four cylinders, and that when test results are marginal or below specification, officials repeatedly crossed out the worst result and used the remaining three to calculate a new, higher average. In several cases, that boosted the average above the minimum strength allowed by the specification.

Besides violating the DOT’s own rules, throwing out the lowest test result is unusual.

Kelley C. Rehm, an engineer and program manager at the national organization representing state transportation agencies, the American Association of State Highway and Transportation Officials, or AASHTO, said she could understand throwing out both the lowest and highest results. But she couldn’t understand throwing out only the low test result.

“I can’t imagine why they would do that,” said Rehm.

Felag and Corrao initially dismissed the crossed-out results as “outliers,” a term used in statistical analysis for items numerically distant from the rest of the data. However, many of the crossed-out test results weren’t distant, just low.

Also, the DOT never crossed out any high “outliers,” just low ones. Why not?

“I’m not really sure I have a good answer for you at this point,” Corrao said.

High test results wouldn’t indicate anything wrong with the testing process, he said, while low ones might. But in disputing Cardi’s assertion about flawed tests, the DOT also quoted the American Concrete Institute to say that testing will erroneously flunk good concrete only “about once in 100 tests.”

In a written statement, the DOT said the practice of testing four cylinders and selectively throwing out the low test result goes back “at least to the late ’60s as far as anyone can remember,” and that its origin is unknown.

“We did not believe we were doing anything improper,” Corrao said. “Looking back, it’s probably not the best practice.”

Felag and Corrao said the DOT now follows its regulations, averaging the results of three cylinders per sample, with no throw outs.

Corrao also said, however, that the agency made the change because of concern that “it would open us up to criticism if we kept doing that.”

Asked for the document in which the agency made that policy change back to its own regulations after 40 years, he and Felag said there is none. “It was verbal, just to our guys,” Felag said during an interview.

Felag, who has worked in the DOT materials testing section for 23 years and headed it for 14 years, said it was his suggestion to revert to the “Blue Book” specifications. That happened around July 4, he said. The DOT amended that to the beginning of the construction season, in the spring.

IN SEVERAL CASES, records show that the DOT left in place concrete that would have been removed and replaced at the contractor’s expense under a specific standard recommended by AASHTO.

That specification refers to HP, or high performance, concrete, which contains additives not used in conventional concrete. Part of their advantage is to make concrete more durable. AASHTO’s standard is uncompromising: HP concrete that tests below the specification “shall be removed and replaced with acceptable concrete.”

There are 17 items in the Route 403 files, most of them structural bridge parts, which contain HP concrete that flunked that test, although some narrowly.

Although Felag is vice chairman of AASHTO’s Executive Committee on Materials and chairs a committee on cement, the DOT doesn’t use that AASHTO standard. (The DOT said Felag’s committee isn’t the one that developed the high-performance concrete specification.)

“That’s not the law,” Corrao said of the AASHTO specification. Corrao also said that a DOT survey of 27 state and Canadian transportation agencies, prompted by Journal inquiries, showed that none of them remove low-strength, high-performance concrete without any further review.

“We haven’t compromised anything, and we haven’t violated anything,” Corrao said.

Short of demolition, the DOT can impose a range of financial penalties and often impose the maximum — 50 percent of the bid price of the concrete.

The Route 403 files show that the DOT imposed financial penalties for low-quality concrete in at least 12 cases, paying only 50 percent of the contract price in every case but one.

But in another eight cases where the average of the tests was below the specification and a penalty should apparently have been imposed, the records show none.

Although it has demanded that contractors replace pre-cast structures, such as manholes and culverts, the DOT virtually never requires a contractor to demolish and replace faulty concrete in structures that are cast in place. Corrao and Felag could recall only one such instance, during the rehabilitation of a Route 295 bridge about five years ago. Most of the important elements of bridges, like abutments, footings and piers, are cast in place.

Removal would be a much more severe penalty, costing the contractor more than double — for demolition and disposal and pouring new concrete.

Corrao said that replacing below-specification concrete hasn’t been necessary, and that it could cause more problems than it solves, delaying the project and possibly damaging the rest of the structure.

Connecticut takes a much tougher line. Connecticut DOT spokesman Kevin Nursick said his agency makes contractors remove bad concrete “several times a year.”

He said officials try to avoid demolition and use a variety of tests before requiring it. Also, he said, that if the concrete, for example, is part of a sidewalk, where the issue is durability, “we would probably dock the contractor.”

However, Nursick said, “If you’re talking about something structural, you would have to take it out and start from scratch.”

That approach could have led to the partial demolition of up to nine bridges in the Route 403 project — if further tests showed that the below-specification concrete was indeed faulty.

The DOT also did not take the next step, a standard industry practice to follow up on suspect concrete: testing a sample of the cured concrete.

When compression testing suggests weak concrete, “then you go to the structure and drill a core,” said Terry C. Collins, a concrete engineer with the Portland Cement Association, an industry group.

The cylindrical test cores drilled from the suspect concrete are tested by crushing them like the test cylinders from the original pour, to determine whether the concrete gained enough strength through more curing. (That still wouldn’t mean it met the contract specification, which applies at 28 days.)

However, Corrao said that cores were drilled in the 403 project only once, after The Journal reported July 30 that concrete in one bridge abutment had tested below specification. Corrao said that cores from the bridge, which carries Frenchtown Road over a highway ramp, showed that the concrete, poured in June, had gained enough strength by October to test well above the 5,000 psi specification.

IN VERY RARE CASES, the records show, lower-level DOT officials have questioned the acceptance of below-standard concrete.

But in January 2002, John C. Lonardo, the engineer in charge of the Route 403 project, vehemently urged rejection of a concrete beam intended for the bridge carrying an off-ramp from Route 403 over Davisville Road.

The specification required a strength of 9,000 pounds per square inch. The four test cylinders, cast Aug. 30, 2001, and tested about four weeks later, averaged 7,611 psi, well below specification. Throwing out the low test wouldn’t have changed anything — even the best test cylinder scored only 7,684 psi.

“In my opinion,” Lonardo wrote at the time on a DOT form titled “Unacceptable Portland Cement,” “[the] beam is not acceptable & should not be used.”

“This is one of only four beams for [the] new bridge & should be rejected out-right!” he wrote.

The beam was used anyway, with the DOT paying Cardi only half-price.

During an interview along with Corrao in September, Lonardo reversed his earlier objections and said it was correct to accept the beam.

“There is no compromise to structural integrity,” he said. His earlier emphatic rejection of the beam, he said, was intended only to put the supplier on notice. He said that the other beams came in well over specification, and that the outcome was a victory for the public because the state got an adequate beam cheap.

blandis@projo.com