A scientist who conducted tests on a controversial insulating product made by a Washington Gas Light Co. subsidiary has said he cautioned officials of the utility and the subsidiary a year ago that his research did not support their claims for the insulation.

William R. Lanier, a project engineer for Artech Corp., said in an interview that only one of 11 tests he ran of Thermlo-K insulation provided results anywhere near the product's advertised insulating values.

Officials of Washington Gas and its subsidiary, Devanport Insulation Inc., have denied Lanier's claims.

Disclosure of the 11 tests came in response to a subpoena issued in connection with a $1.25 million lawsuit filed against the two companies by a man who once ran a small insulating business in Fairfax County.

In his suit, Jack J. Goehring of McLean accused both companies of fraud and of a breach of contract that forced his firm to close. The utility "either knew or should have known to be untrue and unsubstantiated" its claims that the Thermio-K insulation had an insulating value of "R 5.2," his suit said. The effectiveness of insulation is stated in terms of its R-value, a term denoting the insulation's value in keeping houses and buildings warm in the winter and cool in the summer.

Goehring's suit, filed in Arlington Circuit Court, is likely to provide the first legal test of Davenport's claims for the insulation industry and have attracted the interest of congressional committees and government agencies studying the need for tighter regulation of the growing cellulose insulation industry.

Earlier this year, Davenport, the largest home-insulating firm in the Washington area, announced it was creasing production of Thermlo-K and replacing it with Thermlo-K II, a product that claims sharply lower insulating values.

Last year, when The Washington Post published an article on the controversy over Davenport's product, Carol V. Davenport, the company's president replied that two independent laboratories supported his claim. One of those cited was Artech, a small Fairfax County research organization.

Letters recently introduced during congressional hearings raised questions about whether either of the two labs conducted the Thermlo-K insulation tests properly and questioned the safety of Dovenport's other major home insulation product, Cellincraft.

In one letter an official of a Cambridge, Mass, lab that is well known for its tests of insulation made from finely ground newsprint, said his lab had never tested a product with an "R value of 42 or better."

In the other letter, the staff of the House energy subcommittee said Cellincroft insulation had failed four tests conducted by the Energy Research and Development Administration in a 1976 survey of cellulose insulation. Cellincraft, sold by Daveport along the East Coast, was found to be corrosive, to show excessive moisture absorption and fungal growth, and exhibit separation of this retardant chemicals under laboratory tests, the staff letter said.

Davenport said in an interview this week that the ERDA tests were poorly conducted - a point conceded in part by Stephen F. Sims, a special assistant to the House oversight subcommittee, which conducted the insulation hearings. However, Sims said that repeated failures of the Davenport products and the dispute over the R value of its Thermlo-K insulation indicated that "a problem exists and must be corrected."

Washington Gas spomeskan Paul Young said yesterday that the utility "categorically denies that the R value of the insulation was misrepresented and will fully substantiate the R value in court." Davenport alleged that Goehring refused to pay his bills at Davenport and that the company would supply him with insulation on a cash basis prior to the collapse of his company, Thermal-Max Corp.

Lanier, the Artech engineer, said the 11 tests he conducted on Thermlo-K insulation indicated it had an average R value of 3.780. The tests ranged from a high of 5.196 to a low of 3.030.

Lanier said he advised both Robert Bucco, a Davenport vice president, and an official of Washington Gas prior to its axquisition of Davenport in September that the average test scores from his tests and not the high one should be used in the company's advertising. Bucco said he never received such advice and was not aware of 11 separate tests being run by the lab. Davenport said he was aware that Artech conducted more than one test on his product. Lanier said he could not remember the name of the Washington Gas official with whom he talked.