NTSB Safety Recommendation P-72-002

NTSB safety recommendation P-72-002.

THE NTSB RECOMMENDS THAT: THE FEDERAL RAILROAD ADMINISTRATION UNDERTAKE A STUDY OF THE CURRENT METERING PRACTICES IN THE LIQUID PIPELINE INDUSTRY, WITH THE POSSIBLE ASSISTANCE OF QUALIFIED PIPELINE GROUPS, TO DETERMINE THE EXISTING STATE OF THE ART IN DETECTING SMALL PITHOLE-TYPE LEAKAGE BY METER VARIANCE WITH PARTICULAR REGARD TO LARGE DIAMETER PIPELINES OPERATING AT HIGH VOLUMES. THE STUDY SHOULD INCLUDE THOSE PIPELINES WHOSE PUMPING OPERATIONS ARE REGULATED BY THE USE OF RECORDING METERS WHICH MONITOR THE RECEIPTS AND DELIVERIES AND ARE SET TO SHUT DOWN OR OTHERWISE INFORM THE PIPELINE DISPATCHER UPON THE OCCURRENCE OF A SPECIFIED AMOUNT OF INPUT/OUTPUT VARIANCE. THE STUDY SHOULD INCLUDE METER ACCURACIES WITH THE INTENT TO ESTABLISH CERTAIN MINIMUM STANDARDS REGARDING RECEIPT AND DELIVERY VARIANCES WITHIN WHICH LIQUID PIPELINES SHALL OPERATE. BASED UPON THE RESULTS OF THIS STUDY, THE NUMBER OF BARRELS-PER-HOUR VARIANCE ALLOWABLE BETWEEN THE INPUT AND OUTPUT OF LIQUID PETROLEUM PIPELINES SHOULD BE INCLUDED IN 49 CFR 195.

Priority: CLASS II

Overall Status: Closed - Reconsidered

Issued Date: 1971-12-08

Adopted Date: 1971-12-08

Overall Date Closed: 1975-12-30

Synopsis: ON SEPTEMBER 2, 1970, A PIPELINE LEAK OCCURRED IN THE COLONIAL PIPELINE SYSTEM NEAR JACKSONVILLE, MARYLAND. CONTRACTORS WORKED CONTINUOUSLY FOR 20 HOURS TO FIND THE LEAK, AND ON THE NEXT DAY AN EXPLOSION OCCURRED FOLLOWED BY A FIRE. THERE WERE NO FATALITIES, BUT FIVE WORKMEN WERE BURNED.

Probable Cause: The National Transportation Safety Board determines that the probable cause of the leak was a flaw of undetermined origin in the pipe wall, crater-like in appearance, wide at the surface and narrowing down to a thin metal membrane, which failed after a period of constantly fluctuating pumping pressures. The probable cause of the explosion and fires was the ingestion of the gasoline vapor-rich atmosphere by a diesel engine which resulted in the speeding up and backfiring of the engine, igniting the atmosphere.

Ntsbnumber: DCA70MP003

Report Number: PAR-71-02

Addressee Name: FRA

Addressee Status: Closed - Reconsidered

Addressee Date Closed: 1975-12-30

Addressee Acronym: FRA

Addressee Organization Type: G-Federal Government

Communication Date: 1972-04-07

Communication Type: Official Correspondence

Communication Contents: LETTER FROM FRA: THE ABILITY TO CHECK INPUT AND OUTPUT AT THE RECEIPT AND DELIVERY ENDS OF THE PIPELINE CAN BE DONE WITH A HIGH DEGREE OF METER ACCURACY. THE INTERPRETATION OF THE OVER OR SHORT CONDITION OF THE PIPELINE IS WHERE THE STATE-OF-THE-ART HAS NOT BEEN DEVELOPED. MOST PIPELINES ARE CHECKED FOR INPUT AND OUTPUT ON A ONE OR TWO HOUR BASIS BUT UNLESS THE CONDITIONS ARE CONSTANT, A COMPARISON WILL NOT BE AN ACCURATE INDICATION OF WHETHER OR NOT A LEAK EXISTS. THIS CHECK CAN NORMALLY ONLY BE EFFECTIVE OVER AN EXTENDED PERIOD. THE ONLY WAY A PIPELINE CAN BE MONITORED CONTINUOUSLY IS WITH COMPUTER CONTROL AND THERE ARE VERY FEW PIPELINES OR PIPELINE COMPANIES WITH THIS COMPUTER CONTROL CAPABILITY. EVEN WITH COMPUTER CONTROL THE FLOW RATE CHANGES ARE HARD TO INTERPRET AND SMALL LEAKS CAN STILL GO UNNOTICED. WE FEEL THERE ARE TOO MANY UNCONTROLLABLE VARIABLES THAT EXIST IN THE OPERATION OF A PIPELINE TO TRY TO RESTRICT THE OPERATORS TO A GIVEN NUMBER OF BARRELS-PER-HOUR VARIANCE BETWEEN INPUT AND OUTPUT.