Past accidents have demonstrated that vapor cloud explosions (VCEs) are the most severe threat to refining and petrochemical industries. In order to estimate the air blast parameters at any given distance from a possible explosion source, a variety of prediction methods have been developed. A brief description of these prediction methods is first presented. Then, the focus of this paper is on an engineering method based on blast curves, which is the most frequently used prediction method.
Knowledge of blast effects from vapor cloud explosions has greatly improved due to the efforts of investigators in many countries during the last three decades; however, blast curves have not been reexamined. One of the aims of this paper is to present a new set of blast curves which have been validated with available experimental results and analytical solutions. Another aim is to present a comprehensive set of blast curves including blast parameters for both positive and negative phases.
The new blast curves, which are a modification of the Baker-Strehlow curves, improve the prediction for detonations and supersonic deflagrations as well as for subsonic flames. By changing the slope of the blast pressure versus distance curve, the overly consserative blast pressure predictions by the previous Baker-Strehlow curves at large standoff distances for detonations and supersonic deflagrations are reduced. By clarifying the labeling of the flame. Mach number for each blast curve, the overly conservative pressure and impulse preditions for subsonic flames are also avoided.
The new blast curves have been validated against VCE experiments including detonations, fast deflagrations, and slow deflagrations. Particular attention was given to large scale deflagration experiments, which became available only recently. Good agreement is shown between the new blast curves and the experimental blast data for vapor cloud detonations and subsonic deflagrations. For supersonic deflagrations, the blast curve predictions are overly conservative in that blast pressure decays more rapidly with distance in experiments than the calculated curves.