Reliability of bridge wire fuse heads for serial ignition firing applications

Abstract. Serial chains of electric fuseheads are used in industrial and safety applications like detonators in mining or as integral part of pyrotechnic actuators in explosion prevention and fire extingushing systems. Reliable initiation of all fuse heads is a crucial property of serial ignition chains. The risk of misfires caused by interuption of electrical energy supply must be kept as low as possible. The bridge wire can be regarded as critical part. It initiates the primary charge by electro-thermal heating to autoignition temperature. On the other hand it is responsible for the transfer of electric energy to other fuse heads as part of the serial chain. An optimized design with respect to safety against misfire requires a proper selection of bridge wire material and also thermo-chemical properties of the primary charge. The traditional approach for assessing wether a fuse head design is suitable for serial ignition or not, is to perform a number of repetitions of serial ignition tests for the target number of fuse heads and ignition current. The manufacturer afterwards decides from that limited tests that serial ignition can be guaranteed. A new customer demand is to provide an estimation of reliability in discrete numbers. For this purpose a test methodology will be discussed, that provides statistical parameters for initiation of primary charge and current break down. A well known method to estimate failure probalities in structural design of mechanically loaded components is the Load-Resistance Interference Method. If statistical distributions of mechanical loads and structural strength are available, a reliability margin and failure probability can be provided. From literature it is known, that this method can be used to other situations, where the extent of overlap of two distributions determines failure probability. This presentation proposes the estimation of failure probability by use of the Load-Resistance Interference Method with consideration of distributions for the characteristic parameters: time to first reaction and current break down.