Categorical hazard ratings such as NFPA 704 are routinely used as risk surrogates in semiconductor chemical delivery systems, yet across the fifteen precursors examined here they compress a 750-fold range in mass-specific energy release into four levels. This study asks whether a measurement-referenced threshold can replace the category in that role, and what that changes. We define the Critical Impurity Level (CIL), the moisture concentration at which hydrolysis would raise the temperature by 50 K, by inverting the adiabatic temperature rise; couple it to a stochastic ingress model through the exceedance probability P(moisture>CIL); and carry that probability into fault tree analysis and screening safety integrity level (SIL) allocation. The thermodynamic state of the ingressing water is set by the delivery configuration rather than by the substance, which changes the sign of the hydrolysis enthalpy for two of them. One substance, TiCl4, is not classified because that sign is not fixed by the available data. Exceedance probabilities span 0.001 to 0.175; eight of the fourteen classified substances deviate by one or two SIL levels from a LOPA baseline, TDMAH (NFPA F3) ranking above SiH4 (F4). Substance-specific set-points remove 93 to 98% of alarms the thermochemical criterion does not support.
Extreme natural hazards can significantly degrade the survivability of field automation terminals (FATs), thereby affecting post-disaster service restoration (SR) in distribution systems (DSs). However, existing SR studies commonly assume fully available FATs and may consequently overestimate SR performance. To address this issue, this paper proposes a survivability-aware multi-period SR strategy for soft open point (SOP)-based DSs. The cyber-physical survivability of FATs is explicitly modeled through limited switching operations of spring-operated remote-controlled switches and constrained communication availability after faults. A multi-period SR optimization model is then developed by explicitly incorporating FAT survivability into the restoration decision-making process, together with time-varying ZIP load characteristics and distributed energy resource outputs. A nonlinear transformation method is further employed to transform the original optimization problem into a tractable form for efficient solution. Case studies on a modified IEEE 33-node system demonstrate that considering FAT survivability provides a more realistic assessment of SR capability, avoids overestimation of SR performance, and supports more reliable post-disaster decision-making.
