A huge bulk. of coals in a variety centering around the coking coal for steel-making is imported from foreign cOuntries into J<apan The quantity of the coals shares approximately one third of the total amount of seaborne _J transportation of eargo coal in the world. Spontaneous heating accidents of the cargo coals, although small in number, have occurred on coal carriers in the country. Governing factors and mechanism of the spontaneous heating of coals on sea transportation, hcwever, have not been clarified yet In the paper, the primary cause of the ~pontaneous heating is assumed to be elow oxidation of the coals by oxygen in a coal pile in spite of the fact that various factors are related to the heating phenomenon. Considering the balances of heat and oxygen, a theoretical model of the spontaneous heating is expressed with three conservation equations of heat, oxgen and activity of coa] surface, where the main governing factors are heat of reaction, rate constant, activation energy and an initial number of active sites for oxidation of the coals as well as apparent heat conductivity of the coal pile and diffusion coefficient of oxygon molecules in the air. Numerical calculations fcr sinrulating the behavior of temperature increase in the coal pile are conducted on the basis of the model, and the effects of governing factors on the results of simulation are investigated. It results that the maximum tcmperature rise Of the coal pile increases proportionally with an increase of heat of reaction. Rate cOnstant and activation energy are concerned with the rising velocity of the pile temperature. The exponential effect of activation energy on the velocity is especially large. A sTnall initial value of active specific surfaee tends to retard the oxidation even though oxgen molecules remain in the pile. On the other hand, when the nurnber of active sites is large, diffusion coefficient of oxygen related to the oxygen supply into the coal pile controlls the reactiou. Therefore, increases of the latter two factors mainly increase the maximum temperature rlse. Apparent heat conductivity controlls heat release from the coal pile and heat addition into it, ~nd influencee the temperature in the pile.

Numerical　analysis　of　the　sぬock　wave　arisjng　at　electric換1spark　ignition　of　combustible　gas　was　done　by using　the丘nite　di鉦erential　method． Initial　temperature　of　the　spark　kemel　was　set　equal　to2000K，3000K　and4000K　Peak　pressure　of　the　skock wave　varied　depending　on　the　initlal　temperature，but　the　shock　wave　arose　in　all　cases，The　shock　wave　survived in　feW　miCrOSeCOndS　and　prOpagated　abOUt　l　mm・TheSe　VaIUeS　are　bOth　Smaller　than　the　Va1UeS　ObSerVed popularly　in　the　experimental　shock　waves．It　is　considered　to　be　due　to　the　instantaneous　supply　of　thermal energy　in　the　calculation．In　experimental　studies，finite（1uration　time　is　required　for　a　electricalspark、 In　this　study，a　few　tenths　of　thennal　energy　was　converted　to　kinetic　ellergy　of　the　shock　wave、1n　the　time period　in　which　the　shock　wave　survives，chemical　reaction　proceeded　little　and　could　be　ignored　in　calculation、Hot　body　of　the　spark　kemel　expanded　with　the　shock　wave　propagation，but　its　expansion　was　slower　than　the shock　wave　propagation．Temperature　at　the　center　of　the　spark　kemel　decresed　with　time歩and　a　hotter　part apPeared　at　the　outer　edge　of　the　kerneL　Surrounding　gas　of　the　spark　kernel　was　heated　a　little　after　the　shock wave　passed． The　shock　wave　is　usually　ignored　in　a　nurerical　analysis　of　the　electrical　spark　ignition　of　the　combustible gas，The　result　of　thisstudy　revealed　the　presence　ofthe　shock　wave　affectsthe　distribution　ofthe　tkermal　energy．It　should　be　taken　into　account　in　such　an　analysis.

To improve design of systems and items by feliability data, it ls irnportant to estimate correctly how each factor effects on their lifetimes. Fot this purpose, it is necessary to establish the method to evaluate the effect quantitatively and exactly from not many data. As one trial, the authors took notice of Proportional Hazards Model which was developed in medical field and investigated whether this method can be app]ied to the engineering field or not. At first, by computer simulation, it was made char that the reliability can be estimated more correctly with this method tha:n with conventional method. Next, by experiment using lamps, it was made clear that with this method, ( I ) degree of effect of using condition (each factor) on their lifetimes can be evaluated and ( 2. ) the effect of design improvement also can be clarified. Through these investigations, feaeibility of this method to evaluate the effect of competing risk was made clear quantitatively and exact]y.