The objectives of this project are to study the effect of pretreatment methods on the two-stage liquefaction process. In particular, the effects of dispersed catalysts and carbon monoxide atmospheres on a coal liquefaction process. The project is divided into three technical tasks. Task 1 and 2 deal with the analyses and liquefaction experiments, respectively, whereas Task 3 deals with the economic impact of utilizing the pretreatment methods. This quarter we concentrated on Tasks 1 and 2, which are summarized below. Samples of Black Thunder Subbituminous coal and the corresponding recycle solvent were received from Wilsonville. These particular samples were obtained from run [number sign]263. The samples were analyzed for a proximate and ultimate analysis, as well as heat content. We have also received samples of the iron oxide used in the Wilsonville coal liquefaction runs as well as Shell 324 and molyVanL for use in future runs. Base-line screening experiments were conducted in 300 mL autoclaves using 2.5 parts by weight of recycle solvent to 1 part Black Thunder coal at 425[degree]C for one hour. In preliminary results for the comparison of iron oxide and iron carbonyl, the iron carbonyl appears to give higher yields to soluble products.

In logic-based diagnosis, the consistency-based method is used to determine the possible sets of faulty devices. If the fault models of the devices are incomplete or nondeterministic, then this method does not necessarily yield abductive explanations of system behavior. Such explanations give additional information about faulty behavior and can be used for prediction. Unfortunately, system descriptions for the consistency-based method are often not suitable for abductive diagnosis. Methods for completing the fault models for abductive diagnosis have been suggested informally by Poole and by Cox et al. Here we formalize these methods by introducing a standard form for system descriptions. The properties of these methods are determined in relation to consistency-based diagnosis and compared to other ideas for integrating consistency-based and abductive diagnosis.

Batch combustion experiments were performed in a small bubbling fluidized-bed reactor with the objective of establishing the cause of reduced NO{sub x} emissions from pressurized fluidized bed combustion (PFBC). All variables except for pressure were kept constant in the experiments: fuel batch size, for example, was the same in experiments performed at three pressure levels (0.2, 1 and 2 MPa). Two different types of experiments were conducted: one using air diluted with nitrogen (4.5% O{sub 2}) for the purpose of determining the conversion of fuel N to NO{sub x}, and the other with NO-doped diluted air (800 ppM NO, 4.5% O{sub 2}) for the purpose of determining the reduction of bulk-gas NO{sub x} by the burning fuel. A large excess of combustion air was used in all experiments so as to keep the bulk-gas composition relatively unchanged by combustion products. Six different fuels were studied: a bituminous coal, coke prepared from the same coal, three specialty cokes (one of which contained 10 wt % N) and graphite (0%N). The straight-air combustion experiments showed that the conversion of fuel-N to NO{sub x} dropped with increasing pressure (at constant fuel concentration in the bed). The NO-doped combustion experiments showed significantly increased NO{sub …

In logic-based diagnosis, the consistency-based method is used to determine the possible sets of faulty devices. If the fault models of the devices are incomplete or nondeterministic, then this method does not necessarily yield abductive explanations of system behavior. Such explanations give additional information about faulty behavior and can be used for prediction. Unfortunately, system descriptions for the consistency-based method are often not suitable for abductive diagnosis. Methods for completing the fault models for abductive diagnosis have been suggested informally by Poole and by Cox et al. Here we formalize these methods by introducing a standard form for system descriptions. The properties of these methods are determined in relation to consistency-based diagnosis and compared to other ideas for integrating consistency-based and abductive diagnosis.

The objectives of this project are to study the effect of pretreatment methods on the two-stage liquefaction process. In particular, the effects of dispersed catalysts and carbon monoxide atmospheres on a coal liquefaction process. The project is divided into three technical tasks. Task 1 and 2 deal with the analyses and liquefaction experiments, respectively, whereas Task 3 deals with the economic impact of utilizing the pretreatment methods. This quarter we concentrated on Tasks 1 and 2, which are summarized below. Samples of Black Thunder Subbituminous coal and the corresponding recycle solvent were received from Wilsonville. These particular samples were obtained from run {number_sign}263. The samples were analyzed for a proximate and ultimate analysis, as well as heat content. We have also received samples of the iron oxide used in the Wilsonville coal liquefaction runs as well as Shell 324 and molyVanL for use in future runs. Base-line screening experiments were conducted in 300 mL autoclaves using 2.5 parts by weight of recycle solvent to 1 part Black Thunder coal at 425{degree}C for one hour. In preliminary results for the comparison of iron oxide and iron carbonyl, the iron carbonyl appears to give higher yields to soluble products.

None