The fouling of heat transfer surfaces in kraft recovery boilers is a significant concern for the pulp and paper industry. The usual approach to controlling fouling is the use of so-called ``sootblowers,'' that utilize boiler steam to generate supersonic steam jets that are literally used to knock deposits off of the boiler tubes. About 3 to 10\% of the total steam produced in a recovery boiler is used for sootblowing. This high energy cost demands that they be operated as efficiently as possible. It is thus essential to devise improved strategies for maximizing sootblower efficiency and minimizing steam consumption. To achieve this, the behaviour of sootblower jets, and the effects of various parameters on sootblowing, must be well understood. This thesis documents a study of the performance of sootblower jets using numerical simulation; CFDLib 3.02, a CFD code from the Los Alamos National Laboratory, was used for the simulations. This work had two main parts. In the first part, sootblower jets that perform at the design condition (fully-expanded jets) were studied; in the second part, the study was extended to off-design (under/over-expanded) sootblower jets. In the first part, a compressibility-corrected version of CFDLib was validated against a wide range of available experimental data, of subsonic and fully-expanded supersonic free and impinging jets; simulations successfully predicted all of the cases. This compressibility-corrected model was then deemed suitable for modeling the fluid mechanics of fully-expanded sootblower jets, and so was used to study the effects of two parameters on sootblower jets: the lance pressure, and the rate of rotation of a sootblower. To study the effect of the lance pressure, numerical simulation was used to model fully-expanded sootblower jets corresponding to a range of lance pressures. To study the effect of rotation, the equations of motion were modified by adding the Coriolis and centrifugal terms, so that computations could be performed in a rotating frame of reference. Simulations were then run to study a fully-expanded sootblower jet operating at different rotation rates. The results indicate that sootblowers operate more efficiently at lower lance pressures, and that the rate of rotation does not significantly affect the structure of a sootblower jet. In the second part, the study was extended to sootblower jets not operating at the design condition. The compressibility-corrected code failed to properly simulate these under-over/expanded supersonic jets. A wide series of tests was carried out to determine that the problem was due to the turbulence model. The model was then modified to account for turbulence/shock wave interaction, by adding corrections to take into account shock unsteadiness and a realizability constraint. The new model yielded good agreement with some available measurements. The new model was then used to successfully predict some actual sootblower measurements, and to study the interaction of a sootblower jet with geometries similar to tube banks in recovery boilers. A parametric study was carried out to examine the effect of the offset between a sootblower jet and a tube bank, and of deposit size on a sootblower jet. The results indicate that the shock cell structure of a jet is only slightly affected by the offset, but that the size of a deposit strongly affects the pressure exerted by the impinging sootblower jet, which depends both on the jet shock cell structure, and on the location where the interaction occurs.

The fouling of heat transfer surfaces in kraft recovery boilers is a significant concern for the pulp and paper industry. The usual approach to controlling fouling is the use of so-called ``sootblowers,'' that utilize boiler steam to generate supersonic steam jets that are literally used to knock deposits off of the boiler tubes. About 3 to 10\% of the total steam produced in a recovery boiler is used for sootblowing. This high energy cost demands that they be operated as efficiently as possible. It is thus essential to devise improved strategies for maximizing sootblower efficiency and minimizing steam consumption. To achieve this, the behaviour of sootblower jets, and the effects of various parameters on sootblowing, must be well understood. This thesis documents a study of the performance of sootblower jets using numerical simulation; CFDLib 3.02, a CFD code from the Los Alamos National Laboratory, was used for the simulations. This work had two main parts. In the first part, sootblower jets that perform at the design condition (fully-expanded jets) were studied; in the second part, the study was extended to off-design (under/over-expanded) sootblower jets. In the first part, a compressibility-corrected version of CFDLib was validated against a wide range of available experimental data, of subsonic and fully-expanded supersonic free and impinging jets; simulations successfully predicted all of the cases. This compressibility-corrected model was then deemed suitable for modeling the fluid mechanics of fully-expanded sootblower jets, and so was used to study the effects of two parameters on sootblower jets: the lance pressure, and the rate of rotation of a sootblower. To study the effect of the lance pressure, numerical simulation was used to model fully-expanded sootblower jets corresponding to a range of lance pressures. To study the effect of rotation, the equations of motion were modified by adding the Coriolis and centrifugal terms, so that computations could be performed in a rotating frame of reference. Simulations were then run to study a fully-expanded sootblower jet operating at different rotation rates. The results indicate that sootblowers operate more efficiently at lower lance pressures, and that the rate of rotation does not significantly affect the structure of a sootblower jet. In the second part, the study was extended to sootblower jets not operating at the design condition. The compressibility-corrected code failed to properly simulate these under-over/expanded supersonic jets. A wide series of tests was carried out to determine that the problem was due to the turbulence model. The model was then modified to account for turbulence/shock wave interaction, by adding corrections to take into account shock unsteadiness and a realizability constraint. The new model yielded good agreement with some available measurements. The new model was then used to successfully predict some actual sootblower measurements, and to study the interaction of a sootblower jet with geometries similar to tube banks in recovery boilers. A parametric study was carried out to examine the effect of the offset between a sootblower jet and a tube bank, and of deposit size on a sootblower jet. The results indicate that the shock cell structure of a jet is only slightly affected by the offset, but that the size of a deposit strongly affects the pressure exerted by the impinging sootblower jet, which depends both on the jet shock cell structure, and on the location where the interaction occurs.