Search Documents

By Stephen C. Tadolini, Gabriel S. Esterhuizen, Anthony T. Iannacchione

A Major Hazard Risk Assessment (MHRA) was developed in Australia after a series of mine disasters in the 1990’s. A MHRA is used to help prevent major hazards, i.e. fire, explosion, wind-blast, outbursts, spontaneous combustion, roof instability and chemical and hazardous substances, from injuring miners. A MHRA is a structured process that identifies the characteristics of major hazards, assesses and ranks the risk they present, and evaluates engineering and administrative controls to mitigate them. These controls typically consists of a broad spectrum of prevention, monitoring, first response, and emergency response techniques and helps to move an operation from a reactive to a proactive approach towards safety. This paper documents a MHRA performed at an underground mine where strata instabilities and fire hazards may threaten the condition of its escapeways. The objective of this MHRA is to 1) identify what hazards could affect the egress through the mine’s escapeways, 2) determine what unwanted events pose the greatest threat for the mine, and 3) recommend a plan to prevent or recover from the potential disruption of egress through the escapeway. The plan provides information on the key existing controls that should be monitored and audited, and makes recommendation of new potential controls to further reduce related risks. By documenting the use of MHRA to this specific ground control issue, this paper provides a framework for others to judge the merits of this approach and to help design and perform these activities.

By G. W. Tracy, R. D. Carter

Numerical calculations were made to determine the behavior of reservoirs with high-pressure drawdown and wide well spacing where the initial productivity is low and the wells are completed by hydraulic fracturing. The two-phase flow equations were solved for the flow into a single well. This well was assumed to be producing from a reservoir with hydraulically created horizontal fractures (four different systems with fractures were studied). For comparison purposes, additional two-phase flow calculations were made assuming a reservoir with uniform rock properties. The two-phase flow results were also compared with the conventional calculation methods, which do not include the effect of saturation gradients resulting from a simultaneous flow of oil and gas which are normal to this type reservoir. It was found that the conventional methods predict (1) a high and too optimistic value of ultimate recovery, (2) a high producing rate and a high reservoir pressure at a given oil recovery and (3) a low trend of gas-oil ratio with oil recovery. Included in the two-phase flow calculations were provisions to control the oil production rate by an allowable rate and, also, by a gas-oil ratio penalty rule. For the systems with hydraulic fractures, the producing rate was controlled by the gas-oil ratio penalty rule for most of the life. This is in contrast to the system with uniform rock properties which went "on decline" almost immediately. An unexpected characteristic of the systems which included fractures was the early rise in producing gas-oil ratio from 730 cu ft/bbl to approximately 1,200 cu ft/bbl, followed by a "leveling off" before the normally expected gas-oil ratio rise began. Additional features which are a result of hydraulic fracturing are (I) greater ultimate recovery, (2) higher average producing rates and (3) a lower average reservoir pressure at a given oil recovery. INTRODUCTION Some oil fields discovered during the past few years are producing from certain volumetric ally controlled reservoirs (often referred to as solution or internal gas- drive reservoirs) which are characterized by high-pressure drawdown at the wells. Since the available pressure drawdown at a well is limited by the static reservoir pressure and the producing rate is controlled by the available drawdown, wells completed in this type of reservoir usually produce at a rate less than the allowable from the time of completion. Because of this, this type of reservoir is referred to as a low productivity reservoir. Economic considerations require the use of wide well spacing and well stimulation by hydraulic fracturing to make commercial wells in this type of reservoir. Performance predictions for volumetrically controlled reservoirs have been made using a combination of two standard equations. 1. The "Schilthuis" or "Muskat" type material balance equation is used to relate the average reservoir pressure and the cumulative oil recovery. 2. The results from the material balance equation and the productivity factor as described by Pirson' are used to relate the cumulative recovery with producing rate and time. The material balance equation assumed uniform pressure and liquid saturation conditions throughout a reservoir. The steady-state radial flow formula allows for a pressure gradient toward a well but assumes uniform liquid saturation. These calculation methods are adequate for application to reservoirs wherein the drawdown at the well to realize satisfactory producing rates is small compared to the total pressure. In low productivity, volumetrically controlled reservoirs, the pressure drawdown at the well is large cornpared to the total pressure. Although a precise number cannot be given for the magnitude of a large pressure drawdown, values in excess of 1,000 psi would definitely be included. For practical considerations, this usually occurs when the formation flow capacity is less than about 100 md-ft. However, this limit of formation flow capacity will vary with the well producing rate. The low pressure in the neighborhood of the well which results from a high drawdown causes evolution of large volumes of gas. This causes the gas saturation to be higher near the well than at a greater distance—-hence, a non-uniform gas saturation. Also, the relationship between the relative permeability to oil (K/K) and gas saturation is nonlinear but decreases approximately in an exponential way with increases in gas saturation. Because of this, the following chain reaction is established.

By Thomas E. Marshall, Michael Dunn, L. J. Prosser, Anthony T. Iannacchione

During the past few years, the Pittsburgh Research Laboratory of the National Institute for Occupational Safety and Health (NIOSH) examined and characterized conditions at a majority of the underground stone mines in the United States. Observations at these mines revealed a limited degree of roof monitoring beyond visual inspection and sounding. When monitors are used, they typically require the miner to measure movement at the roof. If conditions are unstable, the mine rmay be in a hazardous situation while recording data. Based on this scenario, researchers surmised that a simple, inexpensive monitor with the capability of recording data at a distance from the mine roof would be a safer way to gain the information. Additionally, more widespread use of monitors could potentially lead to better understanding of roof movement in general. A monitor to meet this need was designed, tested, and subsequently improved as experience was gained in its’ use at a number of underground stone mines. The Roof Monitoring Safety System (RMSS) can provide an initial indication of movement in roof beams. By understanding and measuring roof movement in underground mines, the potential for injuries and fatalities to mine workers from falls of ground can be reduced. Also, officials at a mine with a history of data are better prepared to make a decision on remedial actions in the event of ground falls. This paper will outline the evolution of the RMSS and how it can be used in a comprehensive proactive ground control safety program. Also included is a case history describing how the RMSS was used in an evaluation of the effectiveness of a mechanical impact scaling machine at an operating limestone mine.

By Terry G. Montgomery, Eric S. Weiss, John E. Going, Kenneth L. Cashdollar

The Pittsburgh Research Laboratory (PRL) of the National Institute for Occupational Safety and Health (NIOSH) and the Mine Safety and Health Administration (MSHA) conducted joint research on dust explosions by studying post-explosion dust samples. The samples were collected after full-scale explosions at the PRL Lake Lynn Experimental Mine (LLEM), and after laboratory explosions in the PRL 20-L chamber and the Fike I m3 chamber. The dusts studied included both high- and low-volatile bituminous coals. Low temperature ashing for 24 h at 515 °C was used to measure the incombustible content of the dust before and after the explosions. The data showed that the post-explosion incombustible content was always as high as, or higher than the initial incombustible content. The MSHA alcohol coking test was used to determine the amount of coked dust in the post-explosion samples. The results showed that almost all coal dust that was suspended within the explosion flame produced significant amounts of coke. Measurements of floor dust concentrations after LLEM explosions were compared with the initial dust loadings to determine the transport distance of dust during an explosion. All these data will be useful in future forensic investigations of accidental dust explosions in coal mines, or elsewhere.

By Alan Zeni

Blind shaft drilling equipment has not changed substantially since the 1970s. Since it resides in a niche market between conventional shaft sinking and raise boring, the relatively small demand has resulted in some stagnation of development. In the 1990s the technology found a larger foothold in foundation and deep piling work for civil construction. WIRTH GmbH from Erkelenz, Germany has been servicing this market for many years with pile top drilling rigs and tools. Frontier-Kemper Constructors, Inc. recently entered into an agreement with WIRTH to develop a new generation of large diameter blind drilling equipment more specifically aimed at mining needs. The, recently completed, DHI-240 is the first rig to apply the most recent technology. This machine made its commercial debut in the fall of 2008. A description of the technology and a short case study of the first project will be described

By Bill M. Stewart, Phil W. Patton, Curtis C. Clark

Back injuries from handling materials in underground mines continue to be a major safety problem. In spite of the ingenuity of many people and the development of numerous mechanized aids, the number of materials-handling injuries remains second only to the number of roof fall injuries in underground coalmines. Relocation and repositioning of electrical cable, conveyor belt parts, and roof bolt sup-plies in particular are the sources of significant numbers of back injuries. To help reduce such injuries, researchers at the Spokane Research Laboratory of the National Institute for Occupational Safety and Health are examining Mine Safety and Health Administration accident data to determine correlations between materials-handling tasks and the number of back injuries. Also being investigated are new technologies used in underground mines in the United States. Equipment is being developed or modified that would replace the necessity of doing lifting tasks manually. A Coleman manipulator was tested, and modifications were made to make it more suitable for underground mine use. To reduce or eliminate the need to manually clean off materials that commonly plug grizzly openings, a track-guided pincher arm device was developed. Oversized rock can be broken with the pincher arms in the up position, and the arms can be lowered to grab and remove debris. The arm can also be used in a sweeping action to remove cohesive fines that may bridge grizzly openings.

By Deno Pappas, Chrisropher Mark, John Rusnak, Linda McWilliams

Mine planning for a new reserve is based on information obtained from exploratory coreholes. A critical component of an exploration program is the geotechnical evaluation. Poor assumptions about roof conditions greatly add to the risks of mining. Rock mechanics testing is central to a geotechnical exploration program. Typically, 3 to 5 uniaxial compressive strength (UCS) tests are made to characterize a particular roof unit at a given corehole. The average (mean) of these tests is taken as “The UCS” for that location. Isopach contour maps are then used to show spatial trends in roof strength. Two issues are raised by this traditional approach. The first is due to the large variability in UCS values that is typical even within a single unit from a single hole. The average UCS might be higher at Corehole A than it is at Corehole B, but the difference may not be statistically significant. The second issue is whether widely spaced coreholes can identify valid spatial trends in rock strength. The answer depends upon whether rock strength changes over distances that are longer or shorter than the corehole spacing. This is a classical geostatistical problem. While geostatics have been used to investigate many coal quality parameters, they have seldom been used to evaluate rock strength. This paper describes an extensive investigation of these issues conducted by the National Institute for Occupational Safety and Health (NIOSH) in collaboration with Peabody Energy. The study employed the Peabody Rock Mechanics Data Base which contains more than 10,000 individual test results. Data from four important roof units were subjected to statistical analysis: • Brereton Limestone above the Herrin 6 seam (Illinois) • Turner Mine Shale above the No. 9 Seam (Kentucky) • Sandstone above the Eagle Coal (West Virginia) • Shale above the Eagle Coal (West Virginia) The study did not find significant spatial trends in rock strength in any of the cases. Perhaps there are none, or perhaps the exploration coreholes were just too far apart to see them. These results have valuable implications for the design of geotechical exploration programs.

By A. A. Stripling

The Cretaceous limestone wells of the Mara/Maru-caibo Dist. of Venezuela are extremely prolific producers. To maintain production on cessation of natural flow, large scale gas-lift operations were commenced involving high production rates on caring Flow. For the design of these gas-lift installations it was essential to have some knowledge of the pressure gradients involved in casing flow, so that the required injection pressures, optimum gas injection raws, etc., could be forecast. This paper presents a method of calculating these annular pressure gradients. Basically the method makes use of an energy balance equation coupled with an empirical energy loss factor derived from field data. The calculations have been put in a form suitable for "punch card-type" calculating machines. A set of gradient curves for Mara/Maracaibo conditions covering flow rates up to 12,000 B/D, and GOR's of 500 to 2,000 cu fd/bbl is presented. From these curves the flowing BHP can be predicted from surface data and the vertical pow performance under varying conditions of gas injection can be estimated. The method has been applied to La Paz and Mam gas-lift operations for over three years and has given consistently accurate results. INTRODUCTION The Cretaceous limestone fields of the Mara/Maracaibo area in Venezuela are characterized by the presence of widely distributed fissure systems. Under these conditions individual wells frequently have extremely high potentials and such wells are generally flowed on the casing-tubing annulus in order to reduce the pressure loss in vertical flow. To maintain production on cessation of natural flow, an extensive gas-lift system was planned. As a first approach to this gas-lift design problem, an attempt was made to construct a set of empirical flowing gradient curves from the available subsurface pressure data. Such a method has been used successfully for tubing-flow conditions in California by Gilbert'. Unfortunately, the empirical construction of casing-flow gradient curves presents additional difficulties, as pres- sure bombs cannot bc run in the casing-tubing annulus against high rates of flow. Thus, the Full flowing gradients cannot be defined by actual measurement and the best that can be done is to obtain spot pressures at the tubing shoe. Under these conditions a vast accumulation of BHP data is required to construct a set of gradients. To the La Paz and Mara fields the majority of the available flowing pressure measurements fall into a relatively narrow GOR range, in the neighborhood of the solution ratio. While there appeared to be reasonable prospects of constructing gradient curves for this limited range, measurements at higher GOR's, which were of major interest for gas-lift operations, were relatively few and the prospects of constructing empirical curves cor~spondingly remote. As an alternative approach, the possibilities of a mathematical analysis of vertical flow conditions were investigated. May and Laird2fV n 1934 presented an excellent analysis of the vertical-flow conditions in the Anglo-Iranian Oil Co.'s fields. Their analysis was based on an energy-balance equation and calculated the losses due to kinetic energy, friction and slippage separately. By combining these losses with the total calculated available energy, the required depth-pressure traverse could be deduced. On applying the method to La Paz conditions, good results were obtained for tubing flow at high tubing-head pressures. However, for lower wellhead pressures (300 psi and under) the discrepancy between calculated and measured values increased appreciably. It was considered that this was probably due to energy losses caused by slippage or liquid hang-up in a continuous gas phase, for which it was difficult to visualize any analytical approach. In 1952 Poettmann and Carpenter4 presented a method of pressure gradient calculation which, while based on an energy-balance equation similar to that employed by May and Laird, made no attempt to evaluate the various components making up the total energy loss resulting from irreversibilities of flowing fluids. Instead, they proposed a method of analysis which assumed that all significant losses for multiphase flow could be correlated in a form similar to that of the Fanning equation for frictional losses in single-phase flow. They then derived an empirical relationship linking measurable field data with a factor which, when applied to the standard form of the Fanning equation, would enable the energy losses to be determined. The total available

By Robert F. Randolph

Introduction This paper is a descriptive analysis of injuries and fatalities at small surface bituminous coalmining companies in Pennsylvania. The purpose of analyzing a specific industry sector and location is to identify issues that might be missed in a wider analysis. Other studies of small mine safety (e.g., Peters and Fotta, 1994, Randolph and Boldt, 1997) have shown that the relationship between safety and mine size is not a simple one. Some injury types occur at higher rates at small mines, while others occur at lower rates. Consequently, an analysis limited to a specific type of mining can help identify hazards and issues specific to that sector. The results of this analysis will be useful for directing training, intervention, and research efforts towards the most prevalent causal factors and incident types. Procedures and Definitions Data sources The primary sources of data for this analysis were files provided by MSHA containing information collected under Title 30 Code of Federal Regulations, Part 50. These files, commonly referred to as “Part 50 data”, are provided though MSHA’s Program Evaluation and Information Resources directorate. Two primary types of injuries were selected from the Part 50data: Fatalities (MSHA “degree of injury” code 1) and lost-time injuries, otherwise known as nonfatal with days lost (MSHA “degree of injury” codes 2 through 5). Lost-time injuries involve either actual days away from work or days of restricted activity or both. Fatality and lost-time injury incidence rates were calculated using the standard MSHA formula: Number of injuries per 200,000 employee-hours. Additional information on company sizes and mine ownership was provided by the Pennsylvania State University College of Earth and Mineral Sciences. Time span A five-year time span was selected for this study. More than a single year of safety experience is usually needed to capture a broad range of incident types. Many types of injuries, especially fatal injuries, can be extremely rare, and may be missed if too short a time span is selected. However, a time span that extends too far back may begin to suffer historical threats to validity. That is, incidents that occurred several years ago may be caused by hazards that have since been remedied or may be related to mining practices and conditions that no longer exist. A five-year span offers a practical compromise between these two concerns. Consequently, the MSHA Part 50 data files from 1993 through 1997 were used. The 1997 file was in “preliminary” form at the time of the study, so small variations from the final 1997 “close out” file can be expected. Based on past experience, we do not expect the close-out variations to have a significant impact on the findings.

By James P. Rider, Jay F. Colinet

The National Institute for Occupational Safety and Health (NIOSH) has conducted benchmarking surveys at longwall operations across the country to identify current operating practices and the types of controls being used. Gravimetric and instantaneous dust sampling was completed to quantify dust generation from major sources and determine the relative effectiveness of the different control technologies. Results from the underground dust surveys will be presented as an update on longwall dust control technology and operating practices.

By J. D. Rice, S. C. Pitzer, C. E. Thomas

Interpretation of pressure build-up data obtained in the conventional manner has often been difficult because of the deviation from theoretical behavior. Major causes of this deviation have been attributed to damage and afterflow, and to fluid redistribution in the wellbore which, in extreme cases, can result in a pressure hump in the early portion of the build-up curve. Theoretical investigations show that bottom-hole pressure is definitely influenced by phase redistribution in the tubing column during surface shut-in tests, and that the magnitude of the effect is sensitive to the producing gas-oil ratio and stabilized rate of flow in the well. For field experiments a wire-line tubing packer, which can be run in the tubing against a stabilized flow rate, was developed for bottom-hole shut-in tests. By use of this bottom-hole .shut-in method, pressure humps previously observed in surface shut-in tests were completely eliminated and the effects of afterflow minimized. Analog and digital computer studies have been made to obtain theoretical curves for comparison with field curves, and remarkable agreement between the results of bottom-hole shut-in tests and theoretical curves has been obtained. INTRODUCTION Pressure build-up data from shut-in wells have been used by the petroleum industry to determine the permeability of the formation, to estimate wellbore damage, to evaluate the static reservoir pressure and to speculate reservoir volume. Calculation of these factors is based on methods of analysis developed for theoretical systems whose build-up curves have a charac- teristic shape when the wells are shut-in at the sand faCe.1,2,3,5,8,10 However, field curves obtained by conventional surface shut-in methods do not always exhibit this characteristic. Such factors as stratification, rock heterogeneities and irregular reservoir geometry can cause the character of a build-up curve to deviate from that predicted by theory for a simple system. In addition, all field build-up data are affected by the methods utilized in obtaining the measurements. Conventional surface shut-in techniques allow two effects to occur which contribute directly to the manner in which pressure builds up at the sand face. The first, afterflow, has been recognized for some time and methods are available for estimating the magnitude of its effect and for correcting its influence.' The second, phase segregation in the tubing column after shut-in, has been reported only recently and appears to exert a considerable influence on the character of field build-up curves.!' In extreme cases, phase segregation can produce "pressure humps" in the early portion of the build-up curve. Such curves have been considered anomalous and have defied analysis. It is the purpose of this paper to discuss the effect of phase segregation in the tubing string during build-up and to present a bottom-hole shut-in technique for obtaining field build-up data which minimizes the influence of afterflow and eliminates phase segregation effects. It will be shown that reliable reservoir information can be calculated from build-up measurements obtained using this method, even though conventional surface shut-in tests on the same wells yield anomalous data. DISCUSSION OF PHASE REDISTRIBUTION EFFECTS Some of the effects of phase redistribution on pressure build-up curves have been described previously by Matthews and Stegemeier" who have presented evidence that phase segregation is responsible for the pres-

By R. L. Derick

On a regular basis, mine personnel should obtain air samples directly from active underground areas as well as remotely from sealed locations. These samples provide mine officials with data relative to the makeup of the various atmospheres. Should conditions underground change, these results can provide a baseline that can assist officials in determining the extent of change. Relating the concentrations of individual gases to time, applying gases to equations, and examining their change over time are tools that have been successfully used to determine if a heating event exists and, if so, the extent of the emergency. These atmospheric status equations have either been developed for the mining industry or adapted from other industries. They serve as tools to assist personnel in determining the condition of underground mine atmospheres. As long as the problem area continues to be ventilated, certain gases, gas ratios, and equations can be used to determine if a fire exists, what type of material could be burning, and what quantity of coal is burning. Once the mine or the involved section has been temporarily sealed, gases, ratios, and gas-derived equations can assist personnel in determining how conditions within the sealed volume are changing over time. To alleviate potential explosions when fresh air is once again introduced into the sealed volume, equations and a diagram have been developed to predict the likelihood of an explosive atmosphere. Similarly, prior to commencing recovery of a mine or section, equations exist that can assist in determining if the fire will rekindle when air is reintroduced. This paper serves as an overview to remind and/or instruct readers about gas-sampling methodologies and gas analyses to assist in determining the status of underground atmospheres.

By S. J. Schatzel, W. P. Diamond

High methane emissions originating from the active face areas and from the fractured formations overlying and underlying the mined coalbed can adversely affect both safety and productivity in underground coal mines. Since ventilation alone may not be sufficient to control the methane levels in the longwall mining environment, gob gas ventholes have become a standard supplementary methane control option in many mines. As mines progress into deeper and gassier coalbeds, or as longwall panel size increases, ventilation and gob gas ventholes together may not be sufficient to maintain methane levels within statutory limits. To decrease the risk associated with methane emissions under these circumstances, in-seam horizontal methane drainage is often used to reduce the gas content of the coalbed prior to mining. Horizontal methane drainage borehole completion designs, drilling strategies, and degasification lead times may need to be adjusted for site-specific conditions due to mine design, geology, and the gas content of the coalbed. This study investigates different horizontal methane drainage borehole patterns, borehole lengths, and degasification times prior to and during panel extraction to evaluate their effectiveness in reducing methane emissions using a "dynamic" 3D reservoir modeling of a 38 1-m wide longwall panel operating in the Pittsburgh coalbed. Results of this study showed that dual and tri-lateral boreholes are more effective in decreasing emissions and in shielding the entries compared to fewer shorter, cross-panel, horizontal boreholes parallel to the longwall face. Modeling results showed that after 12 months of pre-mining methane drainage, the average longwall face emission rates can be reduced by as much as 10.3 m3/min and 6.8 m3/min using tri- and dual-lateral boreholes, respectively. It was also shown that if pre-mining methane drainage time is short, it is important to continue methane drainage during the panel extraction to maximize reductions in longwall face emissions since additional face emission reductions achieved during this period can be comparable to pre-mining degasification.

By R. G. Gürtunca, J. A. Breslin

This paper briefly describes the progress in mine safety in the United States, with emphasis on recent events that have followed since the explosion at the Sago Mine in January 2006. Legislation following the mine accidents last year and areas of work carried out by the National Institute for Occupational Safety and Health (NIOSH) related to the accidents are discussed. American coal production in 2006 was 1,150 million short tons. Most of this coal is consumed domestically, with annual exports of coal about 50 million tons. One third of annual production is produced at surface mines and the rest from underground. This coal is produced by about 82,000 miners working at 1,400 coal mines. Major improvements in coal mine safety have been made in the US since the Federal Mine Safety and Health Act of 1977. The fatal injury rate in underground coal mines between 1977 and 2004 was reduced by 47.8% to 0.036. The annual number of fatalities dropped from 112 to 14 during that period. Likewise, the non-fatal days-lost injury rate was reduced by 42.6% from 10.87 to 6.24. The improvements in mine safety during this period can be seen in the graph of underground coal mine explosion fatalities and injuries from 1980 to present, which is shown in Figure 1. Between 1984 and 1993 (9 year period) there were 5 explosions and one major mine fire, for a total of 55 fatalities. Over the seven year period from 1993 to 1999 there were no fatalities from coal mine fires and explosions. However, in the six years since 2000 there have been 5 explosions and one fire, which have caused 37 fatalities. [ ]

Pillars with width to height ratios of less than 1.0 are frequently created in underground hard rock mines. The strength of slender pillars can be estimated using empirically developed equations. However, the equations can provide variable results when the width-to-height ratios approach 0.5. This paper investigates some of the issues affecting pillar strength at low width-to-height ratios in hard brittle rock. The investigation includes an evaluation of empirical pillar strength data presented in the literature and observations of pillar performance in underground limestone mines in the eastern United States, supplemented by numerical modeling in which failure processes and sensitivity of slender pillars to variations in rock mass properties are evaluated. The results showed that the strength of slender pillars is more variable than that of wider pillars. The numerical model results demonstrated the increasing role of brittle rock failure in slender pillar strength. The absence of confinement in slender pillars can result in a fully brittle failure process, while wider pillars fail in a combined brittle and shearing mode. The onset of spalling in slender pillars occurs at or near the ultimate strength, while this is not the case for wider pillars. Slender pillars are shown to be more sensitive to the presence of discontinuities than wider pillars, which can partly explain the increased variability of slender pillar strength. Two examples are presented, which illustrate failure initiation by brittle spalling and the sensitivity of slender pillars to the presence of discontinuities.

By L. F. Elkins

Water coming into wells with bottom water present in the Fosterton field, when their oil recovery was only 0.1 to 1.5 per cent of oil in place below the lowest perforation, confirms lack of shale barriers to vertical flow which also are not apparent in cores and electric logy. Low contrast of vertical and horizontal permeability, high mobility ratio betweerz oil and water zone and pos.rible rate effects due to gravity Combine to prevent easy analysis of water invasion pattern and efficiency by analogy or intuition. Theoretical analysis of performance including these factors using a new direct nurnerical .solution of potential distribution in a network led to unanticapted indications that water would under-run the entire oil reservoir with low volumetric coverage if supplied through natural influx or peripheral water injecfion below the oil-water contact. An unusual pattern of water injection is to he adopted in this recently unitized pool to provide more uniform water flooding in that two-thirds of the oil reservoir not underlain by bottom wafer. INTRODUCTION The Fosterton field, Southwestern Saskatchewan, Canada, was discovered in Jan., 1952. It produces at a depth of 3,100 ft from a pinch-out of the Roseray sand at the Lower Cretaceous-Jurassic unconformity. The field, covering 2,500 acres, contains 54 wells on 40-acre spacing. Development was essentially complete in June, 1954, but significant production did not start until May, 1955, when a pipeline outlet became available. Within one year rapid in- creases in gas-oil ratio and early encroachment of water created operating problems in the down-structure eastern half of the field and raised serious questions about reservoir performance and ultimate recovery. This paper describes the unusual behavior of water encroachment—on both well and reservoir bases—and the analyses made to interpret it. The field was unitized July 1, 1958, for water flooding. A somewhat unconventional