Passive seismic measurements were taken before, during, and after a fracture stimulation treatment to monitor the fracture growth and optimize future fracture treatments. The seismic events created by the fracture treatment showed an asymmetrical east-west trend during the treatment, with wide variations in the locations of events. The passive seismic measurements support the previous belief that the fracture orientation for the field is east-west. However, the recorded events showed more complexity to the fracturing process than had been anticipated. The events showed a southwest trend toward a producing well along with the widely scattered events to the east. Neither, the 3D fracture simulator, pressure transient analysis, nor production injection data supports the very large fracture geometry of the passive seismic events.
Fracture lengths and heights from the passive seismic events varied along with the directions. The length of the wing to the southwest showed seismic events over 1200 feet from the well, while the east wing events only reached about 700 feet. The different wings also showed a large variation in the height of events. Although the treatment interval was 4820-4910 feet, seismic events occurred from 4550-4900 feet for the southwest wing and 4600-5008 feet for the east wing. The shorter length to the east is believed to be due to the well being offset to the east, by another injection well. The higher pore pressure, from the water injection, caused the fracture pressure to increase, thereby changing the direction of growth of the fracture.
The passive seismic results show that fully modeling the fracture process will need to incorporate a simulator that allows varying fracture parameters aerially as well as vertically. The 3D model, of the initial 1966 fracture treatment, showed a propped fracture length of 139 feet, with a propped area of 29,000 square feet. The prefracture falloff, which followed running a liner and a cement squeeze, showed an area of 15,000 square feet. The second 1995 treatment model results gave a length of 150 feet and an area of 26,000 square feet. The values for the last job match very closely the pressure transient results following the treatment, of 24,000 square feet, showing that the current model can predict fracture results adequately for most evaluation purposes. However, for nonuniform pressure gradients a more detailed areal model will be needed.