Testing New Submersible Pumps for Proper Sizing and Reduced Costs

Abstract

Summary This paper describes an ongoing program to improve overall submersible pump performance by Thums Long Beach Co., acting as contractor for the City of Long Beach, operator of the Long Beach Unit. Thums Long Beach Co. currently operates 700 submersible pump installations located on four manmade islands and one landfill pier location. The program began with spot testing of submersible pumps for Thums' use. It has evolved to 100% pump testing and the stipulation that only pumps with newly manufactured parts are acceptable. The primary goals of this program are to increase well production and to lower lifting costs. Critical to these goals is increasing the average length of run by using accurate pump-performance data to design equipment and by rejecting defective pumps pump-performance data to design equipment and by rejecting defective pumps before they are run. Increased production is realized from better designs. Lower lifting costs result from using more efficient pumps and a reduced frequency of pulling submersible equipment. Introduction Thums currently produces 65,000 BOPD [10.3 × 10(3) m3/d oil] and 375,000 BWPD [59.6 × 10(3) m3/d water] from the offshore portion of the Wilmington oil field. This production is obtained from nearly 750 wells, of which approximately 700 are produced by the downhole electrical submersible pump. Because of this large use of submersible pumps, a closer look at the performance of submersible pump equipment was initiated. Thums designed its own submersible pump installations and routinely adjusted the designs upward from those generated from vendor-supplied performance curves. Fig. 1 represents a typical submersible pump installation. Even with this adjustment, many designs fell short of reducing the producing bottomhole pressure (BHP) to the desired levels. Coupled with this, many instances of electrical or mechanical failures on startup and high motor loading during the pump run have occurred. These problems traditionally have been attributed to well conditions, but that explanation was considered to he unsatisfactory in many cases. Post failure equipment inspections indicated that a large proportion of submersible motor and cable pothead failures appeared to be pump-related. Besides the obvious pump conditions that resulted pump-related. Besides the obvious pump conditions that resulted in electrical failure, such as a locked pump, radial wear was often found in the hub area of the impeller and diffuser (see Fig. 2), as well as in the pump end bearings. The resulting shaft vibration from this wear was then believed to cause seal failure and eventual electrical failure in the motor or motor pothead. Because of these concerns and the unpact any improvement would have on pump performance and run life, a pilot test program was initiated in June 1984 to test 40 pumps before they were run in wells. What originally was intended as a short-term experiment has resulted in an ongoing project. Test Method Two basic testing methods are currently available to the industry. They involve orienting the pump on a horizontal test bench or in a vertical test well. Besides the orientation difference, other differences exist between individual test configurations, with the major variation being the use of either a speed/torque cell or power-consumption conversion to measure brake horsepower. The speed/torque cell is commonly used in a variety of industries and consists of a strain gauge and toothed-gear impulse counter that together measure the revolutions per minute and front-end torque load. The power-consumption technique uses a plot of a prime-mover's voltage and amperage as generated from calibrated brake-horsepower loads. The resulting plot allows a comparison between the calibrated loads and test loads by measuring kilowatt draw. The speed/torque cell is used in the horizontal test method; both techniques are used in the vertical method.