IN-LINE VORTEX-TYPE SEPARATOR WITH CONTROL SYSTEM BASED ON NEURAL NETWORK AND MOBILE PRE-WATER DISCHARGE UNIT Russian patent published in 2023 - IPC B04C5/00 B01D17/38 

FIELD: oil and gas production industry.

SUBSTANCE: invention can be used as a separation device in preliminary water discharge installations, oil treatment installations, and gas treatment installations. An in-line vortex-type separator with a control system based on a neural network contains a separating vortex chamber, which in the inlet part contains a tangentially located inlet pipe for the input flow, an axially located outlet pipe for the light phase, and on the other side has an outlet pipe for the heavy phase. Control electric valves are installed on both outlet piping lines of the separation vortex chamber. Coriolis flow meters, pressure sensors, temperature sensors and shut-off and control valves are connected to the hydraulic control circuit for in-pipe separation, and the dynamic process of in-pipe vortex separation is controlled by an automated system with an autonomous process mode adjustment module based on an artificial neural network with an intelligent controller based on a three-layer perceptron with a network predictive controller of the PID algorithm for processing high-frequency measurements of the residual water cut of the in-line moisture meter, with adjustable parameters of the control system for the electric drives of the valves of the in-line vortex-type separator. The mobile preliminary water discharge installation contains a pipe vortex sand separator, a pipe vortex gas separator, an in-pipe vortex separator with a control system based on a neural network, a three-phase separator unit, a water settling unit, a reagent dosing unit, connected in series by a pipeline system. The mobile installation is made in form of a single monoblock with a two-tier arrangement of technological blocks placed in frames with the dimensions of 40- and 20-foot cargo containers.

EFFECT: ensuring the process of effective separation of water-oil emulsion with an automatic control system for electric drives of regulatory elements based on an artificial neural network.

2 cl, 2 dwg