SIMULATION OF A HYBRID CONTROLLER FOR CONTROLLING PLASMA WELDING PARAMETERS

Abstract

One of the most common technological operations is welding of individual parts and blocks.
Welding is widely used in shipbuilding, aviation, defense and chemical industries, in the construction
of oil and gas pipelines. At the same time, very strict requirements are imposed on the quality of the
weld in terms of strength, absence of voids and cavities, operability at high pressures (up to 100 kGf /
cm2) and in a wide temperature range (± 50 ° C). Plasma (argon) welding meets these requirements
most fully. A brief analytical review on the research topic was carried out. It is shown that a promising
direction for the development of plasma welding control systems is the use of hybrid regulators
created on the basis of classical automatic control methods and fuzzy control, formalizing the average
knowledge of experts. The fuzzy component (expert knowledge) should be available for quick and
easy input into the controller. A block diagram and a model of a single channel of a hybrid controller
was developed in the Matlab Simulink environment. The current control channel was modeled using
a fuzzy controller from the Fuzzy Logic library, using the Mamdani fuzzy output algorithm. 19 variants
of linguistic and fuzzy variables were set, the surface of the variable membership function was
obtained. It should be noted that it is possible to quickly enter linguistic assessments of experts into
the memory of the hybrid controller. The behavior of hybrid controller models and standard PI and
PID controllers under a single step action was analyzed. The hybrid regulator provides significantly
better quality indicators (2.5-3 times) than standard regulators. The hybrid controller enters the
steady-state mode after 6s, the PID controller – after 13s, the PI controller - after 15s, and the standard
regulators have an overshoot (first emission) of up to 50%. Thus, the real possibility of constructing
a fuzzy hybrid controller with specified characteristics is shown. It is possible to implement a
hybrid controller in the form of an FPGA.