Apr-1999

Operational life improvements in modern hydroprocessing reactors

Major problems that occur during the service life of hydroprocessing reactors can result from existing manufacturing flaws and the high temperature and pressure hydrogen environment

Leslie P Antalffy and Michael B Knowles, Fluor Daniel
Takayasu Tahara, Japan Steel Works
Pran N Chaku, ABB Lummus Global

Article Summary

It was the mid-1920s that heralded the birth of modern hydroprocessing reactors. In Germany, quenched and tempered chrome molybdenum steels were used to fabricate reactors for high pressure hydrogenation plants used to produce transport fuels from coal. However, it was only during the 1960s that modern hydroprocessing reactor technology evolved with the startup of eight quenched and tempered 21/4Cr-1Mo reactors at Chevron’s Richmond, California, refinery.

It is estimated that today well over 1000 hydroprocessing reactors have been fabricated from 21/4Cr-1Mo alloy, some 40 from the new generation vanadium modified 3Cr-1Mo steel and five from the vanadium modified 21/4Cr-1Mo steel. Today, with hydrogen partial pressures ranging as high as 35Mpa in some applications, the new generation vanadium modified 21/4Cr-1Mo and 3Cr-1Mo steels exhibit service life improvements and in many cases cost advantages over the conventional 21/4Cr-1Mo steel in high temperature and high pressure hydroprocessing reactor applications.

Over the years, Japan Steel Works (JSW), a major fabricator of hydroprocessing reactors, undertook a programme of reactor inspection as well as industrial surveys related to hydroprocessing reactor operational damage. Recent information received from JSW (in private communications 1994–1998) quantifies the type of operational damage found in reactors fabricated since 1963 through 1991.

This information, which is categorised in Table 1, indicates that of 164 reactors fabricated between the ‘60s and early ‘90s which were inspected by JSW and others, 116 reactors were found to be damaged and in these there were 122 incidents of damage.

The four most significant damage categories in order of occurrence are; cracking resulting from sigma phase formation in the weld overlay at bed supports as well as at other attachments and at reactor inside shell surface locations; cracks in the main weld seams resulting mainly from existing weld defects which occurred during the initial fabrication process; and sigma phase formation cracking in gasket grooves and weld overlay disbonding. JSW has not yet carried out inspection on newer reactors fabricated after 1991.

Principal causes of damage
From Table 1 it becomes evident that the greatest incident of damage in reactors has resulted from sigma phase formation in the weld overlay. Sigma phase formation results from the transformation of delta ferrite in the weld overlay at service temperatures to an acicular morphology sigma phase resulting in significant reduction in ductility, thereby rendering the weld overlay brittle at temperatures below 150°C. As a consequence, crack formation and propagation in the weld overlay in highly stressed areas can occur during the cool down and subsequent start up (reheat) cycles of the reactor.

The presence of existing flaws and stress raisers in the weld area combined with stress and high temperature hydrogen diffusing into the metal resulting in a loss of ductility due to hydrogen embrittlement can all have a significant effect on crack formation and subsequent propagation.

The damage observed in these reactors generally results from the favourable conditions of hydrogen concentration, stress and defect shape and size to start the initial crack initiation and subsequent propagation. The cyclic effect of multiple shutdowns over the service life of the reactor results in pressure, temperature and stress variations which can further contribute to crack formation and propagation.

Damage considerations
The following sections highlight some of the considerations that play a role in initiating damage mechanisms that can ultimately lead to reactor in-service damage.

Fatigue
Fatigue failure in pressure vessels or components is a failure of the component after a number of stress cycles when the component is stressed to stress levels which are much lower than the ultimate tensile strength of the pressure vessel or its component. In pressure vessels the stresses that produce fatigue failure can be stresses from mechanical loading, such as the pressurisation of a vessel or a component, or a localised loading such as at a bed support; thermally induced stresses based on the temperature difference between any two adjacent points on the vessel or vessel component, or stresses induced by adjacent components having different coefficients of thermal expansion.

In the USA, almost all of the reactors in service are designed in accordance with American Society of Mechanical Engineers (ASME) Section VIII Division 2 Pressure Vessel Code, which gives comprehensive requirements for evaluating and designing for fatigue conditions. Fortunately, in most hydroprocessing reactor applications, the operating parameters of the reactor are such that pressure and thermal cycles resulting from operation over the life of the reactor are low enough in number and severity to preclude damage and subsequent failure from fatigue.

Although fatigue in itself may not be an important consideration in most hydroprocessing reactors, the cyclic nature of mechanical and thermal stresses working in an embrittled or hydrogen charged shell or weld overlay may lead to crack initiation and subsequent propagation at existing flaw sites.

Hydrogen effects
In the past, the 21/4Cr-1Mo alloy has mainly been used for the construction of hydroprocessing reactors because of its strength, toughness, creep rupture properties and its resistance to hydrogen attack as indicated in the American Petroleum Institute Publication 941 (API 941).