Why the same bend fails differently in dilute and dense phase conveying

In dilute phase the particles are suspended in a fast gas stream and travel largely independently of one another. In dense phase the material moves as slugs or a moving bed at much lower velocity, with particles in continuous contact. These are different transport physics, and they damage a pipe wall in different ways: the first by discrete high-energy impacts, the second by sustained sliding contact under load.

Many plants do not know with certainty which regime a given line runs in, particularly where a system was designed for one and has since been rerated, retrofitted or asked to carry a different product. Establishing this is the first technical step, and it is usually cheaper than the bend it saves.

In dilute phase the particles cannot follow the curve. They continue on their original path until they meet the outer wall, and the point where that happens is a short distance beyond the start of the bend, not at its midpoint. The result is a well-defined patch of loss on the outer radius, often surrounded by wall that looks nearly new.

Two consequences follow. Protecting the whole bend uniformly spends most of the budget where nothing is happening. And a wall-thickness survey that samples at the apex of the curve — the intuitive place to measure — can miss the thinning entirely and report a healthy bend shortly before it opens.

Dense phase is gentler on a plain bend: lower velocity means lower impact energy, and the wear spreads over a larger area rather than concentrating. But the moving bed is unforgiving of anything that interrupts it. A step at a flange, a weld protrusion, a mismatched internal diameter or a liner edge that stands proud will collect the attack that the plain wall no longer receives.

This inverts the usual priority. In a dense-phase line, the quality of the internal transitions — flush joints, matched bores, no protruding edges — often matters more than the hardness of the wear face. A cheaper liner installed flush can outlast a harder one installed with a step.

Conveying velocity is rarely what the design assumed for very long. Filters load, valves wear, a compressor is replaced, a product changes bulk density, a line is extended. A system commissioned in dense phase can drift toward dilute, and the wear pattern changes shape before anyone notices the energy bill.

Treat a change in where a bend fails as an instrument reading. If the loss moves from a broad area to a concentrated patch near the tangent, the line has sped up, and the answer is upstream in the air supply rather than in a harder bend. Replacing the bend without asking that question buys the same failure again, sooner.

Wear is symmetrical. The same impacts that erode the wall break the particles, and in friable products the fines generated at speed are a quality problem, a dust-handling problem and sometimes a saleability problem. In those lines the conveying velocity is already constrained by the product, and a wear solution that assumes you can simply accept high velocity is solving the wrong equation.

Where attrition matters, the wear discussion and the product-quality discussion are the same discussion, and they should be held with the same people. A velocity reduction that protects the product usually protects the pipe as well — the rare case where two objectives point the same way.

Record the regime, the product and its friability, the observed wear position on each failed bend, and any change in the air supply since commissioning. That record turns a bend order into a line assessment, and it frequently identifies two or three positions that account for most of the maintenance while the rest of the route needs nothing.

Protect those positions properly and leave the rest alone. Concentrating the specification where the physics concentrates the damage is almost always cheaper than lining a whole route to an average that describes no part of it.