On high points — and on pairs
Building a water tower at the high point works like adding burial depth to the upstream pipeline: it flattens the crest, and a flatter crest is kind to the transient. But it buys this with an operating cost paid for the life of the system. The trade — a one-off investment in deeper burial against decades of running expense — deserves a very deliberate answer.
I am always asked why surge-anticipation valves and other emergency-handling valves come in pairs. My earlier answers never felt quite sufficient. Then the books reminded me: the organs that do real work — eyes, ears, nose, hands and feet — all come in twos. Redundancy is nature’s own choice for tools that matter. Suddenly the answer was easy.
Water hammer is energy, not volume
Does refilling the void after the air pocket collapses — after column separation — solve the water hammer? It does not. Water hammer is a wave, and a wave is energy. Its cure is not volumetric equivalence but energetic equivalence. Remember this. Always.
Which is why protection against a surge farther along the line calls for a larger surge vessel and a higher pre-charge pressure — more stored energy, and naturally a larger first cost. Local treatment near the far tail of a long main deserves proper consideration.
Every profile has its own surge
Lay the pipe one way and you get one surge; lay it another way and you get a different one. Obvious, once it is said.
And the real cure for a hump-backed profile is not to keep bolting on more protection devices — it is to deal with the knee itself. In the real world, of course, that is the hard part.
Protection works downstream. Not upstream.
Most surge protection devices protect what lies downstream of them, and do nothing for what lies upstream. So when a large and complex pump station shows vibration ahead of the protection, look inside the station instead: pipe supports, the way the pipework runs, and where air collects.
Air valves: not every one of them should be anti-surge
Line the whole route with anti-surge air valves and the filling of an empty main becomes painfully slow. Choose ordinary combination air valves at selected intermediate points instead.
A network is not a transmission main
From the surge point of view, supplying a distribution network and supplying a point-to-point main are very different problems. A network distributes: flow is always branching away, and there are many delivery nodes. Transmission is one point to another, over distance, with no draw-off in between.
In energy terms, surge is a transient conversion — the kinetic energy of the moving water turning into potential energy, into pressure. On a transmission main the classic energy wave runs out and back. A network bleeds that wave away: it shunts energy out of the system, and what travels onward keeps on diminishing. The more tangled the network, the less comes back; most often there is no return wave at all, or only a very faint one.
Why a network rarely sends the wave home
Then there is the mass of water. A network is always in the act of delivering. Users keep drawing even when the power fails, unaware the pumps have stopped. So in practice every network empties over a wide area once the pumping station goes down — and the water does not come back to the station.
Even when demand is low at the moment of the trip, the loops of a network send the pressure wave wandering in circles. On the large scale it reads as scattered small ripples in a pond, and it does not trouble the pump station.
A clean high point is a cheap high point
A straight, even line with one clear, regular high point is the easiest surge problem of all: the source is unambiguous and nothing interferes with it. Those schemes usually cost far less to protect.
On bladder vessels feeding a network
A bladder-type surge vessel on a plant that feeds a distribution network is worth a second thought. The network never stops drawing — not even when the plant loses power — so part of the network drains down before anything settles. What volume do you size the vessel for, then?
A hybrid vessel may be the better answer: no bladder, no compressor to make up air, and no sub-atmospheric pressure. It serves low-head systems best, and can also stand in for a two-way surge tower.
Transmission and distribution are different animals
Long-distance transmission is best kept out of distribution. Transmission lives on a steady, nearly constant flow; distribution changes every single day. A storage tank between the two is the clean way to decouple them.
And from the design desk: never run a system designed for full-pipe flow in partial flow — if only to protect yourself. Burst after burst, cavitation, corrosion, a pipeline ageing before its time… the accidents there are not bad luck. They are scheduled.
On high points — and on pairs
Building a water tower at the high point works like adding burial depth to the upstream pipeline: it flattens the crest, and a flatter crest is kind to the transient. But it buys this with an operating cost paid for the life of the system. The trade — a one-off investment in deeper burial against decades of running expense — deserves a very deliberate answer.
I am always asked why surge-anticipation valves and other emergency-handling valves come in pairs. My earlier answers never felt quite sufficient. Then the books reminded me: the organs that do real work — eyes, ears, nose, hands and feet — all come in twos. Redundancy is nature’s own choice for tools that matter. Suddenly the answer was easy.
Water hammer is energy, not volume
Does refilling the void after the air pocket collapses — after column separation — solve the water hammer? It does not. Water hammer is a wave, and a wave is energy. Its cure is not volumetric equivalence but energetic equivalence. Remember this. Always.
Which is why protection against a surge farther along the line calls for a larger surge vessel and a higher pre-charge pressure — more stored energy, and naturally a larger first cost. Local treatment near the far tail of a long main deserves proper consideration.
Every profile has its own surge
Lay the pipe one way and you get one surge; lay it another way and you get a different one. Obvious, once it is said.
And the real cure for a hump-backed profile is not to keep bolting on more protection devices — it is to deal with the knee itself. In the real world, of course, that is the hard part.
Protection works downstream. Not upstream.
Most surge protection devices protect what lies downstream of them, and do nothing for what lies upstream. So when a large and complex pump station shows vibration ahead of the protection, look inside the station instead: pipe supports, the way the pipework runs, and where air collects.
Air valves: not every one of them should be anti-surge
Line the whole route with anti-surge air valves and the filling of an empty main becomes painfully slow. Choose ordinary combination air valves at selected intermediate points instead.
A network is not a transmission main
From the surge point of view, supplying a distribution network and supplying a point-to-point main are very different problems. A network distributes: flow is always branching away, and there are many delivery nodes. Transmission is one point to another, over distance, with no draw-off in between.
In energy terms, surge is a transient conversion — the kinetic energy of the moving water turning into potential energy, into pressure. On a transmission main the classic energy wave runs out and back. A network bleeds that wave away: it shunts energy out of the system, and what travels onward keeps on diminishing. The more tangled the network, the less comes back; most often there is no return wave at all, or only a very faint one.
Why a network rarely sends the wave home
Then there is the mass of water. A network is always in the act of delivering. Users keep drawing even when the power fails, unaware the pumps have stopped. So in practice every network empties over a wide area once the pumping station goes down — and the water does not come back to the station.
Even when demand is low at the moment of the trip, the loops of a network send the pressure wave wandering in circles. On the large scale it reads as scattered small ripples in a pond, and it does not trouble the pump station.
A clean high point is a cheap high point
A straight, even line with one clear, regular high point is the easiest surge problem of all: the source is unambiguous and nothing interferes with it. Those schemes usually cost far less to protect.
On bladder vessels feeding a network
A bladder-type surge vessel on a plant that feeds a distribution network is worth a second thought. The network never stops drawing — not even when the plant loses power — so part of the network drains down before anything settles. What volume do you size the vessel for, then?
A hybrid vessel may be the better answer: no bladder, no compressor to make up air, and no sub-atmospheric pressure. It serves low-head systems best, and can also stand in for a two-way surge tower.
Transmission and distribution are different animals
Long-distance transmission is best kept out of distribution. Transmission lives on a steady, nearly constant flow; distribution changes every single day. A storage tank between the two is the clean way to decouple them.
And from the design desk: never run a system designed for full-pipe flow in partial flow — if only to protect yourself. Burst after burst, cavitation, corrosion, a pipeline ageing before its time… the accidents there are not bad luck. They are scheduled.
