You have a "2-phase" soft starter, although interestingly enough, ABB appears to have now taken the tactic from Siemens' book one step further and no longer tells you this! The reason I know is this statement in their literature:
ABB Brochure said:
Sophisticated algorithm eliminating the DC-component and thereby providing excellent starting performance.
There is no need to make this statement if it is a full 3 phase soft starter, as there is no "DC Component" to be concerned about. But on a 2-phase soft starter, where there are SCRs controlling the voltage on only 2 of the 3 phases, this is a concern. Certainly, the control algorithms developed by Siemens and now apparently copied by ABB help to mitigate this issue as much as possible, but their claim of "eliminating" it is a stretch.
What this means, ESPECIALLY when trying to control the Decel of the motor, is that the power going to the motor is inherently severely unbalanced while ramping, up or down. Ramping up a pump is usually something you want to do as fast as possible, so nobody sees the problem, but ramping down is done to eliminate water hammer. To do that, you must ramp as SLOWLY as possible to give the kinetic energy in the moving water time to dissipate, so that means finding a 'sweet spot" of torque that you hold at until the check valve closes slowly and gently. If you linger there too long, you are heating up the motor disproportionately to the amount of average current going to it. Fortunately for you, you are using a bi-metal thermal OL in the circuit in the form of your adjustable circuit breaker, and IT is picking up this current imbalance, biasing the trip curve, and tripping early, or at least earlier than the soft starter, which most likely DISABLES the OL protection during Decel to try to hide the inherent problem in using that design.
Another potential pitfall of this design concept is that at all times, you ALREADY have 1/2 of an uncontrolled current flow fault to the motor, because one of your 3 phases, usually the center pole, is live at all times (because it's cheaper that way). In a true 6 SR design, you must have 2 SCRs in opposing phases fail shorted before there is uncontrolled current flow to the motor, which has about the same likelihood of happening as a pair welded contacts. But in a 2-phase (4 SCR) design, the "opposing phase" is ALWAYS already hot, so if even ONE of the remaining SCRs shorts, or more likely in this case, fails to gate off, then you IMMEDIATELY have uncontrolled power flowing to / through one winding of the motor, and that may be getting picked up by your breaker as a short circuit. So in other words, you may be seeing the first of a problem inside of the soft starter and it is not turning off one of the SCRs, which will eventually result in a more catastrophic failure. This means changing out the starter.
Welcome to the world of smaller-cheaper-faster.
"Will work for (the memory of) salami" 
