Two protection functions are implemented in the control system: a) Protection for DC fault on the rectifier and b) Protection for commutation failure on the inverter.

DC Fault Protection

A fault in the DC power system is detected on the rectifier side when the following two conditions occur:

The DC voltage falls below a specified threshold for a period of time (adjustable);
No low AC voltage is detected on the rectifier side (characterizing a fault on the AC side).

This detection is done in the “Low AC voltage detection” function described in the control system. When a fault is detected on the DC side, the rectifier α angle is forced to an adjustable value greater than 90^o so that the rectifier operates as an inverter to de-ionize the fault. This forced α is removed after an adjustable period of time. This operation will be repeated until the fault is cleared or the converter is blocked by the protection system after the operation has been repeated a number of times.

This function is implemented for simulating faults on the DC side. The function does not discriminate between the faults on the AC side of the inverter and commutation failures on the inverter, both causing a drop in direct voltage. One way to disable the protection is to assign a very large value to the detection time parameter. No action is taken on the inverter side since it is already in a safe condition.

Protection for commutation failure on the inverter

The protection against commutation failure is provided on the inverter only. The reason is the firing angle on the inverter side is large and, therefore, the inverter is more prone to commutation failures.

Figure 1

The detection of commutation failures is based on the following principle. In the normal state, the currents (in pu) on the AC side and on the DC side are nearly equal. A commutation failure instantaneously creates a short circuit on the DC side and the DC current increases quickly. The inverter then has a null voltage and hence, the current drops on the AC side.

When a commutation failure is detected, the protection decreases the upper limit of alpha by γ_cf . This keeps α away from the area where the risk of commutation failure is high. γ_cf then takes on the form shown in Figure 1. γ_cf increases quickly (adjustable first-order time constant) when a commutation failure is detected for quick protection but decreases more slowly (adjustable first-order time constant) to avoid successive commutation failures.

It is possible to simulate commutation failures by preventing the current from switching from one thyristor to its neighbor. To do so, a given thyristor is prevented from receiving the firing pulse for a given period of time.

Unblocking Sequence

There are two ways to unblock the converters (e.g. following a persistent DC fault): manually or automatically. When the converter is disabled (manually or automatically), the control system cancels the firing pulses to the valves. Additionally, a constant delay angle is imposed permanently.

Manual unblocking of a converter is initiated by the operator and takes effect only when the DC current is lower than a specified threshold (set at 0.10 p.u.). This current must then be decreased by lowering the current reference. All the reference signals have a rate of change limiter which limits their variation speed (adjustable). When the operator varies the reference, the effective reference will follow, but at a speed limited by the rate of change limiters.

The automatic disabling takes effect immediately to protect the system. It is controlled by the DC fault protection system.

Enabling (Start-up)

Unblocking can only be done manually. To enable the converter, the following must be done:

Set the current reference to zero;
Initiate the enabling order. Regular pulses will be sent to the thyristors (instead of pulses from bypass thyristors).
Set the current or voltage reference to the desired value. After unblocking, the converter will adjust to this given value with a certain delay depending on the limits imposed by the rate of change limiters. To maintain the current margin, the current ramp on the rectifier cannot be slower than that of the inverter at start-up.

Browser not supported