DE3219663C2 - - Google Patents

Description

The invention relates to a method for detection breakthroughs in an electrostatic precipitator that has a Rectifier, a high voltage transformer and a Actuator is fed from an AC voltage network, by comparing a voltage value with a reference value, where a deviation from the reference value for control of the actuator is used.

Such a method has become known from DE-OS 25 40 084. In doing so, certain in-phase voltage values successive half-waves with each other compared and deviations that are a fixed predetermined measure used as a criterion for a breakdown. The following formula is used:

U _{i (n)} - U _{i (n -1)} Δ U _R → breakdown.

U _{i (n)} is the voltage value at a specific time i (phase position) of the half-wave n , while U _{i (n -1) is} the voltage value at the corresponding time i (same phase position) of the preceding half-wave (n -1). Δ U _R is a fixed reference value.

This method is considered to be disadvantageous in that it cannot be used over the entire voltage range, because it is not possible to specify a constant reference value Δ U _R which is valid for the entire range. DE-OS 29 49 752 therefore proposes that a fixed partial amount of the voltage measured in each case be used as Δ U _R : Δ U _R = x · U _{i (n)} , with x in the range from 0.05 to 0, 20, ie 5 to 20% of the voltage of the half-wave n to be monitored, which is determined by the phase position i . The procedure can be written as a formula using the indexing system already given as follows:

U _{i (n)} - U _{i (n -1)} x · U _{i (n)} → punch.

The value X is determined when a system is started up, whereby a value of 0.125 has proven to be suitable for an electrical separator.

For a differentiated recording of breakthroughs and a correspondingly coordinated control of the actuator the control device, this type of specification of the reference value but still considered too coarse and inelastic. In particular, not all can be created in this way Take changes in the course of the half-waves into account, that when operating an electrical separator by various Influences can be caused.

There is therefore the task of the last-mentioned method continue to improve, d. H. the detection of breakthroughs to be refined and more sensitive in terms of control technology to respond to it. In particular, too such voltage changes that are not based on punctures based, better identified and their influence on the Voltage control can be switched off.  

To solve this problem, a method of the type mentioned is proposed, which is characterized in that the absolute value of a voltage difference Δ U _j = U _{i +1} - U _i from two successive measurements of the time interval Δ t _j and as a reference value of the voltage value absolute amount of a voltage difference Δ U _{j -1} = U _i - U _{i -1} from two successive measurements of the previous time interval Δ t _{j -1} are used and the absolute amount of the difference between voltage value and reference value is evaluated as a breakdown if a predetermined value x is exceeded and is used to control the control.

Written as a formula, the breakthrough criterion is:

U _{i +1} - U _i | - | U _i - U _{i -1} x → punch.

It is particularly advantageous if one uses a representative, breakdown-free half-wave with the condition
Δ U = a constant time grid with
Δ t ₁. . . Δ t _j . . . Δ t _m determined and this time grid with in-phase start used for all subsequent measurements of the voltage values.

The method can also be applied in an analogous manner to the current Half-waves applied and used for the breakdown detection will.

The invention is based on the consideration that the pulsating DC voltage applied to an electrical separator will have a steady course and that only in the case of breakdowns discontinuities in of the voltage curve.

Using modern measuring and computing devices, it is now possible - as with the known methods - within the  ascending and descending branches of the voltage curve a plurality of measurements and comparisons and at Breakthroughs to respond quickly and accordingly To influence voltage regulation. In the previously proposed There are procedures for the reference values resorted to the previous (undisturbed) half-wave, why according to the number of measurements per Half-wave an equal number of comparison values are stored had to become. In addition, they have become known Procedure still has the disadvantage that one is down the response sensitivity must specify values that can be corrected, but ultimately an arbitrary one Intervention in an otherwise functioning one Represents procedure.

According to the invention, the need for storage is eliminated several reference values. As a comparison, only that the last measured voltage difference is available stand.

With a time grid with a fixed Δ t , you can easily determine from an undisturbed reference curve and depending on the number of measurements per half-wave, which maximum differences between two successive voltage differences can occur and accordingly determine the value x , if necessary with a small safety margin. The larger the number of possible measurements per half-wave, the smaller x can be defined.

A further significant improvement results if one can depart from a time grid with a fixed Δ t and instead covers a representative voltage curve with a grid with a constant Δ U and derives the associated time intervals for the individual measurements.

Because the frequency of the voltage half-waves does not change and also the characteristic course of the voltage curve an electric separator - apart from changing the Amplitude values - practically constant, result the comparisons of successive voltage differences with this time grid for the measurements always 0 as long as no breakthrough takes place.

In this way, an ideal continuity test of the voltage curve of an electrostatic precipitator can be carried out using measurement and computing technology. Theoretically, the value x can be assumed to be 0, in practice one will specify a small, finite value for stabilizing the control system for x , which is either stored as a constant or is coupled in a manner known per se to another characteristic value, e.g. B. with the maximum voltage difference of the previous half-wave.

Further details will be explained with reference to FIGS. 1 and 2. The course of the separator voltage U over time T is shown . Since the separator voltage changes periodically at twice the mains frequency, a full oscillation of the separator voltage is also referred to as a half-wave (the mains voltage). In the present case, the separator voltage is shown in simplified form as a pure sine wave. In practice, the separator voltage deviates more or less from a sine wave. However, this is irrelevant for the following explanations.

The separator voltage changes periodically between U _min and U _max , ie with the amplitude A. Two half-waves are shown with the associated time segments or periods n -1 and n . The voltage curve between U _min and U _max is almost identical in successive periods or half-waves if no breakdowns occur.

According to the invention, successive voltage differences are compared with one another, according to FIG. 1 being based on a time grid with equidistant intervals ( Δ t = const.). With a steady course of the voltage curve and a sufficiently fast measurement sequence, the successive voltage differences differ very little from one another, so that a very small value x can be selected or the response sensitivity can be set very finely. Stored reference values of the previous half-wave n -1 no longer have to be used for the breakdown detection in the half-wave n .

In FIG. 2, a further improvement is illustrated of the inventive method. An undisturbed half-wave n - 1 is first determined with a grid Δ U = const. analyzed in order to obtain a time grid for the voltage measurements at which Δ t can no longer be constant. This time grid is typical for the course of the voltage curve of a certain electrostatic precipitator and can subsequently be retained for the breakdown detection. If the time grid is applied in phase to subsequent half-waves (here n) , voltage differences of equal magnitude inevitably result, so that the value x can be chosen to be zero. In practice, however, a non-zero value x will be chosen so that the control does not become unstable. How small the value x can be selected in this process must be determined when an electrical separator is run in.

Claims (3)

1. A method for detecting breakdowns in an electrical separator, which is fed from an AC voltage network via a rectifier, a high-voltage transformer and an actuator, by comparing a voltage value with a reference value, a deviation from the reference value being used to actuate the actuator, characterized in that that the absolute value of a voltage difference Δ U _j = U _{i +1} - U _i from two successive measurements of the time interval Δ t _j as the voltage value and the absolute value of a voltage difference Δ U _{j -1} = U _i - U _{i -1} as the reference value two successive measurements of the previous time interval Δ t _{j -1} are used and the absolute amount of the difference between the voltage value and the reference value is evaluated as a breakdown when a predetermined value x is exceeded and is used to control the control. 2. The method according to claim 1, characterized in that on the basis of a representative, breakdown-free half-wave with the condition Δ U = constant, a time grid with Δ t ₁. . . Δ t _j . . . Δ t _{m is} determined and that this time pattern with in-phase start is used for all subsequent measurements of the voltage values. 3. The method according to any one of claims 1 and 2, characterized characterized in that the current half-waves in an analogous manner be used to measure the breakdown.