EP0734773A2 - Method for continuous optimization of the operating condition of an electrofilter - Google Patents

Abstract

Continually optimising the operating condition of an electrostatic filter comprises: (a) operating with a direct voltage with superimposed pules, derived by full-wave rectification of the mains voltage, the time spacing between successive pulses being variable by screening a selectable number of full mains waves; (b) operating with a constant number of pulses in the normal phase and subsequent test phase of each of successive operating periods; (c) measuring the residual voltage during the normal and test phases in each of successive cycles (1,2,3...,k,...) at a given limitation of the current or voltage and iteratively varying the limitation, depending on the residual voltage change from cycle, so that the residual voltage tends towards a max.; and (d) operating the test phase of each operating period with a number of pulses differing by 1 from that in the normal phase, measuring and comparing the mean voltage values in the normal and test phases and, depending on the difference, iteratively varying the number of pulses from operating period to operating period such that the mean voltage value tends towards a max..

Description

An electrostatic precipitator has two sets of electrodes: precipitation electrodes and spray electrodes. The precipitation electrodes mostly consist of profiled sheet metal strips, which are put together to form several parallel walls. Two adjacent walls form an alley for the gas flow to be cleaned. The spray electrodes are arranged in the middle of the alley. Often they consist of wires or ribbons that are lined with lace. Most of the time the precipitation electrodes are grounded and the spray electrodes are connected to a high voltage source.

The dust particles to be separated are ionized by electrons which are emitted by the spray electrodes and deflected from the gas flow in the electrostatic field existing between spray electrodes and precipitation electrodes and deposited on the precipitation electrodes. They transfer the electrical charge they carry to the precipitation electrode.

However, if the dust particles to be separated have a very high specific resistance (> 10 ¹¹ Ωcm), under certain operating conditions the electrical charge from the dust layer deposited on the precipitation electrodes cannot flow away as quickly as it is charged by the influx of further charged particles . The result is the so-called back spraying, ie a discharge which is opposite to the discharge occurring at the spray electrodes. The back spraying throws dust back into the gas stream. The degree of separation deteriorates.

The electrostatic separation of high-resistance dusts is optimal at a certain strength of the current that flows between the electrodes of an electrostatic filter. The deposition deteriorates Increasing the current, this is an indication that the back spraying has started. In order to optimize the deposition, it is therefore necessary to limit the current in such a way that the back spraying is just avoided.

The optimal operating point depends on the parameters of the gas flow to be cleaned. If the parameters change, a change in the operating point is generally also necessary. This is illustrated by a few simple examples:

An electrostatic filter can separate the fly ash from the flue gas from a coal-fired boiler. The steam generated in the boiler can be used to operate a generator for generating electrical energy via a turbine. If the demand for electrical energy decreases within the daily load cycle, less steam is required. Less coal is fired to reduce steam generation. Accordingly, the amount of fly ash that the E-filter has to separate decreases. By changing the way the boiler is operated from full load to partial load, an essential parameter changes, namely the volume flow of the fly ash.

To clean the boiler walls, they are blasted with steam during operation. This process is known as soot bubbles. Sootblowing can be done 3 to 4 times a day and each takes half an hour to an hour. The steam leaves the boiler with the flue gas through the deduster. Part of the moisture accumulates on the fly ash particles and changes the electrical properties of the fly ash and the flue gas. Soot blowing also changes the parameters essential for the operation of the electrostatic precipitator.

Load changes and soot bubbles are just examples of quite normal processes in the operation of an electrostatic precipitator, in which the Change the parameters of the gas flow to be cleaned.

From EP 0 097 161 B1 it is known to operate an electrostatic precipitator with a current which corresponds precisely to the point of application of the back spraying. Current and voltage are monitored by increasing or decreasing the excitation to determine the point at which back spraying begins. The source from which the excitation is fed is a 50 Hz network. The current consists of a pulse train and the voltage is a DC voltage with a superimposed AC component. Conventional two-way rectification of the mains voltage results in a frequency of 100 Hz for the current and the AC voltage component.

EP 0 140 855 B1 discloses a method for changing a voltage occurring at the electrodes of an electrostatic dust separator, in which the voltage is generated by a pulse sequence derived from the mains frequency and the change is brought about by the length of the interval between two successive ones individual pulses is varied by hiding an even number of pulses. The number of pulses per second is reduced to 33, 20, 14, 11, etc., depending on the number of hidden pulses. The line voltage is fed via thyristors to a high-voltage transformer, which is connected on the secondary side to a two-way rectifier. The output voltage of the rectifier is at the electrodes of the electrostatic filter. The thyristors are controlled by a control circuit which is switched in such a way that it erases an even number of pulses from the mains voltage between two pulses which are fed to the electrostatic filter.

From EP 0 465 547 B1 it is known to control the power supply of the discharge electrodes of an electrostatic filter in order to achieve a max.

Dust removal to supply current pulses with a given current intensity to the discharge electrodes and to vary the number of pulses per second according to the aforementioned document. Corresponding instantaneous values of the voltage between the discharge and precipitation electrodes are measured for a number of different pulse frequencies, and the current pulse supply is then set to the pulse frequency for which the greatest instantaneous value has been measured. The pulse current is set to a maximum value taking into account the capacity of the power supply unit and any flashovers between the discharge and precipitation electrodes.

The invention has for its object to provide a method for operating an electrostatic filter, in which the set operating point is continuously monitored and tracked, so that the filter works continuously in the vicinity of the optimal operating point. This object is achieved by the features specified in claim 1.

Further advantageous features are specified in the subclaims.

• Figur 1 veranschaulicht den Verlauf von Strom und Spannung für verschiedene Fälle, die sich durch unterschiedliche Länge des Intervalls zwischen zwei aufeinanderfolgenden Pulsen unterscheiden.
• Figur 2 veranschaulicht die Aufeinanderfolge der Betriebsperioden.
• Figur 3 veranschaulicht die Folge der Zyklen innerhalb einer einzelnen Phase.

Figures 1-3 serve to illustrate the invention with the aid of diagrams.

• Figure 1 illustrates the course of current and voltage for different cases, which differ in the length of the interval between two successive pulses.
• Figure 2 illustrates the succession of operating periods.
• Figure 3 illustrates the sequence of cycles within a single phase.

In Figure 1, five diagrams arranged one above the other show the course of current and voltage for conventional operation with mains frequency and two-way rectification (above) as well as in so-called "semi-pulse operation" with different pulse numbers 1 - 4. In two-way rectification, the current profile is just like the voltage profile through a chain of pulses at intervals of 10 ms marked. The voltage pulses are superimposed on a DC voltage. At pulse number 1, the high-voltage part of the power supply between two pulses for a full network oscillation is not activated. The remaining pulses therefore have an increased interval of 30 ms. Each voltage pulse has a steeply rising edge at the beginning. The other flank drops exponentially to a residual voltage until the next pulse begins. The current pulses rise steeply up to a maximum current and then drop steeply down to 0. In the following, "current" always means the maximum current. With pulse number 2, two full network oscillations are not switched through between two pulses. The remaining pulses are therefore 50 ms apart. The corresponding characteristics for the pulse numbers 3 and 4 are readily apparent from FIG. 1.

In Figure 2, several successive operating periods m, m + 1, m + 2, ... are shown schematically. Each operating period includes a normal phase and a subsequent test phase. The normal phase lasts much longer than the test phase. The duration of the normal phase to the duration of the test phase is preferably about 4: 1 to 20: 1. The normal phase lasts e.g. B. 1 h, the test phase 5 - 10 min.

A constant pulse number is used in each normal phase, as well as in each test phase. However, the pulse number of the test phase deviates from the pulse number of the immediately preceding normal phase by ± 1, as will be explained below.

The normal phase and the test phase each comprise a sequence of cycles, which are consecutively numbered 1, 2, 3, ..., k, k + 1, ... The cycles follow one another at intervals of 20-40 s, preferably about 30 s. In each cycle, an upper limit value for the current, ie a current limitation, is entered on the control device of the electrostatic filter. The associated residual voltage is measured and the current limitation for the following cycle is set on the basis of the measured value obtained, as will be explained below with reference to FIG. 3.

In Figure 3, several consecutive cycles k, k + 1, ... are symbolically shown along the horizontal time axis. In this case, the time interval between the individual cycles is 30 s. A normal phase is considered first.

Time 0 in FIG. 3 can be any time during operation, e.g. B. the switch-on time or the beginning of a normal phase. At this point the current limit for cycle k is set to 450 mA. The associated residual voltage is approximately 25 kV according to the upper diagram in FIG. 3. The current limit is then increased to 500 mA in order to test whether a higher residual voltage is now established. The increased current limitation results in a residual voltage of 25.8 kV in cycle k + 1. Since the increase in the current limit has led to an increase in the residual voltage, the current limit is increased again in the following cycle k + 2, this time to 550 mA. There is again an increased residual voltage, namely 26.2 kV. Since the change in the residual voltage has a positive sign again this time, the current limitation is changed again in the same direction, ie increased to 600 mA. There is an increased residual voltage of 26.5 kV in cycle k + 3. Another attempt to achieve an even higher residual voltage by increasing the current limit to 650 mA fails in cycle k + 4. The residual voltage drops to 26.2 kV. This is an indication of the beginning of the Spraying back. Therefore, in the cycle k + 5, the current limitation is set lower again, to 600 mA, with the effect that the residual voltage increases, to 26.5 kV. After the decrease in the current limit in the cycle k + 5 has led to an increased residual voltage, the current limit for the cycle k + 6 is also reduced. This time, however, there is a lower residual voltage of 26.2 kV. Therefore, the current limit for the next cycle is changed in the opposite direction, namely increased to 600 mA. This leads to an increased residual tension in cycle k + 7. Apparently the condition has stabilized from cycle k + 3. The consequence is that from then on the current limitation oscillates between 550 and 650 mA. The residual voltage adjusts to a quasi-stationary value. It fluctuates slightly around 26.4 kV. This state is optimal with the set number of pulses and the instantaneous parameters of the gas stream to be cleaned. However, should the parameters change during the normal phase under consideration, the control device probes analogously to the cycles k to k + 3 to the optimal operating point, which corresponds to the changed parameters. Experience has shown that the changes in the parameters are relatively slow, measured in terms of the cycle duration. The changes in the current limitation from cycle to cycle, each ± 50 mA in the example given, are small in any case compared to the current to which each is limited, preferably about 5 to 15%. Therefore, the voltage changes are relatively small in the steady state, so that the filter operation is not significantly affected by them.

It goes without saying that, taking into account the current-voltage characteristic, which can be determined experimentally, the voltage can also be limited instead of the current.

During the entire duration of the normal phase or at least during a period that spans several cycles at the end of the Normal phase extends, an average of the voltage is calculated and stored by evaluating a large number of instantaneous values. As an average z. B. selected the effective voltage.

A test phase follows the normal phase. In the test phase, it should be tested whether the change can be improved with a changed pulse number.

In the test phase, a pulse number is used which differs from the pulse number of the immediately preceding normal phase by ± 1. Numerous cycles are also run through in the test phase, analogous to the normal phase. After a quasi-steady state has occurred, an average value of the voltage is also calculated and stored in the test phase. This mean is compared with the mean of the associated normal phase.

In the example illustrated in FIG. 2, it follows for the operating period m that the mean value in the test phase is lower than in the normal phase. The change in the number of pulses - in this case from 5 to 6 - did not result in an improvement in terms of increasing the mean value of the voltage. Therefore, the pulse number 5 is used again in the normal phase of the following period m + 1. Since an increase in the number of pulses was unsuccessful in the operating period m, the number of pulses is reduced to 4 in the test phase of the operating period m + 1. The effect again this time is that the mean value of the voltage drops. Therefore, the pulse number is reset to 5 in the normal phase of the operating period m + 2. In the subsequent test phase, the number of pulses 6 is tried again, again with the result that the mean value of the voltage drops. Consequently, the pulse number 5 is again set in the normal phase in the operating period m + 3. In the test phase, the number of pulses 4 is tried again, this time with success, apparently because a parameter of the gas stream to be cleaned has changed in the meantime. It turns out an increased Average voltage. Since the reduction in the number of pulses in the operating period m + 3 was successful, the number of pulses 4 is maintained in the normal phase in the operating period m + 4. In the subsequent test phase, the number of pulses is reduced again, to 3. The effect is negative. Therefore, the number of pulses is reset to 4 in the normal phase of the operating period m + 5.

Since the test phase is relatively short compared to the normal phase and the pulse numbers of the normal phase and the subsequent test phase differ only by ± 1, the fluctuations caused by this are relatively small and have little influence on the quality of the separation in individual cases. If, however, the parameters of the gas stream to be cleaned change sustainably over a longer period of time, the method of operation illustrated in FIG. 2 causes the electrostatic filter to always work in the vicinity of the respective optimum operating point.

Claims (6)

Method for continuously optimizing the operating state of an electrostatic filter with the following features: a) Working with a direct voltage derived from the mains voltage by means of branch-path rectification with superimposed pulses, the time interval between successive pulses being changeable by hiding a selectable number of full mains waves; b) successive operating periods (1, 2, 3, ..., m, ...) each include a normal phase and a subsequent test phase; c) in each operating period, the normal phase and the test phase work with a constant number of pulses; d) during the normal phase and the test phase, the residual voltage is measured in successive cycles (1, 2, 3, ..., k, ...) for a given limitation of the current or voltage and the limitation depending on the change in The residual voltage is iteratively changed from cycle to cycle, so that the residual voltage tends to a maximum. e) in each operating period (1, 2, 3, ..., m, ...) the test phase is operated with a pulse number that has been changed by ± 1 compared to the normal phase, and the mean values of the voltage in the normal phase and in the Test phases are measured and compared with one another, and depending on the difference, the pulse number is changed iteratively from operating period to operating period, so that the mean value of the voltage tends to a maximum. Method according to Claim 1 for the continuous optimization of the operating state of an electrostatic filter with the following features: a) Working with a direct voltage derived from the mains voltage by means of branch-path rectification with superimposed pulses, the time interval between successive pulses being changeable by hiding a selectable number of full mains waves; b) successive operating periods (1, 2, 3, ..., m, ...) each include a normal phase and a subsequent test phase; c) in each operating period, the normal phase and the test phase work with a constant number of pulses; d) the pulse number of the test phase deviates from the pulse number of the immediately preceding normal phase by ± 1, the sign being determined in accordance with the following features h and i; e) the normal phase and the test phase each comprise a sequence of cycles (1, 2, 3, ..., k, ...) according to the following scheme: ea) the current is limited to i _k ; eb) the residual voltage u _k is measured; ec) the current is at a value deviating from i _k $${\text{i}}_{\text{k + 1}}{\text{= i}}_{\text{k}}{\text{+ Δ i}}_{\text{k}}$$

limited, where Δi _{k is} positive or negative and the absolute value | Δi _k | is small compared to i _k ; ed) the residual voltage u _{k + 1} is measured; ee) the difference Δu _k = u _{k + 1} - u _k is determined; ef) the current is at a value deviating from i _{k + 1} $${\text{i}}_{\text{k + 2}}{\text{= i}}_{\text{k + 1}}{\text{± Δi}}_{\text{k}}$$

limited, the sign of Δi _k corresponding to the sign according to ec) only if the sign of Δu _k according to ee) is positive. f) both in the normal phase and in the test phase, an average value of the voltage is determined and stored at least in the end section; g) the mean value of the test phase is compared with the mean value of the associated normal phase; h) if in the period (m) in the test phase the mean value is not greater than in the normal phase, then in the following period (m + 1)
worked in the normal phase with the same number of pulses as
in the normal phase of the period (m)
and when switching to the test phase, the pulse number in reverse
Changed direction as in period (m); i) if in the period (m) the mean value in the test phase is larger than in the normal phase, then in the following period (m + 1)
worked in the normal phase with the same number of pulses as
in the test phase of the period (m)
and at the transition to the test phase the pulse number in the same
Direction changed as in period (m). Method according to Claim 1 or 2, characterized in that the duration of the normal phase to the duration of the test phase is 4: 1 to 20: 1. Method according to one of claims 1 to 3, characterized in that the duration of an operating period is 1 to 2 hours. Method according to one of claims 1 to 4, characterized in that the duration of a cycle is 10 to 30 s. Method according to one of claims 2 to 5, characterized in that the absolute value | Δ i _k | is between 0.05 i _k and 0.15 i _k .

Images