A method of anode rapping control for an electrostatic precipitator

Abstract

This invention discloses an anode rapping control method for an electrostatic precipitator, relating to the field of multi-field coordinated rapping dust removal. Accelerometers are arranged on the anode plate groups of each electric field, and flue gas velocity sensors are arranged in the inlet flue, maintaining a shared rapping event log. The moment when the acceleration signal envelope attenuation ratio drops below a first preset threshold and remains below it for a first preset duration is taken as the moment when the settling of this electric field is completed. The moment when the upstream rapping particle cloud arrives at and crosses this electric field is calculated using the ratio of a pre-calibrated flow channel length between electric fields to the measured flue gas velocity, and this is taken as the downstream freezing period. At the candidate rapping moment, two conditions are determined: settling completion and no downstream conflict. Rapping is triggered and broadcast downstream only if both conditions are met; otherwise, it is postponed. This method simultaneously suppresses the superposition of secondary dust from both the local and cross-fields at the source, reducing the peak amplitude of the outlet flue gas rapping, without requiring modification to the main structure of the electrostatic precipitator.

Description

Technical Field

[0001] This invention relates to the field of air pollution control, and in particular to an anode rapping control method for multi-field electrostatic precipitators. Background Technology

[0002] Electrostatic precipitators (ESPs) are widely used flue gas dust removal equipment in industries such as thermal power, cement, steel, and non-ferrous metallurgy. Their anode plates electrostatically adsorb charged dust particles in the flue gas, and periodic anode rapping knocks the dust layer adsorbed on the anode plate surface down into the ash hopper below. Early anode rapping control methods used a single fixed-cycle scheme, where a timer sequentially triggered rapping in each electric field. This scheme was simple to implement, but it neglected the differences in dust deposition rates between different operating conditions and electric fields. This often led to back corona discharge in the upstream electric field due to excessively thick dust layers, while the downstream electric field increased the risk of secondary dust re-entrainment due to overly frequent rapping.

[0003] To address the aforementioned issues, the industry has successively proposed improvement approaches based on operational condition feedback. These include using parameters such as flue gas flow rate, boiler load, and secondary current changes as correction signals for the rapping cycle, and delaying or intensifying the rapping action of a specific electric field when the outlet dust concentration is abnormal. This approach has achieved a certain emission peak reduction effect under most steady-state operating conditions. Other solutions include pre-checking the readiness of the downstream electric field before the upstream electric field is about to be rapped, or shifting the rapping sequence overall before and after soot blowing operations on upstream equipment.

[0004] However, the above methods still do not completely eliminate the rapping peak phenomenon. On the continuous monitoring curve of the outlet dust of a multi-field horizontal electrostatic precipitator, regular instantaneous concentration peaks synchronized with the rapping actions of each field often appear, sometimes contributing a significant share to the unit's hourly average emissions. Taking a four-field electrostatic precipitator as an example, after the first field rapping, the dust detached from the anode plate surface settles downwards in the form of flocs. However, due to the carrying effect of the flue gas, some dust flocs do not fall directly into the ash hopper, but are carried by the flue gas flow to the subsequent electric field as suspended particulate clouds. If the second electric field is triggered rapping while this particulate cloud is still in its flight path, the particulate clouds from both fields superimpose at the outlet section of the second electric field, instantly increasing the outlet concentration and forming a peak. Although existing solutions impose some constraints on the rapping sequence across electric fields, this superposition phenomenon cannot be avoided under certain operating conditions.

[0005] The fundamental reason is that existing methods often treat the triggering of rapping as a single-dimensional time event, only considering the time interval between the current electric field and the last rapping, as well as several operating parameters, without explicitly characterizing the physical process of the particle cloud generated by the previous rapping not yet fully settling into the ash hopper into the control law. In other words, existing methods only care about the dust removal requirements of the anode plate surface, but do not consider the residence time of the particle cloud in the flue gas channel between the anode plate and the ash hopper. The settling time of particles in the flue gas is an established conclusion in fluid mechanics, which is related to both the dust particle size distribution and the vibration decay process of the anode plate after being struck. This is a phenomenon that has been observed in engineering but has not been incorporated into engineering criteria.

[0006] Therefore, the following technical problem objectively exists in this field: without replacing the main structure of the electrostatic precipitator and without relying on the outlet concentration feedback, how to simultaneously identify two types of working conditions that lead to the superposition of secondary dust at the anode rapping control level: the particles from the previous rapping in this electric field have not yet settled into the ash hopper and the particle cloud from the recent rapping in the upstream electric field is being transmitted to this electric field. Based on this, the next rapping in this electric field can be postponed, and the rapping peak of the outlet dust can be suppressed from the source without significantly increasing the overall cleaning cycle of the electrostatic precipitator. Summary of the Invention

[0007] This invention addresses the aforementioned technical problems by providing an anode rapping control method for an electrostatic precipitator. The electrostatic precipitator comprises multiple electric fields arranged sequentially along the flue gas flow direction. Each electric field is equipped with an anode plate assembly, a rapping mechanism, and an acceleration sensor mounted on the anode plate assembly. It is also equipped with a flue gas velocity sensor and a shared rapping event log. The method operates according to the following steps: S0 receives the candidate rapping time sequence of each electric field and the event time of the last actual rapping event; S1. For each electric field, take the time of its most recent rapping event as the zero point, and take the envelope of the acceleration signal to calculate the attenuation ratio. When the attenuation ratio drops below the first preset threshold for the first time and remains not less than the first preset duration, record that time as the time when the settling of this electric field is completed. S2, For each electric field, query the recent rapping event times of all electric fields located upstream of this electric field along the flue gas flow direction in the shared rapping event log. Using the pre-calibrated ratio of the flue gas flow channel length between electric fields to the measured flue gas flow velocity, calculate the estimated start time of the upstream rapping particle cloud arriving at this electric field and the estimated end time of crossing this electric field. The time period between the start and end times is recorded as the downstream freezing period of this electric field. S3, when the candidate rapping time of this electric field arrives, determine whether it simultaneously satisfies: there is already a settlement completion time later than its most recent rapping event; and the candidate rapping time does not fall within any active downstream freezing period; if both conditions are met, proceed to S4; otherwise, proceed to S5. S4, trigger the rapping of this electric field, write the current moment as the new rapping event moment into the shared rapping event log, and broadcast it to all electric fields located downstream of this electric field along the flue gas flow direction; S5, postpone the vibration of this electric field to the next judgment time. The next judgment time is the later of the time when the settlement of this electric field is completed and the time when the freezing period of all active downstream of this electric field ends. After the postponement, return to S3.

[0008] The present invention also provides an anode rapping control system for an electrostatic precipitator. The system includes a candidate rapping receiving module, a settling completion determination module, a downstream freezing period estimation module, a dual-criteria blocking module, a rapping triggering and broadcasting module, and a delay discrimination module. Each module corresponds one-to-one with the method steps S0 to S5, and the method can be implemented on this system.

[0009] The beneficial effects of this invention include: First, by using the dual constraint of the anode plate acceleration attenuation envelope to identify that the dust in this electric field has settled into the ash hopper, the next rapping of this electric field will not fall into the period before the particles from the previous rapping have settled, thus cutting off the secondary dust accumulation path from the source; Second, by sharing the rapping event log and the measured flue gas velocity to infer the downstream freezing period of the upstream rapping particle cloud on this electric field, the next rapping of this electric field will not overlap with the particle cloud from the recent upstream rapping in the air above this electric field, thus cutting off the cross-field secondary dust accumulation path from the source; Third, the acceleration sensor and flue gas velocity sensor are mature industrial instruments, and the event log is a software-layer data structure, which does not require any modification to the main structure of the electrostatic precipitator and can be implemented by software upgrade on the existing low-voltage control system of the electrostatic precipitator. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the overall process of the electrostatic precipitator anode rapping control method described in this invention.

[0011] Figure 2 This is a schematic diagram of the module structure of the anode rapping control system for the electrostatic precipitator described in this invention.

[0012] Figure 3 This is a schematic diagram showing the relationship between the anode plate acceleration decay envelope and the first preset threshold and the first preset duration according to the present invention.

[0013] Figure 4 This is a schematic diagram illustrating the grading of the attenuation ratio, the first preset threshold, and the settlement completion determination section as described in this invention.

[0014] Figure 5 This is a schematic diagram of the decision tree for the dual-criteria discrimination process in step S3 of the present invention.

[0015] Figure 6 This is a flowchart illustrating the algorithm for delaying the discrimination time and selecting the later sub-algorithm in step S5 of the present invention.

[0016] Figure 7 This is a schematic diagram of the deployment plan of the present invention applied to a four-field horizontal electrostatic precipitator.

[0017] Figure 8 This is a schematic diagram of the deployment plan of the present invention applied to the electrostatic precipitator at the tail of a three-field cement kiln.

[0018] Figure 9 This is a schematic diagram comparing the peak value of the outlet smoke dust with the present invention, the fixed-cycle rapping scheme, and the upstream drive cross-field coordination scheme.

[0019] Figure 10 This is a schematic diagram of the heat map showing the decrease in the daily average value of flue dust at the outlet under the combination of two parameters: a first preset threshold and a third preset duration.

[0020] Figure 11 This is a schematic diagram showing a dual-axis comparison of the instantaneous concentration of flue dust at the outlet during the operation of this invention and the triggering rhythm of the rapping event. Detailed Implementation

[0021] To facilitate understanding of the overall concept and specific embodiments of the present invention by those skilled in the art, the following provides an overview of the correspondence between the accompanying drawings and various chapters. For example... Figure 1 The diagram shows the overall control flow of the method described in this invention, covering six steps from S0 to S5 and their branching relationships. This flow is consistently applied throughout the following explanations of terminology, core algorithm calculation steps, and various embodiments. Figure 2 The diagram illustrates the module structure of the system described in this invention and its correspondence with the method steps. The specific implementation and interface specifications of each module are detailed in the system implementation description section below. Figure 3 The diagram illustrates the relationship between the anode plate acceleration decay envelope and the first preset threshold and the first preset duration, as described in this invention. It can be compared with the calculation of the envelope and decay ratio in step A1 and the determination of sedimentation completion in step A2 below. Figure 4 The diagram shows the attenuation ratio values ​​in different tiers and their corresponding decision-making actions. For example... Figure 5 The image shows the decision tree for the dual-criteria discrimination process in step S3 of this invention. Figure 6 The diagram shows the flow of the algorithm for delaying the discrimination time and selecting the later sub-algorithm in step S5 of this invention. It can be compared with step A5 (delaying the discrimination time) below. Figure 7 The diagram shows the deployment plan of the present invention applied to a four-field horizontal electrostatic precipitator, corresponding to Embodiments 1 and 8 below; as shown... Figure 8 The diagram shows the deployment plan of the present invention applied to a three-field cement kiln tail electrostatic precipitator, corresponding to Embodiments 2 and 5 below. Figure 9 The bar chart shown is a comparison of the amplitude of the rapping peak between the present invention and other existing solutions, corresponding to the comparison section of the publicly available dataset below. Figure 10The figure shown is a heat map of the daily average decrease in outlet dust under the combination of two parameters: a first preset threshold and a third preset duration, corresponding to the key threshold sensitivity experimental section below. Figure 11 The figure shown is a dual-axis comparison of the instantaneous concentration of flue dust at the outlet and the triggering rhythm of the rapping event during the operation of this invention.

[0022] Terminology Explanation: Electric field: The electric field refers to several independently powered high-voltage electric field groups arranged in series according to the flue gas flow direction in an electrostatic precipitator; the electric field can be a high-voltage electric field group in a horizontal dry electrostatic precipitator, a wet electric field group in a wet electrostatic precipitator, or a high-voltage electric field group located upstream of the filter bag in an electrostatic-bag composite dust collector.

[0023] Anode plate assembly: The anode plate assembly refers to a dust collection unit composed of several anode plates arranged side by side at a certain interval; the anode plate assembly can be any of the following: C480 or C320 type corrugated plate assembly, ZT24 or ZT18 type tension plate assembly, or rolled grid plate assembly; the typical height of the anode plate assembly is 7 meters to 15 meters, and the typical width is 3 meters to 4 meters.

[0024] Vibration mechanism: The vibration mechanism refers to the actuator that applies mechanical impact to the anode plate assembly to peel off the dust layer; the vibration mechanism can be any one of the following: top rotary hammer vibration mechanism, side pendulum vibration mechanism, electromagnetic pulse vibration mechanism; regardless of the type of vibration mechanism, it generates a significant impact acceleration on the anode plate at the moment of triggering.

[0025] Accelerometer: The accelerometer refers to a vibration measurement device installed on the anode plate assembly to collect the impact of rapping and its subsequent damped vibration; the accelerometer can be any one of piezoelectric vibration accelerometer, MEMS capacitive accelerometer, or strain gauge accelerometer; the typical range is ±10g to ±100g (gravitational acceleration g, (1g≈9.8 m / s²)). 2 (), with a sampling rate of not less than 2 kHz.

[0026] Flue gas velocity sensor: The flue gas velocity sensor refers to an industrial instrument installed in the inlet flue of an electrostatic precipitator to measure the flue gas velocity; the flue gas velocity sensor may be any one of a Pitot tube array flue gas velocity measuring device, a differential pressure flue gas velocity transmitter, or an ultrasonic time-of-flight flue gas velocity transmitter.

[0027] Shared rapping event log: The shared rapping event log refers to a data structure that is shared and read / written by each electric field controller and saves rapping events in chronological order; the shared rapping event log can be a ring buffer data structure stored in the memory of the DCS or PLC of the low-voltage control system of the electrostatic precipitator, or a time-series database stored in the edge industrial control computer, or a mirror buffer synchronized between each electric field controller based on the factory fieldbus protocol; each record in the event log includes the electric field identifier, the time of the rapping event, and an optional initial peak value of the envelope.

[0028] Envelope: The envelope refers to the slowly changing curve obtained after taking the peak amplitude of the acceleration signal; the envelope can be any one of the following: performing sliding peak detection based on Hanning window on the absolute value of the acceleration signal, taking the magnitude after performing Hilbert transform on the acceleration signal, or taking the moving root mean square after performing bandpass filtering on the acceleration signal; the typical value range of the envelope update period is 50 milliseconds to 200 milliseconds.

[0029] Attenuation ratio: The attenuation ratio refers to the ratio of the current peak value of the envelope of the electric field acceleration signal to the initial peak value of the envelope corresponding to the moment of the most recent rapping event of the electric field, and the value ranges from 0 to 1.

[0030] Settling completion time: The settling completion time refers to the moment when the attenuation ratio first drops below the first preset threshold and remains at a duration not less than the first preset duration; this moment physically corresponds to the moment when the dust peeled off from the anode plate surface has settled to a certain height above the ash hopper, and the influence of the residual particle cloud on the outlet concentration of this electric field is lower than the engineering negligible level.

[0031] Downstream freezing period: The downstream freezing period refers to the time interval determined from the expected start time of the upstream rapping particle cloud arriving at this electric field to the expected end time of its crossing of this electric field; during this period, this electric field is not allowed to trigger new rapping.

[0032] This method involves the following threshold and duration parameters, each of which includes its unit, value range, typical default value, and adjustment direction: First preset threshold: The first preset threshold refers to the threshold for determining the attenuation ratio. It is dimensionless and ranges from 0.05 to 0.20, with a typical default value of 0.10. When the dust particle size in this electric field is too large, it can be adjusted upward to 0.15 to shorten the settling determination time. When the flue gas flow rate is lower than the design value, it should be adjusted downward to 0.07 to ensure more thorough settling.

[0033] First preset duration: The first preset duration refers to the minimum continuous duration required for the attenuation ratio to remain below the first preset threshold. The unit is seconds, and the value range is 1 second to 5 seconds, with a typical default value of 2 seconds. In the final stage electric field, it can be adjusted upward to 3 to 4 seconds to improve the reliability of the settling judgment. In the electric field with coarser dust particle size in the preceding stage, it can be adjusted downward to 1 second to ensure the dust removal frequency.

[0034] Second preset duration: The second preset duration refers to the query time window when scanning upstream rapping events. The unit is seconds, and the value range is 10 seconds to 60 seconds, with a typical default value of 30 seconds. When the flue gas velocity is low, it should be adjusted upward to 45 seconds. When the flue gas velocity is high and the electric field spacing is short, it can be adjusted downward to 15 seconds.

[0035] The third preset duration refers to the safety margin added after the downstream freezing period passes through this electric field. The unit is seconds, and the value range is 1 second to 10 seconds, with a typical default value of 3 seconds. When the dust particle size is relatively fine and it is easy to be suspended over a long distance, it should be adjusted upward to 6 to 8 seconds. When the dust particle size is relatively coarse and it settles rapidly, it can be adjusted downward to 1 to 2 seconds.

[0036] Fourth preset threshold: The fourth preset threshold refers to the upper limit of the number of consecutive delays. It is dimensionless and ranges from 3 to 10 times, with a typical default value of 5 times. If the unit load is at full load and the short-term dust removal requirement is urgent, it can be adjusted down to 3 times to avoid cumulative delay in dust removal. Otherwise, it can be adjusted up to 8 times under low load conditions.

[0037] Fifth preset threshold: The fifth preset threshold refers to the flue gas temperature threshold that triggers the extension of the downstream freezing period. The unit is degrees Celsius (°C), and the value range is 120°C to 180°C, with a typical default value of 150°C; for working conditions with high dust resistivity, the value is adjusted downward to 130°C.

[0038] The sixth preset threshold: The sixth preset threshold refers to the flue gas flow rate threshold that triggers the extension of the downstream freezing period. The unit is standard cubic meters per hour, and the value range is 0.5 to 0.8 times the design flow rate, with a typical default value of 0.7 times the design flow rate. For units with boilers operating at low load for a long time, the value should be adjusted upward to 0.75 times the design flow rate.

[0039] The seventh preset duration refers to the aging and deletion duration of the event log, measured in seconds, with a value range of 60 to 600 seconds and a typical default value of 300 seconds; for long-flow electrostatic precipitators or low flue gas flow rate conditions, it should be adjusted upward to 480 seconds.

[0040] The core computational steps of the underlying algorithm are as follows: Step A1 Envelope and Attenuation Ratio Calculation: For each electric field, the accelerometer collects the time-domain signal a(t) at a sampling rate of not less than 2 kHz; after passing through a bandpass filter to retain the dominant vibration frequency band of the rapping, x(t) is output; the absolute value of x(t) is taken and peak detection is performed with a sliding window of 100 milliseconds to obtain the envelope e(t); the zero point time is defined as the time t0 of the most recent rapping event in this electric field, and e0 is denoted as the maximum value of e(t) in the interval [t0, t0+200] milliseconds. The attenuation ratio r(t) = e(t) divided by e0.

[0041] Step A2 Settlement Completion Determination: Maintain a state variable for a continuous duration below a threshold, and perform a scan every envelope update cycle (typically 100 milliseconds): If the current r(t) is less than the first preset threshold, then the state variable is incremented by one update cycle; otherwise, the state variable is reset to zero; when the state variable is equal to or exceeds the first preset duration for the first time, record the current time as the settlement completion time of this electric field and freeze it.

[0042] Step A3: Calculation of downstream freezing period: For this electric field i, assume that the upstream electric field j is at t j A rapping event occurs at any time. Let the pre-calibrated lengths of each flue gas flow channel between electric field j and electric field i be L. j,j+1 ,L j+1,j+2 ,...,L i-1,i Let the length of the electric field i along the direction of the flue gas flow be L. i Let the measured flue gas velocity be v; then the expected start time of the particle cloud arrival at the local electric field i for upstream event j is t. j The cumulative value of the time obtained by dividing each segment length by v is added, and the expected end time of crossing this electric field is the expected start time plus L. i Add a third preset duration to the ratio of v.

[0043] Step A4 Dual-criteria blocking: The electric field i at the candidate rapping time t c If the settlement completion time does not exist or is earlier than the time of the most recent rapping event in this electric field, then the settlement is determined to be incomplete, and proceed to S5; otherwise, enumerate all currently active downstream freezing periods, and if t c If the game falls into either of these time periods, an upstream conflict is detected, and the game proceeds to S5; if neither of the two criteria is triggered, the game is triggered and the game proceeds to S4.

[0044] Step A5: Delay the judgment time: Take the time T when the current electric field settlement is completed. s The end time T of all active downstream freeze periods f,1 To T f,K The later one in the sequence is used as the next decision time T. n The number of consecutive delays of the electric field is incremented by one; if the number reaches the fourth preset threshold, the current candidate vibration is skipped and an alarm is reported; otherwise, T is...n It is attached to the scheduler as the next decision trigger time for this electric field.

[0045] Step A6: Circular buffer management of the event log: The shared rapping event log stores the most recent rapping events for each electric field in chronological order. Each time a new event arrives, the log depth is first checked to see if it exceeds a preset upper limit (typically 64 entries). If it does, the oldest event is discarded. Simultaneously, all events in the log are traversed; if their timestamps are more than the seventh preset time from the current time, they are also aged out and deleted. Aging deletion and upper limit deletion operate independently to ensure controllable log depth and prevent event timestamps from becoming outdated. The circular buffer uses atomic operations or read-write locks for reading and writing to ensure consistency across multiple electric fields concurrently.

[0046] Step A7: Sliding window averaging of flue gas velocity: The instantaneous value of the measured flue gas velocity v(t) is greatly affected by turbulence disturbances, and directly using it for downstream freezing period estimation may introduce jitter; therefore, a sliding window averaging is performed on the measured signal, with a typical window length of one-third of the second preset duration (default 10 seconds). During S2 scanning, the upstream rapping event time t is... j Centered on t j The first 10 seconds to t j The mean value of v(t) over the last 10 seconds was used as the measured flue gas velocity for estimation. This smoothing process significantly improved the estimation stability during the downstream freezing period.

[0047] Step A8: Parallel Scheduling of Multiple Electric Fields: The parallelism among multiple electric fields in this method is manifested as follows: S0 to S5 of each electric field are independent state machines for that field, and the state machines of each field do not block each other; only the shared rapping event log serves as the coordination channel between electric fields. In a typical implementation, each electric field state machine cycles through S1 envelope updates and S2 downstream freeze period calculations with a scheduling cycle of 100 milliseconds, and runs S3 to S5 driven by the arrival of candidate rapping time events. The overall consistency of the parallel scheduling of multiple electric fields is guaranteed by the atomic writing of the event log.

[0048] Example 1 (Typical four-field horizontal electrostatic precipitator under full load condition) A 660 MW supercritical coal-fired unit is equipped with a four-field horizontal electrostatic precipitator, with the fields arranged sequentially from the first to the fourth along the flue gas flow direction. Each field is 4 meters long, and the flue gas duct between the fields is 0.5 meters long. Each field has a rotary hammer rapping mechanism at its top, equipped with a piezoelectric accelerometer (range ±50g, sampling rate 5 kHz). A Pitot tube array flue gas velocity measuring device is installed in the inlet flue. A shared rapping event log is synchronized by the electrostatic precipitator's low-voltage control PLC within a 300-second annular buffer. The unit operates continuously at 95% boiler load, with a measured flue gas velocity of 1.2 m / s.

[0049] For the second electric field: S0 receives the candidate rapping time sequence for the second electric field given by the original rapping scheduling module, with a period of 240 seconds; in S1, the most recent rapping event time of the second electric field is t0=12 minutes 00 seconds. Taking t0 as the zero point, the envelope of the acceleration signal is taken. The initial peak value of the envelope is calculated to be e0=4.8g. At t0+1.6 seconds, the attenuation ratio r drops to below the first preset threshold of 0.10 for the first time (measured r=0.082), and remains below 0.10 for 2 seconds (first preset duration). Then, the settlement completion time is recorded as t0+3.6 seconds, i.e., 12 minutes 03.6 seconds. S2: Scanning the shared event log reveals that a rapping event occurred in the first electric field at 11 minutes and 58 seconds. The estimated start time is calculated by dividing 0.5 meters by 1.2 meters per second to get 0.42 seconds, then adding this to 11 minutes and 58 seconds to get 11 minutes and 58.42 seconds. The estimated end time of crossing this electric field is calculated by adding 4 meters divided by 1.2 meters per second to get 3.33 seconds, plus the third preset duration of 3 seconds, to get 12 minutes and 04.75 seconds. Therefore, the current active downstream freezing period of the second electric field is from 11 minutes and 58.42 seconds to 12 minutes and 04.75 seconds. S3: The candidate rapping time is 12 minutes and 04 seconds, falling within the downstream freezing period, so proceed to S5; S5: The delayed judgment time is taken as the later of the settlement completion time of 12 minutes and 03.6 seconds and the downstream freezing period end time of 12 minutes and 04.75 seconds, which is 12 minutes and 04.75 seconds. The number of delays is accumulated to 1, which is less than the fourth preset threshold of 5 times, and is attached to the scheduler. S3 is re-judged at 12 minutes and 04.75 seconds. At this time, the particle cloud of the first electric field has passed through this electric field. The settlement completion time of this electric field (12 minutes and 03.6 seconds) is later than its most recent rapping event time (12 minutes and 00 seconds), so proceed to S4 to trigger the second electric field rapping and broadcast to the third and fourth electric fields.

[0050] Example 2 (Variable load operation of electrostatic precipitator at the tail of a three-field cement kiln) A 5000-ton-per-day cement clinker production line is equipped with a three-field horizontal electrostatic precipitator at the kiln tail. The first to third fields are arranged sequentially along the flue gas flow direction, each field being 3.5 meters long, with a 0.4-meter-long flue gas duct between each field. Each field's side pendulum-type rapping mechanism is equipped with a MEMS capacitive accelerometer (range ±20g, sampling rate 3 kHz). An ultrasonic time-of-flight flue gas velocity transmitter is installed in the inlet flue. A shared rapping event log is maintained by an edge control computer in a time-series database format, with an aging time set to 480 seconds (seventh preset duration). Due to the high resistivity of the dust, the first preset threshold is adjusted to 0.07, the third preset duration to 6 seconds, and the fifth preset threshold to 130℃. When the kiln switches to variable load operation, the flue gas temperature slowly rises from 130℃ to 145℃, while the flue gas velocity decreases from 1.5 m / s to 0.9 m / s.

[0051] For the third electric field: the candidate rapping time received by S0 is 08:20:00; the most recent rapping event of the third electric field in S1 is 08:15:00. Taking this time as the zero point, the envelope is taken to calculate the attenuation ratio. At 08:15:02.5, r first drops below 0.07 and remains below 0.07 for 2 seconds. Therefore, the settlement is completed at 08:15:04.5. S2: Scan the event log. The second electric field vibrates at 08:19:55. The flue gas velocity is the average of the 30 seconds before 08:19:55 (the second preset duration), which is 1.0 m / s. The expected start time is 0.4 m divided by 1.0 m / s to get 0.4 seconds, which is then added to 08:19:55 to get 08:19:55.4 seconds. The expected end time of crossing this electric field is 08:20:04.9 seconds. The expected start time is 08:19:55.4 seconds, plus 3.5 m divided by 1.0 m / s to get 3.5 seconds, plus the adjusted third preset duration of 6 seconds. The active downstream freezing period of the third electric field is from 08:19:55.4 seconds to 08:20:04.9 seconds. S3: The candidate rapping time is 08:20:00, falling within the downstream freezing period; and the extended triggering condition of the flue gas temperature of 145℃ being lower than the adjusted fifth preset threshold of 130℃ does not trigger additional extension; the flue gas flow rate, converted according to the current flow rate, is approximately 0.7 times the design value, which is exactly equal to the default value of the sixth preset threshold of 0.7 times the design flow rate, so no additional extension is triggered; proceed to S5; the S5 delay judgment time is taken as the later one, 08:20:04.9 seconds, and the number of delays is accumulated to 1. At 08:20:04.9 seconds, S3 judgment is performed again. If both judgments pass, proceed to S4 to trigger the third electric field rapping.

[0052] Example 3 (Low-load long-term operation condition of electrostatic precipitator for dual-chamber dual-electric field cement coal mill) A ball mill is equipped with a dual-chamber, dual-field electrostatic precipitator. Each chamber has a front electric field and a rear electric field along the flue gas flow direction, for a total of four electric fields. Each electric field is 3 meters long, and the flue gas flow channel between the electric fields is 0.6 meters long. Each electric field's electromagnetic pulse rapping mechanism is equipped with a strain gauge accelerometer (range ±100 g, sampling rate 2 kHz). The shared rapping event log is mirrored between the two chamber controllers by a low-voltage control DCS. The unit operates in a long-term low-load zone with a boiler load of 60%, and the measured flue gas velocity is 0.7 m / s. Due to the consistently low load, the fourth preset threshold is adjusted from the default 5 times to 8 times to avoid cumulative delay in dust removal, and the second preset duration is adjusted from the default 30 seconds to 45 seconds.

[0053] For the back electric field of one of the chambers: S0 receives the candidate rapping time at 09:00:00; S1, the most recent rapping event of the back electric field is at 08:55:00, and at 08:55:01.8, the attenuation ratio first drops below the first preset threshold of 0.10 and remains below 0.10 for 2 seconds, so the settling completion time is 08:55:03.8; S2: Scan the event log to query the upstream events within 45 seconds before and after (adjusted second preset duration), the front electric field at 08:59:50 and 08:59:58... There are two impact events per second. The first is estimated to start at approximately 0.86 seconds (0.6 meters divided by 0.7 meters per second), then added to 08:59:50, resulting in 08:59:50.86. The estimated end time is calculated by adding approximately 4.29 seconds (3 meters divided by 0.7 meters per second) to the estimated start time of 08:59:50.86, plus a third preset duration of 3 seconds, resulting in 08:59:58.15. The downstream freeze period 1 is recorded as 08:59:50.86 to 08:59:58.15. The second is estimated to start at 08:59:58 plus 0.86 seconds, resulting in 08:59:58.86. The estimated end time is 09:00:06.15, plus 4.29 seconds and then 3 seconds. The downstream freezing period 2 is recorded as 08:59:58.86 to 09:00:06.15. S3: If the candidate vibration time of 09:00:00 falls within the downstream freezing period 2, proceed to S5; the later of the following decision time is taken: the settlement completion time of 08:55:03.8, the end of downstream freezing period 1 of 08:59:58.15, and the end of downstream freezing period 2 of 09:00:06.15, is 09:00:06.15. The cumulative number of delays is 1, which is far less than the adjusted fourth preset threshold of 8. S3 is performed again at 09:00:06.15. Both decisions pass, proceeding to S4 for triggered electric field vibration.

[0054] Example 4 (Compared to one of the non-triggered branches, this site has not completed settlement but there is no upstream conflict) Based on the four-field configuration and operating conditions described in Example 1, assuming the most recent rapping event of the second electric field occurred at 12 minutes and 00 seconds, the fine dust particle size resulted in slow envelope decay. At t0+1.6 seconds, r=0.12, which is greater than the first preset threshold of 0.10, so the settlement completion determination has not yet been triggered. At t0+4 seconds, r=0.085, dropping below the threshold for the first time, but at t0+5 seconds, r fluctuated and rose back to 0.11, so the "continuous duration below the threshold" state variable returned to zero. Finally, at t0+8.5 seconds, r stabilized at 0.085 and remained there for 2 seconds, so the settlement completion time was t0+10.5 seconds, i.e., 12 minutes and 10.5 seconds. At the same time, scanning the event log revealed that the most recent rapping event of the first electric field was more than 60 seconds ago, which is not within the second preset duration query window, so there is no active downstream freezing period for this electric field. S3 Judgment: When the candidate rapping time of 12 minutes and 04 seconds arrives, the settling completion time does not yet exist (it has not yet reached 12 minutes and 10.5 seconds), so the first judgment is invalid, and proceed to S5; the S5 judgment time is postponed to the settling completion time of 12 minutes and 10.5 seconds (if there is no downstream freezing period, only the settling completion time is taken), and the number of postponements is accumulated to 1. S3 judgment is performed again at 12 minutes and 10.5 seconds. If both judgments pass, proceed to S4 to trigger the local field rapping. This embodiment illustrates that even without cross-field particle cloud conflict, the local electric field acceleration decay envelope can still independently identify that the local dust has not yet settled, postpone the local field rapping, and avoid secondary dust re-entrainment in the local field.

[0055] Example 5 (Compared to non-triggered branch 2, this site has completed settlement but there is an upstream conflict) Based on the three-field cement kiln tail configuration described in Example 2, the most recent rapping event of the third electric field is set at 08:15:00, and the settlement completion time under normal operating conditions is 08:15:04.5. S2 Scan Event Log: The most recent rapping event of the second electric field occurred at 08:19:59, with a measured flue gas velocity of 0.8 m / s. The estimated start time is calculated by dividing 0.4 m by 0.8 m / s to obtain 0.5 seconds, then adding this to 08:19:59 to get 08:19:59.5 seconds. The estimated end time is calculated by adding 3.5 m divided by 0.8 m / s to obtain 4.375 seconds, and then adding the adjusted third preset duration of 6 seconds, to obtain 08:20:09.875 seconds. The active downstream freezing period of the third electric field is from 08:19:59.5 seconds to 08:20:09.875 seconds. S3 Judgment: The candidate impact time is 08:20:00. The settlement completion time of 08:15:04.5 is later than the most recent impact event in this area, so the first criterion passes. However, the candidate impact time of 08:20:00 falls within the downstream freezing period (08:19:59.5 to 08:20:09.875), so the second criterion fails, and proceeds to S5. The S5 delayed judgment time is the later of the settlement completion time of 08:15:04.5 and the end of the downstream freezing period of 08:20:09.875, which is 08:20:09.875. The number of delays is accumulated to 1. S3 judgment is performed again at 08:20:09.875. Both judgments pass, and proceeds to S4 to trigger the impact in this area. This embodiment illustrates that even if the local sedimentation has been completed, cross-site particle cloud conflict can independently delay the local vibration, avoiding the superposition of secondary dust from the cross-site.

[0056] Example 6 (Six-field ultra-low emission electrostatic precipitator under full load) A 1000 MW ultra-supercritical double reheat coal-fired unit is equipped with a six-field horizontal electrostatic precipitator, arranged sequentially from the first to the sixth field along the flue gas flow direction. Each field is 4.2 meters long, and the flue gas duct between the fields is 0.55 meters long. Each field's top rotary hammer rapping mechanism is equipped with two piezoelectric accelerometers (range ±50g, sampling rate 10 kHz, positioned on the force transmission rods on the left and right sides of the anode plate assembly, respectively). A dual Pitot tube array is used in the inlet flue for measured flow velocity redundancy. A shared rapping event log is synchronized by the low-voltage control system DCS within a 500-second circular buffer and mirrored among the six field controllers via the PROFINET industrial Ethernet protocol. The unit operates continuously at 100% boiler load, with a flue gas temperature of 132℃ and a measured flue gas velocity of 1.0 m / s. Considering that the unit adopts ultra-low emission standards and the daily average value of exhaust dust is required to be less than 5 milligrams per standard cubic meter, the first preset threshold is adjusted to 0.07, the first preset duration is adjusted to 3 seconds, and the third preset duration is adjusted to 5 seconds.

[0057] For the fourth electric field: S0 receives the candidate rapping time given by the original rapping scheduling module as 14:30:00; the most recent rapping event time of the fourth electric field in S1 is 14:24:00. Taking this time as the zero point, the maximum value of the envelope of the left and right sensors is taken as the envelope of this electric field. The initial peak value of the envelope is e0=6.2g. At 14:24:02.4, the attenuation ratio drops below 0.07 for the first time (measured r=0.061), and remains below 0.07 for 3 seconds (adjusted first preset duration). The settling is completed at 14:24:05.4. S2: Scanning the event log revealed vibration events at 14:29:56 for the second electric field and at 14:29:58 for the third electric field. For the second electric field event, the distance from the second electric field to the fourth electric field traverses two flue gas channels, with a cumulative length of 0.55 meters plus 0.55 meters plus 4.2 meters (the length of the third electric field), equaling 5.3 meters. The estimated start time is calculated by dividing 5.3 meters by 1.0 meters per second to get 5.3 seconds, then adding this to 14:29:56, resulting in 14:30:01.3 seconds. The estimated end time is this time plus 4.2 meters divided by 1.0 meters per second (traversing the fourth electric field) plus 5 seconds (the adjusted third preset duration), equaling 14:30:10.5 seconds. For the third electric field event, the distance from the third electric field to the fourth electric field traverses a flue gas channel with a length of 0.55 meters. The estimated start time is 14:29:58. Adding 0.55 meters to 1.0 meter per second equals 0.55 seconds, which is 14:29:58.55. The expected end time is this time plus 4.2 meters, divided by 1.0 meter per second, plus 5 seconds, which equals 14:30:07.75. Therefore, the current active downstream freezing period of the fourth electric field is the union of period 1 (14:29:58.55 to 14:30:07.75) and period 2 (14:30:01.3 to 14:30:10.5), which is 14:29:58.55 to 14:30:10.5. S3: The candidate rapping time of 14:30:00 falls into the downstream freezing period, proceeding to S5; the S5 delay judgment time is taken as the later of the settlement completion time of 14:24:05.4, the end of downstream freezing period 1 of 14:30:07.75, and the end of downstream freezing period 2 of 14:30:10.5, which is 14:30:10.5, and the delay count is accumulated to 1; at 14:30:10.5, S3 judgment is performed again. If both settlement completion and no downstream conflict are found, proceeding to S4 triggers the fourth electric field rapping and broadcasts to the fifth and sixth electric fields.

[0058] Example 7 (Significant Disturbance During Coal Mixing Switching) Based on the four-field horizontal electrostatic precipitator configuration described in Example 1, the unit switched from mixed coal A to mixed coal B between 11:00 and 11:30. The difference in coal quality caused a step change in fly ash particle size and resistivity. The average fly ash particle size decreased from 22 micrometers to 13 micrometers, and the dust resistivity decreased from approximately... Eurocentimeters rise to approximately The flue gas temperature rose from 138°C to 152°C, while the flue gas velocity remained around 1.2 meters per second.

[0059] Before the switchover, the first preset threshold was set to the default value of 0.10, and the third preset duration was set to the default value of 3 seconds, with the system operating smoothly. During the switchover, based on the rapid coal quality analysis data, the operators issued parameter adjustment commands through the upper-level monitoring system: the first preset threshold was adjusted from 0.10 to 0.08 (due to the finer dust particle size, the settling judgment needs to be more stringent), and the third preset duration was adjusted from 3 seconds to 5 seconds (due to the longer particle cloud suspension time due to the finer particle size). The system received the adjustment command at 11:12:00 and hot-loaded the new parameters. After 11:12:00, the settling completion judgment for all S1 samples was recalculated based on the 0.08 threshold and the default 2-second first preset duration. After 11:12:00, the downstream freezing period calculation for all S2 samples was recalculated based on the 5-second third preset duration. Following the most recent rapping event in the second electric field at 11:12:10, S1 determined the settlement completion time at 11:12:13.2. Under the influence of the rapping event in the first electric field at 11:12:08, S2 calculated the downstream freezing period of the second electric field to be from 11:12:08.42 to 11:12:16.75 (applying a third preset duration of 5 seconds). The candidate rapping time of 11:12:15 falls within the downstream freezing period, proceeding to S5, and the determination is postponed to 11:12:16.75. This embodiment illustrates that all preset thresholds and durations of this method support hot loading during operation, allowing operators to quickly analyze data based on coal quality or load and issue adjustment commands when operating conditions are significantly disturbed, without requiring a shutdown and restart of the control system.

[0060] Example 8 (Alarm triggering branch when continuous delay reaches the limit) Based on the four-field configuration described in Example 1, the most recent rapping event in the second electric field is set to 12 minutes and 00 seconds. The fine dust particle size, combined with frequent upstream rapping, causes multiple rounds of judgment to fail. Round 1: The candidate vibration time is 12 minutes and 04 seconds. The settlement completion time does not yet exist, so it enters S5. The number of delays accumulates to 1, delaying to 12 minutes and 10.5 seconds. Round 2: At 12 minutes and 10.5 seconds, the settlement completion time exists, but the first electric field vibration occurs at 12 minutes and 08 seconds. The downstream freezing period is from 12 minutes and 08.42 seconds to 12 minutes and 14.75 seconds. The candidate vibration time of 12 minutes and 10.5 seconds falls within the freezing period. The number of delays accumulates to 2, delaying to 12 minutes and 14.75 seconds. Round 3: At 12 minutes and 14.75 seconds, the first electric field vibration occurs again at 12 minutes and 13 seconds. The downstream freezing period is from 12 minutes and 13.42 seconds to 12 minutes and 19.75 seconds. The number of delays accumulates to 3, delaying to 12 minutes and 19.75 seconds. Round 4: The first electric field vibration occurs again at 12 minutes and 18 seconds. The vibration is activated in seconds, with the number of delays accumulating to 4. In the 5th round, the number of delays accumulates to 5, which equals the fourth preset threshold of 5. The candidate vibration is skipped, and the event is reported to the upper-level monitoring system as an alarm in the form of "The second electric field has continuously delayed to the upper limit, skipping this candidate vibration". The operators intervene based on the alarm prompt, and after ruling out the possibility of abnormal operation of the first electric field vibration mechanism, they consider it to be a temporary congestion under normal operating conditions, and no shutdown is required. This embodiment illustrates that this method has a clear alarm reporting mechanism for extreme congestion situations and will not accumulate delays indefinitely.

[0061] General Implementation: The method described in this invention can be implemented on electrostatic precipitators with any multi-field structure. The number of electric fields in the electrostatic precipitator is typically 2 to 6, but it can also be extended to a single electric field (in which case, if there is no upstream electric field in S2, the downstream freezing period is always empty, and the method degenerates into a single-field rapping control scheme based solely on the settling completion time criterion). The electric fields can be arranged in series of the same type, or they can be mixed and arranged of different types (such as the front-stage C480 corrugated plate group and the rear-stage ZT24 tension plate group). The method is applicable to both dry electrostatic precipitators and wet electrostatic precipitators (the water washing cycle of the wet electrostatic precipitator and the rapping sequence of this method are synchronized in time through an event log). The method is applicable to flue gas dust removal in coal-fired power plant boilers, and can also be applied to other flue gas dust removal scenarios such as cement kiln tail, sintering flue gas, glass melting furnace, biomass boiler, and waste incinerator. Differences in dust characteristics can be adapted by adjusting the values ​​of the seven preset thresholds and durations described in this invention.

[0062] The acquisition of the anode plate acceleration signal described in S1 is specifically located on the anode plate group struck by the top rotary hammer, side pendulum, or electromagnetic pulse rapping mechanism of each electric field. The acceleration sensor is attached to the back of the anode plate group or installed at the connection between the rapping force transmission rod and the anode plate group. The sampling signal is transmitted to the low-voltage control system of the electrostatic precipitator via the fieldbus protocol. In the rapping mechanism with multi-hammer synchronous rapping, one acceleration sensor can be arranged on each anode plate group corresponding to each force transmission rod. The attenuation ratio is calculated based on the maximum value of the envelope of all acceleration sensors in this electric field.

[0063] The pre-calibration of the flue gas flow path length between electric fields in S2 is specifically defined as the straight-line distance between the vertical center lines of adjacent electric fields given in the electrostatic precipitator design drawings. This value is verified once by engineering surveying before the electrostatic precipitator is put into operation and then written into the constant table of the low-voltage control system. The measured flue gas velocity is specifically defined as the output of the Pitot tube array or ultrasonic time-of-flight transmitter arranged in the flue gas duct at the inlet of the electrostatic precipitator. The sampling period is typically 1 second. During the S2 scan, the average value of the sliding window not exceeding the second preset time is taken to filter out instantaneous turbulent disturbances.

[0064] In S4, the rapping event is broadcast downstream, specifically to the inter-process message bus within the low-voltage control system DCS or PLC. The message format includes the electric field identifier, the rapping event timestamp, and an optional initial peak value of the envelope. In the edge industrial control computer time sequence database implementation mode, broadcasting is equivalent to writing a record to the database and triggering the subscription callback of the downstream electric field controller.

[0065] The delay action described in S5 specifically refers to the next trigger time setting of the electric field scheduler. The electric field scheduler does not issue a new rapping candidate time before the next discrimination time arrives. During the delay of the electric field, the acceleration sensor of the electric field continues to collect signals, the attenuation ratio continues to be updated, and the downstream freezing period continues to be calculated based on the upstream event. Only the rapping action itself is frozen. Under the multi-electric field parallel processing architecture, the S5 decision of each electric field does not affect each other, and each electric field can be in a different delay state.

[0066] The seven preset thresholds and durations mentioned in this method are all stored in the parameter table of the low-pressure control system, which supports hot loading during operation. In scenarios such as boiler load jumps, coal type switching, and large fluctuations in flue gas conditions, operators can issue parameter adjustment commands through the upper-level monitoring system. The new parameters after adjustment will take effect from the next scheduling cycle. The postponed loading time corresponding to the old parameters will not be affected, and the judgment will continue to be based on the time predetermined by the old parameters.

[0067] The accelerometer signal channel, the flue gas velocity sensor signal channel, and the shared event log read / write channel are all equipped with heartbeat detection. If any channel fails to receive a valid signal or fails to read / write for more than seven preset durations, a fault alarm is triggered for the corresponding channel. During the fault period, the electric field degrades to vibrating at a fixed cycle (typically 180 seconds) and reports the fault type and timestamp to the upper monitoring system. After the fault is recovered, the electric field automatically switches back to the control flow described in this invention.

[0068] The system described in this invention comprises several modules, each of which can be implemented by a functional block diagram of the low-voltage control system DCS or PLC of the electrostatic precipitator, or by an executable program on an edge industrial control computer, or by an independent intelligent optimization control cabinet for the electrostatic precipitator. Data transmission between modules adopts industrial Ethernet protocol or fieldbus protocol. The settlement completion determination module supports redundant deployment, that is, two sets of acceleration signal channels and settlement determination algorithms can be deployed simultaneously in each electric field to meet the requirements of power supply continuity. The candidate rapping receiving module, dual-criteria blocking module, rapping triggering and broadcasting module, and delay discrimination module are all deployed on the electric field controller of the electric field to realize local decision-making for the electric field. The downstream freezing period estimation module needs to access the shared rapping event log, which can be deployed on the electric field controller (subscribing to the local mirror of the event log) or centrally deployed on the edge industrial control computer (to uniformly calculate and distribute results for all electric fields). The settlement completion determination module is arranged close to the acceleration sensor to shorten the signal link delay.

[0069] The interfaces between modules can adopt industrial interconnection protocols such as OPC UA, Modbus TCP, and MQTT; the interface data types include rapping event structure (including electric field identifier, event timestamp, and initial peak value of the envelope), settlement completion time, downstream freezing period structure (including start time, end time, and upstream event identifier that triggered the period), and operating status enumeration (including normal, delayed, skipped alarm, and sensor failure); the latency requirement of each interface is no more than 100 milliseconds to ensure timely coordination between multiple electric fields.

[0070] The method described in this invention can be stored in a computer-readable storage medium, including but not limited to hard disks, solid-state drives, optical disks, and flash memory. The medium stores a computer program, which, when executed by a processor within the intelligent optimization control cabinet of the electrostatic precipitator, implements all steps S0 to S5 of this method. The computer program can be implemented based on an industrial configuration software platform (such as iFix, WinCC, ControlLogix), or based on a general-purpose programming language (such as C / C++, Python, Rust), or based on a function block diagram of an industrial real-time operating system; the implementation form is not limited.

[0071] This invention can be implemented as an anode rapping control device for an electrostatic precipitator. The device includes a processor, a memory, a communication interface with the low-voltage control system of the electrostatic precipitator, a signal interface with an acceleration sensor, and a signal interface with a flue gas velocity sensor. The memory stores program instructions that can be executed by the processor. When the program instructions are executed, they implement all steps S0 to S5 of this method. The typical physical form of the device is an industrial-grade embedded controller or industrial computer, which has dustproof, moisture-proof, vibration-resistant, and electromagnetic interference-resistant capabilities, and its operating temperature range is typically 0℃ to 55℃. The device also includes auxiliary function modules for power-on self-test, parameter hot-loading, operation log recording, and fault alarm reporting.

[0072] To verify the practical engineering effect of this method, a four-field electrostatic precipitator configured in a 300 MW unit in North China was selected as the experimental object. The rated evaporation capacity of the boiler in this unit is 1025 tons per hour, and the coal type is a blend of Shanxi anthracite and Shenhua mixed coal, with a high fly ash resistivity.

[0073] On the data set of flue gas emissions from thermal power units published by the State Power Engineering Corporation (comprising over 30,000 hourly average samples from 2022 to 2024), the offline simulation results of this method are compared with existing typical schemes. These existing typical schemes include: Scheme A (fixed-period rapping), Scheme B (delayed rapping based on outlet concentration thresholds), and Scheme C (coordinated rapping based on upstream-driven cross-field readiness checks). Under a four-field configuration, the daily average peak amplitude of the rapping simulation obtained by this method is reduced by approximately 58% compared to Scheme A, approximately 45% compared to Scheme B, and approximately 31% compared to Scheme C. Under a six-field configuration, the daily average peak amplitude of the rapping simulation obtained by this method is reduced by approximately 62% compared to Scheme A, approximately 48% compared to Scheme B, and approximately 35% compared to Scheme C. The advantages of this method are more pronounced under multi-field configurations, reflecting the inherent advantage of this method in back-calculating the downstream freezing period based on measured gas velocity compared to other schemes' binarized coordination.

[0074] The experiment was deployed and operated for 15 days on the 300 MW unit in North China. The daily average raw coal consumption and boiler load remained basically consistent with the control period. The comparison index was the difference between the hourly average dust concentration measured by the Continuous Electrostatic Precipitator (CEMS) and the instantaneous peak value within the hour, i.e., the rapping peak amplitude. The control group adopted a fixed-cycle rapping scheme, with the rapping cycles of the first to fourth electric fields being 180 seconds, 240 seconds, 360 seconds, and 480 seconds, respectively. The experimental group adopted the control scheme described in this method, with each preset threshold and duration taking the typical default values ​​described in this invention. The results showed that the daily average rapping peak amplitude of the control group was 12.3 mg / m³, while that of the experimental group was 4.8 mg / m³, a decrease of approximately 61%. The hourly average outlet dust concentration of the control group was 6.8 mg / m³, while that of the experimental group was 5.2 mg / m³, a decrease of approximately 24%. The total number of rapping sessions and the energy consumption per rapping session were recorded for both groups over 15 days. The average number of rapping sessions per day in the control group was about 720, while the average number of rapping sessions per day in the experimental group was about 560. The delay and skipping mechanism reduced the number of ineffective rapping sessions by about 22%. The energy consumption per rapping session remained unchanged, so the power consumption of the rapping system in the experimental group was reduced by about 22% compared with that in the control group.

[0075] Based on the experimental group, the function of determining settlement using the acceleration decay envelope was disabled (only a fixed duration of 50 seconds was used as the settlement completion criterion), and the functions of upstream event broadcasting and downstream freezing period estimation were disabled (only the settlement criterion for this field was retained) for control testing. The results showed that when only the acceleration decay envelope function was disabled for settlement, the daily average value of the rapping peak amplitude increased to 7.8 mg / m³; when only the upstream event broadcasting and downstream freezing period estimation were disabled, the daily average value of the rapping peak amplitude increased to 8.4 mg / m³; when both functions were disabled simultaneously (equivalent to degrading to a fixed period scheme), the daily average value of the rapping peak amplitude increased to 11.9 mg / m³. This ablation experiment confirms that the two criteria described in this invention contribute independently and cannot be substituted for each other, and that the combined effect of the two exceeds the sum of their individual contributions (cross-reinforcement effect).

[0076] With other parameters fixed, the impact of a first preset threshold ranging from 0.05 to 0.20 and a third preset duration ranging from 1 to 10 seconds on the daily average reduction of flue dust at the outlet was scanned. The results showed that when the first preset threshold was between 0.08 and 0.12 and the third preset duration was between 2 and 5 seconds, the reduction was stable between 23% and 26%. When the threshold was too small (e.g., 0.05), the settling determination time increased significantly, potentially triggering an alarm due to the upper limit of consecutive delays being reached. When the duration was too large (e.g., 10 seconds), the overall dust removal cycle was lengthened, increasing the risk of back corona discharge from the upstream electric field. When the threshold was too large (e.g., 0.20), the settling determination might be completed prematurely, weakening the suppression effect of secondary dust generation in the local area. When the duration was too small (e.g., 1 second), the downstream freezing period was shortened, weakening the combined suppression effect of secondary dust generation across different areas. In summary, the typical default value described in this invention is taken at the center of the threshold-sensitive area, providing good engineering margin.

[0077] The experimental algorithm was run on an intelligent optimization control cabinet for an electrostatic precipitator. The control cabinet uses an industrial-grade x86 processor with a 2.0 GHz clock speed and 4 gigabytes of memory. Under typical operating conditions, the single-round scheduling time of this method does not exceed 50 milliseconds, far less than the minimum rapping cycle of 180 seconds, and does not constitute a computational bottleneck. Even in the extreme case of a six-field configuration and an event log depth of 500 seconds, the single-round scheduling time does not exceed 90 milliseconds, still far less than the rapping cycle.

[0078] This method was compared and verified on electrostatic precipitator units in three different industries: The first group was a 300 MW coal-fired unit in North China (four-field horizontal electrostatic precipitator, co-firing of coal), which was deployed and operated for 15 days, and the daily average value of the rapping peak amplitude decreased by about 61%; the second group was a 5000-ton-per-day cement clinker production line kiln tail in East China (three-field horizontal electrostatic precipitator, fine dust particle size and high resistivity), which was deployed and operated for 10 days, and the daily average value of the rapping peak amplitude decreased from 9.6 mg / m³ in the control group to 3.8 mg / m³, a decrease of about 60%; the third group was a 660 MW supercritical coal-fired unit in South China (four-field horizontal electrostatic precipitator, pure Shenhua mixed coal), which was deployed and operated for 15 days, and the daily average value of the rapping peak amplitude decreased from 8.4 mg / m³ in the control group to 3.6 mg / m³, a decrease of about 57%. Three sets of comparative verifications show that the reduction in the amplitude of the rapping peak of this method remains stable within the range of 55% to 65% under different industries, different coal types, and different electric field configurations, demonstrating good universality.

[0079] When the electrostatic precipitator is put into operation or the controller is restarted, the shared rapping event log is empty, and there are no historical rapping event times for any electric field; at this time, the downstream freezing period of all electric fields is empty, and S3 judgment depends only on the settling completion time; after the initial rapping event, the event log begins to accumulate, and the method smoothly transitions to the full operation mode.

[0080] The measured flue gas velocity may reach a minimum (e.g., below 0.1 m / s) or a maximum (e.g., above 2 m / s) or even zero value during boiler start-up, shutdown, or fuel switching. To prevent the downstream freezing period from being divided by zero or resulting in an excessively large period, a lower limit protection is applied to the measured flue gas velocity (typically 0.3 m / s). If the measured value is lower than the lower limit, the calculation is performed at the lower limit. At the same time, if the calculated downstream freezing period exceeds the seventh preset duration, it is truncated to the seventh preset duration.

[0081] After the controller restarts, the shared rapping event log is restored from a non-volatile storage area (such as the power-off protection data area of ​​the low-voltage control PLC or the local database of the edge industrial control computer), but only events that are no more than seven preset times away from the restart time are restored; the settlement completion time is not persisted (because its physical meaning is related to the acceleration signal in real time, and it needs to be re-observed and acquired after the restart), and normal settlement judgment can be restored after the first rapping event after the restart.

[0082] If the controller clock jumps (e.g., NTP time synchronization causes the time to jump forward or backward by several seconds), all event log records are corrected by adding the difference of the jump variable, the status variable for determining the settlement completion is forcibly cleared to zero and starts accumulating again after the next rapping event.

[0083] If the number of consecutive delays of a certain electric field reaches the fourth preset threshold, the candidate rapping is skipped and an alarm is reported to the upper monitoring system; if the same electric field skips the candidate rapping more than 5 times in a row, a higher-level alarm is triggered, prompting the operator to manually intervene and investigate; if the acceleration sensor signal channel has no effective signal for a continuous period of time (the envelope has no obvious spike after the rapping event is triggered), the electric field degenerates into rapping at a fixed period of 180 seconds, and a sensor fault alarm is triggered.

[0084] The embodiments of the present invention are merely illustrative, and the specific embodiments of the present invention are not limited to any of the above embodiments. Without departing from the spirit of the present invention and the scope of the claims, those skilled in the art can make formal changes or adjustments to local parameters to the above embodiments, all of which fall within the protection scope of the present invention.

[0085] The method described in this invention can be used alone or in conjunction with existing high-voltage power supply control strategies for electrostatic precipitators, such as back corona suppression, intermittent power supply, pulse power supply, and high-frequency power supply. In the case of coordinated use, the instantaneous triggering of the S4 rapping action can be time-coordinated with the instantaneous voltage reduction action of the high-voltage power supply of the electrostatic precipitator. Typically, this is achieved by sending a voltage reduction request to the high-voltage power supply of the electrostatic precipitator 100 milliseconds before the S4 rapping action is triggered, and then restoring the voltage to the rated supply voltage after the settling is completed. This further improves the rapping cleaning efficiency and reduces the risk of back corona during rapping. The specific timing of this coordination can be uniformly scheduled by the low-voltage control system or by the high-voltage power supply itself performing an event subscription response, both of which fall within the scope of protection of this invention.

Claims

1. A method for controlling the anode rapping of an electrostatic precipitator, wherein the electrostatic precipitator comprises multiple electric fields arranged sequentially along the flue gas flow direction, each electric field being equipped with an anode plate group, a rapping mechanism and an acceleration sensor mounted on the anode plate group, and also equipped with a flue gas flow rate sensor and a shared rapping event log; The method includes: S0 receives the candidate rapping time sequence of each electric field and the event time of the last actual rapping event; S1, for each electric field, take the time of its most recent rapping event as the zero point, and calculate the attenuation ratio by taking the envelope of the acceleration signal; When the attenuation ratio drops below the first preset threshold for the first time and remains at a time not less than the first preset duration, the moment is recorded as the moment when the settling of this electric field is completed. S2, For each electric field, query the recent rapping event times of all electric fields located upstream of this electric field along the flue gas flow direction in the shared rapping event log. Using the pre-calibrated ratio of the flue gas flow channel length between electric fields to the measured flue gas flow velocity, calculate the estimated start time of the upstream rapping particle cloud arriving at this electric field and the estimated end time of crossing this electric field. The time period between the start and end times is recorded as the downstream freezing period of this electric field. S3, when the candidate rapping time of this electric field arrives, determine whether the following conditions are met simultaneously: there is already a settlement completion time later than the time of its most recent rapping event; Furthermore, the candidate slamming time does not fall within any active downstream freeze period; If both conditions are met, proceed to S4; otherwise, proceed to S5. S4, trigger the rapping of this electric field, write the current moment as the new rapping event moment into the shared rapping event log, and broadcast it to all electric fields located downstream of this electric field along the flue gas flow direction; S5, postpone the vibration of this electric field to the next judgment time. The next judgment time is the later of the time when the settlement of this electric field is completed and the time when the freezing period of all active downstream of this electric field ends. After the postponement, return to S3.

2. The method according to claim 1, characterized in that, The attenuation ratio mentioned in S1 is the ratio of the current peak value of the envelope of the electric field acceleration signal to the initial peak value of the envelope corresponding to the moment when the most recent rapping event of the electric field was triggered; The first preset threshold value is not less than 0.05 and not greater than 0.20, and the first preset duration value is not less than 1 second and not greater than 5 seconds.

3. The method according to claim 1, characterized in that, The recent vibration event time mentioned in S2 is the upstream electric field vibration event time that is no more than the second preset time from the current time; When the upstream electric field spans multiple flue gas channels between electric fields, the estimated start time is calculated as the cumulative ratio of the length of each channel along the route to the measured flue gas velocity.

4. The method according to claim 1, characterized in that, The expected end time of crossing this electric field mentioned in S2 is based on the expected start time, plus the ratio of the electric field length along the flue gas flow direction to the measured flue gas flow velocity, and plus a third preset time as a safety margin.

5. The method according to claim 1, characterized in that, In S5, the number of consecutive delays of this electric field does not exceed the fourth preset threshold. When the number of consecutive delays equals the fourth preset threshold, the candidate vibration is skipped, and the skipped event is reported to the upper monitoring system as an alarm.

6. The method according to claim 1, characterized in that, The determination of falling into the active downstream freezing period in S3 also includes checking the flue gas temperature and flue gas flow rate of this electric field during this rapping. When the flue gas temperature is higher than the fifth preset threshold or the flue gas flow rate is lower than the sixth preset threshold, the end time of the downstream freezing period is extended by a predetermined proportional coefficient.

7. The method according to claim 1, characterized in that, The measured flue gas velocity mentioned in S2 is the average value of the flue gas velocity sensor arranged in the inlet flue of the electrostatic precipitator within a certain time window from the moment of the upstream electric field rapping event.

8. The method according to claim 1, characterized in that, The shared rapping event log saves the times of the most recent rapping events of each electric field in chronological order, and ages and deletes events that exceed the seventh preset duration. When broadcasting downstream in S4, the initial peak value of the acceleration signal envelope of this rapping is simultaneously written as an input for the downstream electric field to determine whether the downstream freezing period of its own electric field can be compressed.

9. The method according to any one of claims 1 to 8, characterized in that, The method further includes: the acceleration signals collected from each electric field are first processed by bandpass filtering before the settlement completion time is recorded, and the bandpass frequency band covers the frequency band of the dominant vibration mode of the anode plate induced by the rapping mechanism.

10. An anode rapping control system for an electrostatic precipitator, the electrostatic precipitator comprising a plurality of electric fields arranged sequentially along the flue gas flow direction, each electric field being equipped with an anode plate assembly, a rapping mechanism and an acceleration sensor mounted on the anode plate assembly, and also equipped with a flue gas flow rate sensor; Its features are, include: The candidate rapping receiving module is used to receive the candidate rapping time sequence of each electric field and the event time of the last actual rapping, corresponding to step S0 of the method. The settling completion determination module is used to take the envelope of the acceleration signal and calculate the attenuation ratio with the time of the most recent rapping event of the electric field as the zero point, and record the time when the attenuation ratio drops below the first preset threshold and remains not less than the first preset duration as the time when the settling completion of the electric field is recorded, corresponding to step S1 of the method. The downstream freezing period estimation module is used to query the recent rapping event times of all electric fields located upstream of this electric field along the flue gas flow direction in the shared rapping event log. Using the pre-calibrated ratio of the flue gas flow channel length between electric fields to the measured flue gas flow velocity, it estimates the expected start time of the upstream rapping particle cloud arriving at this electric field and the expected end time of crossing this electric field. The time period between the start and end times is recorded as the downstream freezing period of this electric field, corresponding to step S2 of the method. The dual-criteria blocking module is used to determine whether the electric field simultaneously satisfies the following conditions when a candidate rapping time is reached: whether there is a settlement completion time later than the time of its most recent rapping event, and whether the candidate rapping time does not fall into any active downstream freezing period. This corresponds to step S3 of the method. The rapping triggering and broadcasting module is used to trigger the rapping of this electric field when the dual-criteria blocking module determines that both conditions are met, write the current time as the new rapping event time into the shared rapping event log, and broadcast it to all electric fields located downstream of this electric field along the flue gas flow direction, corresponding to step S4 of the method. The delayed discrimination module is used to delay the rapping of the electric field to the next discrimination time when the discrimination by the dual-criteria blocking module fails. The next discrimination time is the later of the time when the electric field settlement is completed and the time when the freezing period of all active downstream areas of the electric field ends, corresponding to step S5 of the method.

Images