Low-resistance demister water collecting and process water tank linkage circulation method

Abstract

The application discloses a low-resistance demister water collecting and process water tank linkage circulation method, and particularly relates to the field of wet desulfurization demister water collecting and cooperative control technology, and is used for solving the problem that the existing wet desulfurization demister liquid discharge is disconnected with water tank management, and it is difficult to simultaneously ensure low pressure drop operation and prevent liquid accumulation and secondary liquid carrying. By establishing a demister pressure drop low-resistance boundary based on a flue gas condition in an actual wet desulfurization device, combining with demister structure to divide a water collecting section, and clearly defining the water collecting path of each water collecting section and a process water tank, a water collecting demand index and an acceptance capacity coefficient are calculated according to a pressure drop margin coefficient and water collecting section and water tank state parameters during operation, a water collecting distribution matrix is generated, and each branch liquid discharge is controlled, so that the demister is preferentially ensured to be in a low-resistance safety interval under different loads and conditions, liquid accumulation and secondary liquid carrying are avoided, the downstream water tank liquid level and water quality safety are simultaneously considered, and the overall operation stability and safety of the wet desulfurization system are improved.

Description

Technical Field

[0001] This invention relates to the field of water collection and coordinated control technology for wet desulfurization demisters, specifically a method for linking water collection in a low-resistance demister with the process water tank for circulation. Background Technology

[0002] In wet desulfurization units of coal-fired power plants, the demister at the top of the absorber tower typically plays a crucial role in removing entrained droplets and reducing the risk of secondary liquid carryover. In existing projects, most units only provide empirical ranges for demister pressure drop based on typical flue gas conditions during the design phase. During operation, the demister's water collection capacity is maintained primarily through monitoring the pressure drop, appropriate flushing, and relying on fixed drainage pipeline layouts. The connections between the water collection tray, drainage branches, and downstream absorber tower circulating slurry tank, clear water tank, or intermediate water tank are mostly based on drawings and field experience. Drainage distribution is largely achieved through fixed valve positions or manual intervention, lacking detailed management that matches actual flue gas load, demister pressure drop changes, and the liquid levels and water quality of each tank.

[0003] In existing technical solutions, on the one hand, demister pressure drop control often uses a single upper limit setting or empirical threshold, which makes it difficult to reflect the low-resistance operating boundary under different flow and temperature conditions in a timely manner. This easily leads to contradictions such as compressing drainage capacity to pursue low pressure drop or operating under excessively high pressure drop for a long time to prevent liquid accumulation. On the other hand, there is a lack of unified quantitative and linkage control between the drainage capacity of each water collection zone of the demister and the receiving capacity of multiple downstream process water tanks. Operators often can only make local adjustments based on local liquid level or water quality alarms, and cannot take into account the risks of local liquid accumulation in the demister, water tank overflow, and water quality deterioration on a global scale. Existing systems generally lack a mechanism to uniformly and normally represent multiple source state quantities such as demister pressure drop margin, liquid level or water collection load in each water collection zone, and remaining volume and water quality margin of each water tank, and to dynamically allocate water collection paths and drainage flow rates accordingly. There is also a lack of systematic statistical analysis of long-term operating data to correct pressure drop control boundaries and drainage allocation strategies.

[0004] Therefore, existing wet desulfurization demisters still exhibit a significant disconnect between drainage control and process water tank management. This makes it difficult to effectively mitigate localized liquid accumulation and secondary liquid carryover risks while ensuring low resistance and safe operation of the demister, and to consider the liquid level and water quality constraints of different tanks. Establishing a coordinated control mechanism between pressure drop boundaries, water collection sections, and the topological relationship and operating status of the process water tank under actual operating conditions, so as to match low-resistance operation of the demister with safe drainage and the process water tank's receiving capacity, has become a key technical problem that needs to be addressed by those skilled in the art. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for linking water collection in a low-resistance demister with the process water tank for circulation, thereby solving the problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for linking water collection in a low-resistance demister with the process water tank for circulation, the method being executed by a control system deployed in a wet desulfurization unit, comprising:

[0007] S1. Test the pressure drop of the demister based on the flue gas operating conditions, establish a low resistance boundary where the pressure drop does not exceed the preset envelope, and divide the water collection section according to the structure and water collection path, and determine the process water tank connected to the water collection section.

[0008] S2. Periodically collect flue gas operating conditions, demister pressure difference, water collection volume in the water collection section, water tank level and water quality, and normalize the data to obtain the pressure drop margin coefficient, as well as the state parameters of the water collection section and the state parameters of the water tank.

[0009] S3. Calculate the water demand index of each water collection section based on the pressure drop margin coefficient and the state parameters of the water collection section, and calculate the receiving capacity coefficient of each water tank based on the state parameters of the water tank.

[0010] S4. Construct a water collection distribution matrix based on the water collection demand index of the water collection section and the water tank acceptance capacity coefficient, and convert it into the target distribution flow of the water collection branch.

[0011] S5. Establish a state machine for normal and drainage states. In normal state, smoothly adjust the opening of the water collection branch according to the sliding time window to track the target distribution flow. In drainage state, increase the water collection volume of the water collection section exceeding the threshold and give priority to the water tank with the receiving capacity, and reduce the distribution volume of other water tanks.

[0012] S6. Summarize the pressure drop, water intake, water quality and operating status records according to the statistical period, correct the weights and thresholds of the low resistance boundary, water intake demand index and acceptance capacity coefficient, and obtain updated control parameters for subsequent operation.

[0013] Furthermore, S1 includes:

[0014] Collect flue gas operating conditions and demister pressure drop, align the collected data according to a unified timestamp, and filter out pressure drop interference pulses;

[0015] Divide the range according to the flue gas operating conditions, calculate the upper limit of pressure drop in each range and generate the low resistance boundary by combining it with the allowable pressure drop of the demister. Store the low resistance boundary in the parameter record table with the version identifier.

[0016] The water collection tray and drainage pipeline are divided into water collection sections according to the height and cross-section of the demister. The correspondence between each water collection section and the process water tank is written into the water collection section and process water tank connection table. The process parameter service provides the low resistance boundary and the process water tank number corresponding to each water collection section based on the version identifier.

[0017] Furthermore, S2 includes:

[0018] The control system deployed in the wet desulfurization unit collects flue gas operating conditions, demister pressure difference, water collection volume in the water collection section, water tank level and water quality according to the control cycle;

[0019] Align all measurements using a unified timestamp and replace faulty measurements.

[0020] The above measurements are normalized into operating status parameters, which include pressure drop margin coefficient, water collection section status parameters, and water tank status parameters.

[0021] Write the operating status parameters, along with the device number and timestamp, into the operating status record table and apply an idempotent write constraint.

[0022] The process parameter service returns operating status parameters based on the unit number and timestamp.

[0023] Furthermore, S3 includes:

[0024] The control system deployed in the wet desulfurization unit calculates the water collection demand index for each water collection section within each control cycle based on the pressure drop margin coefficient and the state parameters of each water collection section, according to preset water collection demand weight parameters. The water collection demand index for the i-th water collection section is... satisfy

[0025]

[0026] in, This is the pressure drop margin factor. Let be the liquid level coefficient of the i-th water collection section. Let be the water collection load coefficient of the i-th water collection section. For the pressure drop related weight parameters, For liquid level related weight parameters, For traffic-related weight parameters, This is the liquid accumulation risk correction amount for the i-th water collection section;

[0027] Based on the state parameters of each process water tank, the acceptance capacity coefficient of each process water tank is calculated according to the preset water tank acceptance weight parameters. The acceptance capacity coefficient of the j-th process water tank is... satisfy

[0028]

[0029] in, Let be the remaining volume coefficient of the j-th process water tank. Let be the water quality margin coefficient of the j-th process water tank. Let be the solids margin coefficient of the j-th process water tank. Let be the blowdown margin coefficient of the j-th process water tank. For volume weight parameters, For water quality weighting parameters, For solid weight parameters, For pollution discharge weight parameters;

[0030] Write the water demand index and the receiving capacity coefficient, along with the unit number and control cycle time mark, into the water demand and receiving capacity record table. The process parameter service can then query and return the water demand index and the receiving capacity coefficient by the unit number and control cycle time mark.

[0031] Furthermore, S4 includes:

[0032] The control system determines the allocation weights of the water collection section and the process water tank based on the water collection demand index of the water collection section, the water tank receiving capacity coefficient, and the connection relationship table between the water collection section and the process water tank, and forms a water collection allocation matrix by normalizing the water collection section.

[0033] The control system calculates the target distribution flow rate for each water collection branch based on the water collection distribution matrix and the target discharge rate of the water collection section. Specifically, the target distribution flow rate for the water collection branch from the i-th water collection section to the j-th process water tank is... satisfy

[0034]

[0035] in, The target discharge amount for the i-th water collection section; For the first From the first water collection section to the first The water collection and distribution matrix elements of each process water tank, when the sum of the target distribution flow rates of the branches connected to the same process water tank exceeds the allowable water collection flow rate of that process water tank. At that time, the flow rate allocated to the receiving branch is reduced proportionally, and the target flow rate allocated to the reduced receiving branch is... satisfy

[0036]

[0037] The target flow rate for each water intake branch is then written into the branch allocation record table, whereby... The sum of the target flow rates allocated to all water intake branches connected to the j-th process water tank.

[0038] Furthermore, S5 includes:

[0039] The control system establishes an operating state machine that includes normal state and discharge state at the valve opening control level;

[0040] Within each control cycle, the current state of the state machine is determined based on the pressure drop margin coefficient, the liquid level coefficient of the water collection section, the liquid accumulation risk marker of the water collection section, the remaining volume coefficient of the water tank, and the water quality margin coefficient.

[0041] Under normal conditions, the deviation between the actual flow rate and the target flow rate of the receiving branch is calculated based on a sliding time window. Here, the flow rate deviation of the i-th receiving branch in the current control cycle is... satisfy

[0042]

[0043] in, Let i be the actual flow rate of the i-th water inlet branch. Assign flow rate to the target of the i-th water collection branch; the average flow rate deviation of the i-th water collection branch within the sliding time window. satisfy

[0044]

[0045] Where n is the number of control cycles within the sliding time window. The actual flow rate of the i-th water intake branch within the k-th control cycle. Distribute the flow rate to the target of the i-th water collection branch within the k-th control cycle; and adjust the valve opening of the water collection branch according to the preset opening step size;

[0046] Write the state machine status flags and valve opening adjustment records into the operating status record table and valve opening adjustment record.

[0047] Furthermore, in the draining state, the control system classifies the water collection sections whose level coefficient exceeds the draining start threshold into the draining set based on the water collection section level coefficient and pressure drop margin coefficient.

[0048] Apply a discharge gain coefficient to the water collection branch in the discharge collection section that meets the acceptance reliability threshold for the remaining volume coefficient and water quality margin coefficient, thereby increasing the target distribution flow of the water collection branch;

[0049] Adjustment range of valve opening limits for water collection branch sections in non-drainage collection and water collection zones.

[0050] Furthermore, S6 includes:

[0051] The control system collects and aligns data from the pressure drop record table, branch distribution record table, water tank status record table, and state machine record table according to the statistical cycle;

[0052] The distribution of pressure drop margin coefficient, changes in liquid level coefficient in water collection section, water tank level and water quality margin, and discharge status behavior are statistically analyzed according to the working condition interval.

[0053] Under the premise of meeting mechanical safety constraints and process safety constraints, the low resistance boundary, water collection demand weight, water tank acceptance weight and related thresholds are modified step by step to generate updated control parameters with version identifiers and write them into the parameter update record table.

[0054] Furthermore, the updated control parameters are reviewed and confirmed by engineers through the monitoring interface before taking effect;

[0055] The process parameter service distributes the unit number, parameter type identifier, old version identifier, new version identifier, and parameter update status code through the parameter update interface;

[0056] If data aggregation is incomplete or storage fails during the parameter update process, the original low resistance boundary, water demand weight, and water tank acceptance weight should be kept unchanged, and the reasons for the update failure should be recorded in the parameter update record table and the operation log.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] 1. By establishing a pressure drop margin boundary for the demister based on flue gas operating conditions in an actual wet desulfurization unit, and dividing the water collection section according to the demister structure, clarifying the water collection path between each water collection section and the process water tank, the water collection demand index and receiving capacity coefficient are calculated according to the pressure drop margin coefficient and the state parameters of the water collection section and the water tank during operation. A water collection distribution matrix is ​​generated and the drainage of each branch is controlled. This achieves the effect of prioritizing the demister to be in the low resistance safe range under different loads and operating conditions, avoiding liquid accumulation and secondary liquid carrying, while taking into account the safety of downstream water tank liquid level and water quality, and improving the overall operational stability and safety of the wet desulfurization system.

[0059] 2. By introducing a state machine and parameter version management mechanism into the control system, the system utilizes the state switching rules between normal and discharge states, the branch flow deviation correction within the sliding time window, and the statistical periodic summary and evaluation of pressure drop, water intake, water quality, and status records to automatically correct low resistance boundaries, water intake demand weights, and acceptance capacity thresholds. This reduces manual intervention and suppresses frequent and large valve movements during long-term operation, gradually optimizing the water intake strategies of each section and each water tank. The system then possesses adaptive and multi-condition self-correction capabilities, and reduces the risk of abnormal discharge and overflow. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the process water circulation method of a low-resistance demister and process water tank according to the present invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] Example: Figure 1 A flowchart illustrating a method for linking water collection in a low-resistance demister with process water tank circulation according to the present invention is provided. This method, executed by a control system deployed in a wet desulfurization unit, includes:

[0063] S1. Based on the flue gas operating conditions, test the pressure drop of the demister, establish a low-resistance boundary where the pressure drop does not exceed the preset envelope, and divide the water collection section according to the structure and water collection path, determining the process water tank connected to the water collection section. The specific implementation is as follows:

[0064] In actual wet desulfurization units, taking the flue gas demister at the top of the absorption tower as the object, the flue gas conditions and demister pressure drop are first tested during the stable operation phase of the unit. In this scenario, the flue gas conditions refer to a set of field quantities directly related to the airflow state of the demister, including the flue gas flow rate at the absorption tower inlet, the flue gas temperature at the demister inlet, the flue gas humidity and saturation index at the demister inlet, and the operating load level inside the tower. These quantities are obtained by flow measurement devices, temperature measurement devices, and humidity measurement devices installed in the flue gas duct. The units are preferably set as standard volumetric flow rate per hour, temperature in Celsius, and relative humidity percentage. The acquisition rhythm is preferably set to once every several seconds, for example, once every two to ten seconds, and the time is uniformly marked with the time of the unit control system. The operating load level inside the tower can be divided into several levels according to the unit load percentage or the induced draft fan speed level, and marked in the operation record in the form of enumerated values. The current operating condition is recorded for subsequent analysis of pressure drop distribution under different loads. In this scenario, the demister pressure drop refers to the static pressure difference between the demister inlet and outlet, obtained through static pressure taps and differential pressure measurement transmitters located on the demister inlet and outlet flues. The unit is preferably Pascal. The acquisition rhythm is consistent with the flue gas operating condition. To ensure that the acquired data can be used to establish a stable low-resistance boundary, the above flue gas operating conditions and demister pressure drop are continuously recorded within a preset observation window. Preferably, the observation window is set to 30 minutes to 2 hours. The acquisition interval and window length are coordinated to ensure that each observation window contains at least several hundred sampling points and covers multiple typical operating conditions from low load to high load within a statistical period. The statistical period is preferably set to at least three days, so that the number of pressure drop samples in each operating condition interval reaches several dozen or more to ensure the stability of the statistical results.

[0065] Within this observation window, the flue gas conditions and corresponding pressure drops for each sampling cycle are aligned using time stamps. The time deviation is preferably limited to within half of a single sampling cycle. If the time deviation of individual sampling points is within the allowable tolerance range, linear interpolation is used to supplement the time and value of the sampling point based on the time and value of adjacent valid sampling points. Alignment is achieved in the control system through a unified timestamp field. If the differential pressure signal experiences a jump or drop during a certain period and is inconsistent with the direction of flue gas flow rate change during the same period, the period is determined to be the on-site time based on a preset baseline change rate and consecutive count rule. Interference pulses and the reference rate of change can be obtained by calculating the average amplitude and fluctuation range of the absolute value of pressure drop change within a recent several-minute window. The amplitude threshold is preferably set to two to three times the reference rate of change, and the consecutive number rule is preferably set to three to five consecutive sampling cycles in which the pressure drop change exceeds the amplitude threshold. Thus, isolated points or short strings of points that exceed the threshold and meet the consecutive number condition are identified as interference points, and the values ​​of these points are replaced with the arithmetic mean of the adjacent effective values ​​to achieve noise reduction and boundary filling, thereby obtaining a set of smooth pressure drop curves that match the actual flue gas flow and temperature changes.

[0066] Based on this, several operating condition intervals are divided according to two dimensions: flue gas flow rate and temperature. Preferably, the flue gas flow rate of the entire operating range is divided into several flow intervals based on 10% to 20% of the design operating condition flow rate, and the demister inlet temperature is divided into several temperature intervals based on 5°C to 10°C. For each operating condition interval formed by the combination of flow and temperature intervals, the long-term upper limit value and the normal fluctuation range of pressure drop are statistically analyzed within a statistical period. The long-term upper limit value can be taken as the high quantile value of the pressure drop samples within the operating condition interval, and the normal fluctuation range can be taken as the average value of the pressure drop samples within the operating condition interval plus or minus a certain number of standard deviations. Preferably, the high quantile value is set to the 95th percentile, that is, after sorting all pressure drop samples within the operating condition interval from smallest to largest, the value corresponding to the sample with a cumulative proportion reaching 95% is taken as the long-term upper limit reference value. The standard deviation can be obtained by... The deviation between the pressure drop sample and its average value within the operating condition range is calculated, and the multiplier can be set to an integer between two and three to cover normal fluctuations. Then, the statistical upper limit of the operating condition range is compared with the maximum allowable pressure drop of the demister structure and the mechanical safety margin of the tower top flue and upstream equipment. Preferably, the allowable upper limit of pressure drop is set to the lower of the statistical upper limit and the mechanical limit. Then, an envelope that does not exceed the allowable upper limit of pressure drop is formed for each operating condition range. In this implementation scenario, the envelope is defined as the low resistance boundary where the pressure drop does not exceed the preset envelope. The low resistance boundary is stored in the parameter record table of the control system in the form of a set of allowable pressure drop upper limits corresponding to a set of operating condition ranges. Each set of parameter records includes the generation time, equipment number, operating condition range definition rules, collection batch number and sample quantity, forming a version-locked boundary version identifier, which is saved for subsequent retrieval.

[0067] To ensure idempotency and sequentiality, within the same equipment and operating condition range, only the latest version of the low-resistance boundary is retained as the currently effective version, while historical versions are used as read-only records to form a chain of evidence. For duplicate records appearing under the same equipment number and timestamp combination key, deduplication is achieved by retaining the record with the latest timestamp and discarding other records. When a new trial batch generates a new boundary version, the boundary version switching time point and the number of samples participating in the statistics are recorded in the control system to ensure that those skilled in the art can trace back the boundary parameters used within a certain period of time and judge the reliability of the statistical results based on the number of samples.

[0068] When dividing the demister body into water collection sections, the internal configuration of the demister is first confirmed on-site. In this scenario, the demister can be a blade-type demister or a baffle-type demister. Each set of blades and baffles is installed above the support beam, and several water collection trays and water collection troughs are provided below. The water is then drained into one or more process water tanks through drainage pipes. In this scenario, the water collection section is defined as a group of demister areas that are relatively independent in terms of structure and water collection path. For example, they may correspond to different height levels, different radial areas, and different drainage branches. To ensure that the division results can be referenced in the control system, it is preferable to divide the demister into several levels from top to bottom according to the demister drawings and on-site layout. Each level is further divided into several zones according to the cross-section of the flue. The water collection trays and drainage branches below each zone correspond to a water collection section, and each water collection section is assigned a unique water collection section number.

[0069] When determining the process water tanks connected to the water collection section, on-site engineers checked each drain pipe section by section from the water collection tray downwards. By identifying the confluence points of the drain pipes and the positions of the connecting valves, a one-to-one or many-to-one correspondence was established between each drain pipe and the specific process water tank. In this scenario, the process water tank is usually one of the absorption tower circulating slurry tank, clear water tank, and intermediate water tank. Its physical meaning is a liquid storage unit that can receive the liquid collected by the demister.

[0070] To ensure that this structural relationship can be invoked by subsequent control logic, a connection relationship table between the water collection section and the process water tank is established in the control system. Each record includes the water collection section number, the corresponding water collection tray number, the drain pipe number, the connecting valve identifier, and the target process water tank number. It also records the establishment time, drawing version, and on-site confirmation record number, forming a version identifier for the structural topology. When on-site modifications result in the reconnection of the drain pipe or the addition of a process water tank, the engineering personnel update the connection relationship table and update the version identifier simultaneously. The old version is still retained as a read-only record to form a chain of evidence for the structural topology.

[0071] The low-resistance boundary parameter record table and the connection relationship table between the water collection section and the process water tank are linked through the process parameter service in the control system. This service uses the device number and the current boundary version identifier as query conditions. In each control cycle, it provides the downstream linkage control link with the low-resistance boundary value that the pressure drop under the current flue gas conditions does not exceed the preset envelope, as well as the process water tank number associated with each water collection section. When the process parameter service fails to return a valid result within the agreed time, the calling side resends the request several times according to the agreed waiting time limit, with each resend at a preset time interval. The number of resends is preferably set to three to five times. If it is still unsuccessful after the number of resends, the original boundary and topology parameters remain unchanged in the current control cycle, and a communication fault record is added to the operation record. The record content includes the fault identifier, occurrence time, number of resends, and the conservative strategy adopted.

[0072] To ensure that communication failure does not lead to unsafe conditions, the upper limit of pressure drop control in this situation is jointly limited by the design pressure drop limit and the water tank overflow safety margin. Downstream control links must not push the demister operating pressure drop to a level higher than the design limit. For ease of monitoring and review by higher-level management, a minimum set of fields is agreed upon between the process parameter service and control logic, including the unit number, operating time period, flue gas flow range, flue gas temperature range, low-resistance boundary version identifier, and version identifier of the connection relationship between the water collection section and the process water tank. A status identifier field and a status code field are also agreed upon. The status code can be set to a numeric identifier, for example, zero indicates successful query, one indicates timeout, and two indicates no currently effective version. When communication fails, the returned status identifier is "incomplete," accompanied by a status code and a brief description of the reason, such as "service unavailable" or "query timed out," so that maintenance personnel can verify it during inspections.

[0073] To verify whether the above steps meet expectations in the field, after the low-resistance boundary is established, it is preferable to select at least several representative operating days, such as three to seven consecutive days, and record the fluctuation range of demister pressure drop and flue gas liquid carrying under different loads during these days. In the monitoring of liquid carrying, it is possible to observe whether there are obvious traces of secondary liquid carrying and liquid accumulation through subsequent sampling points in the flue and maintenance ports of downstream equipment. If the measured pressure drop value falls within the low-resistance boundary for most of the time and no obvious liquid accumulation or liquid carrying occurs, then the boundary is considered to be applicable under the current equipment and flue gas operating conditions. If certain operating conditions frequently touch the upper limit of the boundary, the boundary of the operating condition will be appropriately lowered in the next round of statistics to form a new boundary version.

[0074] In a preferred embodiment, under typical operating conditions, the flue gas flow rate of the desulfurization device of a coal-fired unit can be set to several hundred thousand to several million standard volumes per hour, and the demister inlet temperature can be set to fifty to eighty degrees Celsius. After a week of continuous monitoring, it was found that the normal value of the demister pressure drop under medium and high load conditions is about several hectopascals, and the short-term maximum value is about several hectopascals. The upper limit of the low resistance boundary is set to a level that is slightly lower than the short-term maximum value, corresponding to 80% to 90% of the device's design pressure drop. When operating under this boundary, no obvious secondary liquid-carrying stains were observed in the downstream flue gas sampling, and the liquid level in the water collection tray did not frequently overflow, thus confirming the usability of the boundary.

[0075] In another alternative approach, if sufficient field pressure drop records cannot be obtained in the early stages of operation, the pressure drop performance curve and design flue gas conditions provided by the demister manufacturer can be used to estimate the initial low resistance boundary. The pressure drop value under the design conditions can be multiplied by a fixed safety factor, preferably set to between one and two. During subsequent operation, the boundary value can be gradually corrected by comparing it with the measured pressure drop. The correction method follows the aforementioned observation window statistical method and version locking rule, ultimately obtaining a low resistance boundary parameter set with a pressure drop not exceeding the preset envelope, which is equivalent to the aforementioned implementation method.

[0076] Through the above series of steps, those skilled in the art can complete the testing of flue gas operating conditions and demister pressure drop, the establishment of low resistance boundaries, and the division of the structure and water collection path of the water collection section and process water tank in actual production lines, providing a reproducible basis for subsequent linkage control based on low resistance boundaries and water collection section topology.

[0077] S2. Periodically collect flue gas operating conditions, demister pressure difference, water collection volume in the water collection section, water tank level, and water quality. Normalize the data to obtain the pressure drop margin coefficient, as well as the state parameters of the water collection section and the water tank. The specific implementation is as follows:

[0078] Given that the connection between the low-resistance boundary and the water collection section and the process water tank has been established and stored in the control system in the form of a parameter record table, during the long-term operation of the wet desulfurization unit, in order to obtain the pressure drop margin coefficient and the state parameters of the water collection section and the water tank in each control cycle, the control system collects the flue gas conditions, demister pressure difference, water collection volume of the water collection section, water tank level and water quality from the field measuring devices according to a unified clock rhythm in each cycle, and converts these quantities into dimensionless parameters under the same time reference.

[0079] The flue gas operating conditions at this stage still refer to the flue gas flow rate at the inlet of the absorption tower, the flue gas temperature at the inlet of the demister, the humidity and saturation of the flue gas at the inlet of the demister, and the operating load level within the tower. These are obtained through flow measurement devices, temperature measurement devices, and humidity measurement devices arranged near the inlet flue and the top of the tower. The units are preferably standard volumetric flow rate per hour, temperature in Celsius, and relative humidity percentage. The operating load level within the tower is further divided into several levels based on the unit load percentage or the induced draft fan speed level. The current operating condition level is marked in the operation record as an enumerated value for subsequent determination of the operating condition range of the current sampling point. The demister differential pressure is obtained in each control cycle through the aforementioned static pressure tap and differential pressure measurement transmitter. The unit is preferably Pascal.

[0080] In this scenario, the water collection volume of each water collection section is defined as the volumetric flow rate of liquid discharged in the current cycle or the corresponding liquid level height of the collection tray. On-site, flow measurement devices or liquid level measurement devices can be selected to be configured on each drainage pipe or collection tray according to equipment conditions. The preferred unit for flow rate is volumetric flow rate per hour, and the preferred unit for liquid level is meter or relative percentage. When a flow measurement device is configured for a water collection section, the flow rate value is directly read and the time stamp is recorded in each cycle. When a liquid level measurement device is configured, the liquid level value is recorded in each cycle, and the liquid level is converted into a water collection capacity coefficient in the subsequent normalization by using the pre-established correspondence between liquid level and discharge capacity.

[0081] Liquid level measurement devices are installed on the top or side wall of each process water tank, with the unit preferably being meters or a percentage of relative height. Water quality in this scenario primarily considers dissolved salts, conductivity, and suspended solids content within the tank. Conductivity measurement devices and sampling ports are installed at appropriate locations on the circulation pipes or within each process water tank, with the unit preferably being millisiemens per centimeter. Suspended solids content can be calculated using an online turbidity measurement device, with the unit preferably being turbidity units. In some applications where temperature is a significant factor, a temperature measurement device can also be installed within the tank. The acquisition rhythm of the above measurements is preferably consistent with the control cycle of the control system, for example, it can be set to once per second to once every ten seconds, continuously acquiring data within a preset observation window. Each sampling point is accompanied by a uniform timestamp and recorded in the operating buffer.

[0082] To ensure that these quantities can be compared at the same time during normalization, the control system aligns the flue gas conditions, demister pressure difference, water collection volume in the water collection section, water tank level, and water quality records within the same cycle according to timestamps. The time deviation is preferably no more than half of the control cycle. When the timestamp of a certain quantity is slightly earlier or later than the target cycle time point, linear interpolation is performed between two adjacent sampling points of that quantity according to time weight to obtain an estimated value corresponding to the target cycle time point, and the estimated value is used in subsequent calculations. If a certain measurement quantity is missing or significantly deviates from the normal range of change of that quantity within a certain sampling cycle, such as a sudden change in the liquid level value exceeding several times the maximum change in liquid level in the most recent minutes and occurring in multiple consecutive cycles, then the measurement quantity is marked as a fault quantity, and the valid quantity from the previous moment is called to participate in the normalization calculation of this cycle. When the quantity is marked as a fault quantity for several consecutive cycles, a measurement channel fault record is recorded in the control system, and the state parameter corresponding to the measurement channel is marked as invalid. A conservative strategy is used in the downstream logic to restrict the channel.

[0083] In this scenario, normalization refers to converting the above measured values ​​into dimensionless coefficients based on the aforementioned low-resistance boundary, design limit, and rated range of each device within each control cycle. These coefficients are all limited to the range of zero to one. Zero indicates that the corresponding constraint has been fully utilized or exceeded, and one indicates that the corresponding capacity is completely surplus. The values ​​in between reflect the current remaining margin.

[0084] In this scenario, the pressure drop margin coefficient is defined as a dimensionless quantity reflecting the distance between the demister pressure drop and the low resistance boundary under the current operating condition. In each cycle, the control system looks up the aforementioned low resistance boundary parameter record table based on the currently sampled flue gas flow rate, temperature, and load level to determine the operating condition range to which the cycle belongs and the corresponding upper limit of pressure drop. The current measured pressure drop is compared with the upper limit of pressure drop. When the measured pressure drop is lower than the upper limit of allowable pressure drop, the ratio of the difference between the measured pressure drop and the upper limit of allowable pressure drop to the upper limit of allowable pressure drop is converted into a pressure drop margin coefficient and limited to the range of zero to one. When the measured pressure drop is higher than or equal to the upper limit of allowable pressure drop, the pressure drop margin coefficient is directly set to zero.

[0085] In this scenario, the state parameters of the water collection section are defined as a set of dimensionless parameters used to describe the liquid level, water collection volume, and potential liquid accumulation risk of each water collection section in the current cycle. For water collection sections equipped with liquid level measuring devices, the current liquid level value is compared with the designed normal operating liquid level range of the section. Preferably, the designed maximum safe liquid level is used as the upper limit, and the liquid level coefficient is obtained by dividing the current liquid level by the upper limit, which is limited to the range of zero to one. For water collection sections equipped with flow measuring devices, the current water collection flow rate is compared with the designed maximum continuous discharge flow rate of the branch, and the flow rate coefficient is obtained by dividing the current flow rate by the designed maximum flow rate, which is also limited to the range of zero to one.

[0086] Based on a comprehensive consideration of the liquid level coefficient and the flow rate coefficient, the control system can set a set of weights to combine the liquid level coefficient and the flow rate coefficient to obtain the water collection load coefficient in the state parameters of the water collection section. At the same time, a liquid accumulation risk flag is set according to the degree to which the liquid level coefficient is close to one. When the liquid level coefficient exceeds a preset threshold, such as 0.8, the liquid accumulation risk of the water collection section is marked as high, and a record line is added to the operation record indicating that the water collection section has entered a high-risk state.

[0087] In this scenario, the water tank status parameters are defined as a set of dimensionless parameters describing the remaining volume, water quality margin, and sewage discharge capacity margin of each process water tank. In each cycle, the control system compares the current water tank level with the effective height of the tank, subtracts the current level from the effective height to obtain the remaining height, and divides the remaining height by the effective height to obtain the remaining volume coefficient, which is limited to a range of zero to one. Simultaneously, it compares the current conductivity with the upper limit of the allowable conductivity for that water tank, subtracts the current conductivity from the upper limit and then divides by the upper limit to obtain the water quality margin coefficient, which is also limited to a range of zero to one. When the current conductivity is close to the upper limit of the allowable conductivity, the water quality margin coefficient approaches zero. In some cases where solid content needs to be controlled, turbidity can be converted to solid content and compared with the upper limit of the allowable solid content to obtain the solid content margin coefficient. For water tanks with sewage discharge capacity, the current sewage discharge capacity occupancy is estimated based on the opening of the sewage valve and the maximum discharge capacity of the sewage discharge channel, and the sewage discharge margin coefficient is obtained.

[0088] These coefficients constitute the set of tank state parameters for each process water tank. The control system preferably combines each coefficient with a set of weights, for example, giving higher weights to the remaining volume coefficient and water quality margin coefficient, and using the solids margin coefficient and sewage discharge margin coefficient as auxiliary quantities to form a comprehensive tank state coefficient. This coefficient is also limited to the range of zero to one and is used for subsequent allocation logic to determine whether a water tank has sufficient capacity under the current operating conditions.

[0089] At the end of each control cycle, the control system writes the pressure drop margin coefficient, the status parameters of each water collection section and the status parameters of each water tank, together with the timestamp, device number, low resistance boundary version identifier and structural topology version identifier, into an operation status record table. The record table is updated in chronological order and archived in batches by day or shift. Each batch is accompanied by a version number and verification mark to form a chain of evidence.

[0090] To avoid write conflicts and duplicate records under high load, the control system adopts an idempotent strategy for records with the same timestamp and the same water intake section number or the same water tank number. Only the record line that was successfully written the most recently is retained, and the order is constrained by the record number to prevent downstream logic from reading contradictory parameter sets in the same cycle.

[0091] To facilitate downstream logic in calling these normalized parameters, the process parameter service in the control system extends the aforementioned low-resistance boundary and structural topology provision with a set of operational status query functions. The operational status query uses the unit number and current time as conditions, returning the pressure drop margin coefficient, status parameters of each water intake section, and status parameters of each water tank for the most recently completed normalization cycle, along with the batch number and record number of the operational status record table. The delay of the operational status query is preferably limited to less than half of a control cycle. If the normalized result for the current cycle cannot be retrieved from the operational status record table within the specified time, the process parameter service can retry several times, with each retry interval less than one control cycle. The number of retryes is preferably set to two to three. If a new normalized result is still not obtained after the retry, the normalized result of the previous cycle is returned, and the status field is marked as using the parameters from the previous cycle. Simultaneously, a record indicating that the operational status query has not been updated is added to the control system's operational log.

[0092] In the process parameter service's operational status query, it is preferred to define a minimum set of fields including the device number, timestamp, pressure drop margin coefficient, each water collection section number and its corresponding status parameter set, each process water tank number and its corresponding status parameter set, as well as the low resistance boundary version identifier and structural topology version identifier. At the same time, a status code field is defined, and the status code can be set to several discrete values. For example, zero indicates that the normalization result of this cycle is valid and latest, one indicates that the normalization result of the previous cycle is used, and two indicates that the current normalization result is missing and the downstream needs to take a conservative strategy. When the downstream logic receives a return with a status code of two, it should avoid adjusting the water collection allocation strategy and maintain the existing open and closed state, thereby ensuring that the system still operates within the safety boundary when there is a measurement failure or normalization failure.

[0093] To verify the rationality of the on-site performance of the periodic normalization and parameter calculation, several statistical periods can be selected after the device is put into operation. The distribution of pressure drop margin coefficient, water collection section state parameters, and water tank state parameters within these periods can be statistically analyzed. Preferably, it is required that the pressure drop margin coefficient is in the range of 0.2 to 0.8 for a relatively high proportion of the time under normal operating conditions, the liquid level coefficient in the water collection section is below 0.5 for a relatively high proportion of the time, and the remaining volume coefficient of the water tank fluctuates within a certain safe range. These indicators are used to verify whether the normalization rules reflect the actual operating conditions.

[0094] In a preferred embodiment, under medium-to-high load conditions, the upper limit of the pressure drop in the current operating range of a certain unit is a certain number of hectopascals, and the actual pressure drop is between a certain number of tens of Pascals and a certain number of hectopascals. According to the aforementioned rules, the pressure drop margin coefficient fluctuates between 0.3 and 0.7. The liquid level coefficient of the water collection section is mostly below 0.6. Only during short-term high-load switching periods is the liquid level coefficient of a few water collection sections close to 0.9 and marked as high-risk. After process adjustment, it quickly drops back down. The remaining volume coefficient of the water tank is maintained between 0.4 and 0.7, and the water quality margin coefficient corresponding to the conductivity is maintained above 0.3. At this time, the downstream distribution logic can be finely adjusted based on these state parameters.

[0095] In an alternative approach, if it is temporarily impossible to equip all water collection sections and tanks with complete online flow and water quality measurement devices, it is preferable to select key water collection sections and key process tanks to equip them with the aforementioned online measurement devices. For the remaining water collection sections and tanks, the liquid level and water quality are recorded manually during inspections. The manual readings are entered into the control system at the inspection time points, and the control system converts them into state parameters according to the same normalization rules as the online measurements. For the remaining periods, the most recent valid record is used for the unupdated quantities. In the downstream logic, a more conservative allocation strategy is applied to the state parameters from manual records. As the online measurement points are gradually added through upgrades, the system can be fully switched to the aforementioned implementation method.

[0096] Through the above steps, those skilled in the art can establish a periodic normalization mechanism in actual production lines, so that the pressure drop margin coefficient, water collection section state parameters, and water tank state parameters form a structured state description in each control cycle, providing a sufficient and reproducible basis for the subsequent calculation of the section water collection demand index and the water tank acceptance capacity coefficient.

[0097] S3. Calculate the water collection demand index for each water collection section based on the pressure drop margin coefficient and the state parameters of the water collection section, and calculate the receiving capacity coefficient of each water tank based on the water tank state parameters. The specific implementation is as follows:

[0098] Based on the establishment of the aforementioned low resistance boundary and the normalization of the pressure drop margin coefficient, water collection section status parameters, and water tank status parameters in each control cycle, the control system calculates the water collection demand index of each water collection section and the acceptance capacity coefficient of each process water tank in each control cycle according to these normalized parameters during the actual operation of the wet desulfurization unit.

[0099] In this scenario, the water collection demand index is defined as a dimensionless quantity reflecting the urgency and priority of a certain water collection section for discharging liquid in the current cycle. Its value range is limited to zero to one, and a value close to one indicates that the water collection section currently needs to obtain water collection capacity more urgently. The water tank acceptance capacity coefficient in this scenario is defined as the capacity margin of a certain process water tank in terms of volume, water quality and sewage discharge capacity to accept water collected by the demister in the current cycle. It is also limited to the range of zero to one, and a value close to one indicates that the water tank is currently more suitable as a water collection allocation target.

[0100] In practice, at the beginning of each control cycle, the control system queries the most recently normalized operating status record through the process parameter service according to the unit number and the current time. It obtains the pressure drop margin coefficient, liquid level coefficient, flow coefficient, water collection load coefficient and liquid accumulation risk mark of each water collection section, as well as the remaining volume coefficient, water quality margin coefficient, solids margin coefficient, sewage discharge margin coefficient and comprehensive water tank status coefficient of each process water tank. The units of the above quantities have been converted into dimensionless values ​​in the aforementioned normalization step. This step only needs to match the most recent record according to the timestamp.

[0101] To ensure that the calculation results are consistent with the real-time status on site, the process parameter service needs to meet the constraint that the delay does not exceed half of the current control cycle when returning these parameters. Preferably, the operation status query and the calculation of this step are arranged to be completed in the first half of the control cycle. When the operation status record has not been updated in a certain cycle, the process parameter service returns the parameters of the previous cycle according to the aforementioned strategy and attaches a status code to mark that the values ​​of the previous cycle are used. In this case, this step is still calculated according to the received parameters, but the adjustment range is appropriately tightened in the subsequent allocation strategy.

[0102] The calculation of the water demand index is based on the pressure drop margin coefficient of the current cycle and the state parameters of each water collection section. The control system pre-sets a set of water demand weight parameters, including pressure drop related weight, liquid level related weight and flow rate related weight. The weight values ​​are between zero and one, and the sum of the three is preferably close to one. Each set of weights is determined through a period of operation observation during the commissioning phase, and is written into the parameter configuration table with the water demand weight version identifier. The effective time, applicable device number and operation batch number participating in the commissioning are recorded to form version locking and evidence chain traceability.

[0103] Within each control cycle, the control system reads the pressure drop margin coefficient, liquid level coefficient, flow rate coefficient, and water collection load coefficient for each water collection section. The pressure drop margin coefficient, weighted according to pressure drop, is interpreted as an adjustment factor for the overall water collection activism. The liquid level coefficient, weighted according to liquid level, is considered as a factor increasing demand due to the risk of liquid accumulation in that section. The water collection load coefficient or flow rate coefficient, weighted according to flow rate, is considered as the current drainage capacity utilization level of the section. Within the same control cycle, these three factors are superimposed according to their weights and compressed to a range of zero to one through a monotonically increasing mapping. The preferred mapping method is a piecewise linear relationship, which makes the increase in the water collection demand index more sensitive when the liquid level coefficient is high. Specifically, the first... The water demand index for each water collection section is denoted as: The pressure drop margin coefficient is denoted as , No. The liquid level coefficient of each water collection section is denoted as: , No. The water collection load coefficient of each water collection section is denoted as: Then the first The water collection demand index of each water collection section meets the following requirements:

[0104]

[0105] in, For the pressure drop related weight parameters, For liquid level related weight parameters, For traffic-related weight parameters, ; For the first The risk correction amount for each water collection section, when the first When the risk of liquid accumulation in a water collection section is marked as high risk Take the preset increment, when the first... When the risk marker for liquid accumulation in a water collection section is not high risk Zeroing is used to make the pressure drop margin, liquid level status, and water collection load work together to affect the water collection demand index, and limit the final result to the range of zero to one. When the liquid accumulation risk of the water collection section is marked as high risk, the control system adds a fixed penalty increment to the calculation result, but still limits the final result to no more than one, so that the water collection section will get higher priority in subsequent allocation.

[0106] When the pressure drop margin coefficient is zero, it indicates that the current pressure drop of the demister has reached the low resistance boundary. The control system preferentially raises the water collection demand index of all water collection sections to a minimum guaranteed value of no less than a medium level, such as 0.5, so that the forced water collection distribution logic no longer compresses the drainage of any section, thus protecting the demister as a whole from further pressure drop. Specifically, when At that time, the first The water collection demand index for each water collection section is revised as follows:

[0107]

[0108] in, This is a preset minimum value. The above correction ensures that when the demister pressure drop reaches the low resistance boundary, each water collection section maintains a water collection demand index no lower than the minimum value, thus preventing the water collection distribution logic from further compressing the drainage capacity.

[0109] During the above calculation process, if the liquid level measurement or flow rate measurement of a certain water collection section is marked as a fault in the current cycle by the aforementioned normalization stage, the control system can prioritize using the state parameters of the most recent effective cycle to participate in the calculation and add a fault mark. In the subsequent allocation process, the frequency of further reduction and increase of the water collection section is limited. If the state parameters are unavailable for several consecutive cycles, the water collection demand index of the water collection section is maintained at the intermediate value to avoid over-adjustment. At the same time, the time range and handling strategy of the unclear state of the section are recorded in the operation log.

[0110] The calculation of the water tank acceptance capacity coefficient is based on the state parameters of each water tank. The control system pre-sets a set of water tank acceptance weight parameters, including volume weight, water quality weight, solid weight and sewage discharge weight. The weight values ​​are limited to the range of zero to one and the sum is preferably close to one. In desulfurization devices, where both water quality and volume are sensitive, the volume weight and water quality weight are usually set higher. The solid weight and sewage discharge weight are set as auxiliary factors. This set of weights is also written into the parameter configuration table with the water tank acceptance weight version identifier and form a version lock.

[0111] Within each control cycle, the control system, for each process water tank, multiplies the current remaining volume coefficient by its weight to represent the tank's capacity to continue receiving water before overflowing; multiplies the water quality margin coefficient by its weight to represent the tank's ability to continue receiving water with high salinity or high solids content before exceeding water quality standards; multiplies the solids margin coefficient by its weight to represent the tank's leeway in receiving water without significantly increasing solids load; and multiplies the wastewater discharge margin coefficient by its weight to represent the tank's capacity to unload excess pollutants through subsequent wastewater discharge. Within the current cycle, these four weighted results are superimposed and constrained to zero to obtain the tank's receiving capacity coefficient. Specifically, the first... The receiving capacity coefficient of each process water tank is denoted as: , No. The residual volume coefficient of each process water tank is denoted as: The water quality margin coefficient is denoted as The solid margin factor is denoted as The discharge margin coefficient is denoted as Then the first The capacity coefficient of each process water tank meets the following requirements:

[0112]

[0113] in, For volume weight parameters, For water quality weighting parameters, For solid weight parameters, For pollution discharge weight parameters, The above method allows the remaining volume of the process water tank, water quality, solid load, and sewage discharge capacity to collectively influence the receiving capacity coefficient, limiting the result to a range of zero to one. When the tank level approaches the maximum safe level or the water quality margin coefficient approaches zero, the receiving capacity coefficient automatically decreases. When both are very close to their limits, the receiving capacity coefficient is preferably reduced directly to a value close to zero, preventing subsequent distribution from adding more water to the tank. Specifically, when the... When the liquid level in each process water tank reaches the maximum safe level, When the first Water quality margin coefficient of each process water tank season When the liquid level reaches the maximum safe level and the water quality margin coefficient is zero, maintain... The above restrictions ensure that the process water tank, once it reaches its limit, will no longer accept additional water intake.

[0114] To ensure idempotency and sequential consistency in this step, the calculation of the water demand index and the receiving capacity coefficient is performed only once for the same set of operating status parameters within the same control cycle. The calculation results, along with the timestamp, device number, and weight version identifier, are written into a water demand and receiving capacity record table. If duplicate triggering occurs due to communication jitter, this step ensures that only the latest record is retained by checking whether there is already a record in the same cycle before writing, and marking the calculation cycle in the record. This allows downstream users to ensure that the final version can be read through the timestamp and cycle number.

[0115] The process parameter service expands upon the existing operational status query capabilities by adding a water collection demand query. The water collection demand query uses the unit number and the current control cycle's time stamp as conditions, returning the water collection demand index for each water collection section, the acceptance capacity coefficient for each process water tank, the corresponding weight version identifier, and the operational status batch number. It also returns a status code indicating whether these indices and coefficients are calculated based on the latest normalization results. The preferred status code is zero, indicating that both normalization parameter and index calculations are completed within the current cycle; one, indicating that the index calculation is based on the normalization parameters of the previous cycle; and two, indicating that the index calculation cannot be completed. When the downstream allocation logic receives a status code of two, it maintains the allocation matrix of the previous cycle without adjustment according to a safety strategy.

[0116] To ensure time and resource constraints, the calculation of the index and coefficient is preferably scheduled to be performed within a fixed time window within the control cycle. The width of this time window shall not exceed half of the control cycle. When the calculation time exceeds the preset upper limit, the control system shall immediately stop the new calculation of the current cycle, use the index and coefficient of the previous cycle, and write this stop event into the operation log for subsequent analysis.

[0117] During on-site operation, the rationality of the rules can be evaluated by statistically analyzing the distribution of water demand index in each water collection section and the distribution of the acceptance capacity coefficient of each water tank over a period of time. Preferably, it is required that when the pressure drop margin coefficient is large during normal operation, the water demand index in most water collection sections is in the low to medium range. When the pressure drop margin coefficient drops to a low level, the water demand index in the high liquid level section increases significantly, while the acceptance capacity coefficient of the high water level or high conductivity water tank decreases significantly. By comparing the ranking of water demand index and the measured change in liquid level of the collection tray within the same time period, it is possible to check whether the liquid level in the high index section has been effectively reduced, thereby verifying the consistency between the index and the on-site behavior.

[0118] In a preferred embodiment, under medium-to-high load conditions, the current cycle pressure drop margin coefficient of a certain unit is approximately 0.5, the liquid level coefficient of a certain upper water collection section is 0.7, and the water collection load coefficient is 0.6. The weighted version uses a combination of pressure drop-related weight 0.2, liquid level-related weight 0.5, and flow rate-related weight 0.3. After calculation in this step, the water collection demand index of this water collection section is approximately 0.65, which obtains a slightly higher-than-average water collection priority in subsequent allocation. Meanwhile, the remaining volume coefficient of another water tank that is close to overflow is 0. 1. When the water quality margin coefficient is 0.4, the solids margin coefficient is 0.7, and the sewage discharge margin coefficient is 0.5, and the water tank's acceptance weight adopts a volume weight of 0.4, a water quality weight of 0.3, a solids weight of 0.1, and a sewage discharge weight of 0.2, the water tank's acceptance capacity coefficient is approximately 0.35. In subsequent allocations, it is allocated less water for water collection. Statistical results during several hours of operation show that the water level in the water collection section with a high water demand index drops significantly, while the water level in the water tank with a low acceptance capacity coefficient is close to the control limit but does not overflow.

[0119] In another alternative approach, to simplify on-site implementation, a hierarchical rule method can be adopted instead of continuous weighted mapping. For example, the pressure drop margin coefficient, liquid level coefficient, and water collection load coefficient can be divided into multiple intervals. Based on the interval combination, several discrete water collection demand levels can be selected from a lookup table. The water tank remaining volume coefficient and water quality margin coefficient can be classified into levels in a similar way, and the receiving capacity level can be selected from a lookup table. The lookup results are then converted into water collection demand index and receiving capacity coefficient in the range of zero to one according to the corresponding values ​​given in the preset table. This hierarchical rule also records the version identifier in the parameter configuration table and forms an evidence chain. In actual operation, the hierarchical boundary and level corresponding values ​​can be adjusted as experience accumulates. This alternative implementation method is essentially equivalent to the aforementioned weighted mapping method, both forming a dimensionless quantification capability for the water collection section and process water tank status. Those skilled in the art can choose between the above two methods according to the control system software platform and computing power conditions.

[0120] S4. Construct a water collection allocation matrix based on the water collection demand index of the water collection section and the water tank receiving capacity coefficient, and convert it into the target allocation flow of the water collection branch. The specific implementation is as follows:

[0121] With the aforementioned steps having already established the water collection demand index and process water tank acceptance capacity coefficient for each control cycle in the control system, and with these dimensionless parameters and the connection table between the water collection section and the process water tank available for query through the process parameter service, the control system constructs a water collection distribution matrix based on these parameters in each control cycle during the actual operation of the wet desulfurization unit, and calculates the target distribution flow rate of each water collection branch based on this matrix.

[0122] In this scenario, the water collection allocation matrix is ​​defined as a two-dimensional parameter set with water collection sections as rows and process water tanks as columns. Each matrix element corresponds to the relative weight allocated from a water collection section to a certain process water tank. This weight is limited to the range of zero to one. For each row containing a water collection section, the sum of all elements is preferably close to one, and non-zero weights are only assigned to elements with physical connections. The target allocation flow rate of the water collection branch is defined in this scenario as the actual expected volumetric flow rate of each drainage pipeline, preferably in units of volumetric flow rate per hour.

[0123] In practical implementation, after the start of each control cycle, the control system reads the water demand index of the water collection section, the water tank acceptance capacity coefficient, and the connection table between the water collection section and the process water tank recorded after normalization in the previous cycle through the process parameter service. The process parameter service provides these parameters with time stamps and version identifiers to ensure that the water demand index and acceptance capacity coefficient used in this step are consistent with the operating status of the previous cycle. The query delay is preferably limited to within half of the current control cycle. If the latest water demand index and acceptance capacity coefficient cannot be obtained within the preset time, the process parameter service returns the index and coefficient of the previous cycle according to the aforementioned rules and marks it as using the parameters of the previous cycle with a status code. When this status code is received, this step continues to construct the water collection allocation matrix, but appropriately reduces the adjustment range in the valve opening adjustment in subsequent steps to prevent excessive action in the case of information lag.

[0124] In this step, the control system first determines the set of water collection sections and the set of process water tanks in the current unit based on the connection relationship table between water collection sections and process water tanks, as well as the drainage connection relationship between each water collection section and one or more process water tanks. For each water collection section and process water tank combination with a connection relationship, the control system calculates an initial allocation weight based on the water collection demand index of the water collection section and the acceptance capacity coefficient of the process water tank. Preferably, the initial allocation weight can be set as the product of the water collection demand index and the acceptance capacity coefficient, multiplied by a topology availability factor. Specifically, the first... From the first water collection section to the first The original allocation weight of each process water tank is denoted as: Topology availability factor is denoted as Then we have:

[0125]

[0126] Among them, when the first The water collection section and the first When there is a physical connection between the process water tanks When the first The water collection section and the first When there is no physical connection between the process water tanks The above method ensures that the original allocation weights are simultaneously constrained by water collection demand, receiving capacity, and structural topology. When the connection table records that the water collection section is indeed connected to the process water tank through a certain drainage pipeline, the topology availability factor is set to one. When there is no physical connection, the topology availability factor is set to zero, thereby ensuring that no physically non-existent allocation path appears in the water collection allocation matrix.

[0127] After calculating the original allocation weights for all physically connected sections, the control system normalizes the row direction for each water collection section. This involves summing all original allocation weights within the corresponding row of that water collection section. When the sum exceeds a preset minimum value, each non-zero original allocation weight within that row is divided by the sum to obtain the allocation ratio of that water collection section to each connectable process water tank. Specifically, the... From the first water collection section to the first The water collection and distribution matrix elements of each process water tank are denoted as follows: Then we have:

[0128]

[0129] Where the denominator is the first The sum of the original allocation weights of each water collection section leading to all connectable process water tanks. Through the above normalization, the sum of the allocation ratios of each process water tank corresponding to the same water collection section is made one. These allocation ratios constitute the water collection allocation matrix elements of the row containing that water collection section, and are restricted to between zero and one. When the sum of all original allocation weights within a row of a water collection section is less than a preset minimum value, it usually means that the water collection demand index of that water collection section is very low or the acceptance capacity coefficients of all connectable process water tanks are close to zero. In this case, the control system preferably adopts a conservative strategy, restoring the allocation weights of that water collection section to all connectable process water tanks to the allocation weights recorded in the previous cycle, or evenly distributing them to the same small value. Simultaneously, the operation log records the situation where the water collection section failed to generate new reasonable allocation weights based on the state parameters in this cycle, so that the calculation rules for the demand index and acceptance capacity coefficient can be adjusted through subsequent operational analysis.

[0130] To avoid excessively drastic changes in the water collection allocation matrix between adjacent control cycles, which could lead to frequent valve actuation, the control system preferably introduces a short-time sliding observation window. This window smoothly adjusts the allocation weights of the same water collection section to the same process water tank over several consecutive control cycles. For example, the matrix elements can be weighted and averaged over the most recent three to five control cycles. When the weight of the newly calculated weight in the current cycle changes more than the weight of the previous cycle by a preset change range, only the preset maximum change range is adjusted, leaving the remaining change to be gradually realized in subsequent cycles. This reduces control jitter while maintaining responsiveness.

[0131] The water collection allocation matrix generated in this step, along with information such as timestamps, device numbers, and weighted smoothing parameter version identifiers, is written into an allocation matrix record table. This record table is appended in chronological order. If multiple matrix calculation requests occur under the same device number and control cycle time marker, only the last successfully written matrix record is retained. The record number ensures that downstream processes obtain the latest version in chronological order, achieving idempotency and sequential control.

[0132] Based on the above water collection and distribution matrix, the control system further calculates the target distribution flow of each water collection branch. In this scenario, a water collection branch is defined as a physical drainage pipeline from a water collection tray or water collection trough in a certain water collection section to a certain process water tank. Each water collection branch has a unique branch number and corresponding maximum continuous discharge flow in the aforementioned structural topology.

[0133] For each water intake branch, the control system first locates the corresponding allocation weight element in the water intake allocation matrix based on its upstream water intake section number and downstream process water tank number. This allocation weight is then multiplied by the target total discharge of the water intake section for the current cycle to obtain the unconstrained target allocation flow rate for that branch. Specifically, the first... From the first water collection section to the first The target flow rate allocated to the water intake branch of each process water tank is denoted as: , No. The target total discharge for each water collection section is denoted as: Then we have:

[0134]

[0135] The above method converts the corresponding allocation ratio in the water collection allocation matrix into the target allocation flow of specific water collection branches. The target total discharge can be calculated based on the water collection load coefficient and the design maximum continuous discharge flow of the water collection section. For example, when the water collection load coefficient is high in the current cycle, the target total discharge is close to the design maximum continuous discharge flow. When the water collection load coefficient is low, the target total discharge can be set as a part of the design maximum continuous discharge flow. Then, the unconstrained target allocation flow of the branch is compared with the design maximum continuous discharge flow of the branch. When the unconstrained target value exceeds the design maximum value of the branch, the target allocation flow of the branch is limited to the design maximum value, and the target allocation flow of other branches in the same water collection section is slightly reduced proportionally to ensure that the total discharge of the water collection section does not increase indefinitely. When the unconstrained target allocation flow of all branches is within its design range, the unconstrained value can be directly used as the target allocation flow.

[0136] To ensure the robustness of the conversion process, the control system also needs to check the relationship between the sum of the target allocated flow rates of all connected branches of the same process water tank and the current remaining receiving capacity of the water tank in each control cycle. When the sum of the target allocated flow rates of all branches of a water tank exceeds the maximum allowable water intake flow rate corresponding to the receiving capacity of the water tank in the current cycle, the control system uniformly scales the flow rate according to the target allocated flow rate ratio of each branch, so that the sum does not exceed the upper limit corresponding to the receiving capacity of the water tank, while keeping the relative ratio between each branch unchanged. Specifically, the first... The maximum allowable water intake flow rate of each process water tank is denoted as: When accessing the first The sum of the target distribution flow rates of all the water intake branches of each process water tank is greater than At that time, the first From the first water collection section to the first The target flow rate of the reduced-size process water tank and its receiving branch is denoted as: Then we have:

[0137]

[0138] Where, the denominator is the number of connections to the first... The target allocated flow rate is the sum of the target allocated flow rates of all the water intake branches of each process water tank. This method ensures that the original proportional relationship of each branch is maintained after reduction, and the sum does not exceed the maximum allowable water intake flow rate of the process water tank. The operation record records the period in which the water tank experiences capacity limitations and the scaling factor. Preferably, the maximum allowable water intake flow rate of a process water tank in the current period is set as the maximum continuous inflow flow rate of the tank's design multiplied by the capacity coefficient for this period. This allows the flow rate to approach the maximum design inflow flow rate when the capacity coefficient is one, and automatically tightens the upper limit of the tank's allocation when the capacity coefficient is small.

[0139] After this step is completed, the control system writes the target allocated flow rate of each water intake branch, along with the time stamp, device number, allocation matrix version identifier, and topology version identifier, into a branch allocation record table. This record table serves as the basis for subsequent valve opening adjustment logic. The process parameter service provides the downstream valve control module with the target allocated flow rate and status code of each branch in the branch allocation query interface, based on the device number and control cycle time stamp. The status code is preferably set to zero, indicating that the target allocated flow rate of the current cycle is based on the latest matrix and has been verified by capacity; one indicates that the target allocated flow rate is based on the matrix of the previous cycle and is reused; and two indicates that no new target allocated flow rate was generated in this cycle and the valve opening should remain unchanged from the previous cycle. When the branch allocation query times out or the record is missing, the process parameter service can retry two to three times as agreed, with each retry not exceeding half of a control cycle. If the branch allocation record is still not obtained after the retry, the target allocated flow rate of the branch in the previous cycle is returned and the status code is marked as one. At the same time, the reason for the failure of the branch allocation generation in this cycle is recorded in the operation log.

[0140] During on-site operation, the rationality of this step can be verified by statistically analyzing the changes in the water collection allocation matrix and the target allocation flow of the branches over several consecutive days. Preferably, when the pressure drop margin coefficient is at a moderate level, the water collection section with a high water collection demand index should have a larger matrix weight for water tanks with a higher receiving capacity coefficient, and a smaller weight for water tanks with a lower receiving capacity. In the actual collected branch flow, it can be observed that the flow of these branches is indeed high. When checking the changes in water tank level, it can be seen that the water level of high-weight water tanks rises faster and the water level of low-weight water tanks rises slower, thus demonstrating that the matrix and target flow calculation results are consistent with the actual hydraulic behavior.

[0141] In a preferred embodiment, a device allocates water to three water collection sections and two process water tanks within a certain control cycle. The water collection demand indices of the three water collection sections are approximately 0.8, 0.6, and 0.3, respectively, and the receiving capacity coefficients of the two process water tanks are approximately 0.7 and 0.5, respectively. The connection table shows that all three water collection sections can connect to the two process water tanks. After calculating the original allocation weights according to the above rules in this step, and after normalization and smoothing, the allocation weights of the first water collection section to the two water tanks are approximately 0.6 and 0.4, the second water collection section is approximately 0.5 and 0.5, and the third water collection section... Approximately 0.4 and 0.6, the maximum continuous discharge flow rate of the three corresponding branches within this cycle can be set to several cubic meters per hour. After converting the water collection load coefficient and the design flow rate, the target total discharge of each water collection section is obtained, and then converted into the target distribution flow rate of the branch according to the matrix weight. In the subsequent few hours of statistics, it was observed that the flow rate of the branch corresponding to the first water collection section to the first water tank was consistently higher than that to the second water tank branch, and the flow rate of the branch corresponding to the third water collection section to the second water tank was even greater. Moreover, the liquid level growth trend of the first water tank was within an acceptable range, indicating that the distribution rule is consistent with the on-site performance.

[0142] In an alternative approach, to reduce the system's computational requirements, a hierarchical quota method can be used instead of a continuous normalization matrix. Preferably, the water demand index is divided into several levels according to preset thresholds, and the receiving capacity coefficient is also divided into several levels. A pre-configured level quota reference table assigns a discrete allocation quota value to each pair of water demand level and receiving capacity level combinations. The quota value is also limited to the range of zero to one. The quota value within each water collection section row is adjusted according to a simple regularization rule to ensure the sum is close to one. Then, the target allocation flow for each branch is generated according to the same branch capacity verification and scaling rules. This hierarchical quota method is also written to the configuration table with a version identifier and retains modification records. Essentially, it is equivalent to the aforementioned method of generating the water collection allocation matrix by weighted normalization of the index and coefficient. Those skilled in the art can choose between the two methods based on the computational power supported by the control system and the ease of maintenance. Both methods can achieve the conversion function of the water demand index of the water collection section and the water tank receiving capacity coefficient to the target allocation flow of the water collection branch.

[0143] S5. Establish a state machine for normal and drainage states. In normal state, smoothly adjust the opening of the water collection branch according to a sliding time window to track the target distribution flow. In drainage state, increase the water collection volume of the excess water collection section and prioritize its allocation to water tanks with sufficient receiving capacity, while reducing the allocation to other water tanks. Specifically, the implementation is as follows:

[0144] Based on the aforementioned established water collection demand index, water tank acceptance capacity coefficient, water collection distribution matrix, and target distribution flow of each water collection branch, and written into the branch distribution record table, in the actual wet desulfurization unit, the control system establishes an operating state machine that includes normal state and discharge state at the valve opening control level. In each control cycle, the state machine is determined according to the current unit operating signal and normalized state parameters, and different adjustment strategies are given to the valve opening of each water collection branch under different states.

[0145] In this scenario, the operating state machine is defined as the general term for the finite set of states and state transition rules surrounding the opening control of the water intake branch. It includes two states: normal state and discharge state, as well as the transition conditions from the normal state to the discharge state and from the discharge state back to the normal state. The current state of the state machine is recorded in the operating state record table of the control system with a state flag field, along with the state switching time and trigger reason code. The state machine configuration includes parameters such as state transition threshold, continuous cycle number threshold, and opening adjustment step size. After these parameters are determined during the commissioning phase, they are written into the parameter configuration table with the state machine parameter version identifier, recording the effective time, applicable device number, and operating batch number participating in the commissioning, forming version locking and evidence chain traceability.

[0146] To facilitate status assessment, the control system obtains the pressure drop margin coefficient, liquid level coefficient of each water collection section, liquid accumulation risk marker of each water collection section, remaining volume coefficient of each water tank, and water quality margin coefficient of the current or most recent cycle through process parameter services in each control cycle. It also collects the actual flow rate of each water collection branch through on-site flow measurement devices and the current opening degree of each branch regulating valve through valve opening feedback devices. The unit of actual flow rate is preferably volumetric flow rate per hour, and the valve opening degree is expressed as a percentage. The sampling rhythm is consistent with the control cycle, and all sampled values ​​are recorded with a unified timestamp of the control system. Time alignment and removal of obvious outliers are performed in accordance with the time alignment and noise reduction rules used in the aforementioned normalization steps to ensure that the data used for status assessment is aligned with the target flow rate of the branch in time.

[0147] The goal of the state machine under normal conditions is to ensure that the actual flow rate of each water intake branch tracks the target allocated flow rate within the sliding time window, while minimizing frequent and large-amplitude valve movements. To achieve this, the control system establishes a sliding time window of several control cycles for each water intake branch under normal conditions. Within this time window, the deviation between the actual flow rate and the target allocated flow rate for each cycle is recorded. In this scenario, the deviation is defined as the actual flow rate minus the target allocated flow rate, expressed as hourly volumetric flow rate. Within each control cycle, the control system calculates the current deviation and a weighted statistic of the deviation within the sliding time window. Specifically, the... The flow deviation of each water intake branch in the current control cycle is denoted as: , No. The actual flow rate of each water inlet branch is recorded as follows: , No. The target allocated flow rate for each water inlet branch is denoted as: Then we have: The first is determined using the above method. The flow deviation of each water intake branch in the current control cycle, for example, by calculating the average value of the deviation and the average level of the absolute value of the deviation, specifically, the flow deviation of the first water intake branch in the current control cycle. The average flow deviation of each water intake branch within the sliding time window is denoted as . The number of control cycles within the sliding time window is denoted as Then we have:

[0148]

[0149] in, For the first Within the first control cycle, the first The actual flow rate of each water inlet branch. For the first Within the first control cycle, the first The target flow rate for each water intake branch is determined. The deviation changes within the sliding time window are statistically analyzed using the above method. When the absolute value of the deviation consistently exceeds the preset allowable deviation range, the valve opening of that branch is adjusted in the direction of increasing or decreasing the opening, based on the direction of the deviation. The preferred adjustment step size is set to a small proportion of the valve's full stroke within each control cycle. If the deviation changes rapidly within the sliding time window, the single-cycle opening change is limited to a preset maximum change amplitude to prevent system oscillation. When the deviation is within the allowable range, the valve opening remains unchanged or only undergoes minor adjustments.

[0150] Under normal conditions, the state machine does not reconstruct the water collection allocation matrix itself. Instead, it uses the target allocation flow rate of the branch and the deviation within the sliding time window as the basis, and gradually approaches the target allocation flow rate through continuous small-step opening adjustments. All opening adjustment actions are recorded in the valve opening adjustment record with timestamps, branch numbers, old openings, new openings, and reference deviations for subsequent analysis. In this scenario, the drainage state is defined as a temporary state for centralized drainage of water collection sections with significant liquid accumulation risk. Under this state, the water collection volume of the water collection section exceeding the threshold is increased and prioritized for allocation to process water tanks with sufficient receiving capacity. At the same time, the allocation volume to other water tanks is appropriately reduced to lower the liquid level of the high-risk water collection section in a short period of time.

[0151] To determine whether it is necessary to transition from the normal state to the drainage state, the control system checks the liquid level coefficient and liquid accumulation risk marker of each water collection section in each control cycle. When the liquid level coefficient of a water collection section exceeds the preset drainage start threshold and the liquid accumulation risk marker is marked as high risk, and this condition is maintained for several consecutive control cycles, the water collection section is added to the drainage candidate set. When the candidate set contains at least one water collection section and the pressure drop margin coefficient is lower than the preset pressure drop tension threshold, the state machine switches from the normal state to the drainage state, and records the switching time and the list of water collection sections that caused the switch in the operation status record table. When the liquid level coefficient of all water collection sections in the drainage set in the drainage state drops below the drainage stop threshold and remains stable for several consecutive control cycles, and the pressure drop margin coefficient recovers to a medium or higher level, the state machine returns from the drainage state to the normal state.

[0152] There is a certain hysteresis between the discharge start threshold and the discharge stop threshold to avoid frequent switching between the two states. After entering the discharge state, the control system recalculates the temporary adjustment coefficient of the corresponding branch in the discharge set in each control cycle based on the current branch target allocation flow and the water tank acceptance capacity coefficient. Under the premise of meeting the branch design maximum continuous discharge flow and water tank acceptance capacity limits, the target actual opening degree of these branches is increased.

[0153] Specifically, for each water collection section in the drainage collection, the control system selects branches from the branches leading to each process water tank in that section that have a receiving capacity coefficient not lower than the preset receiving reliability threshold and a corresponding water tank remaining volume coefficient higher than a certain safety level. These branches are marked as drainage priority branches. In the current cycle, the target allocation flow of the drainage priority branches is multiplied by a drainage gain coefficient to increase it relative to the target allocation flow under normal conditions. At the same time, the target allocation flow of branches in the same section that lead to water tanks with lower receiving capacity is reduced proportionally to ensure that the total discharge of that section increases while still meeting the requirements that the demister pressure drop does not exceed the boundary and the water tank receiving capacity does not exceed the upper limit.

[0154] When a certain water collection section simultaneously meets the conditions that the liquid level coefficient is higher than a higher safety threshold and the risk of liquid accumulation continues to increase, the drainage gain coefficient can be set to a higher value to accelerate the liquid level drop. If the water tank's capacity is insufficient to handle the entire drainage flow, the control system will avoid the system being in a high-load drainage state for a long time by shortening the duration of the drainage state and increasing the rolling grace time of subsequent cycles. The drainage gain and scaling factor will be recorded in detail in the operation log.

[0155] For the water collection section in the non-drainage collection, the control system preferably limits the valve opening adjustment range in the drainage state. That is, only a small adjustment towards the target distribution flow is allowed within the sliding time window. If necessary, the target distribution flow of these sections can be temporarily reduced slightly proportionally to free up the total water collection capacity for the water collection section in the drainage collection.

[0156] Each valve opening adjustment during the drainage process is also recorded in the valve opening adjustment record. The record includes the current state of the state machine, the number of the drainage collection section, the number of the drainage priority branch, the change in opening, and the actual flow rate and level coefficient of the corresponding branch before and after the adjustment. These records are used to optimize the drainage strategy in the future.

[0157] To ensure the idempotency and sequentiality of the state machine operation, the control system is only allowed to complete one state judgment and one overall opening adjustment decision in each control cycle. After executing the opening adjustment logic of this cycle, the state machine state, the contents of the discharge set, and the comparison results of key thresholds for this cycle are written into the state machine record table. If the state machine logic is repeatedly triggered due to communication jitter or external calls, the timestamp and cycle count are used to limit the application of only the last decision result in this cycle. The downstream valve actuator implements the action based on the last record in the state machine record table.

[0158] Regarding interfaces and error codes, the minimum set of fields agreed upon between the process parameter service and the valve control module includes the device number, control cycle time stamp, number of each water intake branch, target flow rate of the corresponding branch, current state machine status flag, and state machine decision status code. The state machine decision status code is preferably set to several values. For example, one type of value indicates that the smooth adjustment strategy is effective under normal conditions, another type of value indicates that the drainage priority strategy is effective under drainage conditions, and yet another type of value indicates that the state machine decision fails in this cycle and the opening setting logic of the previous cycle needs to be continued. When the process parameter service fails to return the target flow rate of the branch or the water tank acceptance capacity coefficient within the agreed delay limit, it can retry several times. If it still fails after retrying, the state machine will remain unchanged in the previous cycle state in this cycle and prohibit the execution of new drainage state switching. At the same time, the reason for this decision failure will be recorded in the operation log.

[0159] During on-site operation, the state machine logic can be verified by statistically analyzing the percentage of time the state machine is in the drainage state across multiple shifts or operating days, the rate at which the liquid level coefficient of the corresponding water collection section falls back when the drainage state is activated, whether the demister pressure drop remains within the low resistance boundary during the drainage process, and whether the water tank level never exceeds the safety upper limit. Preferably, the drainage state should only be activated when necessary and should bring the liquid level of the high-risk water collection section back to a safe range within a short period of time, while the system should be in a normal state for the vast majority of the time and the actual flow of each water collection branch should stably track the target distribution flow of the branch.

[0160] In a preferred embodiment, when a unit is operating under stable load, the liquid level coefficient of a certain upper water collection section rises from 0.7 to 0.9 within several control cycles and remains thereafter. The risk of liquid accumulation is marked as high risk, and at the same time, the pressure drop margin coefficient drops below 0.2. At this time, the state machine switches from the normal state to the discharge state. The branch leading from the water collection section to the process water tank with a higher remaining volume coefficient and a larger water quality margin is marked as the discharge priority branch. The opening of these branches is increased step by step within several cycles. Under the premise of not exceeding the maximum continuous discharge flow rate of the branch, the discharge capacity of the section is significantly improved. After several control cycles, the liquid level coefficient of the section falls back to around 0.6, and at the same time, the pressure drop of the demister is always within the low resistance boundary. Then the state machine returns to the normal state and continues to smoothly track the target distribution flow according to the sliding time window.

[0161] In another alternative approach, if the on-site control system has difficulty implementing deviation statistics and segmented adjustment logic within the sliding time window, a stepped state machine strategy can be adopted. The deviation level and liquid level coefficient are divided into several levels. Under normal conditions, a fixed opening adjustment step size is given by looking up a table according to the level. Under the drainage state, a fixed drainage gain coefficient is given by looking up a table according to the level. The table lookup results are also limited to the branch design capacity and the water tank's receiving capacity. The level division, step size, and gain configuration of this stepped strategy are also written into the configuration table with parameter version identifiers and modification records are retained. Those skilled in the art can preferentially adopt this method on a control platform with limited computing power. Its overall effect is functionally equivalent to the aforementioned implementation method based on sliding time window and continuous adjustment. Both can achieve the goal of smoothly tracking the target distribution flow under normal conditions and centrally alleviating the accumulation of liquid in high-risk water collection sections under the drainage state.

[0162] S6. Summarize the pressure drop, water intake, water quality, and operating status records according to the statistical period, and adjust the weights and thresholds of the low resistance boundary, water intake demand index, and acceptance capacity coefficient to obtain updated control parameters for subsequent operation. The specific implementation is as follows:

[0163] During the long-term operation of the wet desulfurization unit, in order to ensure that the aforementioned low resistance boundary, water demand index, and water tank acceptance capacity coefficient remain reasonable and adaptable under different seasons, different fuels, and different loads, the control system summarizes and analyzes the pressure drop, water intake, water quality, and operating status records on a long time scale according to statistical cycles. Under safety constraints and manual review constraints, the weights and thresholds of the low resistance boundary, water demand, and acceptance capacity are corrected to generate updated control parameters for subsequent operation.

[0164] In this scenario, the statistical period is defined as a period of operation covering various typical loads and operating conditions. It is preferably rolled out daily or by shift. Each statistical period includes several hundred to several thousand control cycles. After each statistical period ends, the control system automatically initiates a parameter evaluation task. The configuration of the parameter evaluation task includes the length of the statistical period, the device number participating in the statistics, the list of indicators to be analyzed, the expected range of each indicator, and the adjustment step size constraint. These configurations are set by engineers during the commissioning phase and written into the configuration table with the parameter evaluation configuration version identifier. The effective time and the scope of applicable devices are recorded to form version locking and evidence chain traceability.

[0165] During the statistical period, the control system continuously collects raw data from the pressure drop record table, branch allocation record table, water tank status record table, and state machine record table. The pressure drop record includes the flue gas operating conditions and demister pressure drop for each control cycle. The water collection related records include the total discharge of each water collection section, the flow rate of each branch, and the liquid level coefficient of the water collection section. The water quality related records include the conductivity, solid content, and corresponding margin coefficient of each water tank. The status record includes the state machine status of each cycle, the list of discharge collection sections, and the discharge duration.

[0166] To ensure that the statistical results are not affected by short-term interference, the control system aligns the records of multiple tables according to the timestamp before statistical calculation. The time deviation is preferably limited to within half of a single control cycle. For records that are missing or significantly deviate from the previous trend, the aforementioned denoising rules are used to identify interference points and interpolate with the effective values ​​before and after or correct them according to the trend of the most recent cycles to ensure that the data participating in the statistics represent a stable operating state.

[0167] After data processing, the control system first evaluates the rationality of the low-resistance boundary within each statistical cycle. Specifically, this involves statistically analyzing the distribution relationship between the measured pressure drop values ​​and the current upper limit of the low-resistance boundary for each operating range. Within each operating range, the system calculates the percentage of time the pressure drop margin coefficient is within the moderate range, the percentage of time the pressure drop margin coefficient is close to zero, and the frequency of the actual pressure drop value approaching the upper limit. If, in certain operating ranges, the pressure drop margin coefficient remains consistently high for an extended period without significant liquid accumulation risk or drainage events, it indicates that the current upper limit may be set too conservatively. The control system can adjust the upper limit within the allowable range of the design mechanical safety upper limit and the safety margin of upstream equipment. The upper limit of the boundary of the operating condition range is slightly increased, and each increase is preferably limited to a small part of the original boundary value. The reason for the increase and the number of samples included in the statistics are recorded. Conversely, when the pressure drop margin coefficient of certain operating condition ranges frequently approaches zero or even touches the low resistance boundary multiple times, and the state machine records show that the liquid discharge state is frequently activated, or the proportion of the liquid level coefficient of the water collection section that is higher than the safety threshold is large, the control system prefers to lower the upper limit of the boundary of the operating condition range or appropriately refine the granularity of the operating condition division of the operating condition range to give a more conservative boundary value. Both the increase and decrease actions are recorded in the boundary parameter table with the newly generated low resistance boundary version as the identifier, and the original version is retained as a read-only record for comparison.

[0168] When adjusting the weights and thresholds of the water intake demand index and the receiving capacity coefficient, the control system uses changes in liquid level, tank level and water quality, and discharge behavior within the statistical period as evaluation criteria. For the weight of water intake demand, the control system calculates the total time the liquid level coefficient exceeds the warning range, the number of control cycles required for the liquid level in the section to fall back before and after the discharge is initiated, and the changes in the water intake demand index within the sliding time window for each water intake section. When it is found that the water intake demand index of some sections does not increase significantly and the liquid level falls slowly under high liquid level conditions, the weight related to liquid level can be appropriately increased and the weight related to pressure drop can be reduced to make subsequent water intake demand more sensitive to changes in liquid level. When the water intake demand index of some sections is too high when the liquid level is moderately low, resulting in unnecessary discharge, the weight related to liquid level can be appropriately reduced and the weight related to water intake load can be increased to give priority to the discharge capacity of high-load sections under the same liquid level.

[0169] Regarding the weights and thresholds of the water tank acceptance capacity coefficient, the control system determines whether the current weight settings are reasonable by statistically analyzing the relationship between the residence time of the liquid level in each water tank within the safe range, the number of times it approaches the limit, the sewage discharge behavior, and changes in conductivity and solid content. When the liquid level of a water tank frequently approaches the upper limit of safety while the water quality margin is still large, it indicates that the volume weight is insufficient, and the volume weight can be appropriately increased while the water quality weight is decreased. When the conductivity or solid content of a water tank frequently approaches the upper limit while the liquid level margin is large, it indicates that the water quality or solid weight is too low, and the corresponding weight can be appropriately increased, and the sensitivity of water quality-related thresholds can be moderately increased so that the acceptance capacity coefficient decreases rapidly when the water quality approaches the limit.

[0170] When adjusting these weights and thresholds, the control system preferably adopts a step-by-step adjustment strategy, that is, the weight change within each statistical period does not exceed the preset maximum step size, so that the system gradually converges to a parameter combination that takes into account both safety and efficiency. After each adjustment, the weight table and threshold table generate a new version identifier, and the original version identifier used within the statistical period, the statistical sample size, and the key indicators that caused the adjustment are summarized and written into the parameter update record table to meet the requirements of engineering review and operation traceability.

[0171] The updated low-resistance boundary, water intake demand weight, water tank acceptance weight, and related thresholds are confirmed by engineers on the monitoring interface before the start of the next statistical cycle. After confirmation, the control system marks the new version parameters as parameters to be effective and replaces the currently effective parameters in the first control cycle of the next statistical cycle. During the replacement process, the state machine defaults to the normal state and prohibits the initiation of new drainage states within the replacement cycle to avoid additional disturbances caused by the superposition of parameter switching and drainage behavior. If engineers find that some adjustments do not conform to field experience or deviate from the process safety boundary during the review process, they can choose to reject this parameter update, so that the original version continues to be effective, and record the reason for the rejection in the parameter update record.

[0172] Between the parameter update process and the online control logic, the process parameter service connects upstream and downstream through a reserved parameter update interface. The parameter update interface uses a minimal set of fields, including the unit number, parameter type identifier, old version identifier, new version identifier, effective time marker, and parameter update status code. The status code can be set to discrete values ​​such as update effective, awaiting manual confirmation, and update failure. When the parameter update task fails to complete data aggregation or storage failure occurs within the specified time, the parameter update module will not change the currently effective parameters in the current statistical cycle and will return an update failure status code. At the same time, the error reason will be written to the operation log to ensure that the operating parameters will not be mistakenly modified under abnormal conditions.

[0173] During on-site operation, engineers can verify the effectiveness of the automatic correction mechanism by comparing parameter update records, pressure drop margin coefficient distribution, frequency of drainage occurrences, and number of times tank levels and water quality exceed limits across multiple statistical periods. Preferably, after the parameters gradually converge, the proportion of time with the pressure drop margin coefficient in the medium range increases, the proportion of time with extremely low and extremely high margins decreases, the frequency of drainage decreases over time, and the duration of each drainage event shortens, resulting in a significant reduction in events where tank levels and water quality exceed limits.

[0174] In a preferred embodiment, the pressure drop statistics of a certain unit within a statistical period showed that the pressure drop margin coefficient under medium and high load conditions was consistently higher than 0.7 and no significant risk of liquid accumulation occurred. In the next round of parameter evaluation, the system increased the upper limit of the low resistance boundary of the corresponding operating condition range by a small portion of the original value. At the same time, it slightly reduced the liquid level-related weight in the water collection demand weight and increased the water collection load-related weight. In the subsequent statistical period, it was observed that the pressure drop margin distribution was more concentrated in the range of 0.4 to 0.7, the number of liquid discharges did not increase, and the water tank level remained within a safe range, indicating that the adjustment was effective.

[0175] In another alternative approach, if the field control system is not suitable for continuous weight adjustment, a hierarchical parameter table method can be adopted. This involves pre-configuring several sets of candidate weight and threshold combinations, each set corresponding to a different operating strategy label. Within a statistical period, based on indicators such as pressure drop margin distribution, discharge frequency, and number of water quality exceedances, a more suitable set of parameter combinations is selected as the effective parameter combination for the next period by looking up the table. Version management is still performed through parameter version identifiers and parameter update records. This method reduces the amount of online computation and the complexity of weight fine-tuning, but is functionally equivalent to the continuous adjustment method. Those skilled in the art can choose between the two methods based on the computing power of the control system and maintenance habits.

[0176] In the operating scenario shown in this embodiment:

[0177] In a preferred embodiment, the present invention is applied to a wet limestone-gypsum flue gas desulfurization device supporting a coal-fired unit with a rated capacity of 600 MW. Three layers of blade-type flue gas demisters are arranged at the top of the absorber of the device. Below the demisters, multiple sets of water collection trays and drain pipelines are respectively connected to the absorber circulating slurry tank A, the clean water tank B, and the intermediate water tank C. In the initial stage of the device operation, the control system continuously collects the inlet flue gas flow rate of the absorber, the inlet flue gas temperature of the demister, the inlet flue gas humidity and saturation index, and the operating load level inside the tower during the stable operation period when the unit load covers 40% - 100% according to the S1 step. The sampling interval is 5 seconds, the length of a single observation window is 1 hour, and the statistical period is 7 days. At the same time, the demister pressure drop signal is collected through the static pressure tapping ports and differential pressure transmitters arranged on the inlet and outlet flue ducts of the demister. The control system aligns the flue gas operating conditions and pressure drop data according to the time stamp within each observation window, identifies and interpolates and corrects the abnormal points that do not conform to the pressure drop change rate and continuous number rules, obtains a smooth pressure drop curve covering various operating conditions of low, medium, and high loads, divides the operating condition intervals by 10% of the design flow rate of the flue gas flow rate and 5°C for the temperature, statistically calculates the 95th percentile pressure drop as the long-term upper limit within each operating condition interval, and combines the mechanical limit pressure drop provided by the demister manufacturer and the strength margin of the tower top flue duct and upstream equipment, sets the allowable pressure drop upper limit to the lower side of the two, obtains a low-resistance boundary where the pressure drop does not exceed the preset envelope within the full operating condition range, and writes the allowable pressure drop upper limit corresponding to each operating condition interval, together with the equipment number, the definition rule of the operating condition interval, the acquisition batch number, and the sample quantity, into the parameter record table in the form of a boundary version identifier. The on-site engineering personnel divide the three layers of blades and the water collection trays below them into 6 water collection levels and several radial partitions according to the height and the flue duct cross-section, a total of 18 water collection sections with relatively independent structures and water collection paths are divided, a unique number is registered for each water collection section, and the corresponding drain pipelines are checked one by one, establishing a one-to-one or many-to-one connection relationship from each water collection section to the slurry tank A, the clean water tank B, or the intermediate water tank C, recording this connection relationship in the control system in the form of a structure topology version identifier, and at the same time associating the low-resistance boundary parameter record table with the structure topology connection relationship table through the process parameter service to provide the boundary parameter and water collection topology query capabilities for subsequent operation.

[0178] After the unit enters the long-term operation phase, the control system, following step S2, synchronously collects current flue gas conditions, demister pressure difference, water collection volume or level in each water collection section, liquid level in each process water tank, conductivity, and necessary turbidity data every 5 seconds as a control cycle. These data are aligned with timestamps, and linear interpolation is performed on sampling points with slight time deviations. Measurements with sudden changes exceeding recent variation ranges and persisting for multiple cycles are marked as fault quantities, and the effective value from the previous cycle is used to ensure the continuous reliability of the data participating in normalization. Taking a typical medium-to-high load operating day as an example, the unit operates within the 85%–95% rated load range, the flue gas flow rate into the absorber is approximately 800,000–900,000 standard cubic meters per hour, the demister inlet temperature is approximately 60℃–70℃, and the measured demister pressure drop fluctuates between 300Pa and 450Pa. The upper limit of the low-resistance boundary corresponding to this operating condition range is approximately 500Pa. In each cycle, the control system, based on the current flue gas flow rate, temperature, and load level, looks up the upper limit of the allowable pressure drop corresponding to the operating condition range from a table. The difference between the measured pressure drop and the upper limit is divided by the upper limit to obtain the pressure drop margin coefficient, which is truncated to the range of 0 to 1. Simultaneously, for water collection sections with level measuring points, the current level is compared with the designed maximum safe level to obtain the level coefficient; for water collection sections with flow measuring points, the current water collection flow rate is compared with the designed maximum continuous discharge flow rate to obtain the flow coefficient. These are then combined with preset weights to form a water collection load coefficient. When the level coefficient is higher than 0.8, the water collection section is marked as having a high risk of liquid accumulation. For each process water tank, the current level is compared with the effective height to obtain the remaining volume coefficient; the current conductivity is compared with the upper limit to obtain the water quality margin coefficient; the turbidity is converted to solids content and compared with the upper limit to obtain the solids margin coefficient; and the discharge margin coefficient is obtained by combining the drain valve opening and discharge capacity. Finally, a comprehensive water tank state coefficient for each tank is synthesized according to preset weights. The pressure drop margin coefficient, the status parameters of each water intake section, and the status parameters of each water tank are written into the operation status record table at the end of each control cycle, along with the timestamp, device number, low resistance boundary version identifier, and structural topology version identifier. For the same timestamp and section / water tank number, an idempotent strategy is used to remove duplicates. The process parameter service provides an interface to query the most recent normalized result by "device number + current time", and returns the current pressure drop margin coefficient, the status parameters of each water intake section, and the status parameters of each water tank, as well as the corresponding batch number and status code.

[0179] Based on the above, the control system calculates the water collection demand index in each control cycle according to step S3, based on the pressure drop margin coefficient and the state parameters of each water collection section, and calculates the water tank receiving capacity coefficient based on the state parameters of each water tank. Taking a certain period of a medium-high load operation day as an example, when the unit load is about 90%, the pressure drop margin coefficient for the current cycle is about 0.45, the liquid level coefficient of the upper water collection section L1 near the flue gas inlet is about 0.75, and the water collection load coefficient is about 0.65; the liquid level coefficient of the middle water collection section L2 is about 0.55, and the water collection load coefficient is about 0.50; and the liquid level coefficient of the bottom water collection section L3 is about 0.40, and the water collection load coefficient is about 0.35. The control system calls upon the water collection demand weight configuration, employing a set of weights: "pressure drop related weight 0.2, liquid level related weight 0.5, and flow rate related weight 0.3". The pressure drop margin coefficient is regarded as the overall water collection activism adjustment factor, the liquid level coefficient as the liquid accumulation risk amplification factor, and the water collection load coefficient as the current drainage capacity occupancy level. The L1, L2, and L3 sections are weighted and superimposed, and then compressed to the 0-1 range through piecewise linear mapping. The resulting water collection demand indexes are approximately 0.65 for L1, approximately 0.50 for L2, and approximately 0.35 for L3. Among them, L1 is marked as high risk because the liquid level coefficient is close to 0.8. The water collection demand index is added with a fixed increment based on the calculation result, but still limited to not exceeding 1, giving it higher priority in subsequent allocation. Meanwhile, the remaining volume coefficient of slurry tank A is approximately 0.60, the water quality margin coefficient is approximately 0.40, the solids margin coefficient is approximately 0.55, and the sewage discharge margin coefficient is approximately 0.50; the remaining volume coefficient of clear water tank B is approximately 0.35, the water quality margin coefficient is approximately 0.80, the solids margin coefficient is approximately 0.90, and the sewage discharge margin coefficient is approximately 0.40; and the remaining volume coefficient of intermediate water tank C is approximately 0.70, the water quality margin coefficient is approximately 0.30, the solids margin coefficient is approximately 0.45, and the sewage discharge margin coefficient is approximately 0.60. The control system calls the water tank acceptance weight configuration, using a set of weights: "volume weight 0.4, water quality weight 0.3, solid weight 0.1, and sewage weight 0.2". The weighted sum of these coefficients yields the acceptance capacity coefficients for each water tank: slurry tank A is approximately 0.50, clear water tank B is approximately 0.55, and intermediate tank C is approximately 0.49. At this point, clear water tank B has a slightly higher overall acceptance capacity than the other tanks due to its higher water quality margin. The calculated water demand index for each water collection section and the acceptance capacity coefficients for each water tank, along with the weight version identifier, are written into the water demand and acceptance capacity record table. The process parameter service provides an interface to query these indices and coefficients according to the control cycle time stamp and returns a status code indicating whether the calculation is based on the latest normalized result or uses the parameters from the previous cycle.

[0180] Subsequently, the control system constructs a water collection distribution matrix based on the water collection demand index, water tank capacity coefficient, and topological connection relationship in each control cycle according to step S4, and calculates the target distribution flow of each water collection branch. During the above time period, it is assumed that L1 can simultaneously discharge into slurry tank A and clear water tank B, L2 can discharge into slurry tank A and intermediate water tank C, and L3 can discharge into slurry tank A, clear water tank B, and intermediate water tank C. For each physically connected "water collection section-water tank" combination, the control system uses the product of the corresponding water collection demand index and the acceptance capacity coefficient as the original allocation weight, and multiplies it by the topology availability factor. Then, the original weight of each row of water collection section is normalized so that the sum of the effective elements in the row is close to 1. After normalization and smoothing, the following results are obtained: the allocation weights of section L1 to slurry tank A and clear water tank B are approximately 0.4 and 0.6; the allocation weights of section L2 to slurry tank A and intermediate water tank C are approximately 0.5 and 0.5; and the allocation weights of section L3 to slurry tank A, clear water tank B, and intermediate water tank C are approximately 0.3, 0.4, and 0.3. Based on the design maximum continuous discharge flow rate and current discharge load coefficient of each water intake section, the control system sets the target total discharge volume of L1, L2, and L3 to 0.9, 0.7, and 0.5 of the corresponding design maximum continuous discharge flow rate, respectively. The total discharge volume of each section is allocated to specific water intake branches according to the above in-row allocation weights, resulting in the unconstrained target allocation flow rate for each branch. This is then compared with the design maximum continuous discharge flow rate of each branch and truncated as needed. All access branches to the same water tank are summarized and compared with the tank's "design maximum continuous inflow flow rate × receiving capacity coefficient". When the sum of the target allocation flow rates of all branches in a water tank exceeds this upper limit, the target allocation flow rates of each branch are scaled proportionally to ensure that the water tank's receiving flow rate does not exceed the limit. Finally, the target allocation flow rate of each water intake branch, along with the allocation matrix version identifier and topology version identifier, is written into the branch allocation record table. The process parameter service provides the valve control module with the target allocation flow rate and status code of each branch for the current cycle, for subsequent valve opening control.

[0181] At the valve control level, the control system establishes an operating state machine including normal and discharge states according to step S5, and adopts different opening adjustment strategies under different states. During most of the above-mentioned medium-to-high load operating days, the state machine is in the normal state. The control system uses 10 control cycles as a sliding time window to statistically analyze the deviation between the actual flow rate and the target distribution flow rate of each water intake branch. When the absolute value of the deviation of a certain branch is continuously higher than the preset allowable range, the valve opening is slowly adjusted in each cycle by a step size not exceeding 2% of the full valve stroke, so that the actual flow rate gradually approaches the target distribution flow rate under a smooth trend. At a certain period, due to the rapid increase in boiler load and the increase in liquid carried by the upper flue gas, the liquid level coefficient of the L1 section rises from 0.75 to 0.92 within several control cycles, and the liquid accumulation risk is marked as high. At the same time, the pressure drop margin coefficient drops to below 0.18 and remains below it for several consecutive cycles, satisfying the state transition conditions of "liquid level coefficient exceeding the discharge opening threshold, liquid accumulation risk being high, and pressure drop margin coefficient being lower than the pressure drop tension threshold". The state machine switches from the normal state to the discharge state. After entering the drainage state, the control system selects branches from the branches leading to each process water tank in section L1 within each control cycle that have a receiving capacity coefficient higher than the receiving reliability threshold and a corresponding water tank remaining volume coefficient higher than the safety level. In this embodiment, the receiving capacity coefficient of clear water tank B is relatively high, and the L1→B branch is marked as a drainage priority branch. The control system multiplies the target allocation flow of this branch by the drainage gain coefficient, and, without exceeding the maximum continuous discharge flow of the branch design and the upper limit of the receiving capacity of clear water tank B, temporarily increases the total discharge capacity of section L1 by about 20%. At the same time, the target allocation flow of the L1→A branch and other branches leading to water tanks with lower receiving capacities is appropriately reduced, so that the overall water collection increases the drainage capacity of L1 while remaining within the receiving capacity of each water tank and the low resistance boundary of the demister. In the drainage state, the water collection section in the non-drainage set continues to adjust the valve opening slightly according to the sliding time window strategy to avoid significantly compressing its water collection capacity. After several consecutive control cycles, the liquid level coefficient in section L1 dropped from 0.92 to around 0.60, and the pressure drop margin coefficient rose to over 0.35, remaining stable for several subsequent cycles. The state machine detected that the liquid level coefficient in the drainage collection section was below the drainage stop threshold and the pressure drop margin had recovered to a moderate level, thus returning from the drainage state to the normal state. Subsequently, it continued to smoothly track the target flow allocation according to the normal state. Throughout the process, the state machine performed only one state judgment and one valve opening adjustment decision per control cycle. All state switching events, valve opening change records, drainage collection contents, and key threshold comparison results were written to the state machine record table and valve opening adjustment record, ensuring idempotency and traceability of the decisions.

[0182] Over a longer timescale, the control system summarizes pressure drop, water intake, water quality, and status records according to the S6 steps and statistical cycles. Under manual review constraints, it corrects the weights and thresholds of low-resistance boundaries, water intake demand, and acceptance capacity to generate updated control parameters. Taking a 30-day, shift-based rolling statistical cycle as an example, the control system extracts data from pressure drop records, branch allocation records, water tank status records, and state machine records. Within each operating condition interval, it statistically analyzes the distribution of pressure drop margin coefficients and the frequency of pressure drop approaching the low-resistance boundary. In each water intake section, it statistically analyzes the percentage of time the level coefficient exceeds the warning range and the number of control cycles required for level drop before and after the discharge state is activated. For each process water tank, it statistically analyzes the correlation between level and conductivity, the number of times solids content exceeds limits, and the relationship between discharge behavior and water quality changes. Statistical results show that under medium-to-high load conditions, the pressure drop margin coefficient consistently ranges from 0.70 to 0.90, and high liquid levels and drainage events are almost nonexistent in these sections, indicating that the low resistance boundary setting for this operating condition range is conservative. Conversely, under some high-humidity and high-load conditions, the pressure drop margin coefficient repeatedly approaches 0, and drainage occurs frequently. Under the premise of meeting the mechanical limits of the demister and the safety margin constraints of upstream equipment, the control system appropriately raises the upper limit of the low resistance boundary for the former type of operating condition range and appropriately lowers the upper limit of the boundary or refines the granularity of the operating condition classification for the latter type of operating condition range, generating new boundary version identifiers. Simultaneously, in some water collection sections, statistics show that the increase in the water collection demand index is not sensitive enough at high liquid levels, and the liquid level drops slowly. In the new round of parameter evaluation, the control system appropriately increases the weight related to liquid level and decreases the weight related to pressure drop. For some water tanks, statistics show that their liquid levels frequently approach the safety upper limit while the water quality margin remains high. The system appropriately increases the volume weight and decreases the water quality weight. The aforementioned boundary adjustments and weight / threshold corrections all employ a small-step strategy, recording the old version identifier, new version identifier, statistical sample size, and key indicators triggering the adjustment in the parameter update log table. After on-duty engineers review and confirm the parameter updates on the monitoring interface, the system uniformly replaces the currently effective parameters in the first control cycle of the next statistical cycle. During this switching cycle, the state machine is fixed in the normal state, and the triggering of new drainage states is prohibited, thereby avoiding the superimposed disturbances of parameter switching and drainage behavior. After several statistical cycles, the pressure drop margin coefficient distribution gradually converges to the 0.40–0.70 range, the proportion of extremely low and extremely high margin times significantly decreases, the frequency and duration of drainage states decrease, and the number of events exceeding the limits for water level and water quality in each tank significantly decreases. This indicates that the online correction mechanism for low-resistance boundaries and weight thresholds can improve the overall operating efficiency of the system while ensuring safety margins. This embodiment demonstrates that, based on the steps S1 to S6 and parameter configurations provided in this invention, those skilled in the art can construct a complete pressure drop boundary adaptive and water collection linkage control closed loop in an actual wet desulfurization unit, and achieve reproducible engineering applications under different coal types and seasonal operating conditions.

[0183] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0184] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0185] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0186] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0187] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0188] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0189] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0190] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0191] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0192] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for linking water collection in a low-resistance demister with the process water tank for circulation, characterized in that, The method is executed by the control system deployed in the wet desulfurization unit, including: S1. Test the pressure drop of the demister based on the flue gas operating conditions, establish a low resistance boundary where the pressure drop does not exceed the preset envelope, and divide the water collection section according to the structure and water collection path, and determine the process water tank connected to the water collection section. S2. Periodically collect flue gas operating conditions, demister pressure difference, water collection volume in the water collection section, water tank level and water quality, and normalize the data to obtain the pressure drop margin coefficient, as well as the state parameters of the water collection section and the state parameters of the water tank. S3. Calculate the water demand index of each water collection section based on the pressure drop margin coefficient and the state parameters of the water collection section, and calculate the receiving capacity coefficient of each water tank based on the state parameters of the water tank. S4. Construct a water collection distribution matrix based on the water collection demand index of the water collection section and the water tank acceptance capacity coefficient, and convert it into the target distribution flow of the water collection branch. S5. Establish a state machine for normal and drainage states. In normal state, smoothly adjust the opening of the water collection branch according to the sliding time window to track the target distribution flow. In drainage state, increase the water collection volume of the water collection section exceeding the threshold and give priority to the water tank with the receiving capacity, and reduce the distribution volume of other water tanks. S6. Summarize the pressure drop, water intake, water quality and operating status records according to the statistical period, correct the weights and thresholds of the low resistance boundary, water intake demand index and acceptance capacity coefficient, and obtain updated control parameters for subsequent operation.

2. The method for linking water collection in a low-resistance demister with process water tank circulation according to claim 1, characterized in that, S1 includes: Collect flue gas operating conditions and demister pressure drop, align the collected data according to a unified timestamp, and filter out pressure drop interference pulses; Divide the range according to the flue gas operating conditions, calculate the upper limit of pressure drop in each range and generate the low resistance boundary by combining it with the allowable pressure drop of the demister. Store the low resistance boundary in the parameter record table with the version identifier. The water collection tray and drainage pipeline are divided into water collection sections according to the height and cross-section of the demister. The correspondence between each water collection section and the process water tank is written into the water collection section and process water tank connection table. The process parameter service provides the low resistance boundary and the process water tank number corresponding to each water collection section based on the version identifier.

3. The method for linking water collection in a low-resistance demister with process water tank circulation according to claim 1, characterized in that, S2 include: The control system deployed in the wet desulfurization unit collects flue gas operating conditions, demister pressure difference, water collection volume in the water collection section, water tank level and water quality according to the control cycle; Align all measurements using a unified timestamp and replace faulty measurements. The above measurements are normalized into operating status parameters, which include pressure drop margin coefficient, water collection section status parameters, and water tank status parameters. Write the operating status parameters, along with the device number and timestamp, into the operating status record table and apply an idempotent write constraint. The process parameter service returns operating status parameters based on the unit number and timestamp.

4. The method for linking water collection in a low-resistance demister with process water tank circulation according to claim 1, characterized in that, S3 include: The control system deployed in the wet desulfurization unit calculates the water collection demand index for each water collection section within each control cycle based on the pressure drop margin coefficient and the state parameters of each water collection section, according to preset water collection demand weight parameters. The water collection demand index for the i-th water collection section is... satisfy in, This is the pressure drop margin factor. Let be the liquid level coefficient of the i-th water collection section. Let be the water collection load coefficient of the i-th water collection section. For the pressure drop related weight parameters, For liquid level related weight parameters, For traffic-related weight parameters, This is the liquid accumulation risk correction amount for the i-th water collection section; Based on the state parameters of each process water tank, the acceptance capacity coefficient of each process water tank is calculated according to the preset water tank acceptance weight parameters. The acceptance capacity coefficient of the j-th process water tank is... satisfy in, Let be the remaining volume coefficient of the j-th process water tank. Let be the water quality margin coefficient of the j-th process water tank. Let be the solids margin coefficient of the j-th process water tank. Let be the blowdown margin coefficient of the j-th process water tank. For volume weight parameters, For water quality weighting parameters, For solid weight parameters, For pollution discharge weight parameters; The water demand index and receiving capacity coefficient, along with the unit number and control cycle time marker, are written into the water demand and receiving capacity record table. The process parameter service can then query and return the water demand index and receiving capacity coefficient by unit number and control cycle time marker.

5. The method for linking water collection in a low-resistance demister with a process water tank for circulation according to claim 1, characterized in that, S4 include: The control system determines the allocation weights of the water collection section and the process water tank based on the water collection demand index of the water collection section, the water tank receiving capacity coefficient, and the connection relationship table between the water collection section and the process water tank, and forms a water collection allocation matrix by normalizing the water collection section. The control system calculates the target distribution flow rate for each water collection branch based on the water collection distribution matrix and the target discharge rate of the water collection section. Specifically, the target distribution flow rate for the water collection branch from the i-th water collection section to the j-th process water tank is... satisfy in, The target discharge amount for the i-th water collection section; For the first From the first water collection section to the first The water collection and distribution matrix elements of each process water tank, when the sum of the target distribution flow rates of the branches connected to the same process water tank exceeds the allowable water collection flow rate of that process water tank. At that time, the flow rate allocated to the receiving branch is reduced proportionally, and the target flow rate allocated to the reduced receiving branch is... satisfy The target flow rate for each water intake branch is then written into the branch allocation record table, whereby... The sum of the target flow rates allocated to all water intake branches connected to the j-th process water tank.

6. The method for linking water collection in a low-resistance demister with process water tank circulation according to claim 1, characterized in that, S5 include: The control system establishes an operating state machine that includes normal state and discharge state at the valve opening control level; Within each control cycle, the current state of the state machine is determined based on the pressure drop margin coefficient, the liquid level coefficient of the water collection section, the liquid accumulation risk marker of the water collection section, the remaining volume coefficient of the water tank, and the water quality margin coefficient. Under normal conditions, the deviation between the actual flow rate and the target flow rate of the receiving branch is calculated based on a sliding time window. Here, the flow rate deviation of the i-th receiving branch in the current control cycle is... satisfy in, Let i be the actual flow rate of the i-th water inlet branch. Assign flow rate to the target of the i-th water collection branch; the average flow rate deviation of the i-th water collection branch within the sliding time window. satisfy Where n is the number of control cycles within the sliding time window. The actual flow rate of the i-th water intake branch within the k-th control cycle. Distribute the flow rate to the target of the i-th water collection branch within the k-th control cycle; and adjust the valve opening of the water collection branch according to the preset opening step size; Write the state machine status flags and valve opening adjustment records into the operating status record table and valve opening adjustment record.

7. The method for linking water collection in a low-resistance demister with the process water tank for circulation according to claim 6, characterized in that: In the draining state, the control system classifies the water collection section whose level coefficient exceeds the draining start threshold into the draining set based on the water collection section level coefficient and pressure drop margin coefficient. Apply a discharge gain coefficient to the water collection branch in the discharge collection section that meets the acceptance reliability threshold for the remaining volume coefficient and water quality margin coefficient, thereby increasing the target distribution flow of the water collection branch; Adjustment range of valve opening limits for water collection branch sections in non-drainage collection and water collection zones.

8. The method for linking water collection in a low-resistance demister with process water tank circulation according to claim 1, characterized in that, S6 include: The control system collects and aligns data from the pressure drop record table, branch distribution record table, water tank status record table, and state machine record table according to the statistical cycle; The distribution of pressure drop margin coefficient, changes in liquid level coefficient in water collection section, water tank level and water quality margin, and discharge status behavior are statistically analyzed according to the working condition interval. Under the premise of meeting mechanical safety constraints and process safety constraints, the low resistance boundary, water collection demand weight, water tank acceptance weight and related thresholds are modified step by step to generate updated control parameters with version identifiers and write them into the parameter update record table.

9. A method for linking water collection in a low-resistance demister with a process water tank for circulation according to claim 8, characterized in that: Before the updated control parameters take effect, they must be reviewed and confirmed by engineers through the monitoring interface. The process parameter service distributes the unit number, parameter type identifier, old version identifier, new version identifier, and parameter update status code through the parameter update interface; If data aggregation is incomplete or storage fails during the parameter update process, the original low resistance boundary, water demand weight, and water tank acceptance weight should be kept unchanged, and the reasons for the update failure should be recorded in the parameter update record table and the operation log.

Images