Submerged chain conveyor system based on multi-loop automatic water replenishing control

Abstract

The invention relates to the technical field of equipment control data processing, in particular to a submerged chain conveyor system based on multi-loop automatic water replenishing control, which comprises a data acquisition module, a water level analysis module, a water replenishing strategy generation module, a multi-loop control module, a water replenishing execution module, a dynamic adjustment module, an exception handling module and a communication coordination module. The data acquisition module synchronously acquires liquid level, temperature and viscosity data through a multi-mode sensor group, and the data are normalized and denoised by the data fusion unit and then transmitted to the water level analysis module; the water level analysis module adopts an ARIMA time sequence prediction model to generate a water level dynamic trend, and outputs a water replenishing demand level through comparison in combination with a viscosity threshold value; the water replenishing strategy generation module calls the rule base to generate a multi-loop water replenishing strategy, the problems of blockage and environmental protection risks caused by the fact that traditional single-loop control cannot cooperatively regulate and control water quality parameters in a multivariable coupling environment are solved, the water replenishing response precision and the system stability are improved, generation of pipeline hardening substances is restrained, and the equipment shutdown frequency is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment control data processing, and in particular to a slag scooping machine system based on multi-loop automatic water replenishment control. Background Art

[0002] In coal-fired power plant slag water treatment systems, water quality control in the slag extractor system directly impacts equipment operational stability and environmental performance. During long-term operation, existing slag extractor systems experience dynamic changes in the pH value of the circulating water due to the continuous precipitation of alkaline substances from ash. Furthermore, suspended solids concentration fluctuates with coal quality and combustion conditions, creating a multivariable coupled environment. Traditional single-loop water replenishment control lacks real-time, coordinated analysis of multidimensional parameters such as circulating water pH, suspended solids concentration, and liquid level. This makes it difficult to timely perceive the gradual changes in slag water composition, leading to a mismatch between the water replenishment strategy and dynamic changes in water quality. This exacerbates the uncertainty of the formation rate and spatial distribution of aggregates in the slag water system. Especially when water quality parameters suddenly change, the existing control logic is unable to establish a correlated compensation mechanism for these multidimensional parameters, resulting in delayed and over-adjusted water replenishment. This further accelerates crystallization within the circulation pipeline and degrades slag water fluidity, creating a positive feedback loop that increases the risk of system blockage, forcing frequent equipment shutdowns for desilting and creating pressure for environmental disposal. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the present invention provides a slag scraper system based on multi-loop automatic water replenishment control. The present invention solves the problems of system blockage and environmental risks caused by the inability of traditional single-loop water replenishment control to coordinately regulate the pH value, suspended matter concentration and liquid level parameters of circulating water in a multi-variable coupling environment.

[0004] In order to solve the above technical problems, the specific technical solutions of the present invention are as follows: The present invention provides a slag scooping machine system based on multi-circuit automatic water replenishment control, comprising: The data acquisition module is configured to collect real-time water level data, temperature data and slag-water mixing state data in the slag scoop; a water level analysis module, in communication with the data acquisition module, for generating a dynamic water level change trend through a time series prediction model based on the water level data and temperature data, and outputting a water replenishment demand level based on a comparison result of the slag-water mixing state data with a preset viscosity threshold; a water replenishment strategy generation module connected to the water level analysis module and configured to call priority allocation rules and flow allocation rules in a preset water replenishment control rule library based on the water replenishment demand level to generate a multi-circuit water replenishment strategy, wherein the multi-circuit water replenishment strategy includes an opening and closing sequence of at least two independent water replenishment circuits, water replenishment flow allocation parameters, and a water replenishment priority based on the historical response efficiency of the circuit; a multi-circuit control module, connected to the water replenishment strategy generation module, configured to generate a control instruction queue based on the water replenishment flow distribution parameters and the water replenishment priority through a conflict resolution algorithm, and send the control instruction queue to multiple water replenishment execution modules to drive valve opening adjustment of different water replenishment circuits; A water replenishment execution module, comprising a plurality of parallel water replenishment circuits, each of which is equipped with an independent flow sensor and an electric valve, the electric valve receiving instructions from the multi-circuit control module to perform dynamic opening and closing operations; a dynamic adjustment module, connected to the data acquisition module and the multi-loop control module, for correcting the water replenishment flow distribution parameters generated by the water replenishment strategy generation module according to the water level change feedback data collected in real time by the data acquisition module, and updating the water replenishment priority based on the corrected parameters; an exception handling module connected to the data acquisition module and the water replenishment execution module, and configured to trigger a backup circuit switching instruction when the flow sensor of the water replenishment execution module detects that the current circuit flow is lower than a preset threshold or the response time of the electric valve exceeds a preset time, and send a request to the water replenishment strategy generation module to reallocate the water replenishment weights of the remaining circuits; The communication coordination module is connected to all modules to synchronize the data interaction timing between modules through the time division multiplexing protocol, and coordinate the parallel execution of multi-circuit water replenishment actions based on a unified clock signal; Among them, the multi-circuit water replenishment strategy dynamically allocates the water replenishment weight of each circuit to match the water replenishment flow with the viscosity and water level change trend of the slag water state in the slag scooper, and the multi-circuit control module generates a control instruction queue through a conflict resolution algorithm based on the water replenishment priority and circuit load status.

[0005] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention, the water level analysis module further includes: The water level prediction unit is used to output the future water level change range through the ARIMA time series prediction model based on historical water level data and current temperature data; The water replenishment strategy generation module calls the time window constraint rules in the water replenishment control rule library according to the future water level change interval, adjusts the water replenishment priority, and generates a multi-loop water replenishment strategy including a water replenishment action time window.

[0006] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the dynamic adjustment module includes: a feedback compensation unit configured to calculate a water replenishment flow compensation coefficient by a proportional and integral algorithm based on a deviation between real-time water level data and a median of a future water level change interval output by the water level prediction unit after the water replenishment execution module is started; The weight redistribution unit is used to adjust the flow distribution parameters according to the historical response efficiency weights of each water replenishment circuit based on the compensation coefficient, and synchronize the adjusted parameters to the multi-circuit control module through the communication coordination module.

[0007] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the abnormality handling module further includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

[0008] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the abnormality handling module further includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

[0009] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention, the multi-circuit control module further includes: a conflict resolution unit configured to dynamically adjust the execution order of instructions according to the water replenishment priority defined by the water replenishment strategy generation module and the current circuit load status of the water replenishment execution module when there is resource competition among the control instructions of multiple water replenishment circuits; The conflict resolution unit generates an optimal instruction queue by comparing the priority tags of each loop and the execution efficiency score calculated based on the historical response time and the flow deviation value.

[0010] Furthermore, in the slag scoop machine system based on multi-loop automatic water replenishment control described in the present invention, the communication coordination module adopts a time-division multiplexing protocol to distribute water replenishment instructions to different time slices, and dynamically adjusts the time slice length based on the real-time flow fluctuation rate and loop load status fed back by the flow sensor of the water replenishment execution module to match the flow distribution parameters of the water replenishment strategy generation module.

[0011] Furthermore, the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention further includes: The energy consumption optimization module is connected to the communication coordination module and is used to count the cumulative working hours and flow data of each water replenishment circuit, generate an energy consumption evaluation report including the circuit energy efficiency ratio and flow deviation rate, and call the energy consumption optimization rules in the water replenishment control rule library based on the report to adjust the circuit activation order in the subsequent water replenishment strategy.

[0012] Furthermore, in the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention, the data acquisition module includes: A multimodal sensor set configured to simultaneously collect liquid level, temperature, and viscosity data through an ultrasonic water level gauge, an infrared temperature sensor, and a viscosity detector; The data fusion unit is used to normalize the multi-source heterogeneous data collected by the multimodal sensor group, and remove abnormal noise data based on the wavelet transform denoising algorithm of the sliding window before transmitting it to the water level analysis module.

[0013] Furthermore, the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention further includes: A self-learning module is connected to the water replenishment strategy generation module and the dynamic adjustment module, and is used to optimize the parameter thresholds in the water replenishment control rule base through a Q learning algorithm based on the historical water replenishment records of the water replenishment strategy generation module and the flow parameter correction results of the dynamic adjustment module, and generate a priority allocation strategy adapted to the current loop load state.

[0014] Beneficial effects of the present invention: The present invention effectively improves the water quality control capability of the slag scoop system through multi-loop collaborative control and dynamic parameter optimization mechanism. Based on multimodal sensor data fusion and ARIMA time series prediction model, the multi-dimensional coupling state of slag water viscosity, pH value and liquid level parameters is perceived in real time to generate accurate water replenishment demand level; the water replenishment strategy generation module combines the time window constraint rules and the loop historical response efficiency to dynamically allocate multi-loop water replenishment weights, and generates an instruction queue through the conflict resolution algorithm to achieve the timing matching of water replenishment action and water quality change; the dynamic adjustment module uses a closed-loop feedback mechanism to correct the flow parameters to suppress the delayed water replenishment or over-adjustment caused by the deviation of the prediction model; the abnormality handling module ensures the continuity of water replenishment and reallocates weight resources through the fault loop isolation and redundant loop activation mechanism; the self-learning module optimizes the rule base threshold based on the Q learning algorithm to make the priority allocation strategy adapt to equipment aging and environmental interference. The above technical solutions work together to reduce the generation rate and distribution uncertainty of pipeline slabs, reduce the frequency of system shutdown and dredging, and improve the operational stability and environmental compliance of the slag water treatment system of coal-fired power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on the drawings without paying any creative labor.

[0016] Figure 1 A system architecture diagram of a slag scoop system based on multi-loop automatic water replenishment control provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. The technical solutions provided by each embodiment of the present invention are described in detail below in conjunction with the drawings. In order to better understand the purpose of the present invention, the present invention is further described in detail below.

[0018] See also Figure 1 The present invention provides a slag scooping machine system based on multi-circuit automatic water replenishment control, comprising: The data acquisition module is configured to collect real-time water level data, temperature data and slag-water mixing state data in the slag scoop; The data acquisition module uses a multimodal sensor group to achieve synchronous acquisition of the internal parameters of the slag picker. The multimodal sensor group includes an ultrasonic water level meter, an infrared temperature sensor, and a rotary viscosity detector, which are respectively deployed at preset monitoring points on the inner wall of the slag picker. The ultrasonic water level meter is based on the principle of pulse echo ranging. It calculates the real-time liquid level height by the time difference between the emission of high-frequency sound waves and the reception of reflected signals. Its sampling period matches the dynamic change rate of slag water to form a continuous liquid level time series data set. The infrared temperature sensor sets three symmetrically distributed temperature measurement areas along the longitudinal axis of the slag picker. It uses non-contact thermal radiation detection technology to periodically scan the surface temperature data of each area and associate the timestamps to generate a spatiotemporal distribution thermal map. The rotary viscosity detector is embedded in the slag water circulation pipeline. It compares the resistance torque generated by the rotation of the drive shaft with the preset torque threshold, inversely calculates the real-time viscosity value of the slag water mixture, and outputs a digital signal after analog-to-digital conversion.

[0019] The raw data collected by the multimodal sensor group enters the data fusion unit for preprocessing. The preprocessing process first performs a timestamp alignment operation. Based on the sampling frequency and transmission delay parameters of each sensor, a linear interpolation algorithm is used to synchronize the liquid level, temperature, and viscosity data to eliminate timing deviations caused by hardware response differences. Subsequently, the normalization processing unit converts the multi-source heterogeneous data into a unified dimension: the liquid level data is mapped to a percentage scale of the total height of the slag scraper, the temperature data is converted to a standard Celsius value, and the viscosity data is standardized using a preset dimensionless interval to form a matrix data set that can be collaboratively analyzed. This matrix provides the data foundation for multi-parameter coupling analysis in subsequent modules, preventing dimensional differences from interfering with the model input.

[0020] The data fusion unit further uses a sliding window wavelet transform algorithm to denoise the standardized data. The sliding window length is set according to the typical period of slag water state change, and the data in the window is separated into high-frequency noise components and low-frequency effective signals after multi-scale decomposition. The algorithm applies dynamic threshold filtering to the high-frequency components, retains the low-frequency components that represent the actual slag water state, and generates a denoised data set after reconstruction. During the processing process, the data fusion unit monitors the fluctuation range of each sensor data in real time. When it detects that the sampling values of a certain sensor exceed the historical statistical interval for multiple consecutive times, the abnormal marking mechanism is triggered. At this time, the system calls the interpolation values of adjacent sensors to replace the abnormal data segment, and records the fault code and abnormal time window in the maintenance log, and simultaneously sends the sensor calibration instruction to the maintenance terminal.

[0021] The preprocessed data is transmitted to the input buffer of the water level analysis module via a communication protocol. This protocol defines the data packet encapsulation format, including a checksum, timestamp, and data type identification field. An integrity check is performed before transmission. If the check fails, the data fusion unit resends the current window of data to prevent data loss from interfering with subsequent analysis modules. During transmission, the communication coordination module dynamically allocates data transmission bandwidth, prioritizing the real-time performance of high-sampling-rate sensor data and maintaining a synchronized update frequency for water level, temperature, and viscosity data.

[0022] The data acquisition module's exception handling mechanism includes multi-level fault-tolerant logic. When the rotary viscosity tester's drive shaft stalls due to accumulation of impurities in slag water, the system switches to a backup viscosity calculation mode. This mode, based on a correlation model between liquid level change rate and temperature data, estimates the current viscosity value using a regression equation trained with historical data, maintaining basic functionality in the event of a sensor physical failure. Furthermore, the data fusion unit assigns a confidence level to the interpolated data, enabling subsequent modules to dynamically adjust parameter weights during strategy generation, balancing the impact of data reliability on control logic.

[0023] In this technical solution, a multimodal sensor array achieves simultaneous acquisition of multidimensional parameters through physical deployment and signal acquisition mechanisms. The data fusion unit aligns timestamps and normalizes data to eliminate discrepancies between sources. A wavelet denoising algorithm suppresses environmental noise interference. An exception handling mechanism ensures data continuity. Communication protocols and verification logic maintain data transmission integrity. These steps work together to provide highly accurate, multidimensional, and time-consistent data input for water level trend forecasting and recharge strategy generation.

[0024] a water level analysis module, in communication with the data acquisition module, for generating a dynamic water level change trend through a time series prediction model based on the water level data and temperature data, and outputting a water replenishment demand level based on a comparison result of the slag-water mixing state data with a preset viscosity threshold; The water level analysis module establishes a data link with the data acquisition module via a communication interface, receiving preprocessed water level, temperature, and slag-water mixture viscosity data in real time. This communication interface utilizes a time-division multiplexing protocol to divide data transmission channels. Water level and temperature data are assigned to high-frequency channels for time synchronization, while viscosity data is periodically updated via a low-frequency channel to maintain temporal consistency across multiple parameter inputs. Received data first enters the input buffer for timing alignment, eliminating packets with misaligned timestamps due to network latency, thereby forming a complete time series dataset.

[0025] The water level dynamic trend prediction unit performs a stationary test on the historical water level data and eliminates seasonal fluctuations and random noise in the data through differential operations. The order of the differential operation is dynamically adjusted according to the water level change cycle. For example, the first-order difference is used in the acceleration stage of the slag water circulation, and the zero-order difference is used in the stable stage. The data after differential processing is input into the time series prediction model, and the model parameters are dynamically corrected in combination with the current temperature data. The temperature data is embedded in the model input layer through a weighted coefficient, where the proximal temperature data has a higher weight than the historical temperature data to reflect the immediate impact of temperature on water level changes. The model outputs the water level fluctuation range for the future period, including the water level rise rate threshold and the decline inflection point prediction value, providing a trend judgment basis for the generation of water replenishment strategies.

[0026] The slag-water mixing state analysis unit synchronously receives the viscosity detection data and compares it with the preset viscosity threshold interval in real time. The viscosity threshold is generated based on the cluster analysis of the slag scraper operation history data, and includes three levels: normal operating range, warning range and abnormal range. During the comparison process, the viscosity data is first processed by sliding window mean filtering to eliminate instantaneous fluctuation interference, and then the difference is calculated with the threshold interval boundary value. When the real-time viscosity value continuously exceeds the upper limit of the warning interval, the viscosity abnormality mark signal is triggered; if it continues to exceed the abnormal interval, a high-priority water replenishment request is generated. This signal is coupled with the water level prediction result for analysis. When the water level decline trend and the viscosity abnormality mark exist at the same time, the water replenishment demand level is increased to an emergency state.

[0027] The water replenishment demand level generation module quantifies water replenishment demand into multi-level control instructions based on a coupled analysis of water level trend prediction results and viscosity comparison status. The quantification rules are dynamically adjusted based on the slope changes in the water level fluctuation range and the duration of viscosity anomalies: when the water level drop rate exceeds a preset threshold and the viscosity is in the warning range, an intermediate water replenishment instruction is generated; when the water level forecast shows a steep drop trend and the viscosity is continuously abnormal, a high-level water replenishment instruction is output. The water replenishment instruction is encapsulated using a priority encoding protocol, embedded with a timestamp and data source identification field, and transmitted to the water replenishment strategy generation module via the communication coordination module. During transmission, the instruction queue is sorted according to the priority tag, and high-level instructions take priority in occupying the communication bandwidth, shortening the system response delay.

[0028] The water level analysis module features a built-in feedback calibration mechanism that dynamically adjusts the weighting coefficients of the time series prediction model based on actual water level changes after water replenishment. This calibration process utilizes an incremental learning algorithm to convert the deviation between the actual water level and the predicted value into adjustments to the model parameters. For example, when the water level consistently falls below the predicted range, the weighting coefficient for temperature data is increased to enhance the temperature's ability to characterize water level changes. Furthermore, the viscosity threshold intervals are periodically updated based on seasonal operating data. The upper viscosity threshold automatically increases in summer conditions, while the lower threshold dynamically decreases in winter conditions, ensuring that the threshold settings adapt to the impact of ambient temperature on the physical properties of the slag water.

[0029] In this technical solution, the communication interface's timing synchronization mechanism ensures the consistency of multi-source data input. The time series prediction model achieves accurate trend prediction through differential processing and temperature weighting. A stratified comparison of viscosity thresholds enhances the robustness of abnormal state identification. A quantitative rule for water replenishment demand levels couples water level and viscosity decision-making with two factors. A feedback calibration mechanism dynamically adapts model parameters to environmental conditions. These links collaborate through closed-loop data flow and parameter iteration, providing a multi-dimensional decision-making basis for multi-circuit water replenishment control.

[0030] a water replenishment strategy generation module connected to the water level analysis module and configured to call priority allocation rules and flow allocation rules in a preset water replenishment control rule library based on the water replenishment demand level to generate a multi-circuit water replenishment strategy, wherein the multi-circuit water replenishment strategy includes an opening and closing sequence of at least two independent water replenishment circuits, water replenishment flow allocation parameters, and a water replenishment priority based on the historical response efficiency of the circuit; The water replenishment strategy generation module receives the water replenishment demand level code and associated parameters transmitted by the water level analysis module via a data interface. This data interface utilizes a priority queue protocol, prioritizing high-level water replenishment instructions to the processing thread while temporarily storing low-level instructions in a buffer queue to avoid response delays due to instruction backlogs. The received water replenishment demand level is decoded and matched with the preset conditions in the water replenishment control rule library, triggering the corresponding priority allocation rules and flow distribution rule call process. The rule library is stored in a tree-like index structure, with the root node representing the water replenishment demand level classification and the child nodes associated with the priority weight calculation formula and flow distribution ratio parameters.

[0031] The priority allocation rule dynamically calculates the weight value based on the historical response efficiency of the water replenishment circuit. The historical response efficiency is comprehensively evaluated through the average response time, flow deviation rate and fault frequency in the past execution records of the circuit, among which the response time has the highest weight, and the fault frequency is included in the calculation as a de-weighting factor. During the calculation process, additional weight is given to the circuit whose response time is lower than the system average for three consecutive times to form an initial priority sequence. The flow allocation rule is based on the maximum theoretical flow value of the circuit and the current load status, and uses a capacity ratio weighted algorithm to allocate the initial flow parameters. For example, the high-priority circuit is allocated 60% of the total demand flow, and the remaining flow is proportionally distributed to the secondary circuit, while reserving 5% of redundant flow for dynamic adjustment.

[0032] The multi-circuit water replenishment strategy generation unit couples the priority sequence with the flow parameters and generates a control instruction set containing the start-up and closing sequence, flow distribution parameters and priority tags. The start-up and closing sequence is arranged in descending order according to the priority tag. The high-priority loop is started first, and the secondary loop is activated after a set time window to avoid resource competition caused by the simultaneous opening of multiple loops. The flow distribution parameters are associated with the priority weight through a linear interpolation algorithm. For every 10% increase in the weight value, the corresponding loop flow increase ratio is set to 8%, forming a nonlinear mapping relationship to balance efficiency and energy consumption. The priority tag is embedded with a timestamp and loop identification code and distributed to the multi-loop control module through the communication coordination module.

[0033] The water replenishment control rule base has a built-in dynamic update mechanism, which iteratively optimizes the rule parameters based on the feedback data after the water replenishment strategy is executed. The feedback data includes the deviation rate between the actual water replenishment flow and the target value, the fluctuation range of the loop response time, and the fault recovery record. During the optimization process, the rule base dynamically adjusts the weight coefficient of the historical response efficiency evaluation formula. For example, when the system frequently experiences flow deviations, the weight of the flow deviation rate is increased to 50%, and the influence of flow control accuracy on priority calculation is strengthened. At the same time, the flow distribution ratio parameters are periodically revised according to seasonal operating conditions data. The allocation weight of high-flow loops is increased under high-temperature conditions in summer, and the activation frequency of low-flow loops is increased under low-temperature conditions in winter to adapt to the influence of ambient temperature on water replenishment efficiency.

[0034] Before the generated water replenishment strategy is transmitted to the multi-loop control module through the communication protocol, the conflict pre-check logic is executed. The pre-check logic traverses the opening and closing time windows and flow parameters in the instruction set to identify whether there is a risk of time slice overlap or flow exceeding the limit. If it is detected that the time window of the high-priority loop overlaps with the secondary loop, a buffer time slice is automatically inserted; when the flow parameter exceeds the maximum carrying capacity of the loop, the flow redistribution algorithm is triggered to distribute the excess flow to other loops according to the priority weight ratio. The policy instructions after pre-check are encapsulated into a standardized control message. The message header contains a check code and an instruction type identifier, and a redundant check field is attached to the tail to improve the integrity of data transmission.

[0035] In this technical solution, priority allocation rules optimize circuit performance through dynamic calculation of historical efficiency. Flow allocation rules combine capacity weighting and nonlinear mapping to balance system load. Opening and closing sequences and time windows are designed to avoid resource contention. A dynamic update mechanism adapts to environmental changes and equipment status, and conflict pre-check logic eliminates execution risks. Through the synergistic effect of data-driven and rule-based iteration, each link forms a multi-circuit coordinated control strategy that matches water level trends and slag water viscosity.

[0036] a multi-circuit control module, connected to the water replenishment strategy generation module, configured to generate a control instruction queue based on the water replenishment flow distribution parameters and the water replenishment priority through a conflict resolution algorithm, and send the control instruction queue to multiple water replenishment execution modules to drive valve opening adjustment of different water replenishment circuits; The multi-loop control module receives the water replenishment flow allocation parameters and priority tags from the water replenishment strategy generation module via a data bus. This data bus utilizes a priority-weighted round-robin mechanism, where high-priority instructions are prioritized and low-priority instructions are transmitted when bandwidth is available, reducing processing delays caused by instruction backlogs. Received parameters are formatted and stored in the instruction cache. This cache is divided into multiple logical partitions, each storing flow parameters, priority weights, and time window constraints for different water replenishment circuits.

[0037] The conflict resolution algorithm generates an initial control instruction queue based on the parameters in the instruction cache. The algorithm first detects resource competition conditions between multiple loops, including the time window overlap rate, the risk of total flow exceeding the limit, and the valve load balancing status. When it is detected that the execution time windows of two or more loops overlap by more than a preset threshold, a dynamic time slice segmentation mechanism is activated: the high-priority loop retains 70% of the original time window as the core execution segment, and the remaining 30% is allocated as a buffer segment to the secondary loop. The buffer segment length is dynamically adjusted based on the loop load fluctuation rate. When the total flow exceeds the limit, the algorithm reduces the flow allocated to each loop according to the priority weight ratio. The reduction is negatively correlated with the weight value, and the reduction of the high-weight loop is lower than that of the low-weight loop.

[0038] The generated instruction queue is reordered by the instruction optimization unit. This optimization unit uses a greedy algorithm to arrange the execution order of instructions, placing control instructions for high-priority loops first and inserting low-priority instructions into idle time slices. Each instruction in the queue contains a valve opening target value, a gradient rate parameter, and a timeout protection flag. The gradient rate parameter is set based on the historical valve response curve to avoid mechanical shock caused by sudden changes in opening. The timeout protection flag is associated with the monitoring interface of the exception handling module, triggering an alarm and releasing occupied resources when the valve action is not completed within the set time.

[0039] The control command queue is distributed to the water replenishment execution module via the communication coordination module. The communication protocol utilizes a master-slave response model. The master control node sends command packets to each loop slave node, which then returns a verification response signal. The packet encapsulation format includes a loop identification code, command type, and a redundancy check field. A retransmission mechanism is triggered when the check fails, and the maximum number of retransmissions is dynamically set based on the command priority. High-priority commands are allowed three retransmissions, while low-priority commands are only allowed one retransmission, balancing communication reliability and timing constraints.

[0040] During valve opening adjustment, the multi-loop control module receives real-time feedback from flow sensors. This feedback data is compared against the target flow values in the instruction queue in a closed-loop manner. A dynamic correction process is triggered when the deviation exceeds a set threshold. This correction process generates a compensation coefficient based on the direction and magnitude of the deviation and adjusts the flow distribution parameters for unexecuted instructions using a weighted algorithm. For example, when the actual flow rate falls below the target value, the flow parameters for the high-priority loop in subsequent instructions are increased by 5%, while the flow rates for the low-priority loops are proportionally reduced. The deviation values for executed instructions are stored in a historical database and used to optimize the weight calculation model of subsequent conflict resolution algorithms.

[0041] In this technical solution, the data bus's priority transmission mechanism ensures timely command processing. A conflict resolution algorithm eliminates resource competition through time-slicing and traffic redistribution. An instruction optimization unit adapts execution timing to machine characteristics. The communication protocol's master-slave response model improves command reliability. A closed-loop feedback mechanism dynamically corrects execution deviations. These links achieve synergy through data-driven, real-time interaction, enabling precise control of multi-circuit water replenishment operations and improving system stability.

[0042] A water replenishment execution module, comprising a plurality of parallel water replenishment circuits, each of which is equipped with an independent flow sensor and an electric valve, the electric valve receiving instructions from the multi-circuit control module to perform dynamic opening and closing operations; The water replenishment execution module consists of multiple parallel water replenishment circuits, each of which contains an independent flow sensor, electric valve and control unit. The water replenishment circuits are distributed radially on the slag scoop circulation pipeline. The water inlet of each circuit is connected to the main water supply pipeline through a diverter, and the water outlet is connected to the slag-water mixing area. The flow sensor adopts the electromagnetic detection principle and is installed in the straight pipe section of the loop pipeline. It measures the electromotive force generated by the fluid cutting the magnetic flux lines to reversely calculate the real-time flow value, and its output signal is transmitted to the control unit after analog-to-digital conversion. The electric valve adopts a stepper motor to drive the valve core structure, receives the opening instruction issued by the multi-circuit control module, and controls the number of motor rotation steps through pulse signals to achieve gradual adjustment of the valve opening.

[0043] The control unit of each water replenishment circuit has built-in instruction parsing and execution logic. The parsing logic performs format verification on the received control instructions and extracts the valve target opening, adjustment rate, and time window parameters. The execution logic generates the frequency of the stepper motor drive pulse based on the adjustment rate parameters, and the opening change rate is positively correlated with the rate parameter. For example, a high adjustment rate corresponds to a high-frequency pulse signal, allowing the valve to reach the target opening in a shorter time. The time window parameter defines the start and end time of the valve action. The control unit forcibly terminates the unfinished action after the time window exceeds the limit and sends an execution timeout alarm signal to the exception handling module.

[0044] Real-time data collected by the flow sensor is fed back to the multi-loop control module via an independent communication link. This communication link utilizes a differential signaling protocol. Flow data and valve status information are encapsulated into data frames. The frame header includes a loop identification code and a timestamp, and the frame footer contains a cyclic redundancy check (CRC) code. The control unit calculates a checksum on the data frame before transmission. Failure to verify triggers a data retransmission mechanism, with the upper limit set dynamically based on the command priority. Flow data for high-priority loops is allowed three retransmissions, while low-priority loops are allowed only one retransmission, balancing data reliability and transmission efficiency.

[0045] The exception handling mechanism is embedded in the control unit's execution flow. When the flow sensor detects that the current circuit flow rate is continuously below the preset lower threshold, the control unit marks the circuit as abnormal and cuts off the drive power to the electric valve. At the same time, the fault code and abnormal time window information are sent to the multi-circuit control module, triggering the activation command of the backup circuit. Mechanical failures of the electric valve are monitored by the Hall effect sensor to monitor the motor rotor position. If the rotor is detected to be stuck or out of step, the control unit immediately initiates a reverse drive pulse to attempt a reset. If the reset fails, the hardware isolation signal is triggered to prevent equipment damage.

[0046] Coordinated control between parallel circuits is achieved through a communication coordination module. Each circuit's control unit receives a unified clock signal and executes valve adjustment actions in a time-division multiplexing (TDM) protocol. Time slice allocation is dynamically adjusted based on circuit priority. High-priority circuits receive continuous time slices for rapid valve opening adjustment, while low-priority circuits' actions are broken down into multiple discrete time slices. During execution, flow sensor data is uploaded in real time to the dynamic adjustment module, which calculates each circuit's actual flow contribution rate and compares it against the target value in a closed-loop manner to generate adjustment compensation parameters for the next cycle.

[0047] In this technical solution, the parallel circuit structure achieves physical isolation and independent control, electromagnetic flow sensors provide high-precision detection, stepper motors enable precise adjustment of valve opening, command parsing and execution logic ensure accurate timing, exception handling mechanisms enhance system fault tolerance, and a time-division multiplexing protocol coordinates the parallel operation of multiple circuits. Through the coordinated design of hardware deployment, control logic, and communication protocols, each link forms an execution system that dynamically matches the water replenishment strategy.

[0048] a dynamic adjustment module, connected to the data acquisition module and the multi-loop control module, for correcting the water replenishment flow distribution parameters generated by the water replenishment strategy generation module according to the water level change feedback data collected in real time by the data acquisition module, and updating the water replenishment priority based on the corrected parameters; The dynamic adjustment module receives real-time water level change feedback data uploaded by the data acquisition module and the current water replenishment flow allocation parameters from the multi-loop control module through a data interface. This data interface utilizes a dual-channel transmission protocol, with water level data distributed to the primary channel for low-latency transmission. Flow parameters are periodically updated via the secondary channel to maintain the timeliness and consistency of data input. The received water level data is first processed using a sliding window mean filter to eliminate transient fluctuations. A smoothed water level change curve is generated, and deviations are calculated from the water level trend predicted by the water replenishment strategy generation module.

[0049] The deviation calculation unit generates a deviation based on the difference between the smoothed water level curve and the median of the predicted trend. This deviation is converted into a flow compensation coefficient using a proportional-integral algorithm. The proportional term is dynamically adjusted based on the current deviation magnitude, with a higher proportional coefficient indicating a larger deviation. The integral term is calculated using an exponential decay model based on the accumulated historical deviations to emphasize the impact of recent deviations on the compensation amount. The compensation coefficient's sign determines the direction of adjustment: a positive value indicates a need to increase the replenishment flow rate, while a negative value triggers a flow reduction.

[0050] The flow distribution parameter correction process couples the compensation coefficient with the current flow parameters of each circuit. This calculation utilizes a priority-weighted correction strategy. The flow rate of high-priority circuits is adjusted by 120% of the product of the compensation coefficient and the weight, while the flow rate of low-priority circuits is adjusted by 80%, balancing system response speed and stability. The corrected flow parameters are verified against a checksum to ensure they fall within the logical range. If they exceed the maximum theoretical flow rate of the circuit, they are normalized and scaled according to the priority weight ratio to prevent execution anomalies caused by parameter over-limit.

[0051] The replenishment priority update mechanism recalculates weights based on revised flow parameters and historical loop response efficiency data. Historical response efficiency is calculated using a sliding window to calculate the average response time and flow deviation rate of the five most recent executions. The weight of response time is increased to 70%, while the weight of flow deviation rate is reduced to 30%, strengthening the impact of timeliness on priority. The recalculated weights are then used to generate updated priority labels through a dynamic sorting algorithm. High-weight loops are moved forward in the label sequence, while low-weight loops are moved back, forming a priority queue adapted to the current water level.

[0052] The updated parameters and priority queues are synchronized to the water replenishment strategy generation module and the multi-loop control module via the communication coordination module. This synchronization process utilizes a differential transmission protocol, transmitting only the corrected incremental data to reduce communication bandwidth usage. Upon receiving this data, the water replenishment strategy generation module triggers a strategy reconstruction process, regenerating water replenishment instructions within the time window constraints. The multi-loop control module adjusts the input parameters of the conflict resolution algorithm based on the new priority queues, optimizing the execution order of subsequent instructions.

[0053] A closed-loop feedback mechanism continuously monitors the effectiveness of the corrected water replenishment. If the water level curve deviates from the predicted trend and remains below a threshold for three consecutive sampling cycles, the dynamic adjustment module reduces the frequency of compensation coefficient calculations, entering steady-state control mode. If the deviation increases again, high-frequency compensation calculations are resumed. Furthermore, the historical response efficiency database regularly clears out expired data, retaining only execution records from the last 24 hours, ensuring that priority calculations always reflect the latest state of the loop.

[0054] In this technical solution, sliding window filtering and deviation calculation enable effective feature extraction from water level data. A proportional-integral algorithm converts deviations into actionable compensation coefficients. A priority-weighted correction strategy balances the adjustment range between loops. A dynamic weighted sorting mechanism adapts to the system's real-time status. A differential transmission protocol improves data synchronization efficiency. A closed-loop feedback mechanism maintains the dynamic balance of water level control. Through data-driven and parameter iteration, each link develops adaptive regulation capabilities, ensuring precise matching of water replenishment strategies with water level fluctuations.

[0055] an exception handling module connected to the data acquisition module and the water replenishment execution module, and configured to trigger a backup circuit switching instruction when the flow sensor of the water replenishment execution module detects that the current circuit flow is lower than a preset threshold or the response time of the electric valve exceeds a preset time, and send a request to the water replenishment strategy generation module to reallocate the water replenishment weights of the remaining circuits; The exception handling module obtains flow sensor data and electric valve status information from the water replenishment execution module through a real-time monitoring interface. This monitoring interface utilizes an event-driven communication protocol. When flow data falls below a preset threshold or valve response time exceeds a set duration, an interrupt signal is immediately triggered, triggering a jump to the fault handling thread. The flow threshold is dynamically calculated based on historical loop operating data and corrected using a sliding window averaging algorithm combined with the current slag water viscosity. As viscosity increases, the lower threshold automatically increases by 5%-10% to prevent misjudgments due to increased fluid resistance.

[0056] The fault detection unit categorizes and labels abnormal events. If flow sensor data falls below a revised threshold for three consecutive sampling cycles, it is identified as a pipe blockage or pump failure, generating a Level 1 fault code. If the electric valve response time exceeds a preset duration, it is identified as a mechanical jam or drive circuit anomaly, generating a Level 2 fault code. The fault code is associated with a specific loop identifier and timestamp, written to the system status register, and triggers an audible and visual alarm.

[0057] The backup circuit switching instruction generation module starts the corresponding processing flow according to the fault code level. The first-level fault triggers the immediate switching logic: a command to interrupt the operation of the current circuit is sent to the communication coordination module, and the readiness status detection of the backup circuit is activated. The detection process includes the backup circuit flow sensor self-test, valve opening and closing response test and pipeline pressure verification. After all are passed, the backup circuit activation permission signal is generated. The second-level fault adopts a gradual switching strategy: reduce the target flow parameter of the faulty circuit to a safe value, and gradually increase the allocation weight of the backup circuit. The flow switching is completed within a 10-second transition period to avoid sudden changes in the total amount of water replenishment.

[0058] The weight redistribution request is sent to the water replenishment strategy generation module through the data encapsulation protocol. The request message contains the fault circuit identifier, the available capacity data of the current remaining circuits, and the status information of the backup circuit. The available capacity data is calculated by subtracting the real-time flow value from the maximum theoretical flow value of each circuit, and the capacity of the backup circuit is calculated according to the initial preset value. After receiving the request, the water replenishment strategy generation module calls the fault mode-specific algorithm in the weight allocation rule base to recalculate the water replenishment weight based on the response efficiency score and available capacity ratio of the remaining circuits. The response efficiency score is calculated based on the average response time and flow control accuracy in the historical data. Circuits with an available capacity ratio of more than 60% receive additional weight bonuses.

[0059] During redundant loop activation, the exception handling module synchronously updates the command queues of the multi-loop control modules. This updated queue inserts control commands from the backup loop and dynamically adjusts the time window parameters of adjacent loops to avoid resource conflicts caused by overlapping time slices. The activated backup loop enters enhanced monitoring mode, increasing its flow sensor sampling frequency to twice that of a regular loop. Valve status data is transmitted back to the exception handling module in real time, creating a dual-channel monitoring mechanism.

[0060] The fault recovery verification mechanism continues to operate after the switchover is complete. It evaluates the effectiveness of the switchover by comparing the deviation rate between the backup circuit's actual replenishment flow rate and the target value. If the deviation rate remains below 5% for five consecutive sampling periods, the switchover is deemed successful and the faulty circuit is de-identified. If the deviation rate persists beyond the target, a secondary switchover process is triggered, using the secondary backup circuit to replace the faulty device. Historical fault data is stored in a knowledge base and used to optimize subsequent threshold calculation models and switchover strategy parameter settings.

[0061] In this technical solution, a dynamic threshold correction mechanism improves fault identification accuracy, hierarchical fault codes implement differentiated switching strategies, a backup circuit detection process ensures switching reliability, a weight redistribution algorithm optimizes system resource utilization, and a dual-channel monitoring and verification mechanism maintains the stability of the water replenishment process. Through a closed-loop logic of condition monitoring, intelligent decision-making, and execution feedback, each link forms a rapid response to equipment abnormalities and adaptive recovery capabilities.

[0062] The communication coordination module is connected to all modules to synchronize the data interaction timing between modules through the time division multiplexing protocol, and coordinate the parallel execution of multi-circuit water replenishment actions based on a unified clock signal; Among them, the multi-circuit water replenishment strategy dynamically allocates the water replenishment weight of each circuit to match the water replenishment flow with the viscosity and water level change trend of the slag water state in the slag scooper, and the multi-circuit control module generates a control instruction queue through a conflict resolution algorithm based on the water replenishment priority and circuit load status.

[0063] The communication coordination module establishes a star-topology network with each submodule via a master control node. The master control node houses a built-in clock signal generator that broadcasts synchronization pulses to all slave nodes. This clock signal is calibrated using the GPS timing protocol, with deviations compared to an external time source every five minutes to maintain a system-wide clock deviation of less than 1 millisecond. Upon receiving the synchronization pulse, the slave nodes adjust their local timers via a phase-locked loop, eliminating timing errors caused by transmission delays and achieving microsecond-level synchronization accuracy for multi-module operations.

[0064] The time-division multiplexing protocol divides the communication channel into periodic time slices, each corresponding to a data transmission window for a specific module. The time slice lengths are dynamically adjusted based on the module's data volume: the data acquisition module is allocated high-frequency, short time slices, polled every 50 milliseconds; the control instruction module uses low-frequency, long time slices, updated every 200 milliseconds. A time slice allocation table is stored in the master node's buffer memory. When new modules are added, idle time slices are dynamically inserted and the allocation table is updated to avoid channel resource conflicts.

[0065] During the data interaction timing synchronization process, data packets sent by each module are embedded with a timestamp generated by a unified clock. This timestamp, accurate to the millisecond level, includes a date identifier and a cycle count field, which are used to align timing when reassembling data on the receiving end. Data packets are encapsulated in a layered format, with a header containing the source / destination address identifier, a middle section containing the payload data, and a cyclic redundancy check code appended to the end. Data packets that fail verification trigger an automatic retransmission mechanism, with the number of retransmissions set based on the data type: control instruction data is allowed three retransmissions, while status monitoring data is allowed only one retransmission, balancing real-time and reliability requirements.

[0066] Parallel execution of multi-loop water replenishment actions is achieved by coupling priority tags with time slices. Based on the priority tags issued by the water replenishment strategy generation module, the master control node assigns a continuous time slice sequence to high-priority loops and a discrete time slice combination to low-priority loops. Each time slice is embedded with a loop identifier, target flow value, and execution duration parameters. Control instructions are activated immediately when the corresponding time window arrives. If the execution time windows of multiple loops overlap, a conflict resolution algorithm intervenes, reordering the execution order by descending priority tags, and delaying the start of secondary loops until the next available time slice.

[0067] A dynamic weight allocation mechanism adjusts the allocation of time-slice resources based on real-time data collected on slag-water viscosity and water-level change rate. This adjustment process utilizes a sliding window average algorithm, calculating a correlation index between viscosity and water-level data every five minutes. When the index exceeds a preset threshold, the time-slice length for the circuit with the highest viscosity is extended by 20%. For every 0.1 m / h increase in the water-level change rate, the frequency of time-slice allocation for the corresponding circuit is increased by 15%. The adjusted parameters are updated to each slave node via differential encoding, transmitting only the changed portion to reduce bandwidth consumption.

[0068] The exception handling unit monitors clock synchronization status and data transmission integrity. If a slave node's clock deviation is detected to persist for more than 5 milliseconds, the master node triggers a local resynchronization process and sends a forced calibration command to the node. If the packet verification failure rate exceeds 10% for three consecutive cycles, the master node automatically switches to a backup communication channel. All exception events are logged, including the fault code, timestamp, and impact description. These logs are uploaded to the maintenance terminal via a separate alarm channel, establishing a traceable fault management system.

[0069] In this technical solution, a star topology and a unified clock achieve global timing synchronization, dynamic time slice allocation optimizes channel resource utilization, data encapsulation and verification mechanisms ensure transmission reliability, priority tags and conflict resolution algorithms coordinate the parallel execution of multiple loops, dynamic weight adjustment adapts to changing operating conditions, and an exception handling unit maintains system robustness. These links collaborate through protocol layering and real-time interaction, supporting the efficient and stable operation of the multi-loop water replenishment control system.

[0070] The slag scoop system, based on multi-loop automatic water replenishment control, provides dynamic control of water quality parameters through the collaborative operation of multiple modules. The data acquisition module uses an ultrasonic water level meter, infrared temperature sensor, and viscosity meter to synchronously collect liquid level, temperature, and viscosity data. A data fusion unit then normalizes this multi-source, heterogeneous data. Specifically, the data fusion unit uses a sliding window wavelet transform denoising algorithm to remove abnormal noise and transmits the normalized data to the water level analysis module, providing a high-confidence data foundation for subsequent analysis.

[0071] The water level analysis module receives processed data and uses the ARIMA time series forecasting model to model historical water levels and current temperature data, generating forecasts for future water level fluctuations. Simultaneously, the module compares real-time slag-water mixing data with preset viscosity thresholds, integrating water level trends and viscosity deviations to output a water replenishment demand level. This level, divided into multiple quantitative intervals, represents the urgency of the current water replenishment demand and the direction of flow adjustment, providing a basis for decision-making in strategy generation.

[0072] The water replenishment strategy generation module calls the water replenishment control rule library based on the water replenishment demand level and determines the activation and closure order of each circuit according to priority allocation rules. The flow allocation rules, combined with the circuit's historical response efficiency, calculate initial flow parameters and generate a multi-circuit water replenishment strategy with time window constraints. For example, when viscosity exceeds the specified value, the high-flow circuit is prioritized for activation. The water replenishment time window is dynamically adjusted based on future water level forecasts to avoid water replenishment delays or redundancy caused by forecast deviations.

[0073] After receiving policy instructions, the multi-loop control module parses the instruction queue using a conflict resolution algorithm. Based on loop priority tags and real-time load status, this algorithm compares each loop's historical response time and flow deviation to derive an execution efficiency score and generate the optimal instruction sequence. Instructions are allocated to different time slots using the communication coordination module's time-division multiplexing protocol. Time slot lengths are dynamically adjusted based on real-time fluctuations reported by flow sensors, ensuring that high-priority loops receive priority execution resources in the event of resource competition.

[0074] The dynamic adjustment module continuously monitors feedback data on water level changes. When the deviation between the real-time water level and the predicted value exceeds a threshold, the feedback compensation unit calculates the flow compensation coefficient using a proportional-integral algorithm. The weight redistribution unit proportionally adjusts the flow distribution parameters for each circuit based on the compensation coefficient and the circuit's historical efficiency weights. The updated parameters are synchronized with the control module via the communication coordination module, forming a closed-loop correction mechanism to mitigate the impact of parameter drift on system stability.

[0075] The exception handling module monitors the circuit status in real time through flow sensors. If the flow rate continuously falls below a preset lower limit or a valve response times out, the fault circuit isolation unit flags the fault and triggers an incremental command, allowing other circuits to compensate for the flow shortfall. If the fault persists beyond the limit, the redundant circuit activation unit invokes a weight allocation rule to recalculate the remaining circuit capacity, activate the backup circuit, and dynamically allocate water replenishment weights, preventing system paralysis caused by a single point of failure.

[0076] The energy consumption optimization module calculates the operating hours and flow deviation rate of each circuit, generates an energy efficiency evaluation report, and calls upon energy consumption optimization rules to adjust the subsequent activation sequence. The self-learning module uses a Q-learning algorithm to iteratively optimize the rule base thresholds based on historical water replenishment records and dynamic adjustment data. This allows the priority allocation strategy to adapt to equipment aging and environmental interference, enabling the continuous evolution of the system control logic. Through a closed-loop data flow and rule-based iteration mechanism, these modules ultimately achieve a multidimensional balance among pH, suspended solids concentration, and liquid level parameters, effectively mitigating the risk of pipeline hardening.

[0077] Specifically, the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the water level analysis module further includes: The water level prediction unit is used to output the future water level change range through the ARIMA time series prediction model based on historical water level data and current temperature data; The water replenishment strategy generation module calls the time window constraint rules in the water replenishment control rule library according to the future water level change interval, adjusts the water replenishment priority, and generates a multi-loop water replenishment strategy including a water replenishment action time window.

[0078] The water level prediction unit of the water level analysis module in this invention processes historical water level data and current temperature parameters using an ARIMA time series prediction model to establish a mathematical representation of water level variation trends. During implementation, the water level prediction unit performs a stationarity test and differential processing on the historical water level data, eliminating seasonality and random noise in the data. It then constructs an autoregressive integral moving average model parameter matrix to generate predictions for the upper and lower limits of future water level variation. These predictions are output as a time series, representing the range of water level fluctuations within a specific future time period and providing a quantitative basis for planning the time window for water replenishment strategies.

[0079] After receiving the future water level change interval data from the water level prediction unit, the water replenishment strategy generation module invokes the preset time window constraint rules in the water replenishment control rule library. These time window constraint rules calculate the start time and duration of the water replenishment action based on the slope change rate and viscosity test data during the prediction interval. For example, if the prediction interval indicates that the water level drop rate exceeds a threshold and the slag water viscosity increases, the rule library automatically shortens the trigger delay of the water replenishment action and extends the water replenishment duration to the end of the prediction interval, preventing water replenishment delays caused by prediction errors.

[0080] The water replenishment strategy generation module uses a dynamic weighting algorithm to adjust water replenishment priorities based on time window constraints. This algorithm couples historical loop response efficiency data with predicted water level change rates to calculate a real-time priority score for each loop. During the scoring process, loops with response latencies below average are given higher weights. Loops experiencing periods of predicted water level drops are also weighted to generate a priority sequence that matches the time window. This updated priority sequence is embedded in the multi-loop water replenishment strategy, forming the basis for control instructions that incorporate time constraints.

[0081] The generated water replenishment strategy is transmitted to the multi-loop control module via the communication coordination module and cross-validated with real-time water level monitoring data. When the deviation between the actual water level data and the predicted interval exceeds the preset tolerance range, the dynamic adjustment module activates the feedback compensation mechanism to shrink or expand the time window parameters. Simultaneously, the exception handling module continuously monitors the loop execution status. If it detects that a higher-priority loop fails to respond to instructions on time due to equipment failure, the strategy reconstruction process is immediately triggered, and the water replenishment strategy generation module recalculates the time window adaptation parameters for the remaining loops to maintain the temporal continuity of water level control.

[0082] Specifically, the slag scoop system based on multi-circuit automatic water replenishment control according to the present invention, the dynamic adjustment module includes: a feedback compensation unit configured to calculate a water replenishment flow compensation coefficient by a proportional and integral algorithm based on a deviation between real-time water level data and a median of a future water level change interval output by the water level prediction unit after the water replenishment execution module is started; The weight redistribution unit is used to adjust the flow distribution parameters according to the historical response efficiency weights of each water replenishment circuit based on the compensation coefficient, and synchronize the adjusted parameters to the multi-circuit control module through the communication coordination module.

[0083] After the water replenishment execution module is activated, the feedback compensation unit of the dynamic adjustment module obtains the median data of the future water level change interval output by the water level prediction unit in real time. This median data is used as a reference value and the difference between it and the actual current water level measurement value uploaded by the data acquisition module is calculated to generate the water level deviation. The feedback compensation unit uses a proportional and integral composite algorithm to process the deviation. The proportional term is used to quickly respond to the current water level deviation amplitude, and the integral term accumulates historical deviation trends to output a comprehensive compensation coefficient. The compensation coefficient represents the correction amount for the matching degree between the current water replenishment flow and actual demand. A positive coefficient indicates that the water replenishment flow needs to be increased, while a negative coefficient triggers a flow reduction.

[0084] After receiving the compensation coefficient, the weight redistribution unit adjusts the flow parameters based on the historical response efficiency weights of each water replenishment circuit. The historical response efficiency weight is dynamically calculated based on the response delay time and flow control accuracy of the circuit's past execution of the water replenishment command. The higher the weight value, the better the circuit execution efficiency. During the adjustment process, the weight redistribution unit will perform a weighted calculation on the compensation coefficient and the weight of each circuit according to a preset ratio to generate new flow distribution parameters. For example, when the compensation coefficient is positive, the high-weight circuit is allocated a higher flow increase to quickly correct the water level deviation; if the compensation coefficient is negative, the flow output of the low-weight circuit is preferentially reduced.

[0085] The updated flow distribution parameters are transmitted to the multi-loop control module via the communication coordination module's time-division multiplexing channel. Upon receiving the parameters, the multi-loop control module immediately updates the flow constraints in the conflict resolution algorithm and regenerates the control instruction queue. Simultaneously, the dynamic adjustment module continuously monitors water level change data. When it detects a new water level deviation exceeding a threshold, it triggers a new round of compensation coefficient calculation and weight allocation, forming a closed-loop control circuit. This mechanism effectively suppresses parameter drift caused by prediction model errors or environmental interference, maintaining a dynamic match between the replenishment flow rate and the changing state of the slag water.

[0086] Specifically, in the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the abnormality handling module further includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

[0087] The fault circuit isolation unit of the abnormality handling module continuously monitors the real-time flow data of each circuit through the flow sensor of the water replenishment execution module. When it is detected that the flow value of a certain circuit is lower than the lower flow threshold preset in the water replenishment control rule library for three consecutive sampling cycles, the circuit is determined to have entered an abnormal state. The lower flow threshold is dynamically set according to the minimum effective water replenishment flow value in the historical operation data, and is corrected by associating it with the current slag water viscosity data. After the fault is determined, the isolation unit marks the circuit status as a fault code and writes it into the system status register. At the same time, it sends a flow increment instruction to other normal circuits through the communication coordination module. The increment value is 1.2 times the original allocated flow of the faulty circuit, and is shared by the remaining circuits according to the priority ratio.

[0088] The redundant circuit activation unit starts working after the fault state lasts for more than a preset time of 5 minutes. The time setting is based on the statistical median of multiple fault recovery test data. The activation process first retrieves the backup circuit device ID, sends a circuit activation request to the multi-circuit control module, and after receiving the hardware ready signal, calls the weight allocation rule in the water replenishment strategy generation module. The weight allocation rule recalculates the water replenishment weight using a capacity-weighted average algorithm based on the current available capacity of the remaining circuits and the maximum flow parameters of the backup circuit. Specifically, the maximum theoretical flow value of each circuit minus the real-time flow occupancy value is used as the available capacity base, the initial weight of the backup circuit is set to 20% of the total base, and the remaining weights are allocated to the original normal circuits in proportion.

[0089] The recalculated water replenishment weights are synchronously updated to the conflict resolution unit of the multi-loop control module via the communication coordination module. The conflict resolution unit adjusts the priority tags in the instruction queue based on the new weights and reallocates time slice resources. During this process, the dynamic adjustment module continuously monitors the total water replenishment volume. If fluctuations in the total volume exceed a preset tolerance, the feedback compensation unit intervenes and makes fine adjustments to maintain a dynamic balance between the replenishment flow rate and the updated weight allocation strategy, ensuring that the total water replenishment volume remains compatible with changes in the slag water state.

[0090] Specifically, in the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the abnormality handling module further includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

[0091] The fault circuit isolation unit of the abnormality handling module obtains the real-time flow data of each circuit through the flow sensor of the water replenishment execution module, and compares it with the preset lower flow threshold in the water replenishment control rule library. The lower flow threshold is dynamically calculated based on the lowest effective water replenishment flow in the historical operation data, and is dynamically corrected in conjunction with the current slag water viscosity detection value. When it is detected that the flow of a certain circuit is lower than the threshold for three consecutive sampling cycles, it is determined to be a hardware blockage or valve failure. The circuit status is marked as a fault code and written into the system status register. After the instruction is triggered, the communication coordination module sends a flow increment request to the remaining normal circuits. The increment value is 120% of the original allocated flow of the faulty circuit, and is shared and executed by the top three priority circuits according to the weight ratio.

[0092] The redundant circuit activation unit starts after the fault state lasts for more than a preset 5-minute threshold. The duration is set based on the median value of the equipment fault recovery test data. The activation process first retrieves the device identification code of the backup circuit, sends a hardware activation instruction to the multi-circuit control module, and after receiving the ready signal feedback from the backup circuit, calls the weight allocation rule in the water replenishment strategy generation module. The weight allocation rule is calculated based on the available capacity of the current remaining circuits. The available capacity is defined as the difference between the maximum theoretical flow value of the circuit and the real-time flow occupancy value. The initial weight of the backup circuit is set to 20% of the total available capacity base, and the remaining weight is dynamically allocated according to the historical response efficiency score of each circuit.

[0093] After the recalculated water replenishment weights are synchronized to the multi-loop control module via the communication coordination module, the conflict resolution unit reconstructs the instruction queue priority labels based on the updated weight values. The dynamic adjustment module monitors changes in the system's total water replenishment volume in real time. When fluctuations in the total volume are detected outside the preset tolerance range, the feedback compensation unit intervenes to fine-tune the compensation coefficient, dynamically adapting the flow distribution parameters to the updated weight strategy. Furthermore, the self-learning module records parameter adjustments made during fault handling, which are used to optimize the flow lower limit threshold setting logic in the water replenishment control rule base, improving the sensitivity of subsequent fault identification and the scheduling efficiency of redundant loops.

[0094] Specifically, the slag scoop system based on multi-circuit automatic water replenishment control according to the present invention, the multi-circuit control module further includes: a conflict resolution unit configured to dynamically adjust the execution order of instructions according to the water replenishment priority defined by the water replenishment strategy generation module and the current circuit load status of the water replenishment execution module when there is resource competition among the control instructions of multiple water replenishment circuits; The conflict resolution unit generates an optimal instruction queue by comparing the priority tags of each loop and the execution efficiency score calculated based on the historical response time and the flow deviation value.

[0095] The multi-loop control module's conflict resolution unit initiates a resource contention analysis process when it detects that multiple loops in the water replenishment execution module are simultaneously requesting execution instructions. The resource contention trigger condition is based on a comprehensive assessment of the time-slice overlap ratio of control instructions in the instruction queue and the loop load status. A resource contention event is identified when the time-slice overlap ratio exceeds a preset threshold and the loop load reaches 80% of its maximum capacity. At this point, the conflict resolution unit obtains each loop's priority tag from the water replenishment strategy generation module. This tag is a composite of the water replenishment demand level, the loop's historical response efficiency, and the current load status, forming a dynamically updated priority score sequence.

[0096] The conflict resolution unit couples priority tags with a real-time calculated execution efficiency score. The execution efficiency score is calculated based on the average historical response time of the circuit and the flow deviation rate of the three most recent water replenishment operations, with response time weighted 60% and flow deviation rate weighted 40%. After the score calculation is complete, the circuits are sorted in descending order to generate an initial priority queue. Furthermore, based on circuit load status data, the scores of highly loaded circuits are dynamically attenuated to prevent excessive use of system resources.

[0097] The generated initial priority queue is cross-validated with the instruction queue of the multi-loop control module through the communication coordination module. When an overlap between the instruction execution window of a high-priority loop and a low-priority loop is detected, the conflict resolution unit initiates a dynamic time-slicing mechanism. This mechanism compresses the execution window of high-priority instructions to 70% of its original length and inserts a buffer time slice for low-priority instructions. The buffer duration is dynamically adjusted based on the loop load fluctuation rate. The adjusted instruction queue is redistributed to each loop via a time-division multiplexing protocol, and the execution effect is monitored in real time by flow sensors.

[0098] If it is detected that the execution efficiency score deviation in the adjusted queue still exceeds the preset tolerance range, the feedback compensation unit intervenes to generate a compensation coefficient. The conflict resolution unit dynamically corrects the priority label weight according to the compensation coefficient, and re-sorts it to generate an optimized instruction queue. During the correction process, the priority of loops that have been scored below the average for three consecutive times and whose load status has not reached the threshold is downgraded, and the released resources are allocated to loops with a significant upward trend in scores. The final generated instruction queue is issued for execution through the multi-loop control module, and the score calculation parameters are continuously iterated and optimized during the execution process to form a closed-loop control logic.

[0099] Specifically, in the slag scoop machine system based on multi-loop automatic water replenishment control described in the present invention, the communication coordination module adopts a time-division multiplexing protocol to allocate water replenishment instructions to different time slices, and dynamically adjusts the time slice length based on the real-time flow fluctuation rate and loop load status fed back by the flow sensor of the water replenishment execution module to match the flow distribution parameters of the water replenishment strategy generation module.

[0100] The communication coordination module allocates water replenishment instructions to a sequence of discrete time slices using a time-division multiplexing protocol. The initial length of each time slice is calculated based on the flow allocation parameters issued by the water replenishment strategy generation module. In specific implementation, the time slice initialization module allocates instructions for high-priority loops to the front end of consecutive time slices based on each loop's priority label and historical average response time, inserting low-priority instructions into the interval buffer period. After the time slices are divided, the communication coordination module initiates a real-time monitoring mechanism, continuously receiving loop flow fluctuation rate data and load status parameters uploaded by the flow sensors in the water replenishment execution module.

[0101] Flow rate fluctuation data is processed using a sliding window variance algorithm to calculate the standard deviation of each circuit's flow rate within the current period. The circuit load state parameter is also dynamically characterized by the ratio of each circuit's electric valve opening to the theoretical maximum flow rate. When a circuit's flow rate fluctuation exceeds a preset threshold or the load reaches a critical value, the dynamic adjustment module generates a time slice adjustment factor. This factor is calculated by weighting the fluctuation rate increase and the load deviation, with a weighting ratio of 7:3.

[0102] The adjusted time slice lengths are redistributed to the corresponding loops via the communication coordination module's clock synchronization unit, and the timestamps of the instruction queues are updated. For loops with high traffic fluctuations, the time slice length is expanded by the adjustment factor, with the expansion not exceeding 150% of the original length. For loops with load conditions exceeding the safety threshold, the time slice length is compressed in segments, with each compression step being 15% of the original length until the load returns to the normal range. The adjusted time slice sequence is verified by a checksum and transmitted to the multi-loop control module to ensure timing consistency with the instruction queue generated by the conflict resolution unit.

[0103] During the adjustment process, the water replenishment strategy generation module continuously receives updated time slice parameters and invokes the flow allocation adaptation algorithm in the rule base for cross-validation. This validation logic analyzes the match between the total time slice length and the theoretical water replenishment demand. If the deviation exceeds 5%, the strategy reconstruction process is triggered, regenerating the water replenishment priority sequence and updating the instruction queue. Furthermore, the dynamic adjustment module records the triggering conditions and execution results of each time slice adjustment, which serve as a training dataset for the self-learning module to optimize the rule base parameters, thereby improving the accuracy of the subsequent time slice allocation strategy's adaptation to environmental changes.

[0104] Specifically, the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention further includes: The energy consumption optimization module is connected to the communication coordination module and is used to count the cumulative working hours and flow data of each water replenishment circuit, generate an energy consumption evaluation report including the circuit energy efficiency ratio and flow deviation rate, and call the energy consumption optimization rules in the water replenishment control rule library based on the report to adjust the circuit activation order in the subsequent water replenishment strategy.

[0105] The energy consumption optimization module obtains the cumulative operating hours and real-time flow data of each water replenishment circuit through the communication coordination module, and performs data aggregation processing based on preset statistical periods. Cumulative operating hours are calculated based on the difference in opening timestamps of the circuit's electric valves, while real-time flow data is integrated using a sliding window based on the instantaneous values collected by the flow sensors. The module's built-in energy efficiency ratio calculation unit uses the ratio of the total circuit flow to operating hours as the core energy efficiency indicator. The flow deviation rate is dynamically calculated based on the percentage difference between the actual flow rate and the theoretical flow value issued by the water replenishment strategy generation module.

[0106] The generated energy consumption assessment report includes each circuit's energy efficiency ranking, flow deviation rate distribution, and historical trend comparison data. After the report is generated, the energy consumption optimization module invokes the energy consumption optimization rule in the water replenishment control rule library. This rule sets high-priority activation for circuits with energy efficiency ratios above the system average and flow deviation rates below 15%. When the rule is applied, the module compares the current circuit's energy efficiency data with the historical best record. Circuits with energy efficiency ratios that have decreased by more than 10% over three consecutive statistical periods are marked as downgraded and their activation order is shifted two places back.

[0107] The adjusted circuit activation order is synchronized to the water replenishment strategy generation module via the communication coordination module, embedding the priority allocation logic for the multi-circuit water replenishment strategy. During strategy execution, the dynamic adjustment module monitors circuit energy efficiency changes in real time. When it detects that the energy efficiency ratio of a degraded circuit has returned to its historical average level, it triggers a mechanism to reweight the activation order. Simultaneously, the self-learning module records energy efficiency change data after each strategy adjustment and optimizes the threshold parameters in the energy consumption optimization rules using a Q-learning algorithm, adapting the circuit activation order adjustment strategy to the energy efficiency degradation trend caused by equipment aging.

[0108] Specifically, the slag scooping machine system based on multi-circuit automatic water replenishment control according to the present invention, the data acquisition module includes: A multimodal sensor set configured to simultaneously collect liquid level, temperature, and viscosity data through an ultrasonic water level gauge, an infrared temperature sensor, and a viscosity detector; The data fusion unit is used to normalize the multi-source heterogeneous data collected by the multimodal sensor group, and remove abnormal noise data based on the wavelet transform denoising algorithm of the sliding window before transmitting it to the water level analysis module.

[0109] The data acquisition module's multimodal sensor suite uses an ultrasonic water level meter, infrared temperature sensor, and viscosity meter to simultaneously collect liquid level, temperature, and viscosity parameters. The ultrasonic water level meter uses the pulse-echo ranging principle to calculate the real-time liquid level by emitting high-frequency sound waves and receiving the reflected signal. The sampling frequency is set to 10 times per second to match the dynamic rate of change of the slag water. The infrared temperature sensor, based on non-contact thermal radiation detection technology, cyclically scans three pre-set monitoring points on the inner wall of the slag extractor. When collecting temperature data, it simultaneously records the corresponding timestamp, enabling the associated storage of data in the temporal and spatial dimensions.

[0110] The raw data collected by the multimodal sensor array undergoes multi-source heterogeneous processing through a data fusion unit. The processing flow first performs time alignment on the ultrasonic level data, infrared temperature data, and viscosity measurement data to eliminate timing deviations caused by sensor response delays. Subsequently, a normalization unit converts the level data to a percentage scale, normalizes the temperature data to degrees Celsius, and maps the viscosity data to a preset dimensionless range of 0-100, forming a standardized data matrix to address the dimensional inconsistency of multiple sensors.

[0111] The data fusion unit processes standardized data using a sliding window wavelet transform denoising algorithm. The window length is set to 30 seconds to cover the typical period of slag-water state changes. The algorithm performs a multi-scale decomposition of the data within the window, identifying high-frequency noise components and applying threshold filtering to them, retaining the low-frequency, significant signals that represent the true slag-water state. After the denoised data is verified for integrity using a checksum, it is encapsulated into data packets according to a pre-set protocol and transmitted to the input buffer of the water level analysis module, providing a highly reliable data source for subsequent trend forecasting and water replenishment demand analysis.

[0112] During processing, the data fusion unit monitors sensor data for abnormalities in real time. If a sensor's sampled data exceeds its historical fluctuation range for five consecutive times, an anomaly flagging mechanism is triggered. The data for that period is replaced with interpolated values from adjacent sensors, and a fault code is recorded in the system log. Simultaneously, the communication coordination module sends a sensor status warning signal to the maintenance terminal, prompting equipment inspection or calibration to maintain the long-term reliability of the data acquisition system.

[0113] Specifically, the slag scooping machine system based on multi-circuit automatic water replenishment control of the present invention further includes: A self-learning module is connected to the water replenishment strategy generation module and the dynamic adjustment module, and is used to optimize the parameter thresholds in the water replenishment control rule base through a Q learning algorithm based on the historical water replenishment records of the water replenishment strategy generation module and the flow parameter correction results of the dynamic adjustment module, and generate a priority allocation strategy adapted to the current loop load state.

[0114] The self-learning module accesses historical water replenishment records from the water replenishment strategy generation module and flow parameter correction records from the dynamic adjustment module through a data acquisition interface. These historical water replenishment records contain the start and close times, flow distribution parameters, and execution performance evaluation data for each circuit under different operating conditions. The flow parameter correction records store the compensation coefficients and weight distribution results for each dynamic adjustment. The data preprocessing unit extracts features from these records and constructs a state space vector based on the water replenishment demand level, circuit load status, and environmental parameters, which serves as the input dataset for the Q-learning algorithm.

[0115] The Q-learning algorithm iteratively updates the parameter thresholds in the rule base. Its action space is defined as the direction and magnitude of parameter threshold adjustments. The reward function is calculated based on the water level stability, the improvement in the circuit's energy efficiency, and the reduction in the failure rate after the water replenishment strategy is executed. Water level stability is weighted 50%, while energy efficiency and failure rate each account for 25%. After each strategy execution, the algorithm updates the Q-value table based on the reward value of the current state and action, dynamically adjusting the learning rate of the parameter thresholds to adapt the optimization process to the changing rate of the system's operating environment.

[0116] The optimized parameter thresholds are written into the water replenishment control rule library through the communication coordination module, replacing the original threshold parameters. The priority allocation strategy generation unit calculates the real-time priority score for each circuit based on the updated thresholds and the current circuit load status data. The score calculation uses a weighted summation model, assigning weights to the circuit response efficiency, historical fault frequency, and current load ratio in a 4:3:3 ratio to generate a priority sequence that adapts to the current operating conditions.

[0117] After the generated priority allocation strategy is embedded in the water replenishment strategy generation module, the multi-loop control module executes a new round of water replenishment instructions, while the self-learning module initiates a closed-loop verification process. This verification process evaluates the effectiveness of parameter threshold adjustments by comparing the water level fluctuation variance, average loop energy efficiency ratio, and fault interval duration before and after strategy optimization. If the evaluation results fall short of expectations, the self-learning module triggers a reinforcement learning mechanism, increasing the sampling frequency of historical data and adjusting the reward function weight ratio. A second iterative optimization process is then performed until the strategy is effective.

[0118] The specific embodiment of the present invention relates to a slag scoop system based on multi-loop automatic water replenishment control, which is applied to the slag water treatment scenario of coal-fired power plants, and realizes dynamic control of water quality parameters through the collaboration of multiple modules. The data acquisition module synchronously collects liquid level, temperature and viscosity data through ultrasonic water level gauge, infrared temperature sensor and viscosity detector, and the sampling frequency is set to 10 times per second to match the dynamic change rate of slag water. The raw data collected by the multimodal sensor group is time-aligned and normalized by the data fusion unit. The liquid level data is converted to a percentage scale, the temperature data is unified into Celsius units, and the viscosity data is mapped to a dimensionless range of 0-100 to eliminate dimensional differences. The sliding window wavelet transform denoising algorithm uses a window length of 30 seconds to perform multi-scale decomposition on the standardized data, and transmits it to the water level analysis module after filtering out high-frequency noise, providing a high-confidence data source for subsequent analysis.

[0119] The water level analysis module receives processed data and models historical water level and temperature data using an ARIMA time series forecasting model to generate forecasts for future water level fluctuations. This model performs a stationarity check and differential processing on the historical data to eliminate seasonal and random noise, outputting upper and lower limits for water level fluctuations in future time periods. Simultaneously, real-time slag-water mixing data is compared with preset viscosity thresholds. The water level trend and viscosity deviation are combined to generate a water replenishment demand level, which is divided into multiple quantitative intervals to indicate the urgency of water replenishment and the direction of flow adjustment. For example, if the viscosity value exceeds the threshold and the water level forecast indicates a downward trend, a high-level water replenishment demand signal is output.

[0120] The water replenishment strategy generation module calls the water replenishment control rule library based on the water replenishment demand level, and generates a multi-loop water replenishment strategy by combining the time window constraint rules and the historical response efficiency of the loop. The priority allocation rule dynamically calculates the weight based on the historical response time of the loop and the flow control accuracy, and the high response efficiency loop is activated first. The flow allocation rule combines the predicted water level change rate to give priority to high flow loops when the viscosity exceeds the standard, and dynamically adjusts the water replenishment time window. For example, when a sudden drop in water level is predicted, the water replenishment start time is advanced by 20%, and the duration is extended to the end of the predicted interval to avoid water replenishment lag. The generated strategy includes the start-up and closing sequence, flow parameters and priority tags, which are transmitted to the multi-loop control module through the communication coordination module.

[0121] The multi-loop control module uses a conflict resolution algorithm to parse the instruction queue and dynamically adjust the execution order based on the loop load status and execution efficiency score. The execution efficiency score is calculated by combining the historical response time average and the flow deviation rate of the last three times with a weight of 6:4. The scores are arranged in descending order to generate the initial queue. When it is detected that the time slice overlap rate exceeds the threshold and the loop load reaches 80%, the dynamic time slice segmentation mechanism is activated to compress the high-priority instruction time window to 70% of the original length and insert a buffer period for low-priority instructions. The adjusted instruction queue is distributed to each loop through the time division multiplexing protocol. The flow sensor monitors the execution effect in real time. The feedback compensation unit calculates the compensation coefficient based on the water level deviation and dynamically corrects the flow parameters.

[0122] The dynamic adjustment module continuously receives feedback data on water level changes. When the real-time water level deviates from the median of the predicted interval by more than 5%, the compensation coefficient is calculated using the proportional-integral algorithm. The weight redistribution unit adjusts the flow distribution parameters proportionally based on the historical efficiency weight of the loop. For example, when the compensation coefficient is positive, the flow increase of the high-weight loop is increased by 15%. The updated parameters are synchronized to the control module through the communication coordination module, and the conflict resolution unit regenerates the instruction queue to form a closed-loop control logic. The exception handling module monitors the loop flow and valve response time. When the flow of a certain loop is lower than the preset lower limit for three consecutive times, the fault state is marked and the incremental instruction is triggered. The remaining loops share 120% of the flow according to priority. If the fault lasts for 5 minutes, the backup loop is activated and the weight is recalculated based on the available capacity. The initial weight of the backup loop is set to 20% of the total available capacity, and the rest is distributed according to the response efficiency.

[0123] The energy consumption optimization module counts the cumulative working hours and flow deviation rate of the loop and generates an energy efficiency evaluation report. Loops with energy efficiency ratios higher than the average and deviation rates lower than 15% are marked as high-priority activation objects, and loops with energy efficiency drops of more than 10% for three consecutive cycles are downgraded. The self-learning module optimizes the rule base threshold through the Q learning algorithm based on historical water replenishment records and parameter correction data. The reward function iteratively updates the parameters according to the weights of 50% water level stability, 25% energy efficiency ratio, and 25% failure rate. The optimized threshold is embedded in the strategy generation module to generate a priority sequence adapted to the current load state, suppressing the energy efficiency degradation caused by equipment aging. The above-mentioned collaborative mechanism achieves a dynamic balance between pH value, suspended solids concentration and liquid level, effectively reducing the risk of pipeline hardening and improving system operation stability and environmental protection indicators.

[0124] The technical features of the present invention are explained as follows: Data acquisition module: A multimodal sensor group consisting of an ultrasonic water level meter, an infrared temperature sensor and a viscosity detector is responsible for collecting real-time data on the liquid level, temperature and slag-water mixture viscosity in the slag remover.

[0125] Ultrasonic water level meter: The liquid level is calculated based on the time difference of sound wave reflection. The sampling frequency is set to 10Hz to capture dynamic water level changes.

[0126] Viscosity tester: It uses the principle of rotary torque measurement to reversely calculate the fluid viscosity value through the resistance torque of the drive shaft in the slag water.

[0127] Data fusion unit: performs timestamp alignment, dimension normalization (liquid level percentage, temperature standardization, viscosity dimensionless) and wavelet threshold denoising on multi-source heterogeneous data to form cleaned data in a unified format.

[0128] Water level analysis module: used to build water level change trend model and quantitative assessment of water replenishment demand: ARIMA time series prediction model: By performing differential processing on historical water level data to eliminate non-stationarity, an autoregressive (AR), differencing (I), and moving average (MA) combined model is established to output the water level fluctuation range in the future period (such as [1.2m, 1.5m]).

[0129] Viscosity threshold comparison: A preset viscosity threshold range (e.g., 30-50 units) is set. When the real-time viscosity data exceeds the threshold, a viscosity abnormality flag is triggered. Combined with the water level forecast results, a water replenishment demand level (e.g., emergency, high, medium, and low) is generated.

[0130] Water replenishment strategy generation module: A decision engine based on a rule base that generates multi-loop collaborative control strategies: Priority allocation rule: Dynamic weights are assigned based on the historical response efficiency of the circuit (response delay ≤ 2 seconds, flow deviation rate ≤ 5% is a high-efficiency circuit), and high-efficiency circuits are activated first.

[0131] Flow allocation rules: Allocate flow according to the viscosity abnormality level. For example, when the viscosity exceeds the standard, a high flow circuit (≥50m³ / h) is activated, and the water replenishment time window is dynamically adjusted based on the predicted water level change rate (for example, water replenishment is triggered 10% in advance).

[0132] Time window constraint: Set the start time and duration of the water replenishment action based on the slope change of the water level prediction interval (such as the water level drop rate ≥ 0.1m / h).

[0133] Multi-loop control module: realizes instruction optimization scheduling and resource competition elimination: Conflict resolution algorithm: Generates a priority queue based on the execution efficiency score (historical response time weighted 60% + traffic deviation rate weighted 40%). When the loop load is greater than 80%, dynamic time slice segmentation is initiated to compress the time window of high-priority instructions (for example, 70% of the original length).

[0134] Time Division Multiplexing protocol: Allocates instructions to discrete time slices, dynamically expands the time slice length (≤150% of the original length) based on traffic fluctuation rate (standard deviation >10%), and compresses in 15% steps when the load exceeds the limit.

[0135] Dynamic adjustment module: Closed-loop feedback control core components: Proportional-integral compensation algorithm: Based on the deviation between the real-time water level and the predicted median (e.g., deviation > 5%), a compensation coefficient (proportional term Kp=0.8, integral term Ki=0.2) is calculated. A positive coefficient triggers a flow increase, while a negative coefficient reduces the flow.

[0136] Weight redistribution: Adjust the flow parameters according to the historical efficiency weight of the loop (high-efficiency loop weight + 20%). The updated parameters are synchronized in real time through the communication coordination module.

[0137] Exception handling module: Fault tolerance and resource reorganization mechanism: Flow lower limit threshold: Dynamically calculate the historical minimum effective flow value (such as ≥8m³ / h). If the sampling is lower than the threshold for three consecutive times, it is judged as a fault.

[0138] Redundant circuit activation: After a fault persists for 5 minutes, the weight is redistributed based on available capacity (theoretical traffic minus real-time traffic). The initial weight of the backup circuit is set to 20% of the total capacity, and the remaining weight is distributed based on response efficiency.

[0139] Self-learning modules: Parameter optimization system based on reinforcement learning: Q-learning algorithm: The state space is defined as the water replenishment demand level, circuit load, and environmental parameters. The action space is the threshold adjustment direction (±5% step size). The reward function is calculated based on water level stability (50%), energy efficiency (25%), and failure rate (25%).

[0140] Iterative threshold optimization: Dynamically updates rule base parameters based on historical policy execution results (e.g., a 30% extension of the fault interval) to generate a priority allocation policy that adapts to the aging status of devices.

[0141] The modules of the present invention achieve data synchronization through the time-division multiplexing protocol of the communication coordination module, forming a closed-loop control chain of "data acquisition → trend prediction → strategy generation → instruction scheduling → dynamic adjustment → exception handling → self-learning optimization." For example, when viscosity detection is abnormal, the water level analysis module increases the water replenishment demand level, the strategy generation module prioritizes high-flow circuits, the control module allocates resources using a conflict resolution algorithm, the dynamic adjustment module corrects flow parameters based on real-time feedback, the exception handling module switches to a backup circuit in the event of a fault, and the self-learning module continuously optimizes the rule base threshold. This collaborative mechanism effectively solves the problem of water replenishment mismatch in a multivariable coupled environment and has clear technical feasibility.

[0142] ARIMA time series forecasting model (water level analysis module): Principle and application: The ARIMA (Autoregressive Integrated Moving Average) model consists of three parts: autoregression (AR), differencing (I) and moving average (MA), and is used to predict the dynamic change trend of water levels.

[0143] Autoregressive (AR): uses linear combinations of historical water level data to predict future values, capturing short-term dependencies in water level changes.

[0144] Difference (I): Perform difference processing on non-stationary water level data to eliminate seasonal or trend fluctuations and make the data series stable.

[0145] Moving Average (MA): A weighted average of historical forecast errors is used to modify the forecast value and suppress random noise interference.

[0146] Technical Implementation: The model inputs historical water level data and current temperature parameters, determines the differencing order (e.g., d=1) through a stationarity test, constructs the AR (p) and MA (q) parameter matrices, and outputs the future water level fluctuation range (e.g., the water level fluctuation range for the next hour is ±0.3m). This prediction provides a quantitative basis for planning the time window for water replenishment strategies, such as triggering water replenishment actions in advance if the water level is declining.

[0147] Conflict resolution algorithm (multi-loop control module): Principle and Application: Used to solve the resource competition problem in multi-loop parallel control, and generate the optimal instruction queue based on priority tags and execution efficiency scores.

[0148] Priority tag: A dynamic score is generated by combining the water replenishment demand level (urgent / high / medium / low), the circuit's historical response efficiency (response time ≤ 2 seconds is considered high efficiency), and the current load status (load rate ≤ 80%).

[0149] Execution efficiency score: Calculated based on the historical response time average (weight 60%) and the traffic deviation rate of the last three times (weight 40%). A higher score indicates better loop execution efficiency.

[0150] Technical implementation: When it is detected that the time slice overlap rate is greater than 30% and the loop load is greater than 80%, the algorithm starts the dynamic time slice segmentation mechanism: High-priority instructions: Compress the execution time window to 70% of the original length (for example, from 10 seconds to 7 seconds) to ensure fast response.

[0151] Low-priority instructions: insert a buffer time slice (such as 3 seconds), and the buffer time is dynamically adjusted according to the loop load fluctuation rate.

[0152] Q learning algorithm (self-learning module): Principle and application: Based on the reinforcement learning parameter optimization framework, the water replenishment strategy is adapted to equipment aging and environmental interference by iteratively updating the rule base threshold.

[0153] State space: defined as a multidimensional vector of water replenishment demand level, circuit load rate, ambient temperature and viscosity parameters.

[0154] Action space: includes the direction (e.g., up / down) and amplitude (e.g., ±5% step length) of threshold adjustment.

[0155] Reward function: The reward value is calculated based on water level stability (weight 50%), circuit energy efficiency ratio (25%), and failure rate reduction ratio (25%) to drive the model optimization direction.

[0156] Technical Implementation: The model iteratively updates the Q-value table based on historical water replenishment records and dynamic adjustment data, optimizing the threshold parameters in the water replenishment control rule base (for example, adjusting the lower flow threshold from 8 m³ / h to 7.5 m³ / h). These optimized thresholds are embedded in the strategy generation module to generate a priority allocation strategy tailored to the current operating conditions. For example, in scenarios with aging equipment, the load weight threshold for high-efficiency circuits can be reduced.

[0157] Wavelet transform denoising algorithm (data fusion unit): Principle and application: Used to filter out high-frequency noise in multimodal sensor data and retain low-frequency effective signals that characterize the state of slag water.

[0158] Wavelet decomposition: decompose the data into approximate coefficients (low frequency) and detail coefficients (high frequency) according to different frequency scales.

[0159] Threshold filtering: Perform soft threshold processing on high-frequency detail coefficients to suppress random noise (such as electromagnetic interference and mechanical vibration noise).

[0160] Technical Implementation: A sliding window (30-second window length) is used to process data segment by segment. After decomposing the data within the window using a three-layer wavelet decomposition, a threshold (e.g., σ = 1.5 times the standard deviation) is applied to the detail coefficients to reconstruct the denoised signal. The processed data is transmitted to the water level analysis module to ensure the accuracy of trend forecasts.

[0161] Proportional-integral compensation algorithm (dynamic adjustment module): Principle and application: The core algorithm of closed-loop feedback control is used to correct the water replenishment flow distribution parameters.

[0162] Proportional term (P): quickly responds to real-time water level deviations (e.g., if the current water level is 0.2m lower than the predicted value, the proportional coefficient Kp=0.8 triggers a flow increase).

[0163] Integral term (I): cumulative historical deviation trend (e.g., if there are five consecutive negative water level deviations, the integral coefficient Ki = 0.2 triggers cumulative compensation).

[0164] Technical implementation: The compensation coefficient calculation formula is: ; Where e(t) is the real-time water level deviation. For example, when e(t) = +5, the compensation coefficient triggers a 15% increase in the high-weight circuit flow.

[0165] Each model of the present invention realizes data interaction through the time division multiplexing protocol of the communication coordination module to form a closed-loop control chain: Data layer: Wavelet denoising algorithm cleans sensor data, and ARIMA model predicts water level trends.

[0166] Decision layer: The Q learning algorithm optimizes the rule base threshold, and the conflict resolution algorithm generates the instruction queue.

[0167] Execution layer: The proportional-integral algorithm dynamically corrects flow parameters, and the exception handling module ensures system fault tolerance. For example, when the viscosity test value exceeds the standard, the ARIMA model predicts the water level decline trend, the Q-learning algorithm adjusts the lower flow threshold, the conflict resolution algorithm prioritizes high-flow circuits, and the proportional-integral algorithm compensates flow based on real-time deviations, ultimately achieving precise water replenishment control in a multivariable coupled environment.

[0168] The technical synergy of the above models solves the system blockage problem caused by the mismatch of pH value, suspended solids concentration and liquid level parameters in traditional single-loop control, significantly improving the environmental compliance and operating efficiency of the slag water system of coal-fired power plants.

[0169] The technical solution of the present invention solves the environmental adaptability problem of traditional single-loop water replenishment control by constructing a multi-dimensional parameter collaborative analysis framework and a dynamic compensation mechanism. First, the data acquisition module integrates a multi-modal sensor group to synchronously acquire liquid level, pH value and suspended solids concentration data, normalizes the multi-source heterogeneous data through the data fusion unit, and removes abnormal noise based on the sliding window wavelet transform denoising algorithm to form a high-confidence fusion data set. The water level analysis module uses the ARIMA time series prediction model to model the dynamic change trend of the water level, combines the slag-water mixed viscosity threshold comparison results, generates the water replenishment demand level, associates the liquid level fluctuation with the change of water quality parameters, and realizes the real-time perception of the multi-variable coupling state.

[0170] Secondly, the water replenishment strategy generation module calls the water replenishment control rule library based on the water replenishment demand level to generate a multi-loop water replenishment strategy that includes the start-up and closing sequence, flow distribution parameters, and priority. The multi-loop control module parses the instruction queue using a conflict resolution algorithm, dynamically adjusts the execution order based on the loop load status, and allocates time slices in conjunction with the communication coordination module's time-division multiplexing protocol to ensure parallel execution of multi-loop water replenishment actions and eliminate resource competition. The dynamic adjustment module adjusts the flow distribution parameters in real time based on water level feedback data and optimizes priority by calculating compensation coefficients using a proportional-integral algorithm. This forms a closed-loop control logic and reduces the risk of delayed water replenishment or overshoot.

[0171] Finally, the exception handling module monitors the circuit status through flow sensors. When a flow anomaly or valve response timeout is detected, it triggers the backup circuit switching command and redistributes the water replenishment weight. Combined with the redundant circuit activation mechanism, it ensures system continuity. The energy consumption optimization module compiles circuit energy efficiency data to generate an evaluation report and adjusts the activation sequence of subsequent circuits. The self-learning module optimizes the rule base parameters based on the Q-learning algorithm, allowing the water replenishment strategy to adapt to environmental interference and equipment aging. Through this multi-level collaborative mechanism, the system achieves a dynamic balance between pH value, suspended solids concentration, and liquid level parameters, suppressing the formation of aggregates and the risk of pipe blockage, and improving environmental protection indicators and operational stability.

Claims

1. A slag scooping machine system based on multi-circuit automatic water replenishment control, characterized in that: include: The data acquisition module is configured to collect real-time water level data, temperature data and slag-water mixing state data in the slag scoop; A water level analysis module is used to generate a dynamic change trend of the water level through a time series prediction model based on the water level data and temperature data, and output a water replenishment demand level based on a comparison result of the slag-water mixing state data with a preset viscosity threshold; a water replenishment strategy generation module configured to, based on the water replenishment demand level, invoke priority allocation rules and flow allocation rules in a preset water replenishment control rule library to generate a multi-circuit water replenishment strategy, wherein the multi-circuit water replenishment strategy includes an opening and closing sequence of at least two independent water replenishment circuits, water replenishment flow allocation parameters, and a water replenishment priority based on the circuit's historical response efficiency; a multi-circuit control module, configured to generate a control instruction queue based on the water replenishment flow distribution parameters and the water replenishment priority through a conflict resolution algorithm, and send the control instruction queue to multiple water replenishment execution modules to drive valve opening adjustment of different water replenishment circuits; a dynamic adjustment module, configured to modify the water replenishment flow rate distribution parameters generated by the water replenishment strategy generation module according to the water level change feedback data collected in real time by the data acquisition module, and update the water replenishment priority based on the modified parameters; an exception handling module configured to trigger a backup circuit switching instruction and send a request to the water replenishment strategy generation module to reallocate the water replenishment weights of the remaining circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is lower than a preset threshold or the response time of the electric valve exceeds a preset time; The communication coordination module is used to synchronize the data interaction timing between modules through the time division multiplexing protocol, and coordinate the parallel execution of multi-circuit water replenishment actions based on a unified clock signal.

2. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 1 is characterized in that: The water level analysis module also includes: The water level prediction unit is used to output the future water level change range through the ARIMA time series prediction model based on historical water level data and current temperature data; The water replenishment strategy generation module calls the time window constraint rules in the water replenishment control rule library according to the future water level change interval, adjusts the water replenishment priority, and generates a multi-loop water replenishment strategy including a water replenishment action time window.

3. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 2 is characterized in that: The dynamic adjustment module includes: a feedback compensation unit configured to calculate a water replenishment flow compensation coefficient by a proportional and integral algorithm based on a deviation between real-time water level data and a median of a future water level change interval output by the water level prediction unit after the water replenishment execution module is started; The weight redistribution unit is used to adjust the flow distribution parameters according to the historical response efficiency weights of each water replenishment circuit based on the compensation coefficient, and synchronize the adjusted parameters to the multi-circuit control module through the communication coordination module.

4. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 3 is characterized in that: The exception handling module also includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

5. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 1 is characterized in that: The exception handling module also includes: a fault circuit isolation unit configured to mark the circuit as a fault state and trigger the communication coordination module to send a flow increment instruction to other water replenishment circuits when the flow sensor of the water replenishment execution module detects that the current circuit flow is continuously lower than the flow lower limit threshold preset in the water replenishment control rule library; The redundant circuit activation unit is used to activate the backup water replenishment circuit after the fault state lasts for more than a preset time, and recalculate the water replenishment weight according to the available capacity of the current remaining circuit based on the weight allocation rule in the water replenishment strategy generation module.

6. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 5 is characterized in that: The multi-loop control module further includes: a conflict resolution unit configured to dynamically adjust the execution order of instructions according to the water replenishment priority defined by the water replenishment strategy generation module and the current circuit load status of the water replenishment execution module when there is resource competition among the control instructions of multiple water replenishment circuits; The conflict resolution unit generates an optimal instruction queue by comparing the priority tags of each loop and the execution efficiency score calculated based on the historical response time and the flow deviation value.

7. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 1 is characterized in that: The communication coordination module uses a time-division multiplexing protocol to allocate water replenishment instructions to different time slices, and dynamically adjusts the time slice length based on the real-time flow fluctuation rate and loop load status fed back by the flow sensor of the water replenishment execution module to match the flow allocation parameters of the water replenishment strategy generation module.

8. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 7 is characterized in that: Also includes: The energy consumption optimization module is connected to the communication coordination module and is used to count the cumulative working hours and flow data of each water replenishment circuit, generate an energy consumption evaluation report including the circuit energy efficiency ratio and flow deviation rate, and call the energy consumption optimization rules in the water replenishment control rule library based on the report to adjust the circuit activation order in the subsequent water replenishment strategy.

9. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 1 is characterized in that: The data acquisition module includes: A multimodal sensor set configured to simultaneously collect liquid level, temperature, and viscosity data through an ultrasonic water level gauge, an infrared temperature sensor, and a viscosity detector; The data fusion unit is used to normalize the multi-source heterogeneous data collected by the multimodal sensor group, and remove abnormal noise data based on the wavelet transform denoising algorithm of the sliding window before transmitting it to the water level analysis module.

10. The slag scooping machine system based on multi-circuit automatic water replenishment control according to claim 9, characterized in that: Also includes: A self-learning module is connected to the water replenishment strategy generation module and the dynamic adjustment module, and is used to optimize the parameter thresholds in the water replenishment control rule base through a Q learning algorithm based on the historical water replenishment records of the water replenishment strategy generation module and the flow parameter correction results of the dynamic adjustment module, and generate a priority allocation strategy adapted to the current loop load state.

Images