Intelligent temperature control system and method for circulating fluidized bed boiler

Abstract

The invention relates to the technical field of circulating fluidized bed boilers, in particular to an intelligent temperature control system and method for a circulating fluidized bed boiler, accurate regulation and control of boiler operation parameters are achieved through a multi-algorithm cooperative control and intelligent compensation mechanism, a model prediction control unit is based on a multivariable control matrix constructed by an ARX / FIR model, and the temperature control accuracy of the circulating fluidized bed boiler is improved. Dynamic tracks of boiler bed temperature, pressure and oxygen parameters can be predicted in real time, a multivariable coupling problem which is difficult to process in traditional PID control is converted into an optimization solution problem, control precision is improved, a neural network optimization control unit dynamically adjusts the coal feed quantity and air quantity ratio through a reinforcement learning mechanism, boiler thermal efficiency is improved, and the control precision is improved. The fuzzy control unit converts an operator experience rule into a compensation control quantity under a non-linear boundary condition, so that bed temperature overshoot and oxygen abrupt change are effectively inhibited, and the stability of the combustion process can still be kept under a variable working condition.

Description

Technical Field

[0001] The present invention relates to the technical field of circulating fluidized bed boilers, and more specifically, to an intelligent temperature control system and method for a circulating fluidized bed boiler. Background Technique

[0002] Circulating fluidized bed boilers are efficient and low-pollution coal-fired boilers, which have the advantages of wide fuel adaptability, high combustion efficiency, and low pollutant emissions, and are widely used in fields such as electric power, chemical industry, and heating. However, the operation process of circulating fluidized bed boilers is complex and affected by various factors, such as coal type, load, air volume, etc., resulting in unstable boiler operation, large bed temperature fluctuations, and affecting the safety and economy of the boiler.

[0003] As an efficient and clean coal-fired device, the precise control of the operating parameters of circulating fluidized bed boilers directly affects combustion efficiency, pollutant emissions, and operating safety. Existing temperature control systems mostly rely on single control algorithms, such as traditional PID control (the most basic feedback control algorithm in the field of automatic control) or basic fuzzy control, which have problems such as lag in dynamic response, insufficient multi-variable coupling compensation, and poor working condition adaptability. For example, there is a strong non-linear correlation between bed temperature, pressure, and oxygen content parameters, and it is difficult for a single algorithm to decouple and optimize the control quantity in real time; at the same time, traditional systems lack an effective processing mechanism for sensor noise, data redundancy, and transmission delay, resulting in a decrease in prediction accuracy and untimely warning response. In addition, there is no safety verification mechanism during manual intervention by operators, and the adjustment of control strategies depends on offline configuration, making it difficult to meet the real-time optimization requirements of complex working conditions.

[0004] Therefore, developing an intelligent temperature control system for circulating fluidized bed boilers that integrates multiple intelligent control algorithms, has an adaptive learning mechanism, and can achieve multi-modal intelligent control is of great significance for improving the operation stability, safety, and economy of boilers. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent temperature control system and method for a circulating fluidized bed boiler to solve the problems raised in the above background technique.

[0006] To achieve the above purpose, the present invention provides the following technical solutions: including a data acquisition module, an intelligent control module, an execution module, a monitoring and warning module, and a human-computer interaction module: The data acquisition module is used to collect the boiler operation parameters in real time and perform standardized processing, and synchronously transmit the standardized data to the intelligent control module through a redundant communication bus; The intelligent control module consists of a model predictive control unit, a fuzzy control unit, and a neural network optimization control unit. It is used to receive the standardized data transmitted by the data acquisition module in real time, integrate three control algorithms of model predictive control, fuzzy control, and neural network optimization for dynamic calculation to generate the final control instruction, synchronously transmit it to the execution module, and at the same time generate prediction data and transmit it to the monitoring and warning module; The execution module includes a coal feeder, a primary air fan, a secondary air fan, and valves. It receives the final control instruction of the intelligent control module to execute the equipment action, feeds back the execution status data to the data acquisition module, and uploads the action parameters to the monitoring and warning module at the same time; The monitoring and warning module processes the boiler operation parameters of the data acquisition module, the prediction data of the intelligent control module, and the action parameters of the execution module in parallel. It triggers a three-level warning mechanism to generate warning information through pattern matching. The warning information is synchronously pushed to the human-computer interaction module and triggers the protection logic of the intelligent control module; The human-computer interaction module uses three-dimensional visualization technology to display the boiler operation status, the final control instruction parameters, and the warning information in real time. It integrates a two-way control channel and supports the operator to manually correct the final control instruction parameters. The corrected instruction is directly transmitted to the intelligent control module after permission verification.

[0007] A further technical solution of this application: The boiler operation parameters include boiler bed temperature, pressure, and oxygen content parameters; The standardized data contains three fields: timestamp, sensor ID, and original measurement value. The intelligent control module automatically matches the corresponding multi-variable control matrix column vector by parsing the sensor ID in the field; The model predictive control unit constructs a multi-variable control matrix based on the ARX model and the FIR model, predicts the future trajectory of the boiler operation parameters in real time, and generates a multi-parameter coupled reference control quantity including the coal feeding speed reference and the damper opening reference; The fuzzy control unit uses the T-S fuzzy model to convert the operator's experience rules into a compensation control quantity under non-linear boundary conditions, dynamically adjusts the compensation weight by monitoring the bed temperature overshoot and the oxygen content mutation rate, and generates a fusion control instruction after fusing with the multi-parameter coupled reference control quantity; The neural network optimization control unit adopts a parallel fast learning network structure, dynamically adjusts the coal feeding amount and air volume ratio setting value through a reinforcement learning mechanism, and generates an optimization control instruction; The fusion control instruction and the optimization control instruction generate the final control instruction after dynamic calculation.

[0008] A further technical solution of this application: The final control instruction generated by the intelligent control module is encapsulated in JSON format, including four-dimensional information of device address code, action type, target value, and priority. The execution module triggers the control program of the corresponding device by parsing the device address code.

[0009] A further technical solution of this application: When the execution status data fed back by the execution module is transmitted through the redundant communication bus of the data acquisition module, a digital signature and a timestamp are synchronously appended. The monitoring and early warning module calculates the transmission delay compensation amount by comparing the difference between the timestamp and the local clock.

[0010] A further technical solution of this application: The three-level early warning mechanism includes: The early warning information output by the monitoring and early warning module is transmitted through a hierarchical queue. The red early warning directly triggers the audible and visual alarm interface of the human-computer interaction module. The orange early warning is pushed to the operation log pending processing queue. The yellow early warning is displayed as a floating prompt on the interface.

[0011] A further technical solution of this application: The correction instruction contains the operator identity authentication token and the parameter change track. After the intelligent control module verifies the validity of the token, the parameter change track is stored in the control strategy version library.

[0012] A further technical solution of this application: The prediction data generated by the intelligent control module includes the future tracks of the boiler bed temperature, pressure, and oxygen content parameters predicted in real time. The prediction data is exchanged with the monitoring and early warning module in real time in a shared memory manner to avoid repeated transmission of data on the redundant communication bus.

[0013] A further technical solution of this application: The neural network optimization control unit is also provided with a training data buffer for receiving the original boiler operation parameter data transmitted by the data acquisition module. When the data volume in the training data buffer reaches the preset threshold, the offline retraining process of the neural network optimization control is automatically triggered.

[0014] An intelligent temperature control method for an intelligent temperature control system of a circulating fluidized bed boiler, the method comprising the following steps: S1. The data acquisition module obtains the boiler operation parameters in real time and generates standardized data; S2. The intelligent control module receives the standardized data and generates a final control instruction through multi-algorithm fusion calculation; S3. The execution module adjusts the device action according to the final control instruction and feeds back the execution status data; S4. The monitoring and early warning module triggers a three-level early warning mechanism through pattern matching to generate early warning information and synchronously pushes the early warning information to the human-computer interaction module; S5. The human-computer interaction module supports manual correction of the parameters of the final control instruction, and the correction instruction is transmitted to the intelligent control module after passing the permission verification.

[0015] A further technical solution of the present application: The multi-algorithm fusion calculation in step S2 is specifically three control algorithms of model predictive control, fuzzy control and neural network optimization, including a dynamic belief assignment mechanism, which evaluates the output credibility of each control algorithm in real time through the entropy weight method. Among them, the weight coefficient of model predictive control has a negative correlation with the working condition change rate; The execution status data in step S3 includes digital signatures and timestamps. The monitoring and warning module calculates the transmission delay compensation amount by comparing the difference between the timestamp and the local clock; The warning information in step S4 adopts a hierarchical queue transmission mechanism. The red warning directly triggers the audible and visual alarm interface of the human-computer interaction module. The orange warning is pushed to the operation log pending processing queue, and the yellow warning is displayed as a floating prompt on the interface.

[0016] The present invention realizes precise regulation of boiler operation parameters through multi-algorithm collaborative control and an intelligent compensation mechanism. The model predictive control unit based on the multi-variable control matrix constructed by the ARX / FIR model can predict the dynamic trajectories of boiler bed temperature, pressure and oxygen content parameters in real time, convert the multi-variable coupling problem difficult to handle by traditional PID control into an optimization problem, and improve the control accuracy. The neural network optimization control unit dynamically adjusts the ratio of coal feeding amount to air volume through the reinforcement learning mechanism, improving the boiler thermal efficiency. The fuzzy control unit converts the operator's experience rules into compensation control quantities under non-linear boundary conditions, effectively suppressing the overshoot of bed temperature and the sudden change of oxygen content, and maintaining the stability of the combustion process under variable working conditions.

[0017] The present invention constructs a full-link closed-loop monitoring from data collection to execution feedback through a three-level safety prevention and control system. The data collection module ensures data reliability through redundant communication buses and digital signature technologies. The three-level instruction fusion mechanism of the intelligent control module improves control robustness. The innovative hierarchical queue transmission mechanism of the monitoring and warning module shortens the red warning response time, and cooperates with the audible and visual alarm interface to achieve immediate intervention in dangerous working conditions. In addition, the human-computer interaction module based on three-dimensional visualization supports the rapid loading of fault plans, and combines the permission verification mechanism to ensure the safety of manual intervention, reducing the fault recovery time. The built-in energy-saving mode in the system automatically reduces the fan speed during partial load operation, and cooperates with intelligent control algorithms to optimize the energy efficiency under all working conditions. Description of the Drawings

[0018] Figure 1 It is a flow chart of the present invention. Detailed Embodiment

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention. The present invention will be further described below in conjunction with the embodiments.

[0020] Please refer to Figure 1 , in an embodiment of the present application, an intelligent temperature control system for a circulating fluidized bed boiler includes a data acquisition module, an intelligent control module, an execution module, a monitoring and warning module, and a human-computer interaction module. The data acquisition module includes a bed temperature sensor, a pressure sensor, a flowmeter, and an oxygen concentration sensor, which collect the boiler operation parameters in real time and synchronously transmit the standardized data to the intelligent control module through a redundant communication bus. The standardized data includes multi-source heterogeneous parameters with time stamp alignment. The specific implementation of the data acquisition module realizes the full-dimensional perception of the boiler operation parameters through the integration of multiple types of sensors. The bed temperature sensor is vertically inserted into the middle of the dense phase zone using a sheathed thermocouple, and a supporting installation bracket and a compressed air purging device are installed to prevent ash and slag blockage. The pressure sensor is horizontally installed in the inlet flue of the cyclone separator, and a supporting condensate tank is installed to reduce the medium temperature. The flowmeter is installed in the primary air main pipe, and a supporting rectifier is installed to ensure a stable flow field. The oxygen concentration sensor is obliquely inserted into the tail flue, and an electric heating device and a compressed air back-purging system are installed to keep the probe clean. The data acquisition card is configured with a synchronous sampling mode, and a built-in digital filter is used to suppress high-frequency noise. The signal conditioning circuit uses an isolation amplifier and a low-pass filter to improve the signal quality. The communication bus uses a dual-redundant network, follows the standard protocol to encapsulate the time stamp, sensor ID, and raw data, and ensures that the time deviation is less than 1 ms through a hardware clock synchronization module. The redundant design uses a triple configuration at the sensor level, a dual-bus architecture at the communication level, and a hot standby module at the power supply level to ensure the high reliability and real-time performance of data acquisition. The standardized data is transmitted to the intelligent control module through the redundant communication bus, providing a complete and accurate input basis for the system.

[0021] The intelligent control module is composed of a model predictive control unit, a fuzzy control unit, and a neural network optimization control unit. The model predictive control unit constructs a multi-variable control matrix based on the ARX model (AutoRegressive with eXogenous inputs) and the FIR model (Finite Impulse Response), and predicts the future trajectory of the boiler operation parameters in real time, generating a multi-parameter coupled reference control quantity including the coal feeding speed reference and the damper opening reference. The fuzzy control unit adopts an improved T-S fuzzy model (Takagi-Sugeno fuzzy model), converts the operator's experience rules into compensation control quantities under non-linear boundary conditions, dynamically adjusts the compensation weight by monitoring the overshoot of the bed temperature and the mutation rate of oxygen content, and generates a fused control instruction after fusing with the reference control quantity; The neural network optimized control unit adopts a parallel fast learning network structure, and dynamically adjusts the set value of the coal feeding amount and air volume ratio through a reinforcement learning mechanism to generate an optimized control instruction.

[0022] Among them, the specific implementation method of the intelligent control module adopts a multi-algorithm collaborative control architecture, integrating three control algorithms: model predictive control, fuzzy control, and neural network optimization. The model predictive control unit constructs a multi-variable control matrix based on the ARX model and FIR model, online updates the model parameters through the recursive least squares method, and real-time predicts the future dynamic trajectories of the boiler bed temperature, pressure, and oxygen content. The prediction time domain is set to 15 steps, the control time domain is 8 steps, and the rolling time domain optimization algorithm is used to solve the optimal control increment with constraints, generating a multi-parameter coupled control quantity of the coal feeding speed reference and the damper opening reference to solve the optimization control problem under variable coupling and constraint conditions.

[0023] The fuzzy control unit adopts a T-S fuzzy model, which converts the operator's experience rules into compensation control quantities under non-linear boundary conditions. The input variables are selected as the bed temperature deviation and the oxygen content change rate, the output is the compensation control quantity, the membership function adopts a Gaussian distribution, and the compensation weight is dynamically adjusted by monitoring the real-time overshoot of the bed temperature and the mutation rate of oxygen content. The weight function is designed to be positively correlated with the deviation amplitude. The compensation control quantity is non-linearly fused and then superimposed with the reference control quantity to generate a fused control instruction, enhancing the robustness of the system under variable working conditions.

[0024] The neural network optimized control unit adopts a parallel fast learning network structure. The main network is responsible for real-time optimization of set value generation, and the auxiliary network provides historical experience reference. The network consists of 4 layers (8 nodes in the input layer, 16 nodes in each of the two hidden layers, and 4 nodes in the output layer), and the activation function adopts a combination of ReLU and Sigmoid. The online optimization module dynamically adjusts the set value of the coal feeding amount and air volume ratio through a reinforcement learning mechanism. The reward function comprehensively considers the thermal efficiency, NOx emissions, and pressure fluctuations, and the exploration rate decays exponentially to balance exploration and exploitation. The optimized parameters are generated as optimized control instructions after safety verification to achieve global energy efficiency optimization.

[0025] The three-level instruction fusion mechanism realizes multi-instruction collaboration through dynamic weight allocation. The weight of the reference instruction decreases linearly with the bed temperature deviation, the weight of the compensation instruction is positively correlated with the oxygen content change rate, and the weight of the optimized instruction is dynamically adjusted according to the historical optimization effect. The fusion formula adopts a weighted summation algorithm to ensure the smooth transition and real-time performance of the control instruction. The final output is transmitted to the execution module after anti-normalization processing to form a closed-loop control.

[0026] The intelligent control module outputs the final control instruction through a three - level instruction fusion mechanism; The execution module includes a coal feeder, a primary air fan, a secondary air fan and valves. It receives the final control instruction of the intelligent control module to execute the equipment actions, and feeds back the execution status to the data acquisition module. At the same time, it uploads the action parameters to the monitoring and warning module; The monitoring and warning module processes the boiler operation parameters of the data acquisition module, the prediction data of the intelligent control module and the action parameters of the execution module in parallel. It triggers a three - level warning mechanism through pattern matching to generate warning information. The warning information is synchronously pushed to the man - machine interaction module and triggers the protection logic of the control module; The man - machine interaction module uses three - dimensional visualization technology to display the boiler operation status, the final control instruction parameters and the warning information in real time. It integrates a two - way control channel, supports the operator to manually correct the final control instruction parameters, and the corrected instruction is directly transmitted to the intelligent control module after permission verification.

[0027] The specific implementation of the execution module realizes the physical execution of the final control instruction through multi - device cooperation. The coal feeder adopts variable - frequency control to adjust the speed, the primary air fan and the secondary air fan are equipped with moving vane regulating mechanisms, and the valves adopt electric actuators to control the opening. After receiving the final control instruction, the equipment feeds back the actual speed, damper opening and vibration status to the data acquisition module through the Profibus - DP bus, and uploads the action parameters to the monitoring and warning module synchronously. The execution module designs a redundant control loop, which automatically switches to the standby loop when the main loop fails to ensure the action reliability.

[0028] The monitoring and warning module adopts a stream - processing architecture to analyze multi - source data in parallel, and uses the Apache Flink engine to set a sliding window for real - time pattern matching. The warning rule library includes three - level fault characteristics. The red warning matches emergency events such as sudden change of bed temperature and pressure over - limit, directly triggers an audible and visual alarm and pushes it to the man - machine interface; the orange warning corresponds to abnormal conditions such as oxygen content deviation and air volume imbalance, stores them in the pending queue and marks the operation log; the yellow warning identifies warning states such as parameter fluctuation and equipment pre - heating, and reminds the operator in the form of a floating prompt. When a warning is triggered, a protection instruction is synchronously sent to the intelligent control module to activate the equipment safety interlock.

[0029] The human-computer interaction module constructs a three-dimensional visualization interface based on the Unity3D engine and integrates the HoloLens 2 to achieve augmented reality interaction. The main interface renders the three-dimensional model of the boiler in real time. The temperature field is intuitively displayed using gradient coloring. The historical trend curve supports multi-parameter time-axis comparison. The control panel provides sliding adjustment of key parameters. The two-way control channel integrates an access verification module. The operator inputs correction instructions through gestures or voice. After double authentication by fingerprint and dynamic token, the standardized instructions are directly transmitted to the intelligent control module, and the correction trajectory is automatically stored in the audit log for traceability.

[0030] Please refer to Figure 1 , as a preferred embodiment of the present application, the standardized data output by the data acquisition module includes three fields: timestamp, sensor ID, and raw measurement value. The intelligent control module automatically matches the corresponding column vector of the multivariable control matrix by parsing the sensor ID in the field.

[0031] This embodiment is implemented as follows. The standardized data frame output by the data acquisition module is transmitted through the redundant CANopen bus. After receiving it, the intelligent control module first parses the data frame structure. The data frame includes three fields: timestamp (UTC format, accuracy 1ms), sensor ID (2-byte hexadecimal encoding), and raw measurement value (32-bit floating point type). The sensor ID adopts a predefined encoding rule. For example, 0x01 represents the bed temperature sensor, 0x02 represents the pressure sensor, 0x03 represents the flow meter, and 0x04 represents the oxygen concentration sensor, ensuring a one-to-one correspondence with the column vector of the control matrix.

[0032] The model predictive control unit maintains a multivariable control matrix, and the column vectors of the matrix are arranged in the order of sensor ID. When the sensor ID is parsed, it is directly mapped to the matrix column index through the look-up table method. For example, the bed temperature data with ID 0x01 corresponds to the first column of the matrix, and the pressure data with ID 0x02 corresponds to the second column, and so on. After the raw measurement value is subjected to range conversion and standardization processing, it is filled into the current moment row of the corresponding column vector.

[0033] The timestamp is used for data alignment and synchronization compensation. The intelligent control module adopts a sliding time window mechanism (window width 200ms) to update the matrix for the data with the same batch of timestamps. If the timestamp disorder is detected (such as caused by network delay), the current moment value is estimated through the linear interpolation algorithm to ensure the real-time performance of the control matrix.

[0034] For redundant sensor data (such as triple-redundant bed temperature sensors), the raw value is first processed by median filtering during parsing and then mapped to the control matrix. If the sensor ID conflicts or the data is abnormal (beyond the physical range), the data is marked as invalid, triggering a secondary warning of the monitoring and warning module, and temporarily freezing the update of the corresponding matrix column until the fault is restored.

[0035] After the control matrix is constructed, the model predictive control unit performs multi-step prediction by combining the ARX model with the FIR filtering algorithm. The ARX model calculates the future trajectory based on the current matrix column vector, and the FIR filter compensates for the noise of the prediction result. Finally, a multi-parameter coupling control quantity including the coal feeding speed reference and the damper opening reference is generated and transmitted to the instruction fusion mechanism.

[0036] Please refer to Figure 1 , as a preferred embodiment of the present application, the final control instruction generated by the intelligent control module is encapsulated in JSON format, including four-dimensional information of device address code, action type, target value, and priority. The execution module triggers the control program of the corresponding device by parsing the device address code.

[0037] This embodiment is implemented as follows: The final control instruction is encapsulated in JSON format and includes four-dimensional information fields. The device address code uses 4-bit hexadecimal encoding, where the first digit identifies the device type (e.g., 0x1 for coal feeder, 0x2 for primary air fan), and the last three digits are the device serial number. The action type field is defined as an enumeration type, including speed (speed regulation), opening (opening adjustment), position (position control), etc. The target value field sets the physical range according to the device type. For example, the coal feeder speed range is 0 - 1200 r / min, and the damper opening is 0% - 100%. The priority field uses an integer from 1 to 5 for identification, with 1 being the highest priority and 5 being the lowest, and the default value is 3.

[0038] In the generation process, the intelligent control module queries the physical address code through the device mapping table according to the fused control quantity. The action type is automatically determined by the dimension of the control quantity. For example, the coal feeder corresponds to speed regulation, and the damper performs opening adjustment. The target value is converted to a physical unit after amplitude limiting filtering, and the priority is dynamically allocated according to the control mode. The instruction priority is automatically increased in an emergency condition.

[0039] After receiving the instruction through the TCP / IP protocol, the execution module first checks the integrity of the JSON format and the CRC check code. The device address code is parsed using a segmented matching algorithm. The first 4 bits extract the device type, and the last 3 bits match the specific device. The action type triggers the corresponding control program. The speed-type instruction outputs a PWM signal to the frequency converter through a PID controller, and the opening-type instruction drives the position closed-loop control of the electric actuator. The target value is smoothed by linear interpolation to avoid step changes, and the execution cycle is controlled within 50 ms.

[0040] The priority processing mechanism adopts preemptive scheduling. When the priority of a new instruction is higher than that of the currently executing instruction, it immediately interrupts and saves the current state, and then executes the high-priority operation. After the instruction is executed, a status feedback packet is generated, which includes the actual value, execution error, and timestamp. The packet is sent back to the data acquisition module through the redundant bus and synchronously uploaded to the monitoring module to record the operation log.

[0041] In terms of the security mechanism, if the instruction times out and does not respond for more than 200 ms, retransmission is automatically triggered. If three consecutive transmission failures occur, it will be downgraded to the backup control channel. When a device fails, it automatically switches to the safe state (such as the coal feeder stops rotating and the damper is fully open), and a red warning is pushed to the human-machine interface to form a complete closed-loop control process.

[0042] Please refer to Figure 1 , as a preferred embodiment of this application, when the execution status data fed back by the execution module is transmitted through the redundant communication bus of the data acquisition module, a digital signature and a timestamp are synchronously appended. The monitoring and warning module calculates the transmission delay compensation amount by comparing the difference between the timestamp and the local clock.

[0043] This embodiment is implemented as follows: when the execution module generates a status data packet, it first generates a digital signature through the AES-CMAC algorithm. The unique device identifier is selected as the key-associated data, and a hash operation is performed on the status data payload (including the device address, action type, and actual value field) to generate a 128-bit signature value, which is appended to the end of the data packet. The timestamp is generated by a hardware clock synchronized by the PTPv2 protocol, with an accuracy of 1 μs, and the format is a 64-bit Unix timestamp (UTC). The redundant communication bus uses two independent CANopen channels for transmission. The data packets of each channel carry the same signature and timestamp, and the receiving end verifies the transmission integrity by comparing the data consistency of the two channels.

[0044] When the monitoring and warning module receives data, it first extracts the timestamp and calculates the difference with the local PTP clock (synchronization accuracy ±500 ns) to obtain the transmission delay compensation amount Δt. The compensation algorithm uses a moving average filter (window width of 100 cycles) to dynamically adjust the time reference of the prediction model. Digital signature verification uses a pre-stored key library to match the device identifier. If the verification fails, a secondary warning is triggered, and the data packet is discarded. The timestamp Δt of the valid data packet is used to correct the real-time deviation of the prediction model. For example, in the prediction of the bed temperature, it compensates for the state estimation error caused by network delay.

[0045] The exception handling mechanism includes: when the timestamp jump exceeds 1 ms, the local clock calibration process is started, and the data acquisition module is resynchronized through the three-way handshake protocol; if the signature verification fails for three consecutive cycles, the device is marked as an untrusted node, and the backup control channel is switched; when the transmission delay compensation amount exceeds the threshold (such as Δt > 50 ms), a buffer queue is automatically enabled for timing reconstruction to ensure the timing rationality of the control instruction.

[0046] This implementation method ensures data credibility through cryptographic signatures, uses high-precision clock synchronization to achieve nanosecond timestamp alignment, and combines a redundant channel verification mechanism to ensure the integrity, real-time and reliability of state feedback data during transmission, providing an accurate basis for dynamic compensation of the monitoring module.

[0047] See also Figure 1 As a preferred embodiment of the present application: the warning information output by the monitoring and warning module adopts a hierarchical queue transmission mechanism, the red warning directly triggers the sound and light alarm interface of the human-computer interaction module, the orange warning is pushed to the operation log to be processed queue, and the yellow warning is displayed as an interface floating prompt.

[0048] This embodiment is implemented as follows: First, the warning classification and generation logic, the real-time data stream is pattern matched through the Flink stream processing engine, and the rule library predefines three levels of feature thresholds: Red warning: emergency events such as sudden change of bed temperature exceeding 50℃ / s, pressure exceeding ±10% of the rated value, etc., are triggered immediately after matching; Orange warning: abnormal operating conditions such as oxygen content continuously deviating from the set value by ±3%, air volume imbalance exceeding 15%, etc., are generated for three sampling cycles; Yellow warning: Warnings such as parameter fluctuations exceeding the normal range but not reaching the threshold, equipment preheating status, etc. are generated in real time.

[0049] Secondly, the queue transmission mechanism, the red warning queue: uses a memory priority queue (maximum length 100). After the warning is generated, it is directly pushed to the human-computer interaction module through a TCP long connection to trigger the sound and light alarm protocol (including buzzer control instructions and red light flashing frequency encoding).

[0050] Orange warning queue: Use Kafka message queue for temporary storage. The message contains device ID, warning type, timestamp and original data fragment. It is pushed to the operation log interface and marked as pending. Operators can view it by priority.

[0051] Yellow warning queue: Pushed to the front end in real time via WebSocket, the warning information is encapsulated as a JSON object (including device identification, parameter values, and recommended operations), displayed in a floating prompt box, and automatically fades out after 5 seconds.

[0052] Among them, regarding the guarantee of transmission reliability, the red warning adopts a three-way handshake confirmation mechanism. If no confirmation frame is received within 3 seconds, it switches to UDP broadcast mode to ensure that the alarm is not missed; the orange warning message enables persistent storage (retained for 72 hours), and the consumer end confirms the processing status through the ACK mechanism; the yellow warning integrates heartbeat detection, and automatically caches it to the local IndexedDB when the network is interrupted, and resends it after recovery.

[0053] Then comes the linkage control logic. When a red warning is triggered, a protection command (such as emergency coal stop, full-open air damper) is synchronously sent to the intelligent control module, and the command priority is forcibly set to level 1. After the orange warning is activated, the monitoring module automatically retrieves the historical similar case library, generates a list of recommended operations and pushes it to the human-machine interface. If a yellow warning is triggered three times on the same device, it will automatically upgrade to an orange warning and generate an equipment maintenance work order.

[0054] Finally, the interface interaction design is introduced. The human-machine interface integrates the sound and light alarm hardware interface (supports RS485 / Modbus protocol). When a red warning occurs, red light flows around the borders of the screen, and the buzzer emits a rapid pulse sound. The operation log interface adopts the B / S architecture. The orange warning items are displayed with a red border and are pinned to the top. Batch processing operations are supported (confirm / ignore / transfer work orders). The floating prompt is rendered using HTML5 Canvas, supports dragging and dropping, and expands a detailed parameter trend chart after clicking.

[0055] Ultimately, differentiated processing of warning information is achieved through hierarchical queues, and transmission reliability is ensured by combining the stream processing engine and message middleware. The linkage control logic and interface interaction design form a complete closed loop to ensure the timeliness of warning response and convenience of operation.

[0056] See also Figure 1 , as a preferred embodiment of the present application: the correction instruction includes an operator identity authentication token and a parameter change trajectory. After the intelligent control module verifies the validity of the token, it stores the parameter change trajectory in the control strategy version library.

[0057] The specific implementation is as follows: The correction instruction is transmitted through a two-way control channel using the HTTPS protocol, encapsulated in JSON format, and contains the operator authentication token (JWT format), parameter change track (key-value pair array) and digital signature triple fields. The authentication token is generated by the authority management system, including the operator ID, role level (such as administrator / operator) and validity period, and is signed by RSA asymmetric encryption. The digital signature uses the HMAC-SHA256 algorithm and the token key as the secret key to generate the full text of the instruction to ensure that the instruction is tamper-proof.

[0058] After receiving an instruction, the intelligent control module first performs a security check: it parses the JWT token to verify the validity and expiration date of the signature, and queries the pre-stored token blacklist; it compares the digital signature with the locally calculated value. If the verification fails, the instruction is discarded and an audit log is triggered. After passing the verification, it parses the parameter change trajectory. The trajectory fields include the parameter ID, such as the coal feeding PID parameter, the original value, the new value, and the change timestamp, and supports batch modification, such as adjusting the parameters of multiple control loops simultaneously. The control strategy version library adopts a Git-LFS distributed storage architecture. The main library is deployed on redundant servers, and the slave library is synchronously replicated to edge nodes in real time. When a parameter changes, the system automatically generates a new version branch. The version number follows the semantic specification, and the change record is appended to the version log, including the operator ID, the hash value of the change content, and the timestamp. If a parameter conflict is detected, such as the new value exceeding the physical constraint, a conflict resolution mechanism is triggered to prompt the operator to select an overwrite / rollback / cancel operation.

[0059] The storage process includes four-step transactions: 1. Write-Ahead Logging records the change intention; 2. Execute the parameter update to the in-memory control matrix; 3. Persistently store to the main library of the version library; 4. After synchronizing to the slave library, commit the transaction. Version rollback supports quick recovery by specifying the historical version number. The parameter differences are calculated through the diff algorithm and updated in batches.

[0060] The monitoring module listens to the version library change events in real time and generates an operation audit report, including the operator, the changed parameters, the affected scope, and the effective time, and supports multi-dimensional retrieval by time / operator / parameter type. In terms of security mechanisms, modifying critical parameters requires secondary authentication, such as a dynamic verification code. Token permission grading controls the modification depth. For example, an operator can only adjust the set value offset, and an administrator can modify the control algorithm structure.

[0061] Please refer to Figure 1 , as a preferred embodiment of this application: The prediction data output by the intelligent control module is exchanged with the monitoring and early warning module in real time in a shared memory manner, avoiding repeated transmission of data on the redundant communication bus.

[0062] The specific implementation is as follows: A multi-segment shared memory pool architecture is adopted. The main memory area is divided into a circular buffer (with a capacity of 1024 data blocks). Each block contains a timestamp (64 bits), a prediction parameter header (16 bits, identifying bed temperature, pressure, oxygen content, etc.), a floating-point prediction array (with adjustable dimensions, default 128 points), and a checksum (CRC16). An additional control information segment (128 bytes) is allocated to store the positions of the read and write pointers, the version number (32-bit integer, supporting semantic version control), and the synchronization semaphore. After the model predictive control unit generates prediction data, it is directly written into the circular buffer through a memory-mapped file (mmap). Before writing, a mutex lock (pthread_mutex) is applied to ensure atomic operations. The data is stored in little-endian byte order, and the prediction array is optimized and compressed by the SIMD instruction set to reduce memory occupancy. The timestamp is calibrated by the TSC counter + hardware clock, with an accuracy of 0.1 μs. After writing is completed, the write pointer and version number in the control information segment are updated, the mutex lock is released, and a condition variable (pthread_cond_signal) is triggered to notify the monitoring module. The monitoring module polls the shared memory through an independent thread and uses the difference between the dual pointers (read / write) to judge the data validity. When the parity of the version number changes, a full-scale data copy to the local cache is triggered. The reading process uses lock-free programming techniques (CAS atomic operations) to directly map the memory to the user space, avoiding the overhead of kernel-mode switching. Data parsing is processed in parallel by SIMD instructions to improve throughput. For synchronization and fault tolerance processing, the initial synchronization calibrates the clocks of both parties through the PTPv2 protocol, and the compensated offset is stored in the control information segment. During operation, a dynamic time warping algorithm (DTW) is used to compensate for minute clock drifts. When a checksum error occurs, the data is automatically rolled back to the previous version. If three consecutive errors occur, a bus redundancy transmission mechanism is triggered to ensure that critical data is not lost. For performance optimization measures, write combining technology is used to reduce cache line misses, and a pre-allocated memory pool for the circular buffer is used to avoid the overhead of dynamic allocation. The monitoring module implements zero-copy technology and directly accesses the GPU video memory through memory registration to accelerate visual rendering.

[0063] Please refer to Figure 1 , as a preferred embodiment of the present application: The original boiler operation parameter data transmitted by the data acquisition module is simultaneously injected into the training data buffer of the intelligent control module. When the data volume in the buffer reaches a preset threshold, an offline retraining process for neural network optimization control is automatically triggered.

[0064] Specifically, the intelligent control module automatically matches the column vectors of the multivariable control matrix by parsing the standardized data (including three fields: timestamp, sensor ID, and original measurement value) output by the data acquisition module, and generates the final control instruction containing the dynamic weight reference instruction, compensation instruction, and optimization instruction by adopting a three-level instruction fusion mechanism. The instruction encapsulates four-dimensional information of device address code, action type, target value, and priority in JSON format, and is transmitted to the execution module via the TCP / IP protocol to trigger the action execution of the coal feeder, primary air fan, secondary air fan, and valve. When the execution status is returned via the redundant communication bus, a digital signature and timestamp are appended. The monitoring and early warning module adopts a stream processing architecture to analyze multi-source data in parallel, and triggers a three-level early warning mechanism through pattern matching. The red early warning directly triggers the audible and visual alarm interface of the human-computer interaction module and synchronously transmits the protection instruction to the control module. The orange early warning is pushed to the operation log pending processing queue, and the yellow early warning is displayed in the form of a floating prompt. The human-computer interaction module constructs a three-dimensional visualization interface based on the Unity3D engine and integrates the HoloLens 2 to achieve augmented reality interaction, supporting the operator to input the corrected instruction verified by permission through the bidirectional control channel. The correction trajectory is stored in the control strategy version library and triggers the offline retraining process of neural network optimized control. The training data buffer injects the original data using a double-buffer queue architecture. When the data volume reaches the threshold, the asynchronous training process is automatically started. After the new model passes the verification, it is hot-switched to the control system. The standardized data output by the data acquisition module is simultaneously injected into the training data buffer of the intelligent control module. The monitoring module calculates the transmission delay compensation amount by comparing the difference between the timestamp and the local clock. When the execution status data feedback by the execution module is transmitted via the redundant communication bus, a digital signature and timestamp are appended synchronously, forming a full-closed-loop control architecture including data acquisition, intelligent control, device execution, monitoring and early warning, and human-computer interaction. Each module realizes the real-time exchange of predicted trend data through shared memory, uses the AES-CMAC algorithm and digital signature to ensure the security of data transmission, and realizes nanosecond-level clock synchronization through the PTPv2 protocol to ensure the real-time performance and reliability of the control instruction. The overall system adopts an edge computing and cloud collaboration architecture, supports dynamic configuration of the final control instruction parameters and model structure, and realizes the efficient and stable operation and intelligent operation and maintenance management of the industrial boiler.

[0065] A method for an intelligent temperature control system of a circulating fluidized bed boiler, the method comprising the following steps: S1. The data acquisition module obtains the boiler operation parameters in real time and generates standardized data; S2. The intelligent control module receives the standardized data and generates the final control instruction through multi-algorithm fusion calculation; S3. The execution module adjusts the device action according to the final control instruction and feeds back the execution status data; S4. The monitoring and warning module triggers a three - level warning mechanism through pattern matching to generate warning information, and synchronously pushes the warning information to the human - machine interaction module; S5. The human - machine interaction module supports manually correcting the parameters of the final control instruction. After the correction instruction passes the permission verification, it is transmitted to the intelligent control module.

[0066] Furthermore, the multi - algorithm fusion calculation in step S2 is specifically the model predictive control, fuzzy control, and neural network optimization control algorithms, including a dynamic belief assignment mechanism. The output credibility of each control algorithm is evaluated in real - time through the entropy weight method. Among them, the weight coefficient of the model predictive control has a negative correlation with the working condition change rate; The execution status data in step S3 includes digital signatures and timestamps. The monitoring and warning module calculates the transmission delay compensation amount by comparing the difference between the timestamp and the local clock; The warning information in step S4 adopts a hierarchical queue transmission mechanism. The red warning directly triggers the audible and visual alarm interface of the human - machine interaction module. The orange warning is pushed to the operation log pending - processing queue, and the yellow warning is displayed as a floating prompt on the interface.

[0067] Specifically, the data acquisition module obtains the boiler operation parameters (including bed temperature, pressure, oxygen content, coal feeding amount, etc.) in real - time through a redundant sensor network, and generates standardized data containing three fields: timestamp, sensor ID, and measured value after amplitude - limiting filtering and standardization processing of the original data; After receiving the standardized data, the intelligent control module adopts a multi - algorithm fusion calculation framework (integrating model predictive control, fuzzy PID control, and expert rule base), and evaluates the output credibility of each algorithm in real - time through a dynamic belief assignment mechanism. Among them, the entropy weight method dynamically adjusts the weight coefficient according to characteristics such as historical prediction error and working condition volatility. The weight of the model predictive control has a negative correlation with the working condition change rate to ensure the balance of steady - state accuracy and dynamic response; After parsing the final control instruction (encapsulating the device address code, action type, target value, and priority in JSON format), the execution module drives the regulating devices such as the coal feeder frequency converter and the damper actuator to act, and feeds back the execution status data synchronously appended with AES - CMAC digital signatures and timestamps; The monitoring and warning module uses the Flink stream processing engine to analyze multi - source data streams in real - time. Through pattern matching, it triggers a three - level warning mechanism. The red warning (such as the bed temperature suddenly changing by more than 50℃ / s) directly triggers the audible and visual alarm interface of the human - machine interaction module and synchronously sends a protection instruction to the intelligent control module. The orange warning (such as the oxygen content continuously deviating from the set value) is pushed to the operation log pending - processing queue, and the yellow warning (such as the parameter fluctuation not reaching the threshold) is displayed in the form of a floating prompt; The human-computer interaction module constructs a three-dimensional visualization interface based on the Unity3D engine, supports operators to perform augmented reality interactions through HoloLens 2. After the manual correction instructions are authenticated by the JWT token and verified by the digital signature, they are transmitted to the intelligent control module to trigger the Git-LFS distributed update of the control strategy version library. At the same time, when executing the status data feedback, the monitoring module calculates the transmission delay compensation amount dynamically by using the sliding average filtering algorithm by comparing the time difference between the timestamp and the locally synchronized clock of PTPv2, ensuring the real-time performance of the prediction model and the control instructions.

[0068] In summary, the present invention realizes the precise regulation of the boiler operation parameters through the multi-algorithm collaborative control and the intelligent compensation mechanism. The model predictive control unit based on the multi-variable control matrix constructed by the ARX / FIR model can predict the dynamic trajectories of the boiler bed temperature, pressure and oxygen content parameters in real time, transform the multi-variable coupling problem difficult to handle by the traditional PID control into an optimization problem, and improve the control accuracy. The neural network optimized control unit dynamically adjusts the ratio of the coal feeding amount and the air volume through the reinforcement learning mechanism, improving the boiler thermal efficiency. The fuzzy control unit transforms the operator's experience rules into the compensation control amount under the non-linear boundary conditions, effectively suppressing the overshoot of the bed temperature and the sudden change of the oxygen content, and still maintaining the stability of the combustion process under variable working conditions.

[0069] The above schematically describes the present invention and its implementation manners. This description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. Therefore, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments without creative efforts without departing from the spirit of the present invention, they shall fall within the protection scope of the present invention.

[0070] In addition, it should be understood that although this specification is described according to the implementation manners, not each implementation manner only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other implementation manners understandable by those skilled in the art.

Claims

1. An intelligent temperature control system for a circulating fluidized bed boiler, comprising a data acquisition module, an intelligent control module, an execution module, a monitoring and early warning module and a human-computer interaction module, characterized in that: The data acquisition module is used to collect boiler operating parameters in real time and perform standardized processing, and synchronously transmit the standardized data to the intelligent control module through a redundant communication bus; The intelligent control module consists of a model prediction control unit, a fuzzy control unit and a neural network optimization control unit. It is used to receive the standardized data transmitted by the data acquisition module in real time, integrate the three control algorithms of model prediction control, fuzzy control and neural network optimization to dynamically calculate and generate the final control instructions, and transmit them to the execution module synchronously. At the same time, it generates prediction data and transmits it to the monitoring and early warning module. The execution module includes the coal feeder, primary fan, secondary fan and valve. It receives the final control instruction from the intelligent control module to execute the equipment action, and feeds back the execution status data to the data acquisition module, and uploads the action parameters to the monitoring and early warning module. The monitoring and early warning module processes the boiler operation parameters of the data acquisition module, the prediction data of the intelligent control module and the action parameters of the execution module in parallel, triggers the three-level early warning mechanism to generate early warning information through pattern matching, and the early warning information is synchronously pushed to the human-computer interaction module and triggers the protection logic of the intelligent control module; The human-computer interaction module adopts three-dimensional visualization technology to display the boiler operation status, final control instruction parameters and early warning information in real time. It integrates a two-way control channel to support operators to manually correct the final control instruction parameters. The corrected instructions are directly passed to the intelligent control module after authority verification.

2. The intelligent temperature control system for a circulating fluidized bed boiler according to claim 1, characterized in that: The boiler operating parameters include boiler bed temperature, pressure and oxygen content parameters; The standardized data includes a triple field of timestamp, sensor ID and raw measurement value, and the intelligent control module automatically matches the corresponding multivariable control matrix column vector by parsing the sensor ID in the field; The model prediction control unit constructs a multivariable control matrix based on the ARX model and the FIR model, predicts the future trajectory of the boiler operating parameters in real time, and generates a multi-parameter coupling reference control quantity including a coal feeding speed reference and a damper opening reference; The fuzzy control unit adopts the TS fuzzy model to convert the operator's experience rules into compensation control quantities under nonlinear boundary conditions, dynamically adjusts the compensation weight by monitoring the bed temperature overshoot and oxygen content mutation rate, and generates a fusion control instruction after fusing with the multi-parameter coupling benchmark control quantity; The neural network optimization control unit adopts a parallel fast learning network structure, dynamically adjusts the set value of the coal feed and air volume ratio through a reinforcement learning mechanism, and generates optimized control instructions; The fusion control instruction and the optimization control instruction are dynamically calculated to generate the final control instruction.

3. The intelligent temperature control system for circulating fluidized bed boiler according to claim 1 is characterized in that: The final control instruction generated by the intelligent control module is encapsulated in JSON format, including four-dimensional information of device address code, action type, target value and priority. The execution module triggers the control program of the corresponding device by parsing the device address code.

4. The intelligent temperature control system for circulating fluidized bed boiler according to claim 1 is characterized in that: When the execution status data fed back by the execution module is transmitted via the redundant communication bus of the data acquisition module, a digital signature and a timestamp are synchronously added, and the monitoring and early warning module calculates the transmission delay compensation amount by comparing the timestamp with the local clock difference.

5. The circulating fluidized bed boiler intelligent temperature control system according to claim 1, characterized in that: The three-level early warning mechanism includes: The warning information output by the monitoring and warning module is transmitted in a hierarchical queue. The red warning directly triggers the sound and light alarm interface of the human-computer interaction module, the orange warning is pushed to the operation log processing queue, and the yellow warning is displayed as a floating prompt on the interface.

6. The circulating fluidized bed boiler intelligent temperature control system according to claim 1, characterized in that: The correction instruction includes an operator identity authentication token and a parameter change trajectory. After the intelligent control module verifies the validity of the token, it stores the parameter change trajectory in the control strategy version library.

7. The intelligent temperature control system for circulating fluidized bed boiler according to claim 1 is characterized in that: The prediction data generated by the intelligent control module includes real-time predictions of the future trajectory of boiler bed temperature, pressure and oxygen content parameters, and the prediction data is exchanged in real time with the monitoring and early warning module using a shared memory method to avoid repeated transmission of data on a redundant communication bus.

8. The intelligent temperature control system for circulating fluidized bed boiler according to claim 2 is characterized in that: The neural network optimization control unit is also provided with a training data buffer for receiving the original data of the boiler operation parameters transmitted by the data acquisition module. When the data volume in the training data buffer reaches a preset threshold, the offline retraining process of the neural network optimization control is automatically triggered.

9. The intelligent temperature control method applied to the intelligent temperature control system of a circulating fluidized bed boiler according to any one of claims 1 to 8, characterized in that: The method comprises the following steps: S1. The data acquisition module obtains boiler operating parameters in real time and generates standardized data; S2. The intelligent control module receives standardized data and generates the final control instructions through multi-algorithm fusion calculation; S3. The execution module adjusts the device action according to the final control instruction and feeds back the execution status data; S4. The monitoring and warning module triggers the three-level warning mechanism through pattern matching to generate warning information, and simultaneously pushes the warning information to the human-computer interaction module; S5. The human-computer interaction module supports manual correction of the final control instruction parameters. The correction instruction is passed to the intelligent control module after authority verification.

10. The intelligent temperature control method of the circulating fluidized bed boiler intelligent temperature control system according to claim 9, characterized in that: The multi-algorithm fusion calculation in step S2 is specifically three control algorithms: model predictive control, fuzzy control and neural network optimization, including a dynamic credibility allocation mechanism, and the output credibility of each control algorithm is evaluated in real time by the entropy weight method, wherein the model predictive control weight coefficient is negatively correlated with the rate of change of the operating condition; The execution status data in step S3 includes a digital signature and a timestamp, and the monitoring and warning module calculates the transmission delay compensation amount by comparing the timestamp with the local clock difference; The warning information in step S4 adopts a hierarchical queue transmission mechanism. The red warning directly triggers the sound and light alarm interface of the human-computer interaction module, the orange warning is pushed to the operation log processing queue, and the yellow warning is displayed as an interface floating prompt.

Images