Intelligent Optimization Method and System for Wet Ball Mill in Coal-Fired Power Plant

Abstract

This application relates to the field of ball mill optimization technology, and discloses an intelligent optimization method for wet ball mills in coal-fired power plants, including the following steps: S1. Pre-check the wet ball mill and its supporting equipment, and access the distributed control system through Internet of Things technology to obtain real-time operation parameters; also discloses an intelligent optimization system for wet ball mills in coal-fired power plants, including the following modules: a data acquisition layer module for accessing the distributed control system through sensors to obtain the real-time operation parameters of the wet ball mill; a data analysis layer module for cleaning, feature extraction and modeling analysis of the collected real-time operation parameters. The present invention realizes the intelligent prediction and dynamic optimization of wet ball mills through the artificial intelligence Informer large model and IoT technology, reduces the consumption of steel balls and electric energy, improves economic benefits, the system supports multi-mode operation, combines data visualization technology, eliminates the limitations of manual monitoring, and improves management efficiency and safety.

Description

Technical Field

[0001] The present invention relates to the technical field of ball mill optimization, and specifically to an intelligent optimization method and system for wet ball mills in coal-fired power plants. Background Art

[0002] Wet ball mills play a crucial role in the limestone-gypsum wet flue gas desulfurization system of coal-fired power plants, and they achieve desulfurization requirements by grinding limestone slurry. However, the traditional steel ball addition method has significant deficiencies in terms of efficiency, energy consumption, and safety, which restrict the overall performance of wet ball mills.

[0003] In the traditional method, the addition of steel balls depends on the experience of operators, and usually, the decrease in current is used as the basis to judge whether steel balls need to be replenished. However, the operating environment of wet ball mills is complex, and current changes may be caused by various factors. Judging solely based on a single parameter, it is difficult to guarantee the accuracy of steel ball addition. As a result, situations of excessive or insufficient steel ball addition often occur. Excessive addition not only increases energy consumption but also places an additional burden on the equipment; while insufficient steel balls lead to unqualified control of slurry particle size, thus affecting desulfurization efficiency. This human-dominated method has obvious limitations in terms of accuracy and efficiency.

[0004] In addition, wet ball mills have a high proportion of energy consumption during operation, and their operating efficiency is closely related to the amount and timing of steel ball feeding. Traditional manual adjustment cannot dynamically optimize the steel ball replenishment strategy according to real-time working conditions, and it is often difficult to find a balance between energy consumption and efficiency. For example, excessive steel ball input has limited improvement in efficiency but a sharp increase in energy consumption; insufficient steel balls result in low efficiency, directly affecting the achievement of production goals. This contradiction has not been effectively solved for a long time, making it difficult for wet ball mills to achieve economic operation.

[0005] At the same time, the traditional steel ball addition method requires frequent manual intervention, which not only interrupts the continuous operation of the equipment but also increases safety hazards. During the steel ball feeding process, operators need to enter high-temperature and high-humidity environments to work, posing risks of equipment damage or other accidents. In addition, frequent shutdown operations also reduce production efficiency and further increase operation and maintenance costs.

[0006] Therefore, it can be seen that the traditional steel ball addition method faces major challenges in terms of precision control, energy consumption optimization, and safety guarantee. This method not only has difficulty meeting the requirements of modern coal-fired power plants for high-efficiency and low-consumption operation but also restricts the improvement space of wet ball mills in terms of environmental protection and economic goals. Summary of the Invention

[0007] Aiming at the deficiencies of the prior art, the present invention provides an intelligent optimization method and system for wet ball mills in coal-fired power plants, which solves the problems of low efficiency, high energy consumption, frequent manual intervention, and safety hazards existing in the traditional steel ball addition method for wet ball mills in coal-fired power plants.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: An intelligent optimization method for a wet ball mill in a coal-fired power plant, comprising the following steps:

[0009] S1. Pre-inspect the wet ball mill and its supporting equipment, and access the distributed control system through the Internet of Things technology to obtain real-time operating parameters, which include but are not limited to limestone consumption, limestone slurry volume, working current of the wet ball mill, power consumption of the wet ball mill, and operating duration of the wet ball mill;

[0010] S2. Based on big data analysis, clean, extract features, and integratively analyze the collected data to identify the complex relationships between limestone consumption and the operating current, steel ball weight, and steel ball specifications of the wet ball mill;

[0011] S3. Use the artificial intelligence Informer large model to calculate the optimal working state of the wet ball mill, including the optimal current range and steel ball addition strategy;

[0012] S4. Dynamically adjust the operating parameters of the wet ball mill according to the optimization decision, control the current of the wet ball mill within the optimal range, and automatically execute the steel ball addition operation, including the addition time and quantity of the steel balls.

[0013] Preferably, the cleaning of the collected data in step S2 includes the following steps:

[0014] Smooth the working current and power consumption signals of the wet ball mill by the sliding window method;

[0015] Detect outliers in the operating data and mark or remove them;

[0016] Fill in the missing data using the interpolation algorithm to generate a complete data set.

[0017] Preferably, the optimization decision in S4 is based on the following:

[0018] Use a time series prediction model to predict the future change trend of the operating current of the wet ball mill;

[0019] Determine the optimal addition time and quantity of the steel balls through big data analysis, and the addition time and quantity are dynamically calculated according to the limestone consumption, power consumption of the wet ball mill, and current steel ball consumption rate;

[0020] Dynamically adjust the operating current of the wet ball mill to stabilize it within the preset minimum working current range under the condition of meeting the optimal grinding efficiency.

[0021] Preferably, the optimal steel ball addition strategy in step S3 is determined according to the following steps:

[0022] Calculate the relationship between the consumption of limestone and the consumption of steel balls per unit time;

[0023] Based on the maximum ball loading capacity, steel ball specifications and wear rate of the equipment, optimize the calculation of the quantity of steel balls to be added;

[0024] Determine the specific time points for adding steel balls according to the load and working conditions of the wet ball mill predicted by big data.

[0025] An intelligent optimization system for the wet ball mill in a coal-fired power plant, comprising the following modules:

[0026] The data acquisition layer module is used to access the distributed control system through sensors to obtain the real-time operation parameters of the wet ball mill;

[0027] The data analysis layer module is used to clean, extract features and perform modeling analysis on the acquired real-time operation parameters;

[0028] The decision-making layer module is used to generate the optimal working state of the wet ball mill based on the analysis results, including the current range and the steel ball addition strategy;

[0029] The execution control layer module is used to dynamically adjust the operation parameters of the wet ball mill and control the steel ball addition equipment to perform corresponding operations according to the output of the decision-making layer

[0030] The data storage module is used to store the real-time parameters and historical operation data of the wet ball mill operation;

[0031] The alarm module is used to trigger an alarm to notify relevant personnel when the current of the wet ball mill exceeds the set range or an abnormality occurs;

[0032] The data visualization module is used to display the operation trend of the wet ball mill and the steel ball addition statistical information in the form of charts.

[0033] Preferably, the data acquisition layer module includes:

[0034] The sensor interface unit is used to receive the operation parameters of the wet ball mill, including the limestone consumption, the working current of the wet ball mill and the power consumption of the wet ball mill;

[0035] The data transmission unit is used to transmit the real-time data to the distributed control system through industrial wireless communication technology or wired Ethernet;

[0036] The security encryption unit is used to encrypt the acquired and transmitted data to ensure data security.

[0037] Preferably, the data analysis layer module includes:

[0038] A data cleaning unit for smoothing the noise data in the operating parameters of the wet ball mill and interpolating and complementing the missing values;

[0039] A feature extraction unit for extracting the core feature parameters of limestone consumption, the working current of the wet ball mill, and the amount of steel balls added;

[0040] A time series analysis unit for predicting the future operating state of the wet ball mill based on the Informer model.

[0041] Preferably, the decision-making layer module includes:

[0042] A rule generation unit for formulating the optimal current range and steel ball addition strategy for the operation of the wet ball mill;

[0043] An optimization calculation unit for dynamically adjusting the operating current and the steel ball addition time according to the limestone consumption and the operating load of the wet ball mill;

[0044] A decision output unit for sending the optimization result to the execution control layer module.

[0045] Preferably, the execution control layer module includes:

[0046] An operating parameter adjustment unit for dynamically adjusting the operating current of the wet ball mill and controlling it to be stable within the optimal range;

[0047] A steel ball addition control unit for automatically performing the steel ball addition operation according to the optimization result;

[0048] An interlock protection unit for ensuring that the steel ball addition is allowed only when the wet ball mill and related equipment are in the operating state.

[0049] The present invention provides an intelligent optimization method and system for a wet ball mill in a coal-fired power plant. It has the following beneficial effects:

[0050] 1. By introducing the artificial intelligence Informer large model and IoT technology, the present invention realizes the intelligent prediction and dynamic adjustment of the operating parameters of the wet ball mill, successfully controls the current of the wet ball mill within the lowest threshold range that meets the production requirements, optimizes the steel ball addition strategy, greatly reduces the consumption of steel balls and electric energy. Compared with the traditional operation method relying on manual experience, the present invention avoids the problems of untimely or excessive steel ball replenishment, significantly reduces the operating cost, and solves the problems of energy waste and insufficient economic efficiency in the traditional technology.

[0051] 2. The present invention adopts multi-mode operation support, combines deep learning algorithms with data visualization technology, and provides three flexible options: intelligent mode, automatic mode, and manual mode. The system can not only analyze data trends in real time but also automatically adjust the operating state of the device. In traditional methods, the operation relies on manual monitoring and cannot respond in real time. The present invention effectively eliminates the limitations of manual intervention, solves the problems of single operation mode and difficulty in adapting to complex working conditions in the prior art, and greatly improves management efficiency and production safety.

[0052] 3. Through a complete closed-loop control of data acquisition, analysis, decision-making, and execution, the present invention realizes fully automated operation from device operation monitoring to steel ball addition, and has a breakpoint resumption and multiple fault tolerance mechanisms. The TLS encryption protocol and authentication are introduced during the data transmission process to ensure communication security and data integrity. Traditional technologies have insufficient capabilities in handling data loss and network failures. The present invention not only enhances the stability and security of the system but also improves the reliability of device operation, and solves the potential problems of downtime or efficiency decline caused by device failures in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the method flow of the present invention;

[0054] Figure 2 It is an overall system architecture diagram of the present invention;

[0055] Figure 3 It is a network topology diagram of the present invention;

[0056] Figure 4 It is a system framework diagram of the present invention;

[0057] Figure 5 It is a data acquisition layer module of the present invention;

[0058] Figure 6 It is a data analysis layer module of the present invention;

[0059] Figure 7 It is a decision-making layer module of the present invention;

[0060] Figure 8 It is an execution control layer module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0062] Please refer to the appendix Figure 1 The embodiment of the present invention provides an intelligent optimization method for a wet ball mill in a coal-fired power plant, including the following steps:

[0063] S1. Pre-check the wet ball mill and its supporting equipment, and access the distributed control system through Internet of Things technology to obtain real-time operation parameters, including but not limited to limestone consumption, limestone slurry volume, working current of the wet ball mill, power consumption of the wet ball mill, and operation duration of the wet ball mill;

[0064] In this embodiment, first, conduct a comprehensive inspection on the wet ball mill and its supporting equipment to ensure that the equipment is in normal operation. This link includes the integrity inspection of the mechanical system, electrical system, and sensors of the equipment. Specifically:

[0065] For the ball storage bin, motor, speed reducer, sprocket, and transmission mechanism, check whether their connections are firm and whether the wear condition is normal.

[0066] Check whether accessories such as protective covers and vibrators can work properly to avoid affecting the accuracy of subsequent sensor collection.

[0067] Verify the status of the sensors to ensure that they can accurately sense key parameters such as current, voltage, and limestone consumption.

[0068] Generally, the inspection of sensors also includes regular calibration to ensure the accuracy of data collection. For example, verify whether the sensor output value meets the specifications by means of standard signal input.

[0069] After confirming that the equipment is in normal state, access the DCS system through Internet of Things technology to collect and transmit the operation data of the equipment. Specifically, in this embodiment, a variety of industrial protocols and data transmission methods are adopted to ensure the comprehensiveness and stability of data collection.

[0070] In some embodiments, the real-time operation data of the wet ball mill is connected to the distributed control system through ModbusTCP or UDP protocol. As an option, data access can also be achieved by directly connecting to the DPU control cabinet board. Such protocols can achieve efficient on-site device communication.

[0071] Generally, data transmission can be carried out through industrial wireless communication technologies (such as LoRaWAN or NB-IoT) or wired Ethernet. In specific implementation, select an appropriate communication method according to the infrastructure conditions of the power plant. For example, when the wet ball mill is located in an area far from the control room, LoRaWAN is preferably selected to provide long-distance wireless communication.

[0072] Data is encrypted using the TLS or SSL protocol during transmission to prevent data tampering or leakage. As an implementation method, identity authentication can also be implemented for devices accessing sensors to ensure that only authorized devices can send or receive data.

[0073] In this embodiment, the collected operating parameters cover various key process and equipment data of the wet ball mill. Specifically, they include but are not limited to the following categories:

[0074] Equipment operating parameters:

[0075] The working current of the wet ball mill (Unit: ampere).

[0076] The power consumption of the equipment (Unit: kilowatt-hour).

[0077] Operating duration (Unit: one hour).

[0078] Rotation speed R (Unit: revolutions per minute).

[0079] Process indicators:

[0080] Limestone consumption (Unit: ton).

[0081] Limestone slurry volume (Unit: cubic meter).

[0082] Equipment configuration parameters:

[0083] Weight of steel balls (Unit: ton).

[0084] Specification of steel balls (Unit: millimeter).

[0085] Maximum ball loading (Unit: ton).

[0086] Rated power (Unit: kilowatt).

[0087] System frequency f (Unit: hertz).

[0088] In a possible implementation method, the collected data is transmitted to the distributed control system in real time and stored in the central database in a structured manner. The data at each sampling time point contains multiple fields, including equipment status information, real-time operating indicators, and environmental parameters.

[0089] For example, for a certain time point tThe operating state, and the data can be represented as a multi-dimensional vector:

[0090]

[0091] This multi-dimensional data format facilitates the extraction of key features in subsequent data analysis and input into the optimization model.

[0092] To ensure the timeliness of the data, generally, in this embodiment, the data sampling frequency is set to 1 second / time, that is, the key operating parameters of the wet ball mill are collected once per second. As an option, when the operating conditions of the equipment are relatively stable, the sampling frequency can be adjusted to 5 seconds / time or a longer time interval to reduce the data transmission burden.

[0093] In another embodiment, a dynamic sampling strategy can also be set. For example:

[0094] When the operating current of the wet ball mill The change amplitude exceeds the set threshold, the system will automatically increase the sampling frequency to capture the details of the current fluctuation.

[0095] When the equipment is in a non-working state, only the static state data of the equipment is collected to save storage resources.

[0096] In some embodiments, to ensure the integrity of the collected data, the following mechanism is adopted in this embodiment:

[0097] After each sampling, the communication between the sensor and the DCS will return a confirmation signal to avoid data loss.

[0098] If the operating data at certain time points is missing, the system will mark these time points as invalid and notify the relevant maintenance personnel.

[0099] For sudden network interruptions, a breakpoint resumption mechanism is adopted to resume data collection.

[0100] In addition, the system also regularly calibrates the sensor to check whether its measurement error is within the allowable range. In this way, the problem of data deviation caused by sensor aging is avoided.

[0101] S2. Based on big data analysis, clean, extract features and integrate and analyze the collected data to identify the complex relationships among the limestone consumption, the operating current of the wet ball mill, the weight of the steel balls, and the specifications of the steel balls;

[0102] In this embodiment, data cleaning first performs smoothing processing on the noise signal. To ensure the stability of the operating data of the wet ball mill, the sliding window averaging method is used to smooth the collected parameters such as current and voltage. Its calculation formula is:

[0103]

[0104] Among them, is the smoothed data value, n is the window size, k = (n - 1) / 2 . The window is selected and dynamically adjusted according to the operating characteristics of the wet ball mill. In one possible implementation, for the operating stage with large current fluctuations, a smaller window n = 5 is selected; while for the stable operating stage, the window size can be increased to n = 10 to further smooth the data.

[0105] In some embodiments, possible outliers are also marked or removed. Specifically, the Z-score normalization method is used to calculate the dispersion degree of each data point:

[0106]

[0107] Among them, x is the current data point, μ is the mean value of the data, σ is the standard deviation. If , then this data point is considered an outlier and is marked as invalid. In another possible implementation, the outlier is replaced with the mean value of its neighboring data points to ensure the continuity of the data.

[0108] In some implementations, the outlier is replaced with the mean value of neighboring data points:

[0109]

[0110] to ensure the continuity of the data.

[0111] For the case of missing data, linear interpolation is used to complete it. Assuming that the current data at time point t is missing, it is calculated by the following formula:

[0112]

[0113] Among them: is the time of adjacent known data points; 、 are the known values at adjacent time points.

[0114] In this embodiment, after completing data cleaning, feature extraction is performed on the processed data. Generally, the goal of feature extraction is to mine the key indicators of the operating parameters of the wet ball mill and provide a scientific basis for subsequent optimization of the model.

[0115] Specifically, the extracted features include but are not limited to:

[0116] Limestone consumption rate , and its calculation formula is:

[0117]

[0118] Among them, is the consumption of limestone per unit time, and ΔT is the sampling time interval.

[0119] Steel ball wear rate , and its calculation formula is:

[0120]

[0121] Among them, is the reduction in the weight of steel balls per unit time.

[0122] Power consumption per unit time , and its expression is:

[0123]

[0124] Among them, is the power consumption of the wet ball mill, is the operating duration.

[0125] In a possible implementation, statistical features within a rolling window can also be constructed, such as mean, standard deviation, maximum value, and minimum value. These statistics help capture local change trends during the operation of the wet ball mill.

[0126] As an option, feature screening can also be performed based on the correlation between features and the target variable. Specific methods include calculating Pearson correlation coefficients, feature importance ranking, etc. For example, when using a random forest model for feature screening, the contribution of each feature to the target output (such as the optimal current range) can be calculated, and the top 10% of the key features can be retained.

[0127] In this embodiment, after feature extraction, a deep learning algorithm is used to analyze the internal relationships between the operating parameters of the wet ball mill. Specifically, multiple groups of relationship models are constructed using regression models, time series models, and machine learning models.

[0128] In the regression model, predict the limestone consumption and the current of the wet ball mill and the weight of the steel balls between the relationships:

[0129]

[0130] Among them, a, b, c is the regression coefficient, obtained by fitting historical data.

[0131] In the time series model, predict the future current The changing trend. Generally, an LSTM network is used to capture the time-dependent characteristics of the current, and its state update formula is as follows:

[0132]

[0133] Among them, is the hidden state at the current moment, is the input data, are the network weights and biases.

[0134] In the machine learning model, XGBoost or LightGBM is used for feature importance analysis and to optimize the operating parameters of the wet ball mill.

[0135] S3. Use the artificial intelligence Informer large model to calculate the optimal working state of the wet ball mill, including the optimal current range and the steel ball addition strategy;

[0136] In this embodiment, first, the trained Informer large model is used to analyze the operating conditions of the wet ball mill. The Informer model is a time series prediction model based on deep learning and can efficiently process large-scale time series data. Specifically, the model mainly consists of an encoder, a decoder, and a self-attention mechanism.

[0137] In a possible implementation, the features input to the model include the limestone consumption , the working current of the wet ball mill , the weight of the steel balls , the specification of the steel balls and other key operating parameters. The model predicts the future operating state of the wet ball mill based on the input features and outputs the optimal current range and the optimal steel ball addition strategy.

[0138] As an option, the encoder is used to capture the long-term dependencies in the time series and extract features by stacking multiple layers of self-attention mechanisms; the decoder generates the prediction results for future time points based on the output of the encoder. The role of the attention mechanism is to screen the input features that have the greatest impact on the prediction target to improve the prediction accuracy.

[0139] In this embodiment, the optimal working current range of the wet ball mill is determined using the model prediction results. Generally, the optimal working current range meets the following conditions:

[0140] The current When within this range, the grinding efficiency of the wet ball mill is the highest.

[0141] The current fluctuation amplitude is small and can be stably maintained at the minimum value within the set range.

[0142] Calculate the grinding efficiency of the wet ball mill at different currents based on historical operation data, and then fit the functional relationship between the current and the efficiency:

[0143]

[0144] Among them, η is the grinding efficiency of the wet ball mill, is the working current. Find the maximum point of the function by taking the derivative, and set reasonable upper and lower limits in combination with industry experience to obtain the optimal current range.

[0145] Specifically, assume the function is in the form of a parabola, and the maximum point is , then:

[0146]

[0147] Among them, is the allowable current fluctuation range.

[0148] In this embodiment, on the basis of determining the optimal working current range, a steel ball addition strategy is further formulated. Generally, the steel ball addition strategy includes two aspects: the addition time point and the quantity.

[0149] The quantity of steel ball addition is calculated based on the following formula:

[0150]

[0151] Among them, is the consumption of limestone, k is the grinding efficiency coefficient of the steel ball, is the specification (diameter) of the steel ball.

[0152] As an implementation method, the steel ball addition time point can be set as the nearest time point when the limestone consumption reaches a certain threshold. Use the time series model to predict the change trend of the limestone consumption, and find the time point that satisfies the following conditions:

[0153]

[0154] Among them, is the limestone consumption at the time of the last steel ball addition, is the critical value of the limestone consumption.

[0155] In this embodiment, in combination with the optimal working current range and the steel ball addition strategy, the operating parameters of the wet ball mill are optimized through a dynamic adjustment algorithm. Generally, the goal of adjustment is to ensure that the current is stabilized at the lowest value within the range, and steel balls are added on time to ensure the grinding efficiency.

[0156] In one possible implementation, the system monitors the operating current of the wet ball mill in real time . If it exceeds the optimal range , the system automatically adjusts the operating state parameters (such as rotational speed or feed rate) of the wet ball mill to bring it back to the set range.

[0157] In addition, the parameters can be continuously optimized based on the actual operating data. For example, using the reinforcement learning algorithm, according to the feedback of historical adjustment results, the key parameters of the current range and the addition strategy are dynamically updated.

[0158] In this embodiment, to verify the accuracy and optimization effect of the model, multiple evaluation metrics can be used, including mean squared error (MSE), mean absolute error (MAE), and root mean squared error (RMSE). Taking the mean squared error as an example, its calculation formula is: where is the actual value, is the predicted value, n is the number of samples.

[0159] In some embodiments, the actual effect of the model can also be evaluated by comparing the operating data of the wet ball mill before and after optimization (such as limestone consumption and power consumption per unit time). Specifically, the reduction range of power consumption before and after optimization can be statistically calculated:

[0160]

[0161] where is the power consumption before optimization, is the power consumption after optimization.

[0162] S4. Dynamically adjust the operating parameters of the wet ball mill according to the optimization decision, control the current of the wet ball mill within the optimal range, and automatically execute the steel ball addition operation, including the addition time and quantity of steel balls.

[0163] In this embodiment, the adjustment of the operating current of the wet ball mill is the core of the optimization operation. According to the optimal current range determined in step S3 , the system realizes the real-time monitoring and dynamic adjustment of the working current through an intelligent adjustment algorithm.

[0164] Specifically, the real-time collected operating current is compared with the range. If Beyond the range, the control module automatically adjusts the operating parameters of the wet ball mill, including equipment speed, feed rate, etc., until the current returns to the preset range.

[0165] In a possible implementation, the adjustment parameter is calculated through the following control logic:

[0166]

[0167] Where: is the adjusted power input; is the target current value;

[0168] are the proportional, integral, and derivative control parameters.

[0169] Generally, this adjustment method based on PID (Proportional-Integral-Derivative) control is suitable for the rapid fluctuation response of the current during the operation of the wet ball mill. For the operation stage with slow current change, the weight of can be appropriately reduced to reduce the unnecessary adjustment frequency.

[0170] In this embodiment, the steel ball addition operation is automatically completed under the control of the intelligent ball adding machine based on the optimization strategy of step S3. Generally, the system, according to the real-time monitored operation data, performs operations according to the established addition time points and quantities for operation.

[0171] In a possible implementation, the quantity of steel balls added is calculated through the following formula:

[0172]

[0173] Where: is the consumption of limestone per unit time; k is the grinding efficiency coefficient of the steel balls; is the steel ball specification (diameter).

[0174] As an option, the system can also dynamically adjust the quantity of steel balls added. For example, when the operating current of the wet ball mill shows abnormal fluctuations, the ball adding operation can be temporarily suspended and restarted after the current resumes.

[0175] The execution process of steel ball addition involves the following specific steps:

[0176] The ball adding machine takes out the set quantity of steel balls from the ball storage bin;

[0177] The steel balls are sent into the wet ball mill through the chain conveyor device;

[0178] After the ball addition is completed, the system records the ball addition log, including information such as time points and quantities.

[0179] In this embodiment, to ensure the accuracy and safety of the ball addition process, the system supports three operation modes: intelligent mode, automatic mode, and manual mode.

[0180] Specifically:

[0181] Intelligent mode: The system fully automatically executes according to the optimization results output by the decision-making layer, including current regulation and steel ball addition operations.

[0182] Automatic mode: The ball addition operation is controlled by preset rules. For example, the ball addition is executed at fixed time intervals or after reaching a set threshold.

[0183] Manual mode: The user manually triggers the ball addition operation through the touch screen or the remote control interface, which is used for special working conditions or the equipment debugging stage.

[0184] In a possible implementation, the system automatically monitors the working state of the ball adding machine, including the running condition of the chain, the remaining amount of steel balls in the ball storage bin, etc., to ensure the continuity and reliability of the ball addition process.

[0185] In this embodiment, to prevent equipment damage caused by misoperation or abnormal conditions, the system sets up a multiple interlock protection mechanism.

[0186] Specifically:

[0187] During the ball addition process, the system detects the running states of the wet ball mill and the weighing belt feeder. If it is found that both are in the stopped state, the ball addition operation is prohibited and an alarm is triggered.

[0188] For abnormal conditions such as too long ball jamming time and no-ball time, the system automatically pauses the operation and records the relevant logs for maintenance personnel to refer to.

[0189] As an option, the safety protection mechanism also includes overload protection and network communication fault detection. For example, when the running current exceeds the set ratio of the rated current of the equipment , the system triggers an emergency stop and sends an alarm notification.

[0190] In this embodiment, after the ball addition operation is completed, the system records the relevant data in the database and displays it to the user through the data visualization interface. The recorded data includes:

[0191] The time point of each ball addition ;

[0192] The quantity of added balls ;

[0193] The operating status of the equipment after adding balls (such as working current, power consumption, etc.).

[0194] Generally, these data will serve as important inputs for subsequent model training and optimization, further enhancing the accuracy and adaptability of the system.

[0195] The intelligent optimization system for wet ball mills in coal-fired power plants described below can be cross-referenced with the intelligent optimization method for wet ball mills in coal-fired power plants described above.

[0196] Please refer to the appendix Figure 2 - appendix Figure 4 A kind of intelligent optimization system for wet ball mills in coal-fired power plants, characterized by including the following modules:

[0197] The data acquisition layer module is used to access the distributed control system through sensors to obtain the real-time operating parameters of the wet ball mill;

[0198] The data analysis layer module is used to clean, extract features and perform modeling analysis on the collected real-time operating parameters;

[0199] The decision-making layer module is used to generate the optimal working state of the wet ball mill based on the analysis results, including the current range and the steel ball addition strategy;

[0200] The execution control layer module is used to dynamically adjust the operating parameters of the wet ball mill and control the steel ball addition equipment to perform corresponding operations according to the output of the decision-making layer

[0201] The data storage module is used to store the real-time parameters and historical operating data of the wet ball mill operation;

[0202] The alarm module is used to trigger an alarm to notify relevant personnel when the current of the wet ball mill exceeds the set range or an abnormality occurs;

[0203] The data visualization module is used to display the operation trend of the wet ball mill and the steel ball addition statistical information in the form of charts.

[0204] Please refer to the appendix Figure 5 The data acquisition layer module includes:

[0205] The sensor interface unit is used to receive the operating parameters of the wet ball mill, including the limestone consumption, the working current of the wet ball mill and the power consumption of the wet ball mill;

[0206] The data transmission unit is used to transmit the real-time data to the distributed control system through industrial wireless communication technology or wired Ethernet;

[0207] The security encryption unit is used to encrypt the collected and transmitted data to ensure data security.

[0208] Please refer to the appendix Figure 6 , the data analysis layer module includes:

[0209] A data cleaning unit for smoothing the noise data in the operating parameters of the wet ball mill and interpolating and complementing the missing values;

[0210] A feature extraction unit for extracting the core feature parameters of limestone consumption, the working current of the wet ball mill, and the amount of steel balls added;

[0211] A time series analysis unit for predicting the future operating state of the wet ball mill based on the Informer model.

[0212] Please refer to the appendix Figure 7 , the decision-making layer module includes:

[0213] A rule generation unit for formulating the optimal current range and steel ball addition strategy for the operation of the wet ball mill;

[0214] An optimization calculation unit for dynamically adjusting the operating current and the steel ball addition time according to the limestone consumption and the operating load of the wet ball mill;

[0215] A decision output unit for sending the optimization results to the execution control layer module.

[0216] Please refer to the appendix Figure 8 , the execution control layer module includes:

[0217] An operating parameter adjustment unit for dynamically adjusting the operating current of the wet ball mill and controlling it to be stable within the optimal range;

[0218] A steel ball addition control unit for automatically performing the steel ball addition operation according to the optimization results;

[0219] An interlock protection unit for ensuring that the steel ball addition is allowed only when the wet ball mill and related equipment are in the operating state.

[0220] The system of this embodiment can be used to execute the above method embodiment, and its principle and technical effects are similar, which will not be elaborated here.

[0221] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent optimization method for a wet ball mill in a coal-fired power plant, characterized in that: The following steps are involved: S1. Preliminary inspection of the wet ball mill and its supporting equipment, and access to the distributed control system through the Internet of Things technology to obtain real-time operating parameters, including but not limited to limestone consumption, limestone slurry volume, wet ball mill operating current, wet ball mill power consumption and wet ball mill operating time; S2. Based on big data analysis, the collected data is cleaned, features are extracted and integrated for analysis to identify the complex relationship between limestone consumption and wet ball mill operating current, steel ball weight and steel ball specifications; S3. Use the artificial intelligence Informer large model to calculate the optimal working state of the wet ball mill, including the optimal current range and steel ball addition strategy; S4, dynamically adjusting the operating parameters of the wet ball mill according to the optimization decision, controlling the current of the wet ball mill within the optimal range, and automatically performing the steel ball adding operation, including the adding time and quantity of the steel balls; The step S2 of cleaning the collected data includes the following steps: The operating current and power consumption signals of the wet ball mill are smoothed by the sliding window method. Detect outliers in operational data and mark or remove them; Interpolation algorithms are used to fill in missing data to generate a complete data set; The optimal steel ball adding strategy in step S3 is determined according to the following steps: Calculate the relationship between limestone consumption and steel ball consumption per unit time; Optimize the calculation of the number of steel balls to be added based on the maximum ball loading capacity of the equipment, steel ball specifications and wear rate; According to the load and working status of the wet ball mill predicted by big data, the specific time point for adding steel balls is determined.

2. The intelligent optimization method for a wet ball mill in a coal-fired power plant according to claim 1, characterized in that: The optimization decision in S4 is based on the following: Use time series prediction model to predict the future trend of wet ball mill operating current; Determine the optimal time and quantity of adding steel balls through big data analysis, which is dynamically calculated based on limestone consumption, wet ball mill power consumption, and current steel ball consumption rate; Dynamically adjust the operating current of the wet ball mill so that it can be stabilized within the preset minimum operating current range while meeting the best grinding efficiency.

3. An intelligent optimization system for a wet ball mill in a coal-fired power plant, applied to the intelligent optimization method for a wet ball mill in a coal-fired power plant according to claim 2, characterized in that: Includes the following modules: The data acquisition layer module is used to obtain the real-time operating parameters of the wet ball mill through the sensor access to the distributed control system; The data analysis layer module is used to clean, extract features and conduct modeling analysis on the collected real-time operating parameters; A decision-making layer module, which is used to generate the optimal working state of the wet ball mill based on the analysis results, including the current range and steel ball addition strategy; The execution control layer module is used to dynamically adjust the operating parameters of the wet ball mill and control the steel ball adding equipment to perform corresponding operations according to the output of the decision layer; A data storage module, used for storing real-time parameters and historical operation data of the wet ball mill; An alarm module is used to trigger an alarm to notify relevant personnel when the current of the wet ball mill exceeds the set range or an abnormality occurs; Data visualization module for displaying wet ball mill operation trends and ball addition statistics in graphical form.

4. The intelligent optimization system for a wet ball mill in a coal-fired power plant according to claim 3, characterized in that: The data acquisition layer module includes: a sensor interface unit for receiving operating parameters of the wet ball mill, including limestone consumption, wet ball mill operating current, and wet ball mill power consumption; A data transmission unit for transmitting real-time data to a distributed control system via industrial wireless communication technology or wired Ethernet; The security encryption unit is used to encrypt the collected and transmitted data to ensure data security.

5. The intelligent optimization system for a wet ball mill in a coal-fired power plant according to claim 3, characterized in that: The data analysis layer module includes: A data cleaning unit is used to smooth the noise data in the operating parameters of the wet ball mill and to interpolate and complete the missing values; Feature extraction unit, used to extract the core feature parameters of limestone consumption, wet ball mill operating current and steel ball addition; The time series analysis unit is used to predict the future operating status of the wet ball mill based on the Informer model.

6. The intelligent optimization system for wet ball mill in coal-fired power plant according to claim 3, characterized in that: The decision-making layer module includes: A rule generation unit for developing the optimal current range and ball addition strategy for wet ball mill operation; Optimization calculation unit for dynamically adjusting the operating current and steel ball addition time according to the limestone consumption and wet ball mill operating load; The decision output unit is used to send the optimization results to the execution control layer module.

7. The intelligent optimization system for a wet ball mill in a coal-fired power plant according to claim 3, characterized in that: The execution control layer module includes: An operating parameter adjustment unit is used to dynamically adjust the operating current of the wet ball mill and control it to be stable within an optimal range; A steel ball adding control unit, used to automatically perform the steel ball adding operation according to the optimization results; Interlock protection unit is used to ensure that steel balls are added only when the wet ball mill and related equipment are in operation.

Images