A coal powder vertical mill intelligent control system and method based on multi-parameter cooperation

Abstract

The application provides a coal powder vertical mill intelligent control system and method based on multi-parameter cooperation, relates to the technical field of grinding control, and randomly generates parameter combinations of feeding amount, motor output power and air supply amount to form a control scheme, establishes a mapping relationship between the control scheme and excellent indexes in combination with historical data, determines an optimal control scheme by using a particle swarm optimization algorithm with the maximization of excellent indexes as a target, and determines a target control scheme in the optimal control scheme by comprehensively considering adjustment costs and excellent indexes, finally generates a control instruction, and realizes cooperative optimization control of the feeding amount, the motor power and the air supply amount.

Description

Technical Field

[0001] This invention relates to the field of grinding control technology, specifically to an intelligent control system and method for a vertical coal mill based on multi-parameter coordination. Background Technology

[0002] Vertical pulverized coal mills are key equipment in industrial processes such as thermal power generation and cement production, and their operating efficiency and stability directly affect production energy consumption and finished product quality. Traditional control methods often rely on adjusting single parameters, lacking systematic optimization of the synergistic effects of multiple parameters. This leads to problems such as increased vibration, high energy consumption, and large fluctuations in fineness during mill operation. Faced with this challenge, more and more companies are beginning to try applying automation technology to pulverized coal preparation systems. However, due to the complexity and variability of vertical mill systems, as well as the differences between different production lines, existing automation control schemes are often difficult to apply directly to all situations. In addition, existing control strategies often fail to dynamically respond to changes in the internal operating conditions of the mill, resulting in instability in the grinding process and an inability to simultaneously achieve optimal output and energy efficiency. Therefore, there is an urgent need for an intelligent control method that can comprehensively coordinate feed rate, air volume, and main motor power, and achieve dynamic feedback adjustment to improve the overall operating performance of vertical pulverized coal mills.

[0003] In the prior art, publication number CN109847916A discloses an energy-saving optimization method for a cement raw material vertical mill system: Historical operation data is collected to form multiple historical operation models. The historical operation data includes controllable variables, such as mill grinding pressure, mill inlet negative pressure, mill outlet negative pressure, hot air valve opening, cold air valve opening, mill inlet temperature, mill outlet temperature, and fan valve opening. The system is divided into feeding rate zones. Based on the power consumption of a single unit, the historical operations of the same feeding rate zone are sorted to obtain the optimal historical operation record for that zone. The optimal historical operation records of all feeding rate zones are merged to form an optimal recommendation table. Optimal operation suggestions for the equipment are obtained based on real-time operating conditions and the optimal recommendation table.

[0004] The main problems with the above methods are: the basic logic of the above scheme is to recommend the combination of operations that had the lowest single-machine power consumption in the past within the same feed rate range. The optimization objective is single-machine power consumption, while ignoring the impact of other parameters such as output and finished product quality, resulting in poor practicality in actual production; essentially, it establishes a static database of "optimal historical operation records", resulting in poor real-time adaptive adjustment capability of the system; after giving the optimal operation suggestion, it does not consider the cost required to adjust from the current working condition to the optimal operation. If the suggested operation parameters differ greatly from the current parameters, direct switching may lead to a series of failures and affect the safety of the system.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent control system and method for a pulverized coal vertical mill based on multi-parameter coordination, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A multi-parameter coordinated intelligent control system for a pulverized coal vertical mill specifically includes: The control scheme generation module is used to randomly generate multiple feed amounts and motor output powers within the range of feed amount and motor output power of the mill, and arrange and combine them to generate several sets of parameter combinations. For the feed amount in each parameter combination, an air volume is randomly generated from the air volume range corresponding to that feed amount and added to the parameter combination to form a control scheme. The mapping analysis module is used to analyze the historical control schemes and corresponding excellent indicators of the mill in the past operation, and to establish the mapping relationship between the control schemes and the excellent indicators. The control scheme optimization module is used to combine the mapping relationship, with the goal of maximizing the excellent index, and optimize the control scheme based on the particle swarm optimization algorithm to obtain several preferred control schemes. The control command generation module is used to quantify the adjustment cost from the current control scheme to each preferred control scheme, combine the adjustment cost with the superior index values ​​of each preferred control scheme, determine the target control scheme, and determine the control command by combining the current control scheme and the target control scheme.

[0008] Furthermore, the principles upon which the control scheme is based are: Based on the mill's rated parameters, the feed rate range and motor output power range are obtained. From each range, a feed rate value and a motor output power value are selected to generate a parameter combination. For each parameter combination, the corresponding airflow range for the feed rate is determined. The underlying principle is as follows: Under the same fineness standard, obtain all air volume values ​​corresponding to the same feed amount in the mill operation history data. The minimum value is the lower limit of the available air volume for that feed amount, and the maximum value is the upper limit of the available air volume for that feed amount. The range between the lower limit and the upper limit of the available air volume is the air supply range under that feed amount. Based on the obtained parameter combination, the air supply range is determined according to the feed rate range, and an air supply rate is randomly selected and added to the parameter combination to generate a control scheme.

[0009] Furthermore, the excellent performance indicators consist of output efficiency and energy efficiency, based on the following principles: The formula used to calculate output efficiency is: in, This represents the output efficiency of the historical control scheme over its longest effective period. This indicates the longest duration of action of any historical control scheme. This indicates the pulverized coal production during the longest possible duration of the historical control scheme. The formula used to calculate energy efficiency is: in, This represents the energy efficiency of the historical control scheme over its longest operating time. This represents the total energy consumption of the historical control scheme during its longest operating time. The formula used to calculate the excellent performance index is as follows: in, Indicators representing the excellence of historical control schemes.

[0010] Furthermore, the principle underlying the establishment of the mapping relationship between control schemes and optimal indicators is as follows: A deep learning model is established, using historical control schemes as input and corresponding superior indicators as labels to train a performance prediction model; the real-time generated control schemes are then input into the performance prediction model to obtain superior indicators that correspond one-to-one with the control schemes.

[0011] Furthermore, the principle underlying the optimization of the control scheme is as follows: Each particle corresponds to a control scheme, represented as follows: ,in, Indices representing particle indices Indicates the first Feed amount per particle Indicates the first The motor output power of each particle Indicates the first Air volume per particle; The constraints are: , , ,in, Indicates the lower limit of feed amount. Indicates the upper limit of feed amount. This indicates the lower limit of the motor's output power. This indicates the upper limit of the motor's output power. Indicates feed amount The corresponding lower limit of air supply volume, Indicates feed amount The corresponding upper limit of air supply volume; The initial position and initial velocity of each particle are randomly initialized, and the number of iterations is set. The particle velocity is updated in each iteration, based on the following formula: in, Indicates the first The particle in the first Speed ​​at the next iteration An index representing the number of iterations. Represents velocity inertia weight. Indicates the first The particle in the first The speed of each iteration They represent individual learning factors and social learning factors, respectively. , This indicates the random numbers used to avoid the algorithm getting trapped in local optima, and , Indicates the first The historical optimal position of a particle, that is, the position that maximizes the value of the excellent index during historical iterations. Indicates the first The particle in the first Position at the next iteration This represents the global optimal position, which is the position where all particles maximize the value of the performance index in the historical iterations. The principle underlying the acquisition of several optimal control schemes is as follows: During the iteration process, the historical best position and global best position of all particles are counted, and they are sorted in descending order according to the size of the excellent index. The top 5% of particles are retained, and each particle corresponds to an optimal control scheme.

[0012] Furthermore, the principle underlying the calculation of adjustment costs is as follows: Obtain the control scheme of the mill at the current moment and set it as the current control scheme, which includes the current feed rate, the current motor output power and the current air volume; The formula used to calculate the cost of adjusting the feed rate is: in, Indicates the first The cost of adjusting the feed rate for each optimal control scheme Index indicating the preferred control scheme. Indicates the first The optimal control scheme for feed amount Indicates the current feed amount. Indicates the feed rate sensitivity coefficient; The formula used to calculate the cost of adjusting the motor output power is: in, Indicates the first The cost of adjusting the motor output power of a preferred control scheme Indicates the first The motor output power of a preferred control scheme This indicates the current motor output power. Indicates the power penalty coefficient; The formula used to calculate the cost of adjusting the air supply volume is: in, Indicates the first The cost of adjusting the air supply volume for each optimal control scheme Indicates the first The air supply volume of the preferred control scheme. Indicates the current air supply volume. Indicates the first The optimal control scheme for the air-to-powder ratio Indicates the current air-to-powder ratio; The formula for calculating the adjustment cost is: in, Indicates the first The adjustment cost of each preferred control scheme.

[0013] Furthermore, the principles upon which the target control scheme is based are as follows: The formula used to determine the overall priority of the optimal control scheme based on the adjustment cost and the calculation of superior performance indicators is as follows: in, Indicates the first The overall priority of the preferred control schemes Indicates the first The superior performance indicators of the optimal control scheme These represent the weighting coefficients for the superior indicators and the adjustment costs, respectively. and , This indicates adjusting the cost neutralization parameter, and ; Calculate the overall priority of all preferred control schemes, and select the preferred control scheme with the highest overall priority value as the target control scheme; The principle behind generating control commands is as follows: The differences between the parameters of the target control scheme and the current control scheme are calculated to generate the adjustment amount. The formula used is as follows: in, This indicates the amount of feed adjustment. This indicates the feed rate for the target control scheme. This indicates the amount of motor output power adjustment. This represents the motor output power of the target control scheme. This indicates the adjustment amount of the air supply volume. This indicates the air volume required for the target control scheme. Control commands are ; When the adjustment result is positive, it means that the corresponding parameter is increased based on the current control scheme; when the adjustment result is negative, it means that the corresponding parameter is decreased based on the current control scheme.

[0014] This invention also provides an intelligent control method for a pulverized coal vertical mill based on multi-parameter coordination. The method is executed by the aforementioned intelligent control system for a pulverized coal vertical mill based on multi-parameter coordination, and the specific steps include: Step 1: Randomly generate multiple feed rates and motor output powers within the range of mill feed rate and motor output power, and arrange and combine them to generate several sets of parameter combinations. For the feed rate in each parameter combination, randomly generate an air volume from the air volume range corresponding to that feed rate and add it to the parameter combination to form a control scheme. Step 2: Analyze the historical control schemes and corresponding excellent indicators of the mill in previous operations, and establish the mapping relationship between control schemes and excellent indicators; Step 3: Combining the mapping relationship, with the goal of maximizing the excellent index, the control scheme is optimized based on the particle swarm optimization algorithm to obtain several preferred control schemes; Step 4: Quantify the adjustment cost from the current control scheme to each optimal control scheme. Combine the adjustment cost with the superior index values ​​of each optimal control scheme to determine the target control scheme. Combine the current control scheme and the target control scheme to determine the control command.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention maps control schemes to optimal indicators based on historical control schemes. Compared to traditional control methods that often focus on a single objective, this scheme unifies output and energy consumption into a comprehensive evaluation index, guiding the system towards "high output and low consumption." By training the model using historical data, it can capture complex nonlinear relationships and adapt to different operating conditions and equipment states. The optimal control scheme is selected through particle swarm optimization, and the adjustment cost of adjusting from the current control scheme to each optimal control scheme is calculated. The cost mechanism suppresses large and high-risk parameter jumps, ensuring stable equipment operation and preventing problems such as increased vibration and shutdown. It suppresses large and high-risk adjustment behaviors and limits the optimal control scheme from the perspective of operational safety, making control commands smoother and more in line with the actual production rhythm, thus enhancing the feasibility and reliability of the system in actual industrial environments.

[0016] This invention also prioritizes the preferred schemes and selects the one with the highest overall priority as the target control scheme. This reflects the regulatory idea of ​​prioritizing performance while taking into account the adjustment costs. It pursues high performance while also ensuring the safety and smoothness of actual operation. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the system modules in an embodiment of the present invention; Figure 2 This is a fitted curve representing the cost of adjusting the feed rate in an embodiment of the present invention. Figure 3 This is a fitted curve representing the cost of adjusting the motor output power in an embodiment of the present invention. Figure 4 This is a schematic diagram of the method flow of an embodiment of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example: Please see Figures 1 to 4 The present invention provides a technical solution: A multi-parameter coordinated intelligent control system for a pulverized coal vertical mill specifically includes: The control scheme generation module is used to randomly generate multiple feed amounts and motor output powers within the range of feed amount and motor output power of the mill, and arrange and combine them to generate several sets of parameter combinations. For the feed amount in each parameter combination, an air volume is randomly generated from the air volume range corresponding to that feed amount and added to the parameter combination to form a control scheme. In this embodiment, the principle upon which the control scheme is based is: Based on the mill's rated parameters, the feed rate range and motor output power range are obtained. From each range, a feed rate value and a motor output power value are selected to generate a parameter combination. For each parameter combination, the corresponding airflow range for the feed rate is determined. The underlying principle is as follows: Under the same fineness standard, obtain all air volume values ​​corresponding to the same feed amount in the mill operation history data. The minimum value is the lower limit of the available air volume for that feed amount, and the maximum value is the upper limit of the available air volume for that feed amount. The range between the lower limit and the upper limit of the available air volume is the air supply range under that feed amount. During the operation of a vertical mill, the air supply volume plays a crucial role in carrying qualified fine particles out of the mill. Different fineness standards require different air volumes to achieve different screening capabilities. The finer the required fineness, the smaller the air volume is needed to avoid carrying out coal dust exceeding the fineness standard; the coarser the required fineness, the larger the air volume is needed to ensure sufficient airflow velocity to carry the coarser coal dust. When the feed rate increases, the amount of material being crushed also increases accordingly. To ensure that the feed can promptly carry away these crushed coal dust particles, the conveying capacity must be increased, i.e., the air supply volume must be increased. The amount of material and the air supply volume are positively correlated.

[0021] Based on the obtained parameter combination, the air supply range is determined according to the feed rate range, and an air supply rate is randomly selected and added to the parameter combination to generate a control scheme.

[0022] Under the same fineness standard, it means that the fineness screening standard is fixed. The fineness is calculated by comprehensively considering the fineness of all finished coal powder. Therefore, the air supply volume corresponding to meeting a fineness standard is not a fixed value. All air supply volumes that can meet the fineness standard are the air supply volume range corresponding to the feed amount.

[0023] The mapping analysis module is used to analyze the historical control schemes and corresponding excellent indicators of the mill in the past operation, and to establish the mapping relationship between the control schemes and the excellent indicators. In this embodiment, the excellent indicators consist of production efficiency and energy consumption efficiency, and the principle behind this is: The formula used to calculate output efficiency is: in, This represents the output efficiency of the historical control scheme over its longest effective period. This indicates the longest duration of action of any historical control scheme. This indicates the pulverized coal production during the longest possible duration of the historical control scheme. Production efficiency reflects the overall efficiency of a mill in producing qualified pulverized coal per unit time under a given control scheme. This represents the output of pulverized coal per unit time. This indicates that efficiency is adjusted by time; the shorter the effective time of the control scheme and the higher the pulverized coal output within that time, the higher the output efficiency. This reflects the duration of action, making the scheme tend towards long-term stable operation rather than short-term high productivity; the longer the duration of action, the better. The closer the value is to 1, the weaker the correction effect. When the duration of effect is short, The closer the value is to 0, the more unstable the control scheme may be, resulting in a shorter duration of action and thus reduced production efficiency. The longest duration of action represents the longest period of continuous and stable operation of the control scheme in its historical operation. The reason for considering the longest duration of action is to comprehensively evaluate the production performance and operational stability of the control scheme, avoid misleading the system with short-term high-yield but unsustainable schemes, and thus guide the optimization direction towards a long-term, stable, efficient, and safe production mode.

[0024] The formula used to calculate energy efficiency is: in, This represents the energy efficiency of the historical control scheme over its longest operating time. This represents the total energy consumption of the historical control scheme during its longest operating time. Energy efficiency reflects the energy consumed by the mill per unit time. The lower the energy efficiency, the lower the energy consumption per unit time, and the better the performance of the control scheme.

[0025] The formula used to calculate the excellent performance index is as follows: in, Indicators representing the excellence of historical control schemes.

[0026] The excellent index reflects the output efficiency that a certain control scheme can achieve under unit energy consumption. The larger the excellent index, the higher the coal powder output that can be obtained under unit energy consumption, the more economical the control scheme, and the better the operation effect. The smaller the excellent index, the higher the energy consumption and the lower the output, and the worse the operation effect of the control scheme. Pursuing high output alone may lead to a sharp increase in energy consumption, while pursuing low consumption alone may lead to a decrease in output. Therefore, the two are combined by ratio to generate the excellent index, which guides the optimization towards high output and low consumption.

[0027] The principle underlying the establishment of the mapping relationship between control schemes and optimal indicators is as follows: A deep learning model is established, using historical control schemes as input and corresponding superior indicators as labels to train a performance prediction model; the real-time generated control schemes are then input into the performance prediction model to obtain superior indicators that correspond one-to-one with the control schemes.

[0028] The structure of the constructed deep learning model is as follows: Input layer: Contains 3 neurons, used to input the feed rate, motor output power and air volume in the control scheme; The first hidden layer contains 64 neurons, activated using the ReLU activation function; The second hidden layer contains 32 neurons, activated using the ReLU activation function; Output layer: Contains 1 neuron, used to output excellent metrics; Select historical control schemes and obtain the corresponding operating time, pulverized coal production, and total energy consumption. Calculate the best performance indicators. Use the feed rate, motor output power, and air volume of the historical control schemes as input values, and the corresponding best performance indicators as labels. Use the operating data from the past 12-24 months as the training set, and the mean squared error function as the loss function. Minimize the loss function through the backpropagation algorithm. Select the operating data from the most recent 12 months of the historical control schemes as the validation set. Stop training when the loss function of the validation set no longer decreases, and save the model at this point as the performance prediction model.

[0029] The control scheme optimization module is used to combine the mapping relationship, with the goal of maximizing the excellent index, and optimize the control scheme based on the particle swarm optimization algorithm to obtain several preferred control schemes. In this embodiment, the principle underlying the optimization of the control scheme is as follows: Each particle corresponds to a control scheme, represented as follows: ,in, Indices representing particle indices Indicates the first Feed amount per particle Indicates the first The motor output power of each particle Indicates the first Air volume per particle; The constraints are: , , ,in, Indicates the lower limit of feed amount. Indicates the upper limit of feed amount. This indicates the lower limit of the motor's output power. This indicates the upper limit of the motor's output power. Indicates feed amount The corresponding lower limit of air supply volume, Indicates feed amount The corresponding upper limit of air supply volume; The initial position and initial velocity of each particle are randomly initialized, and the number of iterations is set. The particle velocity is updated in each iteration, based on the following formula: in, Indicates the first The particle in the first Speed ​​at the next iteration An index representing the number of iterations. Represents velocity inertia weight. Indicates the first The particle in the first The speed of each iteration They represent individual learning factors and social learning factors, respectively. , This indicates the random numbers used to avoid the algorithm getting trapped in local optima, and , Indicates the first The historical optimal position of a particle, that is, the position that maximizes the value of the excellent index during historical iterations. Indicates the first The particle in the first Position at the next iteration This represents the global optimal position, which is the position where all particles maximize the value of the performance index in the historical iterations. After each iteration, the particle moves to a new position based on the updated velocity, i.e. ,in, Indicates the first In the next iteration, the position of the particle is determined by the performance prediction model, which obtains the best performance index for each position and updates the historical best position of the particle. Each particle corresponds to a historical best position. The historical best positions of all particles are sorted in descending order according to the size of the best performance index, and the top 5% are retained. The purpose is to retain multiple excellent solutions and avoid the algorithm getting stuck in local optima. Moreover, the best performance index is only one of the factors in selecting the final control scheme. Therefore, the retention ratio is set to the top 5%, which fully considers the best performance index while leaving adjustment margin for the adjustment cost to be considered later.

[0030] The principle underlying the acquisition of several optimal control schemes is as follows: During the iteration process, the historical best position and global best position of all particles are counted, and they are sorted in descending order according to the size of the excellent index. The top 5% of particles are retained, and each particle corresponds to an optimal control scheme.

[0031] The control command generation module is used to quantify the adjustment cost from the current control scheme to each preferred control scheme, combine the adjustment cost with the superior index values ​​of each preferred control scheme, determine the target control scheme, and determine the control command by combining the current control scheme and the target control scheme.

[0032] In this embodiment, the principle upon which the adjustment cost is calculated is as follows: Obtain the control scheme of the mill at the current moment and set it as the current control scheme, which includes the current feed rate, the current motor output power and the current air volume; The formula used to calculate the cost of adjusting the feed rate is: in, Indicates the first The cost of adjusting the feed rate for each optimal control scheme Index indicating the preferred control scheme. Indicates the first The optimal control scheme for feed amount Indicates the current feed amount. Indicates the feed rate sensitivity coefficient; The cost of adjusting the feed rate reflects the risks and costs associated with changing the feed rate. Indicates the absolute magnitude of the adjustment. This indicates the maximum allowable range of variation in the feed rate throughout the entire operation. The adjustment must be within the maximum permissible range; and the larger the absolute magnitude of the adjustment, the higher the cost of adjusting the feed rate. The feed rate directly determines the thickness of the material bed within the mill. Significant and rapid changes in the feed rate will drastically disrupt the established solid-gas two-phase flow balance and grinding balance within the mill, easily leading to increased mill vibration, increased slag discharge, and even equipment shutdown. The higher the cost of adjusting the feed rate, the greater the risk of impact on the stable operation of the system. Changes in the feed rate need to be implemented through actuators (such as feeders and valves). Significant adjustments will intensify the amplitude and frequency of these mechanical components' movements, thereby increasing their mechanical wear and fatigue stress. Therefore, the cost of adjusting the feed rate also indirectly reflects the mechanical load cost incurred by the related mechanical equipment. A stable transition from one feed rate to another requires time. The larger the adjustment magnitude, the longer the time required to establish a new stable material bed. During this transition period, the fineness of the product may be unstable. The higher the cost of adjusting the feed rate, the longer the uncontrollable or unstable transition period in the production process. In actual production, there is a preference for smooth and gradual adjustments rather than drastic changes. The higher the cost of adjusting the feed rate, the less smooth the adjustment of the feed rate, and the more problems may arise. Since the feed rate directly determines the thickness of the material bed and the material flow state in the mill, its adjustment has a significant impact on the dynamic balance of the system. Therefore, a coefficient is needed to penalize large or frequent adjustments. Used to characterize the sensitivity to feed rate adjustments, and , The larger the value, the more sensitive the system is to changes in feed rate, and thus the better it can suppress large adjustments in feed rate. The score is determined by experts based on the mill's performance. Table 1 shows the relationship between the cost of adjusting the feed rate and the magnitude of the feed rate adjustment. The current feed rate is taken as 65 t / h (tons / hour), and the range of the feed rate is 40 t / h to 80 t / h. .

[0033] Table 1. Changes in the Cost of Adjusting Feed Rate The formula used to calculate the cost of adjusting the motor output power is: in, Indicates the first The cost of adjusting the motor output power of a preferred control scheme Indicates the first The motor output power of a preferred control scheme This indicates the current motor output power. Indicates the power penalty coefficient; The cost of adjusting motor output power reflects the comprehensive costs associated with adjusting the current motor output power to the target power, including changes in equipment load, operational risks, and energy consumption fluctuations. This is analogous to the cost of adjusting feed rate. This reflects that the larger the absolute range of motor output power adjustment, the higher the cost of motor output power adjustment. The cost of motor output power adjustment is directly proportional to the distance between the motor output power of the current control scheme and the motor output power of the optimal control scheme. The smaller the value, the easier it is to adjust the motor output power from the current control scheme to the output power of the preferred control scheme, and the lower the adjustment cost. Based on this, the following is introduced: To punish the adjustment cost in the high-power range, Reflects the first The distance between the motor output power and the upper limit of the motor output power in the preferred control scheme. The closer , The larger the value, the closer the adjusted motor output power is to the upper limit. Continuing to operate at this power level may lead to system shutdown and wear and tear on mechanical parts. Approaching hour The sharp increase reflects that any solution that pushes the motor's output power into dangerous areas is being prioritized for elimination. Used to adjust the system's sensitivity to high power. ,and The larger, The higher the power, the heavier the penalty; the setting... Based on the motor's rated power and the margin relative to the rated power during long-term stable operation, historical operating data should be retrieved to analyze the equipment failure rate in different power ranges. The failure rate is reflected by the number of motor overheat alarms. Different power ranges are set as below 70% of the rated power, 70% to 85%, 85% to 95%, and above 95% of the rated power. By comparing the motor failure rate in different power ranges in historical operating data, when the motor failure rate suddenly increases significantly, the corresponding power range is lower. The higher the value, the worse the equipment stability, and more power margin must be reserved. In actual production, the specific power margin should be determined based on the above principles and expert scoring method. Table 2 shows the fitted curves of motor output power as a function of the preferred control scheme, with the motor output power ranged from 500kW to 1500kW. The current motor output power is 1000kW. Table 2 shows that as the motor power in the preferred scheme increases, the cost of adjusting the motor output power first decreases and then increases. This is because when the motor output power of the current control scheme is adjusted to the preferred control scheme, it may increase or decrease. When the two are exactly equal, the minimum point of the cost of adjusting the motor output power is reached. After exceeding this minimum point, as the motor output power of the preferred scheme increases, the upward trend of the motor output power intensifies, reflecting the penalty when approaching the upper limit of the motor output power.

[0034] Table 2. Variation of Motor Output Power Adjustment Costs with Optimal Scheme Output Power The formula used to calculate the cost of adjusting the air supply volume is: in, Indicates the first The cost of adjusting the air supply volume for each optimal control scheme Indicates the first The air supply volume of the preferred control scheme. Indicates the current air supply volume. Indicates the first The optimal control scheme for the air-to-powder ratio Indicates the current air-to-powder ratio; Adjusting the airflow volume independently is less accurate; therefore, it must be combined with the feed rate. The airflow volume should be adjusted according to the air-to-powder ratio. If the feed rate changes, the airflow volume should also be adjusted proportionally; otherwise, it will lead to an imbalance in the air-to-powder ratio. Using the air-to-powder ratio difference can more accurately reflect changes in the system's gas-solid matching degree, rather than simply changes in the airflow volume. Even if the absolute change in airflow volume is large, as long as the new air-to-powder ratio is close to the current one, it indicates that the overall conveying and classification environment of the system is similar, and the adjustment cost is low. Conversely, even if the absolute change in airflow volume is small, a large change in the feed rate leading to a drastic change in the air-to-powder ratio will be very costly. This is directly related to the fineness of the finished product and the stability of the mill's internal operating conditions. This reflects the adjustment range of the air-to-powder ratio from the current control scheme to the optimal control scheme. The larger the adjustment range, the more prone the system is to instability, and therefore the greater the cost of adjusting the air supply volume.

[0035] The formula for calculating the adjustment cost is: in, Indicates the first The adjustment cost of each preferred control scheme.

[0036] Adjustment cost This reflects the overall cost of the j-th optimal control scheme considering feed rate adjustment, motor output power adjustment, and air volume adjustment. Increasing the value of any one of these costs—feed rate adjustment, motor output power adjustment, or air volume adjustment—will increase the overall system adjustment cost. and They are all directly proportional.

[0037] The principles upon which the target control scheme is determined are: The formula used to determine the overall priority of the optimal control scheme based on the adjustment cost and the calculation of superior performance indicators is as follows: in, Indicates the first The overall priority of the preferred control schemes Indicates the first The superior performance indicators of the optimal control scheme These represent the weighting coefficients for the superior indicators and the adjustment costs, respectively. and , This indicates adjusting the cost neutralization parameter, and ; The overall priority considers the contributions of both superior performance indicators and adjustment costs. A higher overall priority control scheme is selected more frequently. In actual production, to ensure both operational efficiency and safety, both superior performance indicators and adjustment costs are considered. Superior performance indicators reflect operational efficiency, while adjustment costs reflect operational safety. The system tends to favor schemes with higher superior performance indicators and lower adjustment costs. The larger the value, the higher the output of the scheme under the same energy consumption, or the lower the energy consumption under the same output, and the better the economy and operating efficiency. The smaller the value, the lower the switching cost from the current control scheme to the target control scheme, the higher the system stability, and the smoother the transition. The fundamental optimization goal of the scheme is to maximize output and minimize energy consumption while ensuring product quality. The role of adjustment cost is to prevent frequent and large fluctuations in the system and ensure equipment safety and stable operation. If the adjustment cost weight is too high, the system will tend to choose the scheme with "minimum changes but mediocre performance," failing to achieve a significant improvement in energy efficiency. Although the adjustment cost affects short-term stability, the long-term energy efficiency improvement and cost savings brought by a high-performance control scheme far outweigh the cost of a single adjustment. Therefore, performance should be given priority, and then the scheme with the lower adjustment cost should be selected from schemes with similar performance. ,Pick , , This represents the adjustment cost neutralization parameter, used to avoid a denominator of 0. If the current control scheme is exactly the same as the optimal control scheme, the adjustment cost is 0, indicating that no parameter adjustment is needed. Even if the adjustment cost is not 0, if its value is very small, then... It would be too large and would excessively amplify the impact of the adjustment cost. Therefore, the neutralization parameter is adjusted by adjusting the cost. To mitigate these two extreme cases and make the overall priority more smooth and reasonable, if the system is very sensitive to the cost of adjustment, it is desirable to avoid large-scale adjustments as much as possible. If the value is smaller, and the system is not sensitive to the cost of adjustment and focuses more on performance optimization, then... The value is determined based on an expert scoring method.

[0038] Calculate the overall priority of all preferred control schemes, and select the preferred control scheme with the highest overall priority value as the target control scheme; The principle behind generating control commands is as follows: The differences between the parameters of the target control scheme and the current control scheme are calculated to generate the adjustment amount. The formula used is as follows: in, This indicates the amount of feed adjustment. This indicates the feed rate for the target control scheme. This indicates the amount of motor output power adjustment. This represents the motor output power of the target control scheme. This indicates the adjustment amount of the air supply volume. This indicates the air volume required for the target control scheme. Control commands are ; When the adjustment result is positive, it means that the corresponding parameter is increased based on the current control scheme; when the adjustment result is negative, it means that the corresponding parameter is decreased based on the current control scheme.

[0039] Control commands represent the magnitude and direction of adjustments to various parameters of the current control scheme in order to achieve the target control scheme.

[0040] Please see Figure 4 The present invention also provides an intelligent control method for a pulverized coal vertical mill based on multi-parameter coordination. The method is executed by the aforementioned intelligent control system for a pulverized coal vertical mill based on multi-parameter coordination, and the specific steps include: Step 1: Randomly generate multiple feed rates and motor output powers within the range of mill feed rate and motor output power, and arrange and combine them to generate several sets of parameter combinations. For the feed rate in each parameter combination, randomly generate an air volume from the air volume range corresponding to that feed rate and add it to the parameter combination to form a control scheme. Step 2: Analyze the historical control schemes and corresponding excellent indicators of the mill in previous operations, and establish the mapping relationship between control schemes and excellent indicators; Step 3: Combining the mapping relationship, with the goal of maximizing the excellent index, the control scheme is optimized based on the particle swarm optimization algorithm to obtain several preferred control schemes; Step 4: Quantify the adjustment cost from the current control scheme to each optimal control scheme. Combine the adjustment cost with the superior index values ​​of each optimal control scheme to determine the target control scheme. Combine the current control scheme and the target control scheme to determine the control command.

[0041] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0042] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0043] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0044] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A smart control system for a pulverized coal vertical mill based on multi-parameter coordination, characterized in that, Specifically, it includes: The control scheme generation module is used to randomly generate multiple feed amounts and motor output powers within the range of feed amount and motor output power of the mill, and arrange and combine them to generate several sets of parameter combinations. For the feed amount in each parameter combination, an air volume is randomly generated from the air volume range corresponding to that feed amount and added to the parameter combination to form a control scheme. The mapping analysis module is used to analyze the historical control schemes and corresponding excellent indicators of the mill in the past operation, and to establish the mapping relationship between the control schemes and the excellent indicators. The control scheme optimization module is used to combine the mapping relationship, with the goal of maximizing the excellent index, and optimize the control scheme based on the particle swarm optimization algorithm to obtain several preferred control schemes. During the iteration process, the historical best position and global best position of all particles are counted, and they are sorted in descending order according to the size of the excellent index. The top 5% of particles are retained, and each particle corresponds to an optimal control scheme. The control command generation module is used to quantify the adjustment cost from the current control scheme to each preferred control scheme, combine the adjustment cost and the excellent index values ​​of each preferred control scheme to determine the target control scheme, and combine the current control scheme and the target control scheme to determine the control command. Excellent performance indicators consist of output efficiency and energy efficiency, based on the following principles: The formula used to calculate output efficiency is: in, This represents the output efficiency of the historical control scheme over its longest effective period. This indicates the longest duration of action of any historical control scheme. This indicates the pulverized coal production during the longest possible duration of the historical control scheme. The formula used to calculate energy efficiency is: in, This represents the energy efficiency of the historical control scheme over its longest operating time. This represents the total energy consumption of the historical control scheme during its longest operating time. The formula used to calculate the excellent performance index is as follows: in, Indicators representing the best performance of historical control schemes; The principle underlying the calculation of adjustment costs is as follows: Obtain the control scheme of the mill at the current moment and set it as the current control scheme, which includes the current feed rate, the current motor output power and the current air volume; The formula used to calculate the cost of adjusting the feed rate is: in, Indicates the first The cost of adjusting the feed rate for each optimal control scheme Index indicating the preferred control scheme. Indicates the first The optimal control scheme for feed amount Indicates the current feed amount. Indicates the feed rate sensitivity coefficient. Indicates the upper limit of feed amount. Indicates the lower limit of feed amount; The formula used to calculate the cost of adjusting the motor output power is: in, Indicates the first The cost of adjusting the motor output power of a preferred control scheme Indicates the first The motor output power of a preferred control scheme This indicates the current motor output power. Indicates the power penalty coefficient. This indicates the upper limit of the motor's output power. Indicates the lower limit of the motor's output power; The formula used to calculate the cost of adjusting the air supply volume is: in, Indicates the first The cost of adjusting the air supply volume for each optimal control scheme Indicates the first The air supply volume of the preferred control scheme. Indicates the current air supply volume. Indicates the first The optimal control scheme for the air-to-powder ratio Indicates the current air-to-powder ratio; The formula for calculating the adjustment cost is: in, Indicates the first The adjustment cost of a preferred control scheme; The principles upon which the target control scheme is determined are: The formula used to determine the overall priority of the optimal control scheme based on the adjustment cost and the calculation of superior performance indicators is as follows: in, Indicates the first The overall priority of the preferred control schemes Indicates the first The superior performance indicators of the optimal control scheme These represent the weighting coefficients for the superior indicators and the adjustment costs, respectively. and , This indicates adjusting the cost neutralization parameter, and ; Calculate the overall priority of all preferred control schemes, and select the preferred control scheme with the highest overall priority value as the target control scheme; The principle behind generating control commands is as follows: The differences between the parameters of the target control scheme and the current control scheme are calculated to generate the adjustment amount. The formula used is as follows: in, This indicates the amount of feed adjustment. This indicates the feed rate for the target control scheme. This indicates the amount of motor output power adjustment. This represents the motor output power of the target control scheme. This indicates the adjustment amount of the air supply volume. This indicates the air volume required for the target control scheme. Control commands are ; When the adjustment result is positive, it means that the corresponding parameter is increased based on the current control scheme; when the adjustment result is negative, it means that the corresponding parameter is decreased based on the current control scheme.

2. The intelligent control system for a pulverized coal vertical mill based on multi-parameter coordination according to claim 1, characterized in that: The principle upon which the control scheme is formed in the control scheme generation module is as follows: Based on the mill's rated parameters, the feed rate range and motor output power range are obtained. From each range, a feed rate value and a motor output power value are selected to generate a parameter combination. For each parameter combination, the corresponding airflow range for the feed rate is determined. The underlying principle is as follows: Under the same fineness standard, obtain all air volume values ​​corresponding to the same feed amount in the mill operation history data. The minimum value is the lower limit of the available air volume for that feed amount, and the maximum value is the upper limit of the available air volume for that feed amount. The range between the lower limit and the upper limit of the available air volume is the air supply range under that feed amount. Based on the obtained parameter combination, the air supply volume range is determined according to the feed volume range, and an air supply volume is randomly selected and added to the parameter combination to generate a control scheme.

3. The intelligent control system for a pulverized coal vertical mill based on multi-parameter coordination according to claim 1, characterized in that: The principle underlying the establishment of the mapping relationship between control schemes and optimal indicators in the mapping analysis module is as follows: A deep learning model is established, using historical control schemes as input and corresponding superior indicators as labels to train a performance prediction model; the real-time generated control schemes are then input into the performance prediction model to obtain superior indicators that correspond one-to-one with the control schemes.

4. The intelligent control system for a pulverized coal vertical mill based on multi-parameter coordination according to claim 1, characterized in that: The principle underlying the optimization of the control scheme in the control scheme selection module is as follows: Each particle corresponds to a control scheme, represented as follows: ,in, Indices representing particle indices Indicates the first Feed amount per particle Indicates the first The motor output power of each particle Indicates the first Air volume per particle; The constraints are: , , ,in, Indicates the lower limit of feed amount. Indicates the upper limit of feed amount. This indicates the lower limit of the motor's output power. This indicates the upper limit of the motor's output power. Indicates feed amount The corresponding lower limit of air supply volume, Indicates feed amount The corresponding upper limit of air supply volume; The initial position and initial velocity of each particle are randomly initialized, and the number of iterations is set. The particle velocity is updated in each iteration, based on the following formula: in, Indicates the first The particle in the first Speed ​​at the next iteration An index representing the number of iterations. Represents velocity inertia weight. Indicates the first The particle in the first The speed of each iteration They represent individual learning factors and social learning factors, respectively. , This indicates the random numbers used to avoid the algorithm getting trapped in local optima, and , Indicates the first The historical optimal position of a particle, that is, the position that maximizes the value of the excellent index during historical iterations. Indicates the first The particle in the first Position at the next iteration This represents the global optimal position, which is the position where all particles maximize the value of the performance index in the historical iterations.

5. A method for intelligent control of a pulverized coal vertical mill based on multi-parameter coordination, characterized in that: The method is executed by the intelligent control system for pulverized coal vertical mill based on multi-parameter coordination as described in any one of claims 1-4, and the specific steps include: Step 1: Randomly generate multiple feed rates and motor output powers within the range of mill feed rate and motor output power, and arrange and combine them to generate several sets of parameter combinations. For the feed rate in each parameter combination, randomly generate an air volume from the air volume range corresponding to that feed rate and add it to the parameter combination to form a control scheme. Step 2: Analyze the historical control schemes and corresponding excellent indicators of the mill in previous operations, and establish the mapping relationship between control schemes and excellent indicators; Step 3: Combining the mapping relationship, with the goal of maximizing the excellent index, the control scheme is optimized based on the particle swarm optimization algorithm to obtain several preferred control schemes; Step 4: Quantify the adjustment cost from the current control scheme to each optimal control scheme. Combine the adjustment cost with the superior index values ​​of each optimal control scheme to determine the target control scheme. Combine the current control scheme and the target control scheme to determine the control command.

Images