Powder grinding device and control method thereof

Abstract

The present application relates to the technical field of grinding equipment, and provides a powder grinding device and a control method thereof.The powder grinding device comprises a rack, a driving roller, a driven roller and a controller, the driving roller and the driven roller are rotatably installed on the rack, the controller is used for acquiring a network heat value of at least one defect keyword related to a grinding process of a target object, acquiring a reference roller speed ratio based on the network heat value, and controlling the relative rotation speed between the driving roller and the driven roller according to the reference roller speed ratio to grind the target object.The present application can collect and analyze relevant historical data based on network technology, can realize adaptive roller speed difference control, and can optimize the control strategy, thereby solving the problems that the existing roller mill lacks big data analysis and control strategy optimization capability and is difficult to realize adaptive roller speed difference control, and being beneficial to grinding and screening of powders such as sodium carboxymethyl cellulose, thereby improving the water-soluble effect.

Description

Technical Field

[0001] This invention relates to the field of grinding equipment technology, and more specifically, to a powder grinding device and its control method. Background Technology

[0002] For certain types of powders, grinding and pulverizing them before water dissolution can improve the dissolution effect, ensuring rapid and uniform dispersion in water to form a stable solution and avoid defects in use. Specifically, for powders such as sodium carboxymethyl cellulose, the grinding effect depends on the precise control of the roller gap and roller speed difference. If the roller speed difference is too small, the shearing force is insufficient, which not only fails to effectively pulverize the powder, resulting in lumps or flakes, but also easily causes the material to adhere to the roller surface, reducing the uniformity of particle size. If the roller speed difference is too large, the shearing force is too strong, which not only easily leads to over-grinding, producing a large amount of fine powder and agglomeration, but also easily causes local overheating of the roller surface and a sudden increase in equipment load. Therefore, the setting of the roller speed difference is crucial when controlling the grinding and pulverizing of powders using a roller mill.

[0003] Existing roller mills often use fixed speed difference or manual adjustment when setting roller speed difference. In essence, they are still based on local PLC control and adjustment. They do not use historical data collected by the network for analysis, lack the ability of big data analysis and control strategy optimization, and it is difficult to achieve adaptive roller speed difference control. Their automation and intelligence need to be further improved. Summary of the Invention

[0004] Based on this, in order to address the problem that existing roller mills lack the ability to perform big data analysis and control strategy optimization, making it difficult to achieve adaptive roller speed difference control, this invention provides a powder grinding device and its control method, the specific technical solution of which is as follows:

[0005] A powder grinding apparatus includes a frame, a drive roller, and a driven roller, both of which are rotatably mounted on the frame. The powder grinding apparatus also includes a controller, which is used to acquire the network heat value of at least one defect keyword related to the grinding process of the target object, acquire a reference roller speed ratio based on the network heat value, and control the relative rotational speed between the drive roller and the driven roller according to the reference roller speed ratio, so as to grind and sieve the target object.

[0006] The powder grinding device obtains the network heat value of defect keywords related to the grinding process of the target object, and controls the relative rotational speed between the active and driven rollers based on the reference roller speed ratio obtained from the network heat value. It can collect and analyze relevant historical data based on network technology, realize adaptive roller speed difference control, optimize control strategy, and solve the problem that existing roller mills lack the ability to perform big data analysis and control strategy optimization, making it difficult to achieve adaptive roller speed difference control. It is beneficial for grinding and screening powders such as sodium carboxymethyl cellulose, thereby improving their water solubility.

[0007] Preferably, the powder grinding apparatus further includes:

[0008] The particle size detection mechanism is used to obtain the real-time particle size of the target object after grinding and feed it back to the controller;

[0009] The controller is also used to obtain the particle size deviation based on the real-time particle size and the target particle size, obtain the adjustment amount of the basic roller speed based on the particle size deviation, and control the rotational speed of the driven roller based on the adjustment amount of the basic roller speed.

[0010] Preferably, the powder grinding apparatus further includes:

[0011] A temperature detection mechanism is used to acquire the real-time roller surface temperature and feed it back to the controller;

[0012] The controller is also used to obtain the temperature change rate based on the real-time roller surface temperature, adjust the proportional gain of the PID according to the proportional increment that is positively correlated with the network heat value, adjust the integral gain of the PID according to the integral increment that is positively correlated with the particle size deviation, and adjust the derivative gain of the PID according to the exponential decay value of the temperature change rate.

[0013] Preferably, the powder grinding apparatus further includes:

[0014] A current detection mechanism is used to acquire the real-time motor current of the driving roller and / or driven roller and feed it back to the controller;

[0015] The controller is also used to obtain the current change rate based on the real-time motor current, and when the current change rate is greater than the current change rate threshold, the real-time roller spacing is increased by using a number of times the minimum roller spacing adjustment accuracy as the spacing increment.

[0016] Preferably, the sieve disc includes a main sieve and a bottom sieve, and the powder grinding device further includes:

[0017] Screen discs, mounted on the frame and located below the drive roller and driven roller, include the main screen and the bottom screen;

[0018] The oversize detection mechanism is used to detect the percentage of oversize material and feed it back to the controller.

[0019] The undersize material detection mechanism is used to detect the percentage of undersize material and feed it back to the controller.

[0020] The controller is also used to reduce the real-time roller spacing by a first preset distance when the proportion of material on the screen is greater than a first preset threshold, and to increase the real-time roller spacing by a second preset distance when the proportion of material under the screen is greater than a second preset threshold.

[0021] A method for controlling a powder grinding apparatus, applied to the powder grinding apparatus, comprising:

[0022] Obtain the online popularity value of at least one defect keyword related to the grinding process of the target object;

[0023] The reference roller speed ratio is obtained based on the network popularity value;

[0024] The relative rotational speed between the driving and driven rollers is controlled according to the reference roller speed ratio in order to grind and screen the target object.

[0025] Preferably, the control method further includes:

[0026] Obtain the real-time particle size of the target object after grinding;

[0027] The particle size deviation is obtained based on the real-time particle size and the target particle size.

[0028] The adjustment amount of the base roller speed is obtained based on the particle size deviation, and the rotational speed of the driven roller is controlled based on the adjustment amount of the base roller speed.

[0029] Preferably, the control method further includes:

[0030] Obtain the real-time roller surface temperature and calculate the temperature change rate based on the real-time roller surface temperature.

[0031] The proportional gain of the PID is adjusted based on the proportional increment that is positively correlated with the network popularity value.

[0032] The integral gain of the PID controller is adjusted based on the integral increment that is positively correlated with the particle size deviation.

[0033] The derivative gain of the PID controller is adjusted based on the exponential decay of the rate of temperature change.

[0034] Preferably, the control method further includes:

[0035] Obtain the real-time motor current of the driving roller and / or driven roller;

[0036] The rate of change of current is obtained based on the real-time motor current.

[0037] When the rate of change of current is greater than the threshold of the rate of change of current, the real-time roller spacing is increased by using a number of times the minimum roller spacing adjustment accuracy as the spacing increment.

[0038] Preferably, the method for obtaining the reference roller speed ratio includes:

[0039] Calculate the normalized weights of the online popularity values ​​of multiple defective keywords respectively;

[0040] Define the default roll speed ratio, the priority of multiple normalized weights, and the range of roll speed ratios when each normalized weight is greater than the corresponding weight threshold;

[0041] If there is at least one normalized weight greater than the corresponding weight threshold, the reference roll speed ratio is obtained according to the roll speed ratio range corresponding to the highest priority normalized weight; otherwise, the default roll speed ratio is used as the reference roll speed ratio. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the overall structure of a powder grinding device according to an embodiment of the present invention. Figure 1 ;

[0043] Figure 2 This is a schematic diagram of the overall structure of a powder grinding device according to an embodiment of the present invention. Figure 2 ;

[0044] Figure 3 This is a schematic cross-sectional view of a powder grinding device according to an embodiment of the present invention;

[0045] Figure 4 This is a schematic diagram of the overall process of a powder grinding device control method according to an embodiment of the present invention;

[0046] Figure 5 This is a flowchart illustrating a method for obtaining a reference roller speed ratio in one embodiment of the present invention.

[0047] Figure 6 This is a schematic flowchart of a powder grinding device control method according to another embodiment of the present invention. Figure 1 ;

[0048] Figure 7 This is a schematic flowchart of a powder grinding device control method according to another embodiment of the present invention. Figure 2 ;

[0049] Figure 8 This is a schematic flowchart of a powder grinding device control method according to another embodiment of the present invention. Figure 3 ;

[0050] Figure 9 This is a flowchart of the coarse adjustment stage in one embodiment of the present invention;

[0051] Figure 10 This is a flowchart of the fine-tuning stage in one embodiment of the present invention.

[0052] Explanation of reference numerals in the attached figures:

[0053] 1. Frame; 2. Drive roller; 3. Driven roller; 4. Drive motor; 5. Screen plate; 6. Feed inlet. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to its embodiments.

[0055] Before describing the embodiments of the present invention, a brief introduction to the prior art will be given. Existing roller mills for grinding and pulverizing powders have shortcomings in roller speed difference control, particularly in the failure to fully utilize network data for defect analysis and optimization. These shortcomings are mainly reflected in the following aspects:

[0056] 1. Existing roller mills generally employ fixed-speed proportional or simple PID control strategies. These simplistic control strategies cannot dynamically adapt to complex operating conditions such as material characteristics and load variations. This static control strategy fails to consider real-time changes in parameters such as material particle size, hardness, and moisture content, and cannot dynamically adjust roller speed differences using network data, resulting in low grinding efficiency.

[0057] 2. Most roller mills still use local PLC control and are not connected to a factory-level IoT platform, making remote monitoring and data sharing impossible. Traditional PLC control systems, lacking network connectivity, require manual switching of control modes in abnormal operating conditions, leading to efficiency losses.

[0058] 3. Traditional roller mill control systems rely on manual experience for adjustment and cannot achieve adaptive roller speed difference control by training models using historical data. They do not utilize machine learning algorithms to train network data to achieve adaptive roller speed difference control. For example, they do not consider common problems in grinding powders, especially sodium carboxymethyl cellulose powder (adhesion, overheating, particle size unevenness, etc.) mentioned in user feedback or online discussions, making it difficult to automatically adjust the roller speed difference based on market and industry feedback to improve grinding efficiency.

[0059] For sodium carboxymethyl cellulose powder, the core function of pulverizing and grinding it before water dissolution is to optimize the dissolution process. The ultimate goal is to ensure that it disperses quickly and evenly in water to form a stable solution, avoiding defects in its use. The specific functions and purposes mainly include the following:

[0060] 1. Increased specific surface area and accelerated dissolution rate: After pulverization, the particle size of sodium carboxymethyl cellulose decreases, the total surface area increases significantly, and the contact area with water molecules expands accordingly. This allows for faster absorption of water and swelling and dissolution, greatly shortening the overall dissolution time and avoiding the problem of conventional large particles requiring long stirring times to dissolve.

[0061] 2. Prevent clumping and ensure uniform dissolution: When uncrushed sodium carboxymethyl cellulose granules (especially when the granules are large or clustered) are added directly to water, the surface will quickly absorb water and form a viscous "film," hindering the contact between the internal granules and water. This leads to clumping, where the granules are dry inside and stick together on the outside. Crushing the granules results in finer particles with better dispersibility, allowing for uniform contact with water, effectively preventing clumping, ensuring a consistent solution concentration, and eliminating localized viscous or undissolved particles.

[0062] 3. Improved solution stability and reduced defects: If sodium carboxymethyl cellulose is not dissolved uniformly (containing undissolved particles or clumps), subsequent applications are prone to system stratification, precipitation, or localized viscosity abnormalities, affecting product quality (e.g., uneven texture in food, or impaired film-forming properties in coatings). The homogeneous solution formed after pulverization maintains stable viscosity and dispersion, meeting the requirements of subsequent processing or use.

[0063] 4. Facilitates accurate metering and feeding: The pulverized sodium carboxymethyl cellulose particles are finer and have better flowability (over-pulverization should be avoided to prevent dust from flying), making it easier to accurately feed through metering equipment (such as weighing and conveying devices), reducing metering errors caused by uneven particle size and ensuring accurate formula ratios.

[0064] To address the lack of big data analysis and control strategy optimization capabilities in existing roller mills, which makes adaptive roller speed difference control difficult, such as... Figure 1 , Figure 2 as well as Figure 3 As shown, one embodiment of the present invention provides a powder grinding apparatus, including a frame 1, a drive roller 2, and a driven roller 3, both of which are rotatably mounted on the frame. Each drive roller and driven roller is equipped with a drive motor 4, which is connected to the corresponding drive roller and driven roller via a coupling.

[0065] The powder grinding apparatus also includes a controller, which acquires the network heat value of at least one defect keyword related to the grinding process of the target object, obtains a reference roller speed ratio based on the network heat value, and controls the relative rotational speed between the driving roller and the driven roller according to the reference roller speed ratio to grind and sieve the target object. The controller includes, but is not limited to, a PLC, and can be installed on the frame or in the central control room.

[0066] As a preferred technical solution, the target material is sodium carboxymethyl cellulose. Defect keywords related to the grinding process of the target material include, but are not limited to, roller adhesion, particle size unevenness, and overheating degradation. Here, a real-time data acquisition module connected to the controller can capture defect keywords from social media, forums, and customer feedback and feed them back to the controller. The controller converts the defect keyword-related data into online popularity values ​​and then maps them to a reference roller speed ratio. For example, if the adhesion heat is high, the controller automatically adjusts and reduces the roller speed ratio.

[0067] Generally, the online popularity value of user feedback on defects can, to some extent, characterize the grinding pain points of the target object and can serve as a reference for adjusting the grinding process parameters of the grinding equipment. The controller quantifies the weight of online public opinion on roller speed decisions by constructing a mapping model between the online popularity value and the reference roller speed ratio, which is beneficial for driving the adjustment of roller speed strategies. Roller surface adhesion can be obtained from industry forums, equipment work orders, and sensor anomaly logs; particle size unevenness can be collected and analyzed from data sources such as quality inspection report sharing platforms and customer complaint texts; and overheating degradation can be collected and analyzed based on data sources such as industry forums, infrared temperature sensor data clouds from quality inspection report sharing platforms, and energy consumption monitoring sharing platforms. For example, the online popularity value of roller surface adhesion = logarithm of daily discussion volume × failure rate; the online popularity value of particle size unevenness = (number of complaints / total output) × screening exceedance rate; and the online popularity value of overheating degradation = (duration of roller surface temperature above 50℃ / total working hours) × 0.6 + energy consumption increase × 0.4. It should be noted that the formula for obtaining the network heat value can be adjusted according to different factors such as the model of the grinding device, the initial particle size of the target object, and the feed flow rate. Its form includes, but is not limited to, linear functions, quadratic functions, indicator functions, and piecewise functions.

[0068] For example, the normalized weights of the network popularity values ​​of multiple defective keywords can be calculated separately, and the roller speed ratio range can be set based on the magnitude of the normalized weights. A reference roller speed ratio is then obtained based on the roller speed ratio range corresponding to the highest priority normalized weight. Here, the priority of roller surface adhesion is greater than that of particle size unevenness, and the priority of particle size unevenness is greater than that of overheating degradation. If the normalized weight corresponding to roller surface adhesion is greater than 0.4, the roller speed ratio range is set to [1.3, 1.5], and the reference roller speed ratio is set to the median of the range, 1.4. If the normalized weight corresponding to particle size unevenness is greater than 0.5, the roller speed ratio range is set to [1.2, 1.4], and the reference roller speed ratio is set to the median of the range, 1.3. If the normalized weight corresponding to overheating degradation is greater than 0.4, the roller speed ratio range is set to [1.1, 1.3], and the reference roller speed ratio is set to the median of the range, 1.2. The upper and lower limits of the roller speed ratio range can be adjusted based on actual conditions or by technical personnel based on experience, and are not limited here. In this way, by obtaining the reference roller speed ratio through network popularity values, multi-source, unstructured / semi-structured data (such as forum discussions, work orders, and sensor logs) can be transformed into comparable and calculable network popularity values. This solves the ambiguity problem of relying on experience to judge the severity of defects in traditional processes. Furthermore, it transforms ambiguous defects into quantitative indicators and quantitative indicators into precise roller speed ratio control commands. Ultimately, it transforms the roller grinding process of the target object from experience-driven to data-driven, optimizes the control strategy, and improves the grinding quality.

[0069] After obtaining the reference roller speed ratio, the rotational speeds of the driving and driven rollers can be adjusted based on the preset base roller speed, thereby controlling the relative rotational speeds between them. Specifically, the rotational speed of the driving roller is adjusted to the base roller speed, and then the rotational speed of the driven roller is adjusted according to the product of the base roller speed and the reference roller speed ratio. Of course, to make the rotational speeds of the driving and driven rollers more closely match the actual production scenario, the product of the base roller speed and the reference roller speed ratio can be used as the reference rotational speed of the driven roller. Then, based on the reference rotational speed, the actual rotational speed of the driven roller is gradually adjusted by a few revolutions per second or per minute. While adjusting the actual rotational speed of the driven roller, the actual grinding situation is observed, and the final rotational speed of the driven roller is determined based on the actual grinding quality and efficiency.

[0070] In summary, the powder grinding device obtains the network heat value of defect keywords related to the grinding process of the target object, and controls the relative rotational speed between the driving roller and the driven roller based on the reference roller speed ratio obtained from the network heat value. It can collect and analyze relevant historical data based on network technology, realize adaptive roller speed difference control, optimize the control strategy, and solve the problem that existing roller mills lack the ability to perform big data analysis and control strategy optimization, making it difficult to achieve adaptive roller speed difference control. This is beneficial for grinding and screening powders such as sodium carboxymethyl cellulose, thereby improving their water solubility.

[0071] As a preferred technical solution, the powder grinding device further includes a particle size detection mechanism. The particle size detection mechanism is used to acquire the real-time particle size of the target material after grinding and feed it back to the controller. The controller is also used to acquire the particle size deviation based on the real-time particle size and the target particle size, acquire the adjustment amount of the base roller speed based on the particle size deviation, and control the rotational speed of the driven roller based on the adjustment amount of the base roller speed. This particle size detection mechanism includes, but is not limited to, an online laser particle size analyzer and an online ultrasonic particle size analyzer, which can perform real-time particle size detection by collecting the target powder from the outlet 6 of the sieve plate 5.

[0072] In a sense, setting the roll speed ratio range based on normalized weights and obtaining a reference roll speed ratio belongs to the outer loop of strategy decision-making. The adjustment amount of the base roll speed obtained based on particle size deviation in this embodiment belongs to the inner loop of real-time optimization. Assuming the particle size deviation is expressed as e(t) = target particle size - real-time particle size, the adjustment amount ΔV(t) of the base roll speed can be expressed as...

[0073] K p (t), K i (t), K d (t) represent the proportional gain, integral gain, and derivative gain, respectively. Generally, K p (t), K i (t), K dThe initial values ​​of (t) can be set to 0.8, 0.5, and 0.5 respectively. The driving roller speed = the base roller speed, and the driven roller speed = the base speed × ΔV(t). For proportional gain, integral gain, and derivative gain, in some cases, parameter boundary constraints can be applied to limit the proportional gain to between 0.3 and 1.2, the integral gain to between 0.2 and 1.0, and the derivative gain to between 0.1 and 1.0.

[0074] Here, the adjustment amount of the basic roll speed is obtained by the particle size deviation. This is combined with the roll speed ratio range obtained based on normalized weight and the reference roll speed ratio to achieve synchronous optimization of roll speed difference and absolute rotational speed.

[0075] As a preferred technical solution, the powder grinding apparatus further includes a temperature detection mechanism. The temperature detection mechanism is used to acquire the real-time roller surface temperature and feed it back to the controller; wherein, the temperature detection mechanism includes, but is not limited to, an infrared thermal imager. The controller is also used to acquire the temperature change rate based on the real-time roller surface temperature, adjust the proportional gain of the PID controller based on a proportional increment positively correlated with the network heat value, adjust the integral gain of the PID controller based on an integral increment positively correlated with the particle size deviation, and adjust the derivative gain of the PID controller based on the exponential decay value of the temperature change rate.

[0076] For example, adjusting the proportional gain of the PID controller based on a proportional increment positively correlated with the network popularity value specifically involves adjusting the proportional gain of the PID controller based on a proportional increment positively correlated with the normalized weight of the network popularity value. For instance, the proportional gain can be adjusted based on a proportional increment positively correlated with the normalized weight corresponding to particle size unevenness. The adjusted proportional gain = proportional gain reference value × (1 + 0.2 × normalized weight corresponding to particle size unevenness). The adjusted integral gain = integral gain reference value × (1 + 0.3 × absolute value of particle size deviation), and the adjusted derivative gain = derivative gain reference value × exp(-0.05 × temperature change rate). The proportional gain reference value, integral gain reference value, and derivative gain reference value can be set to 0.8, 0.5, and 0.5, respectively.

[0077] In this embodiment, the target material is preferably sodium carboxymethyl cellulose. The glass transition temperature of sodium carboxymethyl cellulose refers to the temperature at which it transitions from a glassy state to a rubbery state or vice versa during heating or cooling. The glass transition temperature of sodium carboxymethyl cellulose is related to its molecular structure and chemical composition, and is also affected by processing conditions and measurement methods. Generally, the glass transition temperature of sodium carboxymethyl cellulose is relatively low, typically in the range of 50°C to 100°C.

[0078] In summary, the proportional gain responds to the thermal properties of particle size defects, primarily addressing the challenge of tough breakage in sodium carboxymethyl cellulose (CMC). The integral gain correlates with the absolute value of particle size deviation to overcome viscoelastic hysteresis. The derivative gain suppresses the rate of temperature rise, mitigating the risk of glass transition in CMC. For CMC, the PID parameter structure is dynamically reshaped using quantifiable material response characteristics, transforming polymeric material properties into control parameter constraints. This enables adaptive adjustment of the PID parameters, thereby improving the quality and efficiency of CMC powder grinding.

[0079] As a preferred technical solution, the powder grinding apparatus further includes a current detection mechanism. This current detection mechanism is used to acquire the real-time motor current of the driving roller and / or the driven roller and feed it back to the controller; the controller is also used to acquire the current change rate based on the real-time motor current, and when the current change rate is greater than a current change rate threshold, to increase the real-time roller spacing by using a multiple of the minimum roller spacing adjustment accuracy as the spacing increment.

[0080] The powder grinding apparatus also includes a roller spacing adjustment mechanism for adjusting the roller spacing. This mechanism can adjust the roller spacing via hydraulic drive, cylinder drive, or worm gear drive. For example, the driven roller can be used as a fixed roller, and the driving roller as a movable roller. The roller spacing adjustment mechanism drives the driving roller to move horizontally, thereby adjusting the roller spacing. Since the structure of the roller spacing adjustment mechanism is a conventional technique in this field, it will not be described in detail here.

[0081] Current surge is a key indicator of overload. When the rate of change of the drive motor current exceeds the threshold, such as 10%, the inter-roller pressure is considered excessive. This could be due to material accumulation or a sudden change in the viscoelasticity of the target object. In this case, increasing the real-time roll gap by using a multiple of the minimum roll gap adjustment precision as the gap increment can quickly alleviate overload while avoiding excessive relaxation. For example, assuming the minimum adjustment precision of the grinding device's roll gap is 0.01mm, when the rate of change of the drive motor current is greater than 10%, the gap increment can be set to 0.02mm to achieve a safe redundancy in the roll gap and provide overload protection for the drive motor. After increasing the roll gap by 0.02mm, run for several minutes, such as 2 or 3 minutes, and then check the rate of change of current again. If the rate of change of current is less than 5%, the current roll gap can be maintained; otherwise, the roll gap can be increased by 0.02mm again.

[0082] When adjusting the roll gap, the roll speed must be adjusted synchronously to avoid parameter conflicts. Specifically, when reducing the roll gap, the base roll speed can be reduced first, and the reference roll speed ratio can be slightly increased to compensate for the extrusion energy, enhance the shearing force, and prevent adhesion. When increasing the roll gap, the base roll speed can be increased first while maintaining the reference roll speed ratio to compensate for the material throughput in the roll gap and avoid material blockage caused by material accumulation.

[0083] As a preferred technical solution, such as Figure 1 as well as Figure 3 As shown, the powder grinding device also includes a sieve disc 5, an oversize detection mechanism, and an undersize detection mechanism. The sieve disc includes a main sieve and a bottom sieve, which are mounted on the frame and located below the drive roller and driven roller. The sieve disc is generally located directly below the drive roller and driven roller, and is used to sieve the target material after it has been ground and pulverized by the drive roller and driven roller. The aperture or mesh size of the main sieve and bottom sieve in the sieve disc can be set according to actual conditions. Also, the structure of the sieve disc is a conventional technical means in this field and will not be described in detail here. The oversize detection mechanism is used to detect the proportion of oversize material and feed it back to the controller, and the undersize detection mechanism is used to detect the proportion of undersize material and feed it back to the controller. Here, the oversize proportion refers to the proportion of target material that flows in from the feed inlet 6, passes through the gap between the rollers, and remains on the main sieve after being filtered by the main sieve, that is, the proportion of coarse particles that did not pass through the target particle size screen. If the aperture of the main sieve is 0.15mm, the oversize proportion = (weight of powder remaining on the 0.15mm main sieve screen, i.e., weight of oversize material) / total feed weight × 100%. The percentage of material passing through the sieve refers to the proportion of the target material after passing through the main sieve and the bottom sieve in sequence. In other words, it represents the proportion of the finer target material passing through the sieve with the minimum allowable particle size. If the bottom sieve aperture is 0.045mm, the percentage of material passing through the sieve = the weight of powder remaining on the 0.045mm bottom sieve (i.e., the weight of material passing through the sieve) / the total feed weight.

[0084] ×100%.

[0085] The oversize material detection mechanism includes, but is not limited to, multiple weight detection modules (such as industrial-grade load cells) installed at the bottom of the main screen. These modules collect the total weight of the main screen and the oversize material in real time, and the weight of the oversize material is obtained by subtracting the unloaded weight of the main screen from the total weight. The total feed mass can be obtained by a total feed rate detection module installed at the feed inlet. This module collects the total material flow rate entering the grinding device in real time, and the total feed weight is obtained by accumulating the data. The undersize material detection mechanism includes, but is not limited to, multiple weight detection modules (such as load cells) installed at the bottom of the bottom screen. These modules collect the total weight of the bottom screen and the undersize material in real time, and the weight of the undersize material is obtained by subtracting the unloaded weight of the bottom screen from the total weight. Because the screen disc is vibrating, the weights of the oversize and undersize materials obtained by the oversize material detection mechanism and the undersize material detection mechanism will fluctuate. Therefore, the weights of the oversize and undersize materials can be collected at a certain sampling frequency, and the final weights can be obtained by calculating the average weights of the oversize and undersize materials. Of course, a moving average algorithm (taking the average of 5 consecutive sampling points) can also be used to eliminate the periodic vibration interference caused by the sieve plate and improve the accuracy of the weight of the material on the sieve and the weight of the material under the sieve.

[0086] Alternatively, one can first obtain the weight of the undersize material using a beta-ray sensor and the beta-ray absorption method, then obtain the total feed weight, and finally determine the undersize percentage based on the weight of the oversize material and the total feed weight. In this case, the undersize detection mechanism is a beta-ray sensor. Since measuring powder weight using the beta-ray absorption method is a conventional technique in this field, it will not be elaborated upon here.

[0087] The controller is further configured to reduce the real-time roller spacing by a first preset distance when the proportion of oversize material is greater than a first preset threshold, and to increase the real-time roller spacing by a second preset distance when the proportion of undersize material is greater than a second preset threshold. If the proportion of oversize material is greater than the first preset threshold and the proportion of undersize material is greater than the second preset threshold, the real-time roller spacing is reduced by the first preset distance.

[0088] Specifically, when the proportion of oversize material exceeds a first preset threshold, such as 5%, the target material flowing out of the screen outlet may contain particles with excessively large sizes, which can easily reduce its solubility and, to some extent, indicate that the real-time roller spacing is too large and / or the shearing force is insufficient. In this case, it is necessary to trigger a real-time roller spacing reduction action. For example, when the proportion of oversize material is between 5% and 8%, the first preset spacing can be set to 0.01mm, 0.02mm, or 0.03mm, and the real-time roller spacing is reduced accordingly. When the proportion of oversize material exceeds 8%, the first preset spacing is set to 0.05mm, the real-time roller spacing is reduced by 0.05mm, and the base roller speed is increased by 5% to enhance the shearing force.

[0089] When the proportion of undersize material exceeds the second preset threshold, it can easily trigger a disaster in the ultrafine dust process. A surge in ultrafine dust can lead to dust explosions and also result in excessively fine target particle sizes, making them prone to agglomeration. Furthermore, it can cause over-grinding of the powder and / or an imbalance in the roller speed ratio. In this case, an emergency adjustment is necessary. For example, when the undersize material proportion is between 30% and 35%, the second preset gap can be set to 0.01 mm or 0.02 mm, and the real-time roller gap can be increased accordingly. Simultaneously, the reference roller speed ratio can be reduced (e.g., by 0.1) to decrease the number of compression cycles and weaken the shearing force. If the undersize material proportion exceeds 35%, the machine should be stopped immediately for cleaning to interrupt the dust chain reaction and prevent a dust explosion.

[0090] An embodiment of the present invention also provides a control method for a powder grinding apparatus, applied to the powder grinding apparatus, such as... Figure 4 As shown, it includes the following steps:

[0091] S1. Obtain the network popularity value of at least one defect keyword related to the grinding process of the target object. Here, defect keywords include, but are not limited to, roller adhesion, uneven particle size, and overheating degradation. Data sources for roller adhesion include, but are not limited to, industry forums (technical discussions), equipment work orders (fault records), and sensor anomaly logs (real-time vibration / current). The network popularity value of roller adhesion = lg(daily discussion volume + 1) × failure rate, which integrates "common industry risks" and "actual faults in the plant", and uses multi-source cross-validation to prevent misjudgment. Daily discussion volume can be understood as the average daily occurrence of forum keywords "adhesion" and "scaling", reflecting the popularity of common industry problems and avoiding reliance solely on industry data while deviating from reality. Failure rate = number of adhesion-related work orders this month ÷ total number of days the equipment is running, used to quantify the actual frequency of occurrence in the plant. The logarithmic function lg can avoid extreme values ​​interfering with the judgment and smooth the fluctuation of discussion volume. Assuming that the discussion volume on a certain day is 100 and the failure rate is 0.05, then the network popularity value of roller adhesion = lg(101) × 0.05 ≈ 0.1.

[0092] The online popularity score for particle size unevenness is calculated as (number of customer complaints in the current month / total output of qualified products) × screening exceedance rate. This score reflects both external customer dissatisfaction and the degree of process control failure. Data sources for particle size unevenness include the quality inspection report sharing platform (hourly screening data) and customer complaint text analysis (using NLP to extract keywords such as "particles" and "caking"). The screening exceedance rate can be understood as the percentage of batches with a sieve oversize ratio >5% or a sieve undersize ratio >30% in real-time testing, directly reflecting the degree of process control failure. Assuming 5 complaints / month, output = 10 tons, and exceedance rate = 0.08, then the online popularity score for particle size unevenness is (5 / 10) × 0.08 = 0.04.

[0093] The network heat value of overheating degradation = (duration of roller surface temperature above 50℃ / total working hours) × 0.6 + energy consumption increase × 0.4. This is mainly used to balance physical and chemical damage through a weighted sum, comprehensively monitor overheating risks, and avoid misjudgment based on a single indicator. Total working hours refer to the equipment operating time. 50℃ can be understood as the preset maximum grinding roller surface temperature. The duration of roller surface temperature above 50℃ / total working hours is used to quantify the proportion of time at which thermal damage risk occurs. Energy consumption increase = (current unit energy consumption - baseline energy consumption) / baseline energy consumption × 100%. For example, if the duration of roller surface temperature above 50℃ is 30 minutes, the total duration is 60 minutes, the current unit energy consumption - baseline energy consumption is 5.2 kWh, and the baseline energy consumption is 40 kWh, then the network heat value of overheating degradation = 0.6 × (30 / 60) + 0.4 × (5.2 / 40) = 0.3 + 0.052 = 0.352. Here, the target object refers to sodium carboxymethyl cellulose.

[0094] S2, obtain the reference roller speed ratio based on the network heat value.

[0095] As a preferred technical solution, such as Figure 5 As shown, in step S2, the method for obtaining the reference roller speed ratio includes the following steps:

[0096] S21, calculate the normalized weights of the online popularity values ​​of multiple defective keywords. The formula for the normalized weights is expressed as follows: Among them, W k α represents the normalized weight of the k-th type of defect keyword. k H represents the weighting coefficient of the online popularity value of the k-th type of defect keyword. k This represents the online popularity value of the keyword for the k-th type of defect. For example, the keywords for the 1st, 2nd, and 3rd types of defects are roller adhesion, particle size unevenness, and overheating degradation, respectively, with online popularity weight coefficients of 0.35, 0.4, and 0.25 for these three categories, respectively. In the normalized weighting formula, the numerator is mainly used to couple the static weight coefficient α. k The denominator is normalized to the coupling result of the numerator and the real-time network heat value (reflecting the cost of long-term loss) and the instantaneous risk level.

[0097] S22 defines the default roll speed ratio, the priority levels of multiple normalized weights, and the range of roll speed ratios when each normalized weight exceeds its corresponding weight threshold. The default roll speed ratio can be set by technicians based on experience, such as 1.1 or 1.2.

[0098] For example, the normalized weight of roller surface adhesion has a higher priority than the normalized weight of particle size unevenness, and the normalized weight of particle size unevenness has a higher priority than the normalized weight of overheated degradation. Roller speed ratio range ΔV range Defined as Where W1, W2, and W3 are the normalized weights of the network heat values ​​corresponding to roller surface adhesion, particle size unevenness, and overheating degradation, respectively. When max(W k When the value exceeds a preset normalized weight threshold, such as 0.6 or 0.7, an alarm is triggered, and engineers are notified via SMS or other means. Alternatively, more resources may be allocated to monitor the grinding process of the target object. This allocation of resources includes, but is not limited to, increasing the frequency of grinding parameter (such as real-time motor current and real-time roller surface temperature) acquisition and increasing the memory of the controller used for analyzing and processing grinding parameters. By enhancing the monitoring intensity, the grinding process of the target object can be monitored to avoid false alarms and missed detections, thereby improving the response speed to grinding defects.

[0099] S23. If there is at least one normalized weight greater than the corresponding weight threshold, then the reference roll speed ratio is obtained according to the roll speed ratio range corresponding to the highest priority normalized weight; otherwise, the default roll speed ratio is used as the reference roll speed ratio. Here, obtaining the reference roll speed ratio according to the roll speed ratio range corresponding to the highest priority normalized weight specifically means obtaining the reference roll speed ratio according to the lower limit, upper limit, or median of the roll speed ratio range.

[0100] S3 controls the relative rotational speed between the driving roller and the driven roller according to the reference roller speed ratio in order to perform grinding treatment on the target object.

[0101] In summary, the powder grinding device control method obtains the network heat value of defect keywords related to the grinding process of the target object, and controls the relative rotational speed between the driving roller and the driven roller based on the reference roller speed ratio obtained from the network heat value. It can collect and analyze relevant historical data based on network technology, realize adaptive roller speed difference control, optimize the control strategy, and solve the problem that existing roller mills lack the ability to perform big data analysis and control strategy optimization, making it difficult to achieve adaptive roller speed difference control. This is beneficial for grinding and screening powders such as sodium carboxymethyl cellulose, thereby improving their water solubility.

[0102] As a preferred technical solution, such as Figure 6 As shown, the control method further includes:

[0103] S4, Obtain the real-time particle size of the target object after grinding. Here, the real-time particle size can be obtained by sampling and analyzing the target object at the screen outlet using an online particle size analyzer.

[0104] S5 calculates the particle size deviation based on the real-time particle size and the target particle size. Particle size deviation = target particle size - real-time particle size. The target particle size can be set according to actual production conditions / customer requirements.

[0105] S6: Obtain the adjustment amount of the base roller speed based on the particle size deviation, and control the rotational speed of the driven roller based on the adjustment amount of the base roller speed.

[0106] For example, PID regulation of the base roller speed based on particle size deviation specifically involves: according to the function The adjustment amount of the base roller speed is obtained, and the driven roller speed is obtained based on the adjustment amount. Driven roller speed = base roller speed × adjustment amount, driving roller speed = base roller speed. In this embodiment, according to the function... The adjustment amount of the base roll speed is obtained, and combined with the roll speed ratio range obtained by normalization weight and the reference roll speed ratio, it can simultaneously optimize the roll speed difference and absolute rotational speed between the driving roll and the driven roll.

[0107] The method for obtaining the base roller speed specifically includes the following steps:

[0108] The first step is to collect historical data and standardize it. Historical data includes historical feed flow rate, historical target capacity, and base roller speed. To reduce noise and fluctuations, the historical feed flow rate can be sampled as an average over several minutes, such as one minute.

[0109] The second step is to construct the feed flow rate Q, target capacity η, and basic roller speed V. base The relationship model is established. For example, the relationship model is expressed as: Base roller speed = Target capacity / (Feed flow rate × Calibration constant). The calibration constant is obtained by fitting the relationship model using the least squares method.

[0110] The third step is to obtain the real-time feed flow rate and then obtain the basic roller speed based on the fitted relationship model.

[0111] The aforementioned model for obtaining the basic roller speed is simple, efficient, and computationally inefficient, making it suitable for real-time control and highly adaptable. However, considering that the calibration constant is typically affected by roller diameter, roller gap, and material grindability index, and that the relationship between target capacity, basic roller speed, and feed flow rate is usually non-linear, the aforementioned model for obtaining the basic roller speed may contain some errors. To improve the accuracy of the basic roller speed, a non-linear model can be constructed between the target capacity, basic roller speed, and feed flow rate. For example, the non-linear model is expressed as follows: Here, α and β represent the roll speed index and the feed index, respectively. A nonlinear model is fitted based on collected historical data to obtain the roll speed index, feed index, and calibration constant. Then, the base roll speed is obtained based on the fitted and calibrated nonlinear model. This nonlinear model... It overcomes the limitations of linear models and can obtain more accurate basic roll speed values.

[0112] As a preferred technical solution, such as Figure 7 As shown, the control method further includes:

[0113] S7, obtain the real-time roller surface temperature T, and obtain the temperature change rate based on the real-time roller surface temperature. Preferably, the target object refers to sodium carboxymethyl cellulose.

[0114] S8 adjusts the proportional gain of the PID controller based on the proportional increment that is positively correlated with the network popularity value. Specifically, the proportional gain of the PID controller is adjusted based on the proportional increment that is positively correlated with the normalized weight of the network popularity value.

[0115] In classical PID theory, the proportional gain is proportional to the system's response to deviations. Addressing the core issue of uneven particle size in grinding, an increase in the normalized weight corresponding to this unevenness indicates, to some extent, an excess of coarse particles on the screen, necessitating increased roller shear force to accelerate crushing. However, excessively increasing the proportional gain can easily lead to vibration or overheating of the roller surface. Therefore, based on the principle of linear superposition, a preset proportional gain reference value K can be used... p0 This is based on a proportional increment that is positively correlated with the normalized weights corresponding to particle size unevenness. For example, the proportional gain K... p (t)=K p0 ·(1+β p •W2). The scaling gain baseline is typically set to 0.8. Experimental data shows that for every 0.1 increase in the normalized weight W2 corresponding to particle size inhomogeneity, the scaling gain needs to increase by 2%-4% to significantly improve the particle size distribution. Therefore, the coefficient β... p The range can be set to 0.2-0.4. Of course, the coefficient β... p The value can also be determined by technicians based on experience, according to actual production conditions.

[0116] S9 adjusts the integral gain of the PID based on the integral increment that is positively correlated with the particle size deviation.

[0117] The integral term is used to eliminate steady-state error, but the viscoelastic memory effect of sodium carboxymethyl cellulose leads to the following: with large deviations, the integral is too strong → continuous accumulation of roller pressure → material agglomeration; with small deviations, the integral is insufficient → particle size fluctuations cannot converge. Therefore, it is necessary to dynamically suppress the integral gain based on the absolute value of the particle size deviation. For example, the integral gain K... i (t)=(1+γ i ·|e(t)|)·K i0 Among them, K i0 This represents the baseline value of the integral gain, typically set to 0.5. The coefficient γ i This is an empirical value, which can be calibrated by technicians based on experience, and is generally set to 0.3. Assuming a particle size deviation of 1.0 mm, the coefficient γ... i If the integral gain is 0.3, then the integral gain is 1.3 times the baseline integral gain, and its integral strength is increased by 30%.

[0118] S10 adjusts the differential gain of the PID controller based on the exponential decay of the temperature change rate.

[0119] The glass transition temperature of sodium carboxymethyl cellulose (SMC) refers to the temperature at which it transitions from a glassy state to a rubbery state or vice versa during heating or cooling. The glass transition temperature of SMC is related to its molecular structure and chemical composition, and is also affected by processing conditions and measurement methods. Generally, the glass transition temperature of SMC is relatively low, typically in the range of 50°C to 100°C. Here, the differential action primarily suppresses dynamic disturbances. However, based on the glass transition characteristics of SMC, the differential action needs to be weakened when the roller surface temperature rises rapidly to avoid molecular chain breakage caused by sudden changes in roller speed. When the temperature stabilizes, the differential gain reference value is restored to improve the response speed.

[0120] For example, differential gain Among them, K d0 Represents the differential gain reference value, with coefficient λ. d This is an empirical value and is generally set to 0.05. The greater the rate of temperature change, the stronger the attenuation of the differential gain, thereby suppressing high-frequency oscillations of the roller speed, protecting the material structure, and preventing overheating and degradation. When the rate of temperature change is zero, the differential gain equals the differential gain reference value.

[0121] In summary, the proportional gain responds to the thermal properties of particle size defects, primarily addressing the challenge of tough breakage in sodium carboxymethyl cellulose (CMC); the integral gain correlates with the absolute value of particle size deviation to overcome viscoelastic hysteresis; and the derivative gain is used to suppress the rate of temperature rise, mitigating the risk of glass transition in CMC. By dynamically reshaping the PID parameter structure through quantifiable material response characteristics, the properties of the polymer material are transformed into control parameter constraints, enabling adaptive adjustment of the PID parameters and thus improving the quality and efficiency of CMC powder grinding.

[0122] As a preferred technical solution, such as Figure 8 As shown, the control method further includes:

[0123] S11, Obtain the real-time motor current of the drive roller and / or driven roller. Here, the real-time motor current can be the average of the real-time current of the drive roller drive motor and the real-time current of the driven roller drive motor, or it can be the real-time current value of either the drive roller or driven roller drive motor.

[0124] S12, obtain the current change rate based on the real-time motor current.

[0125] S13, when the rate of change of current is greater than the threshold of the rate of change of current, the real-time roller spacing is increased by using a number of times the minimum roller spacing adjustment accuracy as the spacing increment.

[0126] Assuming the minimum adjustment accuracy of the grinding device's roller spacing is 0.01mm, when the rate of change of the drive motor current exceeds 10%, the spacing increment can be set to 0.02mm to achieve a safety redundancy in the roller spacing and provide overload protection for the drive motor. After increasing the roller spacing by 0.02mm and running for several minutes, such as 2 or 3 minutes, the rate of change of current is checked again. If the rate of change of current is less than 5%, the roller spacing can be maintained at this point; otherwise, the roller spacing can be increased by 0.02mm again. In this way, by monitoring the real-time rate of change of the motor current, overload protection for the drive motor can be provided, preventing the drive motor from burning out due to overload.

[0127] As a preferred technical solution, the adjustment of the roller gap includes a coarse adjustment stage and a fine adjustment stage. The coarse adjustment stage aims to initially eliminate serious defects affecting the normal operation of the equipment (such as material slippage and brittleness), laying the foundation for precise optimization in the subsequent fine adjustment stage. Figure 9 As shown, during the coarse adjustment stage, the state of the material (target object) is monitored in real time. If the material slips and does not get caught in the roller gap, it indicates that after contacting the working roller, the material cannot be driven into the equipment by the friction of the roller surface due to insufficient clamping force between the rollers, resulting in sliding and accumulation. In this case, the roller gap needs to be reduced, such as by 0.05-0.1mm each time. If the material is compressed into lumps, it indicates that the material is excessively squeezed by the working roller, exceeding its physical tolerance limit, forming lumps with excessive density and irregular shape. In this case, the roller gap needs to be increased, such as by 0.03mm to 0.08mm each time. After completing a single coarse adjustment of the roller gap, a stable operation observation period is entered. The observation time can be set to 5 minutes. This step is to avoid misjudgment due to instantaneous adjustment and to verify long-term stability.

[0128] Instantaneous adjustment misjudgment refers to the lag in material movement within the equipment after a single spacing adjustment. Immediately judging the effect may lead to misjudgment because the material has not fully flowed through the working area. For example, when the spacing is immediately reduced, the material that was previously accumulated may still be slipping. It takes 5 minutes for the new material to be fully entrained before it can be confirmed that the slippage has been eliminated. Verifying long-term stability refers to the fact that some materials, especially highly hygroscopic powders such as sodium carboxymethyl cellulose, may initially be normal and then temporarily clump together. A 5-minute operation period can preliminarily verify whether the adjusted parameters can continuously adapt to the material handling requirements, rather than just solving instantaneous problems.

[0129] After 5 minutes of stable operation, the system enters the effect verification stage to determine whether the current adjustment has eliminated the defects, thus forming a closed-loop logic. If the defects are not eliminated, it indicates that the magnitude of the single roller spacing adjustment is insufficient or there is a deviation in direction. The roller spacing is then adjusted further until the defects are eliminated. If the defects are eliminated, it means that the current spacing parameters have been initially adapted to the material characteristics, and defects that seriously affect equipment operation (such as slippage and agglomeration) have been resolved. The coarse adjustment stage ends, and the equipment can enter the fine adjustment stage.

[0130] like Figure 10 As shown, the fine-tuning stage is the core precision optimization step following the coarse-tuning stage. Its core objective is to control the material particle size within a preset qualified range (usually requiring a qualified rate ≥90%) by making minute and dynamic adjustments to key operating parameters such as roller spacing, after eliminating serious defects such as "slippage and clumping" in the coarse-tuning stage. This ultimately achieves stable and efficient production of the grinding device. The fine-tuning stage mainly involves real-time monitoring of the screening results and drive motor current, specifically real-time monitoring of the proportion of material on the screen, the proportion of material under the screen, and the rate of change of current, to make minute adjustments to the roller spacing.

[0131] When the oversize content exceeds 5%, the target material flowing out of the screen outlet may contain particles with excessively large sizes, which can reduce its solubility and, to some extent, indicate that the real-time roller gap is too large and / or the shear force is insufficient. In this case, it is necessary to trigger a real-time roller gap reduction action. For example, the roller gap can be reduced by 0.01 mm, 0.02 mm, or 0.03 mm each time to slightly increase the extrusion / grinding intensity, ensuring that large-size materials are fully processed and reducing the oversize content.

[0132] When the undersize content exceeds 30%, it can easily lead to a disaster in the ultrafine dust process. A surge in ultrafine dust can potentially cause a dust explosion, and it also results in an excessively fine final particle size, making it prone to agglomeration. Furthermore, it can contribute to over-grinding of the powder and / or an imbalance in the roller speed ratio. In such cases, the roller gap can be increased by 0.01mm or 0.02mm each time to slightly reduce the inter-roller compression / grinding intensity, prevent excessive material breakage, and reduce the generation of undersize material.

[0133] When the current change rate of the drive motor is greater than 10%, the roller spacing can be increased by 0.02mm in one go to achieve safe redundancy of the roller spacing and provide overload protection for the drive motor.

[0134] After a single minor adjustment, the equipment needs to enter a 10-minute stable operation period. This step aims to eliminate hysteresis and verify stability. After the 10-minute stable operation period, the screening results and current change rate are re-tested. For example, if the pass rate is ≥90%, it indicates that the current roller spacing and other parameters have accurately adapted to the material characteristics, and the particle size control meets production requirements. At this point, the roller spacing needs to be locked to ensure that the roller spacing and other parameters do not change arbitrarily in subsequent production, and the fine-tuning stage is officially over. If the pass rate is less than 90%, it indicates that the single minor adjustment has not completely solved the problems of excessive oversize material ratio, undersize material ratio, or current change rate. It is necessary to return to the screening results and current detection stage until the pass rate reaches the standard.

[0135] In this way, by working together with the coarse and fine adjustment stages, the process parameters of the powder grinding device can be better adjusted, which is conducive to quickly and stably obtaining powder that meets the qualification rate requirements and ultimately optimizing the powder water solubility effect.

[0136] As a preferred technical solution, the control method further includes calculating the mill health based on the proportion of oversize material and the proportion of undersize material. For example, mill health = (0.9 - Relu(W' - 5%)) / (1 + f(W2)), where... W1 and W2 represent the percentage of material oversize and undersize, respectively. Specifically, the ideal range for the percentage of material oversize is 4%-6%. If it exceeds a certain value, such as 8%, it indicates wear on the roller gaps or insufficient shearing force. The percentage of material undersize is less than 28%. If it exceeds a certain value, such as 35%, it indicates over-grinding, a sharp increase in energy consumption, and a risk of dust explosion.

[0137] In the roller mill health model, the numerator (0.9 - Relu(W1 - 5%)) quantifies the particle tolerance, linearly penalizing excessive coarse particles. 5% can be understood as the target value for the percentage of material remaining on the screen. 0.9 is a preset baseline health value, which can be adjusted according to actual conditions, such as changing it to 0.85 or 0.95. In Relu(W1 - 5%), Relu is the activation function. When the value of the percentage of material remaining on the screen (-5%) is greater than zero, it returns the value of the percentage of material remaining on the screen (-5%); otherwise, it returns zero. In other words, negative inputs are truncated to 0, while positive inputs retain their original values. When the value of the percentage of material remaining on the screen (-5%) is greater than zero, it means that for every 1% of the target percentage of material remaining on the screen, the health value is reduced by 0.1.

[0138] The denominator amplifies the excessive fine powder content through an exponential effect, applying a non-linear penalty to this excessive fine powder. Essentially, this simulates the exponential increase in energy consumption and the risk of equipment overheating caused by over-grinding. 30% is a preset value and can be adjusted according to actual conditions, for example, changing it from 30% to 28%. When the undersize content is less than 30%, the proportion of excessive fine powder meets the standard, and no further health assessment is applied. When the undersize content is 30%, the denominator is 2, and the health assessment decreases by 50%. When the undersize content is 42%, the denominator is approximately 3, and the health assessment decreases by 67%.

[0139] The health status of the roller mill can be mapped to the equipment status to construct a health grading response mechanism. The specific mapping relationship is shown in the table below:

[0140]

[0141]

[0142] The roller mill health model transforms complex mechanical conditions into quantifiable process parameter functions, enabling the assessment of the grinding unit's mechanical condition based on the proportion of oversize and undersize materials. It pioneers a mapping relationship between oversize and undersize materials and equipment health indicators. When the roller mill health falls below a certain value, such as 0.6, a maintenance work order can be triggered to prevent damage to the roller mill due to the continuous accumulation of mechanical wear.

[0143] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A powder grinding apparatus, comprising a frame, a drive roller, and a driven roller, wherein the drive roller and the driven roller are rotatably mounted on the frame, characterized in that, The powder grinding apparatus also includes a controller, which is used to acquire the network heat value of at least one defect keyword related to the grinding process of the target object, acquire a reference roller speed ratio based on the network heat value, and control the relative rotation speed between the driving roller and the driven roller according to the reference roller speed ratio in order to perform grinding processing on the target object; The target object is sodium carboxymethyl cellulose. Defect keywords related to the grinding process of the target object include roller surface adhesion, particle size unevenness, and overheating degradation. The defect keywords in social media, forums, and customer feedback are captured and fed back to the controller. The controller converts the defect keyword-related data into online popularity values. Methods for obtaining the reference roll speed ratio include: Calculate the normalized weights of the online popularity values ​​of multiple defective keywords respectively; Define the default roll speed ratio, the priority of multiple normalized weights, and the range of roll speed ratios when each normalized weight is greater than the corresponding weight threshold; If there is at least one normalized weight that is greater than the corresponding weight threshold, the reference roll speed ratio is obtained according to the roll speed ratio range corresponding to the highest priority normalized weight; otherwise, the default roll speed ratio is used as the reference roll speed ratio. After obtaining the reference roller speed ratio, the rotational speeds of the driving roller and the driven roller are adjusted based on the preset base roller speed, thereby controlling the relative rotational speeds between the driving roller and the driven roller. Specifically, the rotational speed of the driving roller is adjusted to the base roller speed, and then the rotational speed of the driven roller is adjusted according to the product between the base roller speed and the reference roller speed ratio.

2. The powder grinding apparatus as described in claim 1, characterized in that, The powder grinding apparatus further includes: The particle size detection mechanism is used to obtain the real-time particle size of the target object after grinding and feed it back to the controller; The controller is also used to obtain the particle size deviation based on the real-time particle size and the target particle size, obtain the adjustment amount of the basic roller speed based on the particle size deviation, and control the rotational speed of the driven roller based on the adjustment amount of the basic roller speed.

3. The powder grinding apparatus as described in claim 2, characterized in that, The powder grinding apparatus further includes: A temperature detection mechanism is used to acquire the real-time roller surface temperature and feed it back to the controller; The controller is also used to obtain the temperature change rate based on the real-time roller surface temperature, adjust the proportional gain of the PID according to the proportional increment that is positively correlated with the network heat value, adjust the integral gain of the PID according to the integral increment that is positively correlated with the particle size deviation, and adjust the derivative gain of the PID according to the exponential decay value of the temperature change rate.

4. The powder grinding apparatus as described in claim 3, characterized in that, The powder grinding apparatus further includes: A current detection mechanism is used to acquire the real-time motor current of the driving roller and / or driven roller and feed it back to the controller; The controller is also used to obtain the current change rate based on the real-time motor current, and when the current change rate is greater than the current change rate threshold, the real-time roller spacing is increased by using a number of times the minimum roller spacing adjustment accuracy as the spacing increment.

5. The powder grinding apparatus as described in claim 4, characterized in that, The powder grinding apparatus further includes: Screen discs, mounted on the frame and located below the drive roller and driven roller, include the main screen and the bottom screen; The oversize detection mechanism is used to detect the percentage of oversize material and feed it back to the controller. The undersize material detection mechanism is used to detect the percentage of undersize material and feed it back to the controller. The controller is also used to reduce the real-time roller spacing by a first preset distance when the proportion of material on the screen is greater than a first preset threshold, and to increase the real-time roller spacing by a second preset distance when the proportion of material under the screen is greater than a second preset threshold.

6. A method for controlling a powder grinding apparatus, applied to the powder grinding apparatus as described in any one of claims 1-5, characterized in that, The control method includes: Obtain the online popularity value of at least one defect keyword related to the grinding process of the target object; The reference roller speed ratio is obtained based on the network popularity value; The relative rotational speed between the driving roller and the driven roller is controlled according to the reference roller speed ratio in order to grind and screen the target object. The target object is sodium carboxymethyl cellulose. Defect keywords related to the grinding process of the target object include roller surface adhesion, particle size unevenness, and overheating degradation. The defect keywords in social media, forums, and customer feedback are captured and fed back to the controller. The controller converts the defect keyword-related data into online popularity values. Methods for obtaining the reference roll speed ratio include: Calculate the normalized weights of the online popularity values ​​of multiple defective keywords respectively; Define the default roll speed ratio, the priority of multiple normalized weights, and the range of roll speed ratios when each normalized weight is greater than the corresponding weight threshold; If there is at least one normalized weight that is greater than the corresponding weight threshold, the reference roll speed ratio is obtained according to the roll speed ratio range corresponding to the highest priority normalized weight; otherwise, the default roll speed ratio is used as the reference roll speed ratio. After obtaining the reference roller speed ratio, the rotational speeds of the driving roller and the driven roller are adjusted based on the preset base roller speed, thereby controlling the relative rotational speeds between the driving roller and the driven roller. Specifically, the rotational speed of the driving roller is adjusted to the base roller speed, and then the rotational speed of the driven roller is adjusted according to the product between the base roller speed and the reference roller speed ratio.

7. The control method for a powder grinding apparatus as described in claim 6, characterized in that, The control method further includes: Obtain the real-time particle size of the target object after grinding; The particle size deviation is obtained based on the real-time particle size and the target particle size. The adjustment amount of the base roller speed is obtained based on the particle size deviation, and the rotational speed of the driven roller is controlled based on the adjustment amount of the base roller speed.

8. The control method for a powder grinding apparatus as described in claim 7, characterized in that, The control method further includes: Obtain the real-time roller surface temperature and calculate the temperature change rate based on the real-time roller surface temperature. The proportional gain of the PID is adjusted based on the proportional increment that is positively correlated with the network popularity value. The integral gain of the PID controller is adjusted based on the integral increment that is positively correlated with the particle size deviation. The derivative gain of the PID controller is adjusted based on the exponential decay of the rate of temperature change.

9. The control method for a powder grinding apparatus as described in claim 8, characterized in that, The control method further includes: Obtain the real-time motor current of the driving roller and / or driven roller; The rate of change of current is obtained based on the real-time motor current. When the rate of change of current is greater than the threshold of the rate of change of current, the real-time roller spacing is increased by using a number of times the minimum roller spacing adjustment accuracy as the spacing increment.

Images