An intelligent control method for inhibiting scaling and compensating heat in a clean coal drying process

Abstract

This invention discloses an intelligent control method for scaling inhibition and thermal compensation in the coal drying process, belonging to the field of coal drying technology. The method includes: real-time acquisition of operating parameters such as heat medium temperature, material temperature, wall temperature, motor current, feed rate, moisture content, and heat medium flow rate; calculation of equivalent fouling thermal resistance and determination of scaling levels I, II, and III; applying periodic disturbances to the cylinder rotation speed to remove the initial adhesion layer at level I; increasing the heat medium temperature at level II, utilizing the near-wall vapor film effect to inhibit adhesion and perform thermal compensation; triggering pulsed thermal shock and injecting scale-inhibiting medium at level III; using feedforward-feedback composite control to correct heat medium parameters based on the discharge moisture content and superimposing scaling thermal compensation increments; establishing a scaling growth rate model to predict scaling trends and provide early warnings. This invention achieves closed-loop control of online scaling perception, graded inhibition, and thermal compensation, reducing the number of downtime cleanings, stabilizing product moisture content, and lowering drying energy consumption.

Description

Technical Field

[0001] This invention relates to the field of coal drying technology, and in particular to an intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process, applicable to the intelligent operation control of continuous clean coal drying equipment such as drum-type clean coal dryers and steam pipe rotary clean coal dryers. Background Technology

[0002] Clean coal is a clean coal product with low ash content, low sulfur content, and high calorific value obtained after washing and processing raw coal. It is widely used as coking coal blending, blast furnace injection fuel, and chemical raw material. The moisture content of flotation clean coal and filter press clean coal is typically as high as 20% to 30%, with high viscosity and poor flowability. Before storage, transportation, and blending, it must be dried to reduce the moisture content to below 8% to 12% to meet the stringent requirements of coking coal input and subsequent processes. Clean coal drying commonly employs continuous drying equipment such as rotary drum dryers (direct hot air heat exchange) and steam tube rotary dryers (indirect steam heat exchange). These equipment have large processing capacities and can operate continuously, making them key equipment in coal preparation plants and coking plants.

[0003] However, due to the fine particle size of flotation clean coal (a high proportion of particles smaller than 0.5 mm), the presence of a certain amount of clay minerals, and the complex surface properties, wet clean coal is extremely prone to adhering to and gradually accumulating hard during the drying process on the inner wall of the drying cylinder, the surface of the lifting plates, and the surface of the heat exchange tubes, forming a scale layer. The formation mechanism of the scale layer mainly includes: the adhesion and deposition of wet clean coal particles under the action of liquid bridging force and van der Waals force on the wall surface, the dehydration and hardening of the adhesion layer at high temperature, and the continuous accumulation of fine particles on the receiving surface of the lifting plates. Among them, the liquid bridging force is the dominant force for the adhesion of wet fine particles to the wall surface. When there is a capillary water film between the particles and the wall surface, the surface tension and Laplace pressure difference of the water film cause the particles to be firmly adsorbed to the wall surface. As the wall temperature rises and the water evaporates, the adhesion layer gradually loses water and hardens, and the bonding strength with the wall surface further increases, forming a stubborn scale layer that is difficult to remove. Furthermore, clay minerals in clean coal (such as kaolinite and montmorillonite) swell upon contact with water and exhibit strong binding properties, significantly exacerbating the scaling tendency. The higher the coal's bonding index (G value), the stronger the bonding between coal particles and between coal particles and the wall surface, resulting in a faster scaling rate. Once scaling forms, it significantly increases the thermal resistance of the cylinder wall or heat exchange tubes, drastically reduces the heat transfer coefficient, decreases drying capacity, and increases energy consumption per ton of coal. On the other hand, it reduces the effective volume of the cylinder, increases the rotational load on the cylinder, leading to increased current in the drive motor. In severe cases, this can cause motor overload tripping, discharge port blockage, and forced shutdown for manual cleaning. Manual cleaning is labor-intensive, carries high safety risks, and results in significant downtime losses, becoming a major bottleneck restricting the continuous and efficient operation of clean coal drying systems.

[0004] To address the issues of material sticking and adhering to the walls of rotary drum dryers, various structural improvement solutions have been developed in the prior art. Chinese patent document CN109870000A discloses a self-cleaning device and system for rotary drum dryers of high-moisture, high-viscosity materials. This system comprises, sequentially along the axial direction within the rotating drum, a feeding guide zone, an anti-sticking crushing and lifting zone, an anti-sticking forming lifting zone, a general lifting zone, and a discharge guide zone. Movable grate-type lifting plates, movable lifting plates, and fixed lifting plates are arranged alternately, with chains suspended outside the fixed lifting plates. The mechanical structure crushes and lifts the material to form a material curtain, preventing material adhesion to the inner wall of the drum. This solution is a purely mechanical anti-sticking method. Although it can reduce initial adhesion to some extent, it has the following shortcomings: First, it cannot detect the degree of scaling on the cylinder wall in real time, and the scaling status can only be judged by experience or by stopping the machine for inspection; Second, once the adhesion layer is formed and hardened, it still requires manual cleaning after stopping the machine, and cannot be removed online; Third, it does not consider the heat transfer deterioration caused by scaling, lacks corresponding operating condition compensation methods, and the drying thermal efficiency continues to decrease as the operating time increases.

[0005] Regarding temperature control during the drying process, Chinese patent document CN101726171A discloses a method for temperature regulation and uniform control of a multi-pass drum dryer. This method involves setting up a pressure equalization air distribution ring, an air distribution throttling group, and an external / middle / internal mixing air duct at the hot air inlet. A closed-loop regulation system is constructed using a temperature sensor group, an industrial control computer, a servo amplifier, and a servo motor to achieve separate regulation and uniform control of the heating airflow temperature in each pass. This scheme focuses on the uniformity of hot air temperature, but does not address the sensing of scale buildup on the drum wall, nor does it include heat compensation methods based on the degree of scale buildup. Therefore, it cannot address the deterioration of heat transfer performance caused by scale buildup. Chinese patent document CN104482751A discloses a method and apparatus for eliminating the thermal inertia of a hot air dryer. This method adjusts the ratio of flue gas volume to drying medium using a proportional valve. When the temperature is low, the opening of the proportional valve is increased to increase the flue gas volume and decrease the drying medium volume, thereby eliminating the thermal inertia of the hot air dryer and improving temperature control sensitivity. This solution only addresses the lag in temperature response; its control objective remains temperature itself, without establishing any connection with scaling inhibition actions.

[0006] Regarding scale removal, Chinese patent document CN106867555A discloses a method for cleaning deposits inside a coal moisture-conditioning steam rotary dryer. This method, upon determining an increase in deposits between the heat exchange tubes of the dryer, stops supplying coal and cuts off steam. First, it idles at a frequency of 50Hz to discharge automatically discharged coal. Then, it adds a spiral reverse discharge port, rotating slowly at 30Hz while injecting cleaning fluid to achieve automatic cleaning. Finally, it supplements this with manual cleaning of residues and preheating to resume production. A single cleaning cycle requires 8-10 hours. This solution is a passive, reactive shutdown cleaning method, resulting in long downtime, significant capacity loss, and the inability to inhibit scale formation online. Scale is often not addressed in its early stages, ultimately requiring shutdown maintenance.

[0007] Regarding online fouling monitoring, Chinese patent document CN1903746A discloses a real-time online monitoring and control device and method for fouling thermal resistance. This method calculates the total heat transfer coefficient and fouling thermal resistance of the heat exchanger in real time based on the heat transfer temperature difference method, thus achieving online monitoring of fouling thermal resistance. However, this solution is geared towards general applications such as shell-and-tube heat exchangers, only providing fouling monitoring and alarm functions. It does not couple fouling thermal resistance with the graded fouling suppression action of the drying process, does not involve heat compensation control based on the degree of fouling, and does not consider the prediction of fouling trends by coal quality parameters. Therefore, it is difficult to directly apply to fouling suppression in the clean coal drying process.

[0008] Furthermore, Chinese patent document CN106403563A discloses a drying system that intelligently controls the coal feed rate based on the outlet temperature, adjusting the coal feed rate according to a single parameter of outlet temperature. This scheme uses a single control parameter and cannot comprehensively reflect the drying conditions under the coupling of scaling and heat transfer deterioration, making it difficult to ensure stable control of the output moisture content.

[0009] In summary, existing technologies have the following shortcomings: First, they lack online quantitative characterization methods for the scaling state of the heat exchange surface of the clean coal drying cylinder; the degree of scaling can only be judged by shutdown inspection or experience. Second, scaling inhibition mainly relies on mechanical anti-sticking mechanisms or post-operation shutdown cleaning, lacking online graded inhibition strategies and failing to select matching inhibition methods based on the degree of scaling. Third, temperature control and thermal compensation are not coupled with the scaling state, making it impossible to provide targeted thermal compensation to maintain product moisture stability when scaling leads to a decrease in the heat transfer coefficient; instead, it may cause overheating, coking, or even spontaneous combustion of the clean coal due to blind heating. Fourth, they lack scaling trend prediction and early warning methods based on coal quality parameters and operating data; cleaning plans are mostly formulated on a fixed cycle and cannot adapt to changes in scaling rate caused by fluctuations in coal quality. Therefore, there is an urgent need for an intelligent control method for clean coal drying that can achieve online sensing, graded inhibition, thermal compensation, and closed-loop control of trend prediction for scaling. Summary of the Invention

[0010] To address the shortcomings of existing technologies, the purpose of this invention is to provide an intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process. This method solves the technical problems in existing technologies, such as the inability to quantitatively detect scaling on the heat exchange surface of the clean coal drying cylinder online, the lack of graded inhibition methods, the lack of coupling between thermal compensation and scaling status, and the lack of scaling trend prediction and early warning. It achieves closed-loop intelligent control of online detection, graded inhibition, thermal compensation, and trend prediction of scaling in the clean coal drying process, reducing the number of shutdowns for cleaning, stabilizing the output moisture content, reducing drying energy consumption, and improving the safety and economy of system operation.

[0011] The above-mentioned objective of this invention is achieved through the following technical solutions:

[0012] To achieve the above objectives, the present invention provides an intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process, comprising the following steps:

[0013] Step S1: Real-time acquisition of operating parameters of the clean coal drying system. The operating parameters include at least the inlet temperature of the heat medium T1, the outlet temperature of the heat medium T2, the inlet temperature of the material Tm1, the outlet temperature of the material Tm2, ​​the cylinder wall temperature Tw, the drive motor current I, the feed rate G, the feed moisture content M1, the outlet moisture content M2, and the heat medium flow rate Qh.

[0014] Step S2: Calculate the equivalent fouling thermal resistance Rf of the heat exchange surface of the cylinder based on the operating parameters collected in step S1, and determine the current fouling level based on the equivalent fouling thermal resistance Rf and its rate of change over time dRf / dt. The fouling level includes Level I light fouling, Level II moderate fouling and Level III heavy fouling.

[0015] Step S3: When the scaling level is determined to be Class I mild scaling, a mild scaling inhibition strategy is implemented: keep the inlet temperature T1 of the hot medium constant, apply periodic disturbance to the cylinder rotation speed n at the first disturbance frequency, and synchronously adjust the speed of the dispersing shaft according to the preset ratio, and use the material curtain inside the cylinder to periodically flush the wall surface to remove the initial soft adhesive layer.

[0016] Step S4: When the scaling level is determined to be Class II moderate scaling, a moderate scaling suppression strategy is implemented: Based on the mild scaling suppression strategy, the inlet temperature T1 of the heat medium is increased to the first compensation temperature range [Tc1, Tc2], so that the wall temperature Tw of the cylinder is higher than the boiling point of water in the wet clean coal. The steam layer generated by the rapid vaporization of the near-wall water film inhibits the adhesion of wet clean coal to the wall. At the same time, the heat compensation increment ΔQ is calculated according to the equivalent fouling thermal resistance Rf, and the heat compensation increment ΔQ is used to feedforward compensation of the heat medium flow rate Qh.

[0017] Step S5: When it is determined that the scaling level is level III severe scaling, a severe scale inhibition strategy is executed: trigger a pulse thermal shock program, so that the heat medium inlet temperature T1 rises to the second compensation temperature range according to a preset pulse curve within a preset pulse period and then falls back, while reducing the feed rate G to 60%~70% of the rated feed rate, and injecting a scale inhibition medium into the cylinder.

[0018] Step S6: Taking the discharge moisture content M2 as the target variable, a feedforward-feedback composite control is adopted to correct the heat medium inlet temperature T1 and the heat medium flow rate Qh in real time. The feedforward term of the composite control is calculated according to the change rates of the feed rate G and the feed moisture content M1, the feedback term is obtained through PID operation according to the deviation between the discharge moisture content M2 and the target moisture content set value Mset, and the total output of the composite control is superimposed on the basis of the thermal compensation increment ΔQ applied in steps S3 to S5.

[0019] Step S7: A scaling growth rate model is established according to coal quality parameters and historical operation data, and the scaling growth rate model is used to predict the predicted time point when the scaling level reaches level II moderate scaling within a future preset period. When the predicted time point is earlier than the preset cleaning cycle, an early warning message is generated and the scale inhibition strategy corresponding to the predicted scaling level is executed in advance.

[0020] Further, in step S2, the equivalent fouling thermal resistance Rf is calculated as follows: the actual heat transfer coefficient K of the heat exchange surface of the cylinder is calculated according to the operating parameters, and the equivalent fouling thermal resistance Rf is calculated according to the formula Rf=1 / K-1 / K0, where K0 is the calibrated reference heat transfer coefficient in a clean state; the criteria for determining the scaling level are: when Rf<R1 and dRf / dt<v1, it is determined as level I mild scaling; when R1≤Rf<R2 and dRf / dt<v2, it is determined as level II moderate scaling; when Rf≥R2 or dRf / dt≥v2, it is determined as level III severe scaling; where R1 is the first threshold, R2 is the second threshold, v1 is the first change rate threshold, v2 is the second change rate threshold, and R1<R2, v1<v2.

[0021] Further, the actual heat transfer coefficient K is calculated according to the formula K=Q / (A·ΔTlm), where Q is the actual heat transfer capacity of the heat exchange surface of the cylinder, A is the heat exchange area of the cylinder, and ΔTlm is the logarithmic mean temperature difference between the heat medium and the material; the actual heat transfer capacity Q is calculated according to the heat medium flow rate Qh, the heat medium density ρ, the heat medium specific heat capacity cp, and the temperature difference between the heat medium inlet temperature T1 and the heat medium outlet temperature T2 according to the formula Q=Qh·ρ·cp·(T1-T2).

[0022] Furthermore, in step S3, the first disturbance frequency is 0.2~0.5 times / minute, the disturbance amplitude of the cylinder rotation speed n is ±3%~±8% of the rated cylinder rotation speed, the adjustment amplitude of the dispersing shaft rotation speed is ±5%~±12% of the rated dispersing shaft rotation speed, the duration of a single disturbance is 2~5 minutes, and the periodic disturbance continues before the scaling level changes.

[0023] Further, in step S4, the first compensation temperature range [Tc1, Tc2] is a temperature range 20~40℃ higher than the current inlet temperature T1 of the heat medium, and the lower limit Tc1 of the first compensation temperature range satisfies the requirement that the cylinder wall temperature Tw is higher than 120℃; the heat compensation increment ΔQ is calculated according to the formula ΔQ=Rf·A·ΔTlm·k, where k is the heat transfer loss conversion factor, with a value of 0.8~1.2; when the heat compensation increment ΔQ is used to feedforward compensate the heat medium flow rate Qh, the compensation increment ΔQh of the heat medium flow rate is calculated according to the formula ΔQh=ΔQ / (ρ·cp·(T1-T2)).

[0024] Further, in step S5, the preset pulse curve is a trapezoidal pulse curve, the heating period of the trapezoidal pulse curve is 3-5 minutes, the inlet temperature T1 of the heat medium is raised to the second compensation temperature range, the second compensation temperature range is a temperature range 50-80°C higher than the current inlet temperature T1 of the heat medium, the holding time is 2-3 minutes, the cooling period is 3-5 minutes, and the number of pulse cycles is 2-4 times; the scale inhibitor is fine-grained limestone powder or water-based scale inhibitor, and the injection amount of the scale inhibitor is 0.5%-1.5% of the effective volume of the cylinder.

[0025] Further, in step S6, the feedforward term Ff is calculated according to the formula Ff=Kf1·dG / dt+Kf2·dM1 / dt, where Kf1 is the feedforward coefficient for the change in feed rate and Kf2 is the feedforward coefficient for the change in feed moisture content; the feedback term Fb is calculated according to the incremental PID algorithm; the total output Tset of the composite control on the inlet temperature of the heat medium is calculated according to the formula Tset=Ts0+ΔTf+ΔTb+ΔTcomp, where Ts0 is the basic setpoint of the inlet temperature of the heat medium, ΔTf is the feedforward temperature correction, ΔTb is the feedback temperature correction, and ΔTcomp is the scaling heat compensation temperature. The total output Qhset of the composite control for the heat medium flow rate is calculated according to the formula Qhset=Qh0+ΔQf+ΔQb+ΔQh_comp, where Qh0 is the basic setpoint for the heat medium flow rate, ΔQf is the feedforward flow correction, ΔQb is the feedback flow correction, and ΔQh_comp is the scaling heat compensation flow correction. The scaling heat compensation flow correction is calculated from the heat compensation increment ΔQ. The composite control also includes upper temperature limit protection. When the heat medium inlet temperature T1 exceeds the safe upper limit temperature Tmax, the feedback term gain is reduced and the growth rate of the heat medium flow rate Qh is limited.

[0026] Further, in step S7, the scaling growth rate model takes coal quality parameters and operating parameters as inputs and scaling index FI as output. The scaling index FI is the ratio of the equivalent fouling thermal resistance Rf to the second threshold R2, i.e., FI = Rf / R2. The coal quality parameters include ash content Ad, total moisture Mt, particle size distribution, and adhesion index G. The operating parameters include the inlet temperature of the heat medium T1, the cylinder rotation speed n, and the feed rate G. The scaling growth rate model adopts a multiple linear regression model or a BP neural network model, uses historical operating data for offline training to obtain initial model parameters, and updates the model parameters online using a rolling window method during operation. The predicted time point is the time point obtained by extrapolating the scaling index FI within a future preset period to the time corresponding to the judgment threshold of level II moderate scaling.

[0027] Furthermore, the heat medium is flue gas or steam; when the heat medium is flue gas, the inlet temperature T1 of the heat medium is adjusted by adjusting the mixing ratio of flue gas and cold air, and the flow rate Qh of the heat medium is adjusted by adjusting the frequency of the induced draft fan or the opening of the mixing valve; when the heat medium is steam, the inlet temperature T1 of the heat medium and the flow rate Qh of the heat medium are adjusted by adjusting the opening of the steam regulating valve.

[0028] Furthermore, the method also includes a dynamic adjustment step for the cleaning cycle: based on the scaling trend output by the scaling growth rate model, the preset cleaning cycle is dynamically adjusted to Tclean=α·R2 / (dRf / dt), where α is a safety factor with a value of 0.6~0.9; when the predicted time point is earlier than the adjusted cleaning cycle Tclean, a planned cleaning early warning information is generated, and the scaling distribution prompt on the cylinder wall is attached to the cleaning early warning information. The scaling distribution prompt is determined based on the distribution characteristics of the cylinder wall temperature Tw along the cylinder axis.

[0029] In a preferred embodiment of the present invention, the above-mentioned intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process is executed by an intelligent control system. The intelligent control system includes a data acquisition module, a scaling assessment module, a scaling inhibition execution module, a thermal compensation control module, and a prediction and early warning module. Each module is uniformly scheduled by an industrial control computer and communicates with sensors and actuators through a fieldbus. The data acquisition module is used to execute step S1, the scaling assessment module is used to execute step S2, the scaling inhibition execution module is used to execute steps S3 to S5, the thermal compensation control module is used to execute step S6, and the prediction and early warning module is used to execute step S7.

[0030] In summary, compared with the prior art, the present invention includes at least one of the following beneficial technical effects:

[0031] First, this invention is the first to use the equivalent fouling thermal resistance as an online control variable in the clean coal drying process. Through real-time acquisition of operating parameters and calculation of heat transfer equations, it realizes online quantitative characterization of the fouling state on the heat exchange surface of the cylinder. The degree of fouling is transformed from "stop inspection and experience judgment" to "online perception and quantitative classification", providing a reliable basis for graded fouling suppression. This is a technical means that is not available in the existing pure structural anti-sticking scheme (such as CN109870000A) and post-shutdown cleaning scheme (such as CN106867555A).

[0032] Secondly, this invention establishes a graded scale inhibition strategy linked to the scale level. For Grade I light scaling, the initial soft adhering layer is removed by periodically agitating the cylinder rotation speed. For Grade II moderate scaling, adhesion is inhibited by forming a near-wall vapor film through heating, with simultaneous thermal compensation. For Grade III heavy scaling, the hardened scale layer is removed through pulsed thermal shock and injection of scale-inhibiting media. This achieves online graded scale inhibition, avoiding the passive situation of scale accumulation leading to mandatory shutdown for cleaning. Experiments show that after adopting the method of this invention, the annual shutdown cleaning frequency of the clean coal drying system is reduced from once a month to 1-2 times per year, and the single online scale inhibition treatment time does not exceed 30 minutes, significantly extending the system's continuous operating time.

[0033] Third, this invention couples thermal compensation with scaling conditions. It calculates the thermal compensation increment based on the equivalent fouling thermal resistance and compensates for the heat transfer medium flow rate using a feedforward approach. Simultaneously, it employs a feedforward-feedback composite control to correct the heat transfer medium parameters in real time, using the discharge moisture content as the target variable. This not only compensates for heat transfer losses caused by scaling but also suppresses the interference of feed fluctuations and inlet moisture fluctuations on the discharge moisture content. The discharge moisture content fluctuation is controlled within ±0.5%, significantly improving the product qualification rate. In contrast, existing temperature control schemes (such as CN101726171A and CN104482751A) only use temperature as the target variable, failing to detect the heat transfer deterioration caused by scaling and unable to provide targeted compensation.

[0034] Fourth, this invention establishes a scaling growth rate model based on coal quality parameters, which can predict future scaling trends and provide early warnings and interventions before scaling reaches medium to heavy levels. This transforms the cleaning plan from a fixed cycle to a dynamic adjustment based on the scaling rate, adapting to changes in scaling rate caused by fluctuations in coal quality and further improving the intelligence level of system operation. Simultaneously, the increase in heat medium temperature is constrained by a safe upper temperature limit, avoiding the risks of overheating and coking, and spontaneous combustion of clean coal caused by blind heating, thus ensuring the safe operation of the system. Attached Figure Description

[0035] Figure 1 A flowchart of the intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process provided in an embodiment of the present invention. Detailed Implementation

[0036] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0037] Example 1: System Composition and Overall Process

[0038] The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process provided in this embodiment is applied to a clean coal drum dryer system. This system includes a drum dryer body, a hot air furnace (or steam supply system), a feeding device, a dispersing device, a dust removal and induced draft device, and an intelligent control system. The intelligent control system consists of an industrial control computer, a programmable logic controller, a data acquisition module, a scaling assessment module, a scaling inhibition execution module, a thermal compensation control module, and a prediction and early warning module. The data acquisition module connects to sensors located throughout the system via a fieldbus, including: a hot medium inlet temperature sensor and a hot medium flow sensor located at the hot air furnace outlet or steam inlet pipe; a hot medium outlet temperature sensor located at the dryer exhaust pipe; a material inlet temperature sensor and an online feed moisture content detector located at the feed chute; a material outlet temperature sensor and an online discharge moisture content detector located at the discharge chute; cylinder wall temperature sensors (which can be non-contact infrared thermometers or wall-mounted thermocouples) arranged axially at multiple locations on the outer wall of the drum; and a current transformer for detecting the operating current of the drum drive motor.

[0039] See Figure 1 The overall process of this embodiment is as follows: First, after the system starts, it enters the data acquisition stage, that is, it collects the above-mentioned operating parameters in real time; second, it calculates the equivalent fouling thermal resistance Rf of the heat exchange surface of the cylinder based on the collected operating parameters, and determines the current scaling level based on the equivalent fouling thermal resistance Rf and its rate of change; then, based on the branch judgment result of the scaling level, it executes the mild scaling suppression strategy, moderate scaling suppression strategy, or severe scaling suppression strategy corresponding to Level I mild scaling, Level II moderate scaling, and Level III severe scaling, respectively; while executing the scaling suppression strategy, the discharge moisture content is used as the basis for calculation. Feedforward-feedback composite control is executed for the target variable, and the inlet temperature and flow rate of the heat medium are corrected in real time. Subsequently, the scaling growth rate model is updated and the scaling trend within a preset time period is predicted. It is then determined whether the predicted time point for scaling to reach Level II (moderate scaling) is earlier than the preset cleaning cycle. If it is earlier than the preset cleaning cycle, an early warning message is generated and the scaling inhibition strategy corresponding to the predicted scaling level is executed in advance. Then, the process returns to the data acquisition stage to continue the loop. If it is not earlier than the preset cleaning cycle, the current strategy is maintained and the process returns to the data acquisition stage to continue the loop. The entire process is executed cyclically with a fixed sampling period (e.g., 10 seconds), realizing closed-loop intelligent control of scaling inhibition and thermal compensation in the clean coal drying process.

[0040] Specifically, the sampling period for each operating parameter during the data acquisition phase can be set according to the parameter's variation characteristics: the sampling period for temperature and flow parameters is set to 10 seconds, the sampling period for moisture content parameters is set to 30 seconds, and the sampling period for motor current parameters is set to 1 second. All sampled data are filtered and then stored in a real-time database, which retains historical data for the most recent 720 hours for use in scaling assessment, thermal compensation control, and scaling trend prediction.

[0041] It should be noted that this embodiment uses a drum-type clean coal dryer as an example, but this method is also applicable to steam tube rotary clean coal dryers. When a steam tube rotary dryer is used, the heat medium is steam, the heat medium inlet temperature T1 is the steam inlet temperature, the heat medium outlet temperature T2 is the steam condensate outlet temperature, and the heat exchange surface of the cylinder refers to the outer surface of the heat exchange tubes. The calculation of its equivalent fouling thermal resistance and the application of the graded fouling suppression strategy are the same as those of the drum-type dryer. The only difference is that the temperature and flow rate actuators are different (the steam regulating valve replaces the mixing valve and the induced draft fan). This adaptive adjustment is a conventional choice for those skilled in the art and will not be elaborated further.

[0042] Example 2: Calculation of equivalent fouling thermal resistance and determination of fouling level

[0043] This embodiment details the calculation method for the equivalent fouling thermal resistance and the process for determining the fouling level in step S2, which is executed by the fouling assessment module.

[0044] First, the scaling assessment module calculates the actual heat transfer Q of the cylinder's heat exchange surface based on the collected operating parameters. For a direct heat exchange drum dryer using flue gas as the heat medium, the actual heat transfer Q is calculated using the formula Q=Qh·ρ·cp·(T1-T2), where Qh is the flue gas flow rate (unit: m³ / h), ρ is the flue gas density (unit: kg / m³), cp is the flue gas specific heat capacity (unit: kJ / (kg·℃)), T1 is the inlet temperature of the heat medium (unit: ℃), and T2 is the outlet temperature of the heat medium (unit: ℃). For example, when the flue gas flow rate Qh = 60000 m³ / h, the flue gas density ρ = 0.81 kg / m³, the flue gas specific heat capacity cp = 1.05 kJ / (kg·℃), the inlet temperature of the heat medium T1 = 420℃, and the outlet temperature of the heat medium T2 = 105℃, the actual heat transfer is Q = 60000 × 0.81 × 1.05 × (420 - 105) ≈ 1.607 × 10⁻¹⁰ 7 kJ / h≈4464kW.

[0045] Next, calculate the logarithmic mean temperature difference ΔTlm. For the countercurrent heat exchange between flue gas and material, ΔTlm is calculated according to the formula ΔTlm=(ΔT1-ΔT2) / ln(ΔT1 / ΔT2), where ΔT1 is the temperature difference at the inlet end of the hot medium (i.e., the difference between the hot medium inlet temperature T1 and the material outlet temperature Tm2), and ΔT2 is the temperature difference at the outlet end of the hot medium (i.e., the difference between the hot medium outlet temperature T2 and the material inlet temperature Tm1). Following the above example, assuming the material inlet temperature Tm1=15°C and the material outlet temperature Tm2=75°C, then ΔT1=420-75=345°C, ΔT2=105-15=90°C, ΔTlm=(345-90) / ln(345 / 90)=255 / 1.343≈189.9°C.

[0046] Next, calculate the actual heat transfer coefficient K=Q / (A·ΔTlm). Assuming the cylinder heat exchange area A=180m², then K=4464×1000 / (180×189.9)≈130.6W / (m²·°C). It should be noted that the units of the actual heat transfer quantity Q, the heat exchange area A and the logarithmic mean temperature difference ΔTlm shall be converted uniformly. In engineering calculation, conversion can be carried out according to the commonly used unit system, and the present invention does not impose special limitation on this.

[0047] Then, calculate the equivalent fouling thermal resistance Rf=1 / K-1 / K0. Wherein, K0 is the reference heat transfer coefficient calibrated in a clean state. After the dryer is overhauled or thoroughly descaled, it is operated under rated working conditions and the above calculation process is repeated, and the average value of 2 hours of continuous stable operation is taken as K0. For example, if the calibrated K0=185W / (m²·°C), then the current equivalent fouling thermal resistance Rf=1 / 130.6-1 / 185≈0.00766-0.00541=0.00225m²·°C / W.

[0048] Finally, the fouling grade is determined. A first threshold R1=0.0015m²·°C / W, a second threshold R2=0.0035m²·°C / W, and a first rate-of-change threshold v1=1.5×10 -5 m²·°C / (W·h), and a second rate-of-change threshold v2=5.0×10 -5 m²·°C / (W·h). The determination rules are: when Rf<R1 and dRf / dt<v1, it is determined as grade Ⅰ mild fouling; when R1≤Rf<R2 and dRf / dt<v2, it is determined as grade Ⅱ moderate fouling; when Rf≥R2 or dRf / dt≥v2, it is determined as grade Ⅲ severe fouling. Wherein, the rate of change dRf / dt is obtained by performing moving average on the time series of the equivalent fouling thermal resistance and then calculating the difference, and the length of the moving window is 1 hour. For example, the current Rf=0.00225m²·°C / W, which satisfies R1≤Rf<R2, if dRf / dt=8.0×10 at this time -6If the concentration of fouling material (m²·℃ / (W·h)) is less than v2, it is classified as Level II moderate scaling. It should be noted that the above thresholds can be calibrated experimentally based on different coal qualities and machine models. R1 is typically taken as 15%~25% of the fouling thermal resistance corresponding to K0, and R2 is typically taken as 40%~60% of the fouling thermal resistance corresponding to K0. Specific values ​​are determined through on-site calibration, and this invention is not limited to the above specific values.

[0049] Furthermore, to avoid calculation errors under low-load conditions using a single heat transfer method, the scaling assessment module introduces an auxiliary criterion: the increment of the drive motor current I, ΔI = I - I0, where I0 is the reference current under clean conditions with the same feed rate. When ΔI exceeds 8% of the reference current I0, it indicates a significant increase in the cylinder rotation load, and the scaling level determination result is weighted and corrected, i.e., upgraded by one level based on the heat transfer method determination result. For example, if the heat transfer method determines it as Level I (light scaling) but ΔI exceeds 8% of I0, the overall determination is Level II (moderate scaling). This auxiliary criterion improves the robustness of scaling level determination.

[0050] Example 3: Level I Mild Scale Inhibition Strategy

[0051] This embodiment details the execution process of the mild scale inhibition strategy in step S3 when there is only mild scaling of level I. This process is controlled by the scale inhibition execution module.

[0052] When the scaling assessment module determines that the current scaling level is Grade I (mild scaling), the scaling inhibition execution module implements a mild scaling inhibition strategy. The core idea of ​​the mild scaling inhibition strategy is to change the movement state of the material curtain inside the cylinder by periodically micro-disturbing the cylinder rotation speed without changing the thermal conditions or affecting the drying output. This causes the material curtain to periodically scour and shear the initial soft adhesive layer on the wall, thereby peeling off the unhardened soft adhesive layer from the wall and achieving online scaling inhibition "prevention is better than cure".

[0053] Specifically, the execution parameters of the mild scale inhibition strategy are as follows: the first disturbance frequency is 0.2~0.5 times / minute, that is, a disturbance cycle of "speed increase-hold-decrease" is completed every 2~5 minutes; the disturbance amplitude of the cylinder speed n is ±3%~±8% of the rated cylinder speed. For example, if the rated speed of a drum-type clean coal dryer is 4.5 r / min, then the disturbance amplitude is ±0.14~±0.36 r / min, that is, the speed changes according to the preset curve within the range of 4.14~4.86 r / min. The duration of a single disturbance is 2~5 minutes, and the disturbance waveform adopts a trapezoidal wave, with a 30-second acceleration phase, a 1~3-minute holding phase, and a 30-second deceleration phase. The adjustment range of the dispersing shaft speed is ±5% to ±12% of the rated speed of the dispersing shaft. The adjustment of the dispersing shaft speed and the disturbance of the cylinder speed are synchronized according to a preset ratio. For example, the adjustment range of the dispersing shaft speed is 1.5 times the disturbance range of the cylinder speed, so as to enhance the rolling and dispersing effect of the material inside the cylinder and make the material curtain more fully cover the wall surface.

[0054] During the implementation of the mild scale inhibition strategy, the inlet temperature T1 of the hot medium remains constant, and the feed rate G is maintained at the rated value, thus the drying output is unaffected. This strategy continues until the scale level changes: if the equivalent fouling thermal resistance Rf drops to the clean state level (Rf≤0.2×R1) and remains stable, it indicates that the soft adhesive layer has been removed, and disturbance can be paused to resume the rated speed operation of the cylinder; if the scale level upgrades to level II or III, the corresponding scale inhibition strategy is switched to. Experiments show that implementing the mild scale inhibition strategy during the level I mild scale stage can effectively remove more than 90% of the initial soft adhesive layer, preventing it from hardening and developing into a stubborn scale layer.

[0055] Example 4: Level II Moderate Scale Inhibition Strategy and Thermal Compensation

[0056] This embodiment details the execution process and thermal compensation method of the moderate scale inhibition strategy during Level II moderate scaling in step S4.

[0057] When the scaling level upgrades to Class II moderate scaling, simple speed disturbance is no longer sufficient to completely remove the partially hardened adhesive layer. At this point, a moderate scaling inhibition strategy is implemented. The moderate scaling inhibition strategy includes two parallel technical means: one is temperature gradient suppression, and the other is thermal compensation.

[0058] The principle of temperature gradient suppression is as follows: the adhesion force of wet coal to the wall mainly comes from the liquid bridging force between particles and between particles and the wall. When the wall temperature is higher than the boiling point of water, the water film at the interface between the wet coal and the wall rapidly vaporizes, forming a continuous vapor film between the wall and the material (i.e., near-wall vapor film effect). This vapor film significantly weakens the liquid bridging force, making it difficult for the wet coal to adhere stably to the wall. The material that has already adhered will also become loose under the support of the vapor film, making it easy to be washed away and peeled off by the material curtain. Therefore, the moderate scale inhibition strategy raises the inlet temperature T1 of the heat medium to the first compensation temperature range [Tc1, Tc2]. This range is a temperature range 20~40℃ higher than the current inlet temperature T1 of the heat medium, and the lower limit Tc1 of the first compensation temperature range should make the cylinder wall temperature Tw higher than 120℃ (higher than the boiling point of water at normal pressure by 100℃, with a certain margin to overcome local heat loss at the wall). For example, if the current inlet temperature of the heat medium is T1=380℃, then the first compensation temperature range is [400℃, 420℃]. At the same time, the cylinder wall temperature Tw is monitored to ensure that Tw>120℃. If Tw is still lower than 120℃ after being increased to 400℃, then it is increased to 420℃.

[0059] The principle of heat compensation is as follows: scaling increases thermal resistance and decreases the heat transfer coefficient. To maintain a stable discharge moisture content, it is necessary to compensate for the heat loss due to scaling. The heat compensation increment ΔQ is calculated according to the formula ΔQ=Rf·A·ΔTlm·k, where k is the heat loss conversion factor, with a value of 0.8~1.2. This factor is used to correct the calculation errors of the heat transfer area A and the logarithmic mean temperature difference ΔTlm. The specific value of k is determined by on-site calibration, and is generally taken as 1.0. Using the data from Example 2, Rf=0.00225m²·℃ / W, A=180m², ΔTlm=189.9℃, and taking k=1.0, then ΔQ=0.00225×180×189.9×1.0≈76.9kW, that is, an additional compensation of about 77kW of heat transfer is required.

[0060] Then, the heat medium flow rate Qh is feedforward compensated with the heat compensation increment ΔQ. The compensation increment ΔQh is calculated according to the formula ΔQh=ΔQ / (ρ·cp·(T1-T2)). Using the above data, ρ=0.81kg / m³, cp=1.05kJ / (kg·℃), T1=400℃ (after compensation), T2=105℃, then ΔQh=76.9×3600 / (0.81×1.05×(400-105))≈276840 / 250.9≈1103m³ / h, that is, the flue gas flow rate needs to be increased by about 1103m³ / h on the original basis to compensate for the heat transfer loss caused by scaling. This feedforward compensation amount is maintained until the scaling level changes, and is added to the total output of the composite control in step S6 as the heat compensation increment ΔQ.

[0061] It is important to emphasize that the increase in the inlet temperature of the heat medium in the moderate scale inhibition strategy is constrained by the safe upper limit temperature Tmax. The volatile matter content of the clean coal is relatively high, and excessively high flue gas temperature inside the cylinder poses a risk of spontaneous combustion and explosion of the clean coal. Therefore, the upper limit Tc2 of the first compensation temperature range must not exceed the safe upper limit temperature Tmax (e.g., Tmax = 450℃). When T1 + 40℃ exceeds Tmax, Tmax is used as the actual upper limit of the execution temperature, and heat transfer losses caused by insufficient temperature increase are mainly compensated through feedforward compensation of the heat medium flow rate. During the execution of the moderate scale inhibition strategy, the discharge moisture content is corrected in real time by the composite control in step S6 to ensure stable product moisture content.

[0062] Example 5: Level III Heavy Scale Inhibition Strategy

[0063] This embodiment details the execution process of the heavy scale inhibition strategy in step S5 when there is severe scaling of level III.

[0064] When the scaling level upgrades to Level III severe scaling (Rf≥R2 or dRf / dt≥v2), the scale layer on the cylinder wall or heat exchanger tube surface has partially hardened, and it cannot be removed simply by speed disturbance and temperature gradient. At this time, a severe scaling inhibition strategy is implemented. The severe scaling inhibition strategy includes three technical means: pulsed thermal shock, reduced load operation, and injection of scale inhibition media. The three work together to remove the hardened scale layer.

[0065] The principle of pulsed thermal shock is as follows: There is a difference in the coefficient of thermal expansion between the hardened scale layer and the metal wall surface. When the wall surface temperature rises rapidly, thermal stress is generated between the scale layer and the wall surface. Repeated heating and cooling cycles cause the direction of thermal stress to change alternately, leading to micro-cracks in the scale layer that gradually expand and peel off. Specifically, after triggering the pulsed thermal shock program, the inlet temperature T1 of the heat medium changes according to a preset trapezoidal pulse curve: the heating phase lasts 3-5 minutes, raising the inlet temperature T1 from its current value to the second compensation temperature range, which is a temperature range 50-80°C higher than the current inlet temperature T1; the holding time is 2-3 minutes; the cooling phase lasts 3-5 minutes, lowering the inlet temperature T1 back to its current value; the number of pulse cycles is 2-4. For example, if the current inlet temperature of the heat medium is T1 = 380℃, then the second compensation temperature range is [430℃, 460℃]. The heating phase lasts 4 minutes to reach 440℃, is held for 2 minutes, and then cools down to 380℃ in 4 minutes. This cycle is repeated 3 times, for a total duration of approximately 30 minutes. Similarly, the upper limit of the second compensation temperature range must not exceed the safe upper limit temperature Tmax, and monitoring of the tail gas CO concentration and cylinder wall temperature should be strengthened during pulse thermal shock to prevent local overheating and spontaneous combustion of the clean coal.

[0066] Reduced load operation: During the pulse thermal shock program, the feed rate G is reduced to 60%~70% of the rated feed rate. The purpose of reducing the load is twofold: firstly, to reduce the amount of wet coal entering the cylinder, thereby reducing the continuous coverage of the wall by wet material and allowing the thermal stress generated by the pulse thermal shock to act directly on the scale layer; secondly, to reduce the density of the material curtain inside the cylinder, allowing the scale debris to be smoothly discharged with the material flow and avoiding secondary accumulation. During the reduced load period, the system automatically adjusts the frequency setting of the feeder inverter to 60%~70% of the rated frequency. For example, if the rated feed rate is 50t / h, the feed rate will be adjusted to 30~35t / h.

[0067] Scale inhibitor injection: Scale inhibitor is injected into the cylinder simultaneously with the pulsed thermal shock procedure. The scale inhibitor is either fine-grained limestone powder or a water-based scale inhibitor. When using fine-grained limestone powder, the particle size is 0.05~0.15mm. After entering the cylinder with the material flow, it acts as a "soft abrasive" in the tumbling of the material curtain, grinding the scale layer on the wall and assisting in the removal of the scale layer. At the same time, the surface of the limestone powder is weakly alkaline, which can neutralize the adhesive effect of clay minerals and organic acids in the clean coal on the wall surface, improving the wall wetting state. When using a water-based scale inhibitor, the water-based scale inhibitor is prepared by mixing water, surfactant, and dispersant in a mass ratio of 100:2:1. It is sprayed into the cylinder in a mist form through an atomizing nozzle at the feed end. Its function is to penetrate and soften the scale layer and reduce the interfacial bonding force between the scale layer and the wall surface. The injection volume of the scale inhibitor is 0.5% to 1.5% of the effective volume of the cylinder. For example, if the effective volume of the cylinder is 30 m³, the injection volume is 0.15 to 0.45 m³. After the scale inhibitor is injected, keep the cylinder rotating for 5 to 10 minutes to allow the scale inhibitor to fully act on the scale layer, and then restore the feed rate to the rated value.

[0068] After the heavy scale inhibition strategy is completed, the scale assessment module reassesses the scale level: if the equivalent fouling thermal resistance Rf drops below level II, the corresponding scale inhibition strategy is switched to continue operation; if Rf still shows no downward trend after two complete pulse cycles, it indicates that the scale layer has been severely compacted and the online scale inhibition method is ineffective. The system generates a planned cleaning warning message, prompting the operators to arrange manual cleaning at the nearest planned shutdown window, and recalibrates the baseline heat transfer coefficient K0 after cleaning.

[0069] Example 6: Feedforward-Feedback Composite Control

[0070] This embodiment details the specific implementation of the feedforward-feedback composite control in step S6, which is executed by the thermal compensation control module.

[0071] The feedforward-feedback composite control uses the discharge moisture content M2 as the target variable, and the outputs are the setpoints for the inlet temperature and flow rate of the heat medium. The feedforward term compensates for measurable disturbances (feed rate fluctuations, inlet moisture content fluctuations), the feedback term eliminates residual deviations, and the scaling compensation term compensates for heat transfer losses caused by scaling. These three terms are combined to form the overall output of the composite control.

[0072] The feedforward term Ff is calculated according to the formula Ff=Kf1·dG / dt+Kf2·dM1 / dt, where Kf1 is the feedforward coefficient for changes in feed rate, Kf2 is the feedforward coefficient for changes in feed moisture content, dG / dt is the rate of change in feed rate, and dM1 / dt is the rate of change in feed moisture content. Kf1 and Kf2 are tuned through step response tests; for example, in a certain system, Kf1 is calibrated to be 1.8℃ / (t·h). -1 ·min -1Kf2 = 3.5℃ / (%·min) -1 The feedforward term outputs a temperature correction of 1.8℃ for every 1t / h increase in feed rate (a change rate of 1t / (h·min)); and a temperature correction of 3.5℃ for every 1% increase in feed moisture content (a change rate of 1% / min). The feedforward term's function is to "predict in advance": when the feed moisture content increases, it raises the inlet temperature of the heat medium before the outlet moisture content is affected, overcoming the large hysteresis characteristic of the drying system. The feedforward temperature correction ΔTf is obtained by first-order inertial filtering of the feedforward term Ff.

[0073] The feedback term Fb is calculated using an incremental PID algorithm. Its input is the discharge moisture content deviation e = Mset - M2, where Mset is the target moisture content setpoint (e.g., 10%), and M2 is the real-time detected discharge moisture content. The output of the incremental PID algorithm is ΔFb = Kp·(e(k) - e(k-1)) + Ki·e(k) + Kd·(e(k) - 2e(k-1) + e(k-2)), where Kp, Ki, and Kd are the proportional, integral, and derivative coefficients, respectively, and e(k), e(k-1), and e(k-2) are the deviation values ​​for the current and the previous two sampling periods, respectively. Typical tuning values ​​are, for example, Kp = 0.9, Ki = 0.15, Kd = 0.3, with a sampling period of 30 seconds. The feedback temperature correction ΔTb is obtained by accumulating ΔFb, and anti-integral saturation measures are set: when the inlet temperature of the heat medium T1 reaches the safe upper limit temperature Tmax, the integral term is frozen to prevent integral saturation from causing overshoot.

[0074] Scaling Compensation Term: The scaling thermal compensation temperature correction ΔTcomp is calculated based on the current equivalent fouling thermal resistance Rf. ΔTcomp is obtained by converting the thermal compensation increment ΔQ to the temperature domain, i.e., ΔTcomp = ΔQ / (Qh·ρ·cp), representing the increase in the equivalent temperature of the heat medium required to compensate for scaling heat transfer losses. When the scaling level is I, ΔTcomp = 0; when the scaling level is II or III, ΔTcomp is calculated in real time according to the above formula and continuously output. The scaling compensation term is the embodiment of the thermal compensation increment ΔQ applied in steps S3 to S5 in the total output of the composite control.

[0075] The total output Tset of the composite control is calculated according to the formula Tset=Ts0+ΔTf+ΔTb+ΔTcomp, where Ts0 is the basic setpoint for the inlet temperature of the hot medium (set by the process card, e.g., 380℃). Tset is sent to the actuator as the final setpoint for the inlet temperature of the hot medium. When the hot medium is flue gas, the actuator is a mixing valve for flue gas and cold air and an induced draft fan frequency converter: the inlet temperature of the hot medium is adjusted by changing the opening of the mixing valve to change the proportion of cold air; the flow rate of the hot medium is changed by adjusting the frequency of the induced draft fan or the opening of the mixing valve. When the hot medium is steam, the actuator is a steam regulating valve, which simultaneously regulates the inlet temperature and flow rate of the hot medium by adjusting the valve opening. The composite control also includes temperature limit protection: when the inlet temperature T1 of the heat medium exceeds the safe upper limit temperature Tmax, the feedback term gain is reduced (e.g., Kp is reduced to 60% of its original value) and the growth rate of the heat medium flow rate Qh is limited (e.g., the rate of change of the induced draft fan frequency is limited to no more than 2 Hz / min) to prevent the system from overheating and coking or spontaneous combustion of coking coal due to blindly raising the temperature when the scaling is severe.

[0076] Example 7: Scaling Growth Rate Model and Trend Prediction

[0077] This embodiment details the process of establishing the scale growth rate model and predicting the scale trend in step S7, which is executed by the prediction and early warning module.

[0078] The scaling growth rate model takes coal quality parameters and operating parameters as inputs and the scaling index FI as the output. The scaling index FI is defined as the ratio of the equivalent fouling thermal resistance Rf to the second threshold R2, i.e., FI = Rf / R2. When FI = 1, it reaches the threshold for Class III severe scaling, and when FI = FI2 = R1 / R2 (approximately 0.4~0.6), it reaches the threshold for Class II moderate scaling. Coal quality parameters include ash content Ad, total moisture Mt, particle size distribution (expressed as the proportion of particles smaller than 0.5 mm), and caking index G. Among them, ash content Ad and caking index G reflect the mineral composition and caking characteristics of clean coal. The higher the ash content, the larger the G value, the higher the proportion of fine particles, and the stronger the scaling tendency. Operating parameters include the inlet temperature of the heat medium T1, the cylinder rotation speed n, and the feed rate G. In addition, the model input also includes the feed moisture content M1 and the historical equivalent fouling thermal resistance sequence.

[0079] The model was constructed using either a multiple linear regression model or a backpropagation (BP) neural network model. The expression for the multiple linear regression model is FI(t) = β0 + β1·Ad + β2·Mt + β3·φ + β4·G_exponent + β5·T1 + β6·n + β7·G_feed + β8·t, where φ is the proportion of particles smaller than 0.5mm, t is the running time, and β0~β8 are the regression coefficients. The regression coefficients were obtained through offline training using historical operating data: daily data from a complete scaling cycle (from clean to Level III heavy scaling) under self-cleaning conditions were collected, and the initial regression coefficients were obtained by fitting the data using the least squares method. The BP neural network model adopted a three-layer structure: an input layer with 8 nodes (corresponding to the 8 input variables mentioned above), a hidden layer with 12 nodes (using the Sigmoid function as the activation function), and an output layer with 1 node (corresponding to the scaling index FI). It was trained using historical data until the mean square error was less than 1×10^6. -4 .

[0080] Online model parameter updates: During system operation, model parameters are updated online using a rolling window method. The length of the rolling window is the data for the most recent 30 days. The model parameters are retrained every 24 hours using the latest data within the rolling window to adapt the model to scaling characteristic drift caused by coal quality fluctuations and changes in operating conditions. When coal quality parameters change significantly (e.g., the ash content Ad changes by more than 2 percentage points), a model parameter update is immediately triggered.

[0081] Scaling Trend Prediction: Utilizing a scaling growth rate model, the system inputs current coal quality parameters and operating parameters to extrapolate the scaling index FI for a future preset time period (e.g., the next 72 hours), obtaining a predicted curve of FI over time. The predicted time point t_pred corresponding to the scaling index FI reaching the Class II moderate scaling threshold FI2 is then calculated. When the predicted time point t_pred is earlier than the preset cleaning cycle (e.g., the preset cleaning cycle is 30 days, and Class II is predicted to be reached in 8 days), the prediction and early warning module generates an early warning message, which includes a suggested scaling suppression strategy (determined based on the predicted scaling level). After operator confirmation, the system automatically executes the corresponding scaling suppression strategy in advance, mitigating scaling risks at an early stage. For example, when the predicted scaling level is Class II, the early warning message indicates, "The scaling level is expected to reach Class II in 8 days; it is recommended to implement a moderate scaling suppression strategy in advance."

[0082] Dynamic adjustment of cleaning cycle: Based on the scaling trend output by the scaling growth rate model, the preset cleaning cycle is dynamically adjusted to Tclean = α·R² / (dRf / dt), where α is a safety factor, ranging from 0.6 to 0.9. For example, R² = 0.0035 m²·℃ / W, and the current dRf / dt = 8.0 × 10⁻⁶. -6m²·℃ / (W·h), taking α=0.8, then Tclean=0.8×0.0035 / (8.0×10 -6 The cleaning cycle is approximately 350 hours, or 14.6 days, meaning the fixed 30-day cycle is dynamically adjusted to about 15 days. It should be noted that when dRf / dt is less than the first rate of change threshold v1, it indicates that the current scaling rate is extremely low. To avoid excessively extending the cleaning cycle and affecting routine equipment maintenance, the cleaning cycle remains unchanged at this time. When dRf / dt increases, the cleaning cycle automatically shortens according to the above formula, achieving adaptive matching between the cleaning plan and the scaling rate. When the predicted time point t_pred is earlier than the adjusted cleaning cycle Tclean, a planned cleaning early warning message is generated, which includes a scale distribution indicator on the cylinder wall. The scale distribution is determined based on the distribution characteristics of the cylinder wall temperature Tw along the cylinder axis. Sections where the cylinder wall temperature Tw is significantly lower than the average value are identified as severely scaled sections (due to insulation by the scale layer). For example, if 10 wall temperature measuring points are arranged along the cylinder axis, and the temperature at measuring points 3-5 is more than 15°C lower than the average value, it indicates that the scale is concentrated in the front section of the cylinder (near the feed section), and this section should be the focus of the cleaning. This scale distribution indicator provides precise operational guidance for planned cleaning, reducing the workload of blind cleaning.

[0083] Example 8: Complete Application Case

[0084] This embodiment provides a complete application example of the method of the present invention in a coal preparation plant's clean coal drum drying system to illustrate the implementation effect of the method.

[0085] This system processes flotation clean coal with a rated capacity of 50 t / h. The feed moisture content is approximately 24%, and the target discharge moisture content is 10%. The rotary drum dryer has a diameter of 2.4 m and a length of 18 m, with a rated rotational speed of 4.5 r / min. It is equipped with a dispersing shaft, and the heat source is flue gas generated by a hot blast furnace with a rated flue gas volume of 60,000 m³ / h and a rated flue gas inlet temperature of 420℃. The system is equipped with the intelligent control system described in this invention.

[0086] On the first day of system operation, under clean conditions, the calibrated baseline heat transfer coefficient K0 = 185 W / (m²·℃) and baseline current I0 = 85 A. On the 15th day of operation, the fouling assessment module calculated the equivalent fouling thermal resistance Rf = 0.0009 m²·℃ / W and dRf / dt = 4.0 × 10⁻⁶. -6The temperature (m²·℃ / (W·h)) is below the first threshold R1=0.0015m²·℃ / W, indicating a level I mild scaling condition. A mild scaling suppression strategy is implemented: the cylinder rotation speed is agitated at a frequency of 0.3 times / minute within the range of 4.5±0.3r / min, and the speed of the dispersing shaft is adjusted synchronously by ±8%. On the 20th day of operation, Rf rises to 0.0020m²·℃ / W, indicating a level II moderate scaling condition. A moderate scaling suppression strategy is then implemented: the flue gas inlet temperature is increased from 380℃ to 410℃ (the lower limit of the first compensation temperature range, and the cylinder wall temperature Tw rises to 135℃, which is higher than 120℃). The heat compensation increment is calculated to be approximately 68kW according to ΔQ=Rf·A·ΔTlm·k, and the flue gas volume feedforward compensation increases by approximately 980m³ / h. At the same time, the composite control adjusts the flue gas parameters in real time with a target discharge moisture content of 10%, and the discharge moisture content is stabilized within the range of 9.6%~10.4%.

[0087] On the 32nd day of operation, the ash content of the raw coal entering the washing plant increased from 12% to 16%, resulting in a significantly faster scaling rate, with dRf / dt rising to 6.8 × 10⁻⁶. -5 When the scale concentration (rf) reaches 0.0036 m²·℃ / W, exceeding the second threshold R2 = 0.0035 m²·℃ / W, it is determined to be Class III severe scaling. A severe scaling inhibition strategy is implemented: the flue gas temperature is cyclically adjusted between 380℃ and 440℃ three times according to a trapezoidal pulse curve (heating for 4 minutes, holding for 2 minutes, and cooling for 4 minutes); the feed rate is reduced to 35 t / h (70% of the rated value); and 0.3 m³ of 0.1 mm limestone powder is injected. After the scaling inhibition program is completed, Rf drops to 0.0012 m²·℃ / W, the scaling level returns to Class I, and the system switches back to a mild scaling inhibition strategy and resumes normal feeding.

[0088] Regarding scaling trend prediction, the prediction and early warning module extrapolates the scaling index FI based on the parameters of this coal type (ash content 16%, total moisture 24%, particle size below 0.5mm 78%, adhesion index G=75). It predicts that the time point for reaching Level II moderate scaling will be 18 days later, 30 days earlier than the preset cleaning cycle. The system generates an early warning and implements a moderate scaling inhibition strategy preheating intervention on the 10th day in advance. In actual operation, the scaling level is consistently controlled below Level II. Throughout the entire assessment period (3 months), the system did not experience any unplanned shutdowns due to scaling. The qualified rate of the output moisture content increased from 86% in the original process to 97%, the energy consumption for drying ton of coal decreased by 9.5%, and the annual planned cleaning frequency decreased from once a month in the original process to twice a year. This fully verifies the technical effectiveness of the method of this invention in scaling inhibition, thermal compensation, and intelligent prediction.

[0089] Example 9: Application of Steam Pipe Rotary Coal Dryer

[0090] This embodiment illustrates the application of the method of the present invention in a steam tube rotary coal dryer. The steam tube rotary dryer employs an indirect steam heat exchange method. Multiple sets of heat exchange tubes are arranged axially within the cylinder. Steam flows and condenses within the heat exchange tubes, releasing heat. Wet clean coal is heated and dried within the cylinder by the heat exchange tubes. The scaling problem mainly manifests as scale adhering to the outer surface of the heat exchange tubes, leading to increased thermal resistance and discharge blockage. This is the same scaling mechanism as that of a drum dryer, and both are suitable for the method of the present invention.

[0091] When applied to a steam pipe rotary dryer, the data acquisition differences are as follows: the inlet temperature of the heat medium, T1, is the steam inlet temperature, measured by a pressure transmitter and temperature sensor on the steam inlet pipe; the outlet temperature of the heat medium, T2, is the condensate outlet temperature; the actual heat transfer, Q, is calculated using the formula Q=Qh·r, where Qh is the steam condensate (mass flow rate), and r is the latent heat of vaporization of the steam at the corresponding pressure; for superheated steam, the actual heat transfer should also include the sensible heat portion of the steam as it cools from the superheated temperature to the saturation temperature, i.e., Q=Qh·[cp_v·(T1-Ts)+r], where cp_v is the specific heat capacity of the steam at constant pressure, and Ts is the saturation temperature at the corresponding pressure. In the calculation of the logarithmic mean temperature difference ΔTlm, the temperature on the heat medium side is taken as a reasonable average of the steam temperature and the condensate temperature. The calculation of the equivalent fouling thermal resistance Rf, the determination of the fouling level, and the rules for determining the graded fouling inhibition strategy are completely consistent with those for a drum dryer.

[0092] The actuator of the steam tube rotary dryer is adapted to simultaneously regulate the steam flow rate and steam pressure entering the heat exchange tubes by adjusting the opening of the steam regulating valve, thereby regulating the inlet temperature T1 and the flow rate Qh of the heat medium; the output of the feedforward-feedback composite control is executed through the steam regulating valve. Since the response speed of steam-side temperature regulation is faster than that of flue gas mixing regulation, its feedforward compensation lag of the heat compensation increment ΔQ is smaller, resulting in better control performance. It should be noted that when the steam tube rotary dryer implements the Level III heavy-duty scale inhibition strategy, the pulse thermal shock program achieves rapid rise and fall of the heat exchange tube wall temperature by periodically opening the steam regulating valve. The number of pulse cycles and the heating rate should take into account the fatigue life of the heat exchange tubes, and the heating rate should preferably not exceed 15℃ / min; the preferred injection position of the scale inhibition medium is above the heat exchange tube bundle, so that the scale inhibition medium fully contacts the outer surface of the heat exchange tubes during the material curtain movement.

[0093] Application tests on a coal moisture conditioning steam rotary dryer in a coking plant showed that before adopting the method of this invention, scaling on the heat exchange tubes caused poor material discharge, requiring a monthly shutdown for cleaning, with each shutdown lasting approximately 10 hours. After adopting the method of this invention, through online determination and graded scaling inhibition of scaling levels, the equivalent fouling thermal resistance of the heat exchange tubes was consistently controlled below the Class II moderate scaling threshold. The number of annual shutdowns for cleaning was reduced to twice a year, steam consumption was reduced by 7.8%, and the fluctuation of output moisture content narrowed from ±1.2% to ±0.4%, achieving significant economic benefits.

[0094] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent control of scaling inhibition and thermal compensation in the process of drying clean coal, characterized in that, The method is applied to a clean coal dryer, which includes a cylinder, a dispersing shaft disposed inside the cylinder, a drive motor for driving the cylinder to rotate, and a heat medium supply device. The method includes the following steps: Step S1: Real-time acquisition of the operating parameters of the clean coal drying system. The operating parameters include at least the inlet temperature of the heat medium T1, the outlet temperature of the heat medium T2, the inlet temperature of the material Tm1, the outlet temperature of the material Tm2, ​​the cylinder wall temperature Tw, the drive motor current I, the feed rate G, the feed moisture content M1, the outlet moisture content M2, and the heat medium flow rate Qh. Step S2: Calculate the equivalent fouling thermal resistance Rf of the heat exchange surface of the cylinder based on the operating parameters collected in step S1, and determine the current fouling level based on the equivalent fouling thermal resistance Rf and its rate of change over time dRf / dt. The fouling level includes Level I light fouling, Level II moderate fouling and Level III heavy fouling. Step S3: When the scaling level is determined to be Class I mild scaling, a mild scaling inhibition strategy is implemented: keep the inlet temperature T1 of the hot medium constant, apply periodic disturbance to the cylinder rotation speed n at the first disturbance frequency, and synchronously adjust the speed of the dispersing shaft according to the preset ratio, and use the material curtain inside the cylinder to periodically flush the wall to remove the initial soft adhesion layer. Step S4: When the scaling level is determined to be Class II moderate scaling, a moderate scaling inhibition strategy is implemented: Based on the mild scaling inhibition strategy, the inlet temperature T1 of the heat medium is increased to the first compensation temperature range [Tc1, Tc2], so that the wall temperature Tw of the cylinder is higher than the boiling point of water in the wet clean coal. The steam layer generated by the rapid vaporization of the near-wall water film inhibits the adhesion of wet clean coal to the wall. At the same time, the heat compensation increment ΔQ is calculated according to the equivalent fouling thermal resistance Rf, and the heat compensation increment ΔQ is used to feedforward compensation of the heat medium flow rate Qh. Step S5: When the scaling level is determined to be Class III severe scaling, a severe scaling inhibition strategy is executed: trigger the pulse thermal shock program, so that the inlet temperature T1 of the heat medium rises to the second compensation temperature range according to the preset pulse curve within the preset pulse cycle and then falls back. At the same time, the feed rate G is reduced to 60%~70% of the rated feed rate, and the scaling inhibition medium is injected into the cylinder. Step S6: Using the discharge moisture content M2 as the target variable, a feedforward-feedback composite control is used to correct the inlet temperature T1 and flow rate Qh of the heat medium in real time. The feedforward term of the composite control is calculated based on the rate of change of the feed rate G and the feed moisture content M1. The feedback term is obtained by PID calculation based on the deviation between the discharge moisture content M2 and the target moisture content setpoint Mset. The total output of the composite control is superimposed on the heat compensation increment ΔQ applied in steps S3 to S5. Step S7: Establish a scaling growth rate model based on coal quality parameters and historical operating data. Use the scaling growth rate model to predict the predicted time point when the scaling level reaches Level II moderate scaling within a preset time period. When the predicted time point is earlier than the preset cleaning cycle, generate early warning information and execute the scaling suppression strategy corresponding to the predicted scaling level in advance.

2. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In the step S2, the equivalent fouling thermal resistance Rf is calculated as follows: the actual heat transfer coefficient K of the heat exchange surface of the cylinder is calculated according to the operating parameters, and the equivalent fouling thermal resistance Rf is calculated according to the formula Rf=1 / K-1 / K0, where K0 is the calibrated reference heat transfer coefficient in a clean state; the judgment basis for the fouling grade is: when Rf<R1 and dRf / dt<v1, it is judged as grade I mild fouling; when R1≤Rf<R2 and dRf / dt<v2, it is judged as grade II moderate fouling; when Rf≥R2 or dRf / dt≥v2, it is judged as grade III severe fouling; where R1 is the first threshold, R2 is the second threshold, v1 is the first change rate threshold, v2 is the second change rate threshold, and R1<R2, v1<v2.

3. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 2, characterized in that, The actual heat transfer coefficient K is calculated according to the formula K=Q / (A·ΔTlm), where Q is the actual heat transfer amount of the heat exchange surface of the cylinder, A is the heat exchange area of the cylinder, and ΔTlm is the logarithmic mean temperature difference between the heat medium and the material; the actual heat transfer amount Q is calculated according to the heat medium flow rate Qh, the heat medium density ρ, the heat medium specific heat capacity cp, and the temperature difference between the heat medium inlet temperature T1 and the heat medium outlet temperature T2 according to the formula Q=Qh·ρ·cp·(T1-T2).

4. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In the step S3, the first disturbance frequency is 0.2 to 0.5 times per minute, the disturbance amplitude of the cylinder rotation speed n is ±3% to ±8% of the rated rotation speed of the cylinder, the adjustment amplitude of the breaking shaft rotation speed is ±5% to ±12% of the rated rotation speed of the breaking shaft, the single disturbance duration is 2 to 5 minutes, and the periodic disturbance is continuously performed before the fouling grade changes.

5. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In the step S4, the first compensation temperature interval [Tc1, Tc2] is a temperature interval 20 to 40 °C higher than the current heat medium inlet temperature T1, and the lower limit Tc1 of the first compensation temperature interval satisfies that the cylinder wall surface temperature Tw is higher than 120 °C; The thermal compensation increment ΔQ is calculated according to the formula ΔQ=Rf·A·ΔTlm·k, where k is the heat transfer loss conversion coefficient, with a value of 0.8 to 1.2; when the heat medium flow rate Qh is subjected to feedforward compensation with the thermal compensation increment ΔQ, the compensation increment ΔQh of the heat medium flow rate is calculated according to the formula ΔQh=ΔQ / (ρ·cp·(T1-T2)).

6. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In the step S5, the preset pulse curve is a trapezoidal pulse curve, the heating period time of the trapezoidal pulse curve is 3 to 5 minutes, the heat medium inlet temperature T1 is raised to a second compensation temperature interval, the second compensation temperature interval is a temperature interval 50 to 80 °C higher than the current heat medium inlet temperature T1, the heat preservation time is 2 to 3 minutes, the cooling period time is 3 to 5 minutes, and the number of pulse cycles is 2 to 4 times; the scale-inhibiting medium is fine-grained limestone powder or a water-based scale inhibitor, and the injection amount of the scale-inhibiting medium is 0.5% to 1.5% of the effective volume of the cylinder.

7. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In step S6, the feedforward term Ff is calculated according to the formula Ff=Kf1·dG / dt+Kf2·dM1 / dt, where Kf1 is the feedforward coefficient for the change in feed rate and Kf2 is the feedforward coefficient for the change in feed moisture content; the feedback term Fb is calculated according to the incremental PID algorithm; the total output Tset of the composite control on the inlet temperature of the heat medium is calculated according to the formula Tset=Ts0+ΔTf+ΔTb+ΔTcomp, where Ts0 is the basic setpoint of the inlet temperature of the heat medium, ΔTf is the feedforward temperature correction, ΔTb is the feedback temperature correction, and ΔTcomp is the scaling heat compensation temperature correction. The total output Qhset of the composite control for the heat medium flow rate is calculated according to the formula Qhset=Qh0+ΔQf+ΔQb+ΔQh_comp, where Qh0 is the basic setpoint for the heat medium flow rate, ΔQf is the feedforward flow correction, ΔQb is the feedback flow correction, and ΔQh_comp is the scaling heat compensation flow correction. The scaling heat compensation flow correction is calculated from the heat compensation increment ΔQ. The composite control also includes upper temperature limit protection. When the heat medium inlet temperature T1 exceeds the safe upper limit temperature Tmax, the feedback term gain is reduced and the growth rate of the heat medium flow rate Qh is limited.

8. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, In step S7, the scaling growth rate model takes coal quality parameters and operating parameters as inputs and scaling index FI as output. The scaling index FI is the ratio of the equivalent fouling thermal resistance Rf to the second threshold R2, i.e., FI = Rf / R2. The coal quality parameters include ash content Ad, total moisture Mt, particle size distribution, and adhesion index G. The operating parameters include the inlet temperature of the heat medium T1, the cylinder rotation speed n, and the feed rate G. The scaling growth rate model adopts a multiple linear regression model or a BP neural network model, uses historical operating data for offline training to obtain initial model parameters, and updates the model parameters online using a rolling window method during operation. The predicted time point is the time point obtained by extrapolating the scaling index FI within a future preset period to the time corresponding to the judgment threshold of Class II moderate scaling.

9. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 1, characterized in that, The heat medium is either flue gas or steam. When the heat medium is flue gas, the inlet temperature T1 of the heat medium is adjusted by adjusting the mixing ratio of flue gas and cold air, and the flow rate Qh of the heat medium is adjusted by adjusting the frequency of the induced draft fan or the opening of the mixing valve. When the heat medium is steam, the inlet temperature T1 and the flow rate Qh of the heat medium are adjusted by adjusting the opening of the steam regulating valve.

10. The intelligent control method for scaling inhibition and thermal compensation in the clean coal drying process according to claim 8, characterized in that, It also includes a dynamic adjustment step for the cleaning cycle: based on the scaling trend output by the scaling growth rate model, the preset cleaning cycle is dynamically adjusted to Tclean=α·R2 / (dRf / dt), where α is a safety factor with a value of 0.6~0.9; when the predicted time point is earlier than the adjusted cleaning cycle Tclean, a planned cleaning early warning information is generated, and the scaling distribution prompt on the cylinder wall is attached to the cleaning early warning information. The scaling distribution prompt is determined based on the distribution characteristics of the cylinder wall temperature Tw along the cylinder axis.

Images