Electrolytic manganese combination reaction rate optimization method based on intelligent control

Abstract

The invention discloses an electrolytic manganese combination reaction rate optimization method based on intelligent control, and relates to the technical field of electrolytic manganese production process optimization. Multiple types of sensors are arranged on the reaction kettle and associated equipment, temperature, pressure, liquid level and other parameters are collected in real time, and data reliability is ensured through filtering and complementing of industrial control software; a reaction rate optimization model is constructed, parameters are optimized by using a BP neural network in combination with historical data, and the prediction error is low; the rotating speed of multiple layers of blades is adjusted according to model output, and a gas distributor is matched to enhance mixing; a heating or cooling device and a pressure regulating valve are driven, and the optimal working condition is maintained; injecting a high-efficiency catalyst in sections; rate comparison and parameter optimization are carried out regularly; and when the monitoring parameters exceed the standard, triggering early warning and executing an emergency instruction. The device improves the material mixing uniformity and the temperature and pressure control precision, accelerates the reaction rate, improves the utilization rate of the raw materials and the catalyst, reduces the energy consumption and pollutant discharge, guarantees the stable production, and has both economic benefits and environmental benefits.

Description

Technical Field

[0001] This invention relates to the field of electrolytic manganese production process optimization technology, and in particular to a method for optimizing the electrolytic manganese combination reaction rate based on intelligent control. Background Technology

[0002] In the chemical reaction stage of electrolytic manganese production, the reaction rate and mass transfer efficiency directly determine the production cycle and efficiency. Traditional processes have significant shortcomings in this stage. Traditional chemical reaction vessels often use single-layer or simple double-layer paddle agitators with fixed paddle shapes and angles, which cannot adapt to the material distribution characteristics in different areas of the reactor. Material tends to accumulate in the upper layer and settle in the lower layer, resulting in low uniformity of reactant mixing. Simultaneously, the reaction system lacks auxiliary disturbance methods, and mass transfer efficiency is limited by the single effect of mechanical stirring, making it difficult for raw materials and catalysts to quickly and fully contact each other. Furthermore, traditional temperature and pressure control relies on manual adjustment or simple temperature controllers, resulting in low temperature monitoring accuracy and slow pressure control response. Large temperature fluctuations and pressure deviations from the optimal range during the reaction further restrict reaction activity, prolonging the chemical reaction cycle, hindering production efficiency, and failing to meet the capacity requirements of large-scale production.

[0003] The utilization of raw materials and the control of impurities are equally prominent issues in the electrolytic manganese compounding workshop. On the one hand, the raw material pretreatment process often employs simple physical screening or water washing, resulting in incomplete removal of magnetic impurities and chemical impurities such as carbonates. These impurities not only reduce the reactivity of the raw materials but also participate in side reactions during the reaction process, generating useless products that are mixed into the final product and affect the quality of electrolytic manganese. On the other hand, there is a lack of targeted impurity suppression methods during the reaction process, and harmful impurities cannot be effectively fixed. Consequently, some raw materials cannot participate in the main reaction due to binding with impurities and are ultimately discharged with the waste residue, resulting in resource waste. At the same time, traditional processes treat the waste residue generated by the compounding reaction in a simple manner, only landfilling or discarding it, without recovering the residual manganese and valuable metals, which increases production costs and exacerbates the environmental burden.

[0004] Traditional electrolytic manganese compounding workshops also suffer from significant deficiencies in production control and environmental and energy utilization. The production process relies heavily on manual operation and simple automated equipment, lacking a comprehensive sensor monitoring network. Key data such as reaction temperature, pressure, raw material flow rate, and product composition cannot be collected and analyzed in real time, leading to large operational errors and difficulty in accurately matching production parameters to process requirements. The lack of real-time monitoring of equipment operation status makes it difficult to detect and locate malfunctions promptly, requiring shutdowns for troubleshooting and further reducing production continuity. Regarding environmental and energy aspects, the exhaust gas generated in the workshop is treated only by simple spraying, resulting in incomplete removal of harmful gases such as sulfur dioxide and nitrogen oxides. Furthermore, the waste heat contained in the exhaust gas is not recovered, leading to direct emissions and energy waste. Simultaneously, existing catalysts exhibit low catalytic activity and poor selectivity, not only failing to effectively accelerate the reaction rate but also increasing the probability of side reactions, leading to higher product impurity content and further impacting production efficiency and product quality. Summary of the Invention

[0005] The present invention proposes an intelligent control-based method for optimizing the electrolytic manganese combination reaction rate to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for optimizing the electrolytic manganese combination reaction rate based on intelligent control, comprising: Data acquisition and preprocessing steps: A sensor network is deployed in the electrolytic manganese reaction vessel and related equipment, including temperature sensors, pressure sensors, liquid level sensors, stirring speed sensors, raw material flow sensors, and reaction product composition analyzers. The sensors collect parameters in real time, such as temperature T, pressure P, material liquid level H, stirring speed n, raw material feed flow rate Q, and manganese ion concentration C in the product. The collected data is transmitted to the industrial control software via Ethernet. The industrial control software filters the raw data, uses the Kalman filter algorithm to remove abnormal fluctuations caused by electromagnetic interference, and uses linear interpolation to complete missing data. Industrial control software modeling and analysis steps: The industrial control software constructs a reaction rate optimization model based on big data analysis and PID control algorithm. The model input is the collected T, P, n, Q, and C parameters, and the output is the stirring speed adjustment value Δn, temperature control value ΔT, pressure control value ΔP, and catalyst addition amount Δm. The industrial control software calculates the current reaction rate v in real time. The reaction rate v is evaluated by the increase in manganese ion concentration in the product per unit time, i.e., v=ΔC / Δt. At the same time, a reaction rate-process parameter correlation database is constructed by combining historical production data, and the model parameters are optimized by BP neural network algorithm. Dynamic control steps of the stirring system: Based on the stirring speed adjustment value Δn output by the industrial control software, the speed of the impeller stirring device in the reactor is controlled. The stirring device includes an upper wide-blade impeller, a middle inclined-blade impeller, and a lower propeller. The speed n1 of the upper wide-blade impeller, the speed n2 of the middle inclined-blade impeller, and the speed n3 of the lower propeller are adjusted in linkage according to a preset ratio. The proportional coefficient is dynamically corrected according to the material liquid level H. When H is higher than the preset threshold H0, the proportion of the speed of the lower propeller is increased. When H is lower than H0, the proportion of the speed of the upper wide-blade impeller is increased. At the same time, the flow rate G of the inert gas introduced into the gas distributor at the bottom of the reactor is controlled. The gas flow rate G is positively correlated with the stirring speed n. Temperature and pressure control steps: The industrial control software drives the heating or cooling device of the outer jacket of the reactor to operate based on the ΔT and ΔP output by the reaction rate optimization model. The heating device uses an electric heating tube, and the cooling device uses circulating water cooling. Temperature control is achieved by adjusting the heating power W or the cooling water flow rate V. Pressure control is achieved through the pressure regulating valve at the top of the reactor. Adjusting the valve opening K maintains the reaction pressure stable within the optimal range. The optimal pressure range is preset according to the raw material grade. Catalyst addition steps: Based on the catalyst addition amount Δm calculated by the industrial control software, the catalyst is injected into the reactor through a metering pump. The catalyst is prepared by the sol-gel method with transition metal oxides as the main active components. The catalyst is added in stages: 60% of the total amount is added at the initial stage of raw material feeding, 30% is added in the middle stage of reaction, and 10% is added in the later stage of reaction. The catalyst exerts catalytic activity in each stage of the reaction.

[0007] Furthermore, it also includes: Real-time feedback and parameter iteration steps: Every 5 minutes, the industrial control software compares the current reaction rate v with the preset target reaction rate v0 and calculates the rate deviation Δv = v0 - v. If Δv > 0, that is, the current rate is lower than the target value, the industrial control software re-optimizes the model parameters and adjusts the stirring speed n, reaction temperature T, pressure P and catalyst addition amount m. If Δv ≤ 0, the current process parameters are maintained, and the industrial control software updates the reaction rate-process parameter correlation database in real time. Abnormal warning and emergency handling steps: The industrial control software monitors in real time whether each collected parameter exceeds the safety threshold. When the temperature T exceeds the preset safety range [Tmin, Tmax], the pressure P exceeds [Pmin, Pmax], or the stirring speed n fluctuates abnormally, the industrial control software triggers an audible and visual warning and outputs an emergency handling command. When the temperature exceeds the standard, the cooling water flow rate V is automatically increased. When the pressure exceeds the standard, the pressure regulating valve opening K is automatically increased. When the stirring fails, the system automatically switches to the backup stirring motor. Dynamic calibration steps for stirring speed: During the dynamic control of the stirring system, the industrial control software calibrates the stirring speed n in real time based on the material liquid level H and the product manganese ion concentration C. The calibrated stirring speed ncal satisfies ncal=n0+k1×(H-H0)+k2×(C0-C), where ncal is the calibrated stirring speed, n0 is the base stirring speed, k1 is the liquid level influence coefficient, H is the current liquid level, H0 is the optimal liquid level, k2 is the concentration influence coefficient, C0 is the target manganese ion concentration, and C is the current manganese ion concentration. Through dynamic calibration, the stirring speed is adapted to the material distribution and reaction process under different liquid levels.

[0008] Temperature cumulative deviation compensation steps: During the temperature control process, the industrial control software calculates the temperature cumulative deviation ΔTsum within the reaction time and adjusts the control quantity of the heating or cooling device. The temperature cumulative deviation ΔTsum satisfies ΔTsum=∫(t0 to t1)(Topt-T(t))dt, where ΔTsum is the temperature cumulative deviation within the time period from t0 to t1, t0 is the temperature control start time, t1 is the current time, Topt is the optimal reaction temperature, and T(t) is the actual temperature inside the reactor at time t.

[0009] Furthermore, in the data acquisition and preprocessing steps, a raw material flow sensor is deployed at the outlet of the raw material preprocessing device. The raw material preprocessing device includes a magnetic separator, a screening device, and a preprocessing solution soaking tank. The raw material is first passed through a magnetic separator to remove magnetic impurities. The magnetic field strength of the magnetic separator is set to 1.2T. Then, it is passed through a screening device to remove particulate impurities with a particle size >5mm. Finally, it is soaked in a dilute sulfuric acid preprocessing solution. The soaking time is dynamically adjusted according to the impurity content of the raw material.

[0010] Furthermore, in the catalyst addition step, the catalyst preparation process includes: dissolving manganese salt and iron salt in an ethanol organic solvent at a mass ratio of 3:1, using manganese sulfate as the manganese salt and ferric sulfate as the iron salt; adding citric acid as a complexing agent, with the amount of complexing agent added being 20% ​​of the total mass of the metal salt; adding silica as an auxiliary agent, with the amount of auxiliary agent added being 5% of the total mass of the metal salt; stirring at 600 r / min for 30 minutes to form a uniform sol; controlling the sol temperature at 60-70℃ and the pH value at 3.5-4.5 and allowing it to stand for 2 hours to complete gelation; drying the gel in a 105℃ forced-air drying oven for 12 hours; then calcining it in a muffle furnace at 500℃ for 2 hours; and after natural cooling, pulverizing it to a particle size of 100 mesh to obtain a catalyst with a specific surface area ≥100 m² / g.

[0011] Furthermore, it also includes: Catalyst addition optimization steps: The industrial control software calculates the optimal catalyst addition amount mopt based on the raw material grade γ and the current reaction rate v. mopt satisfies mopt=m0×(γ0 / γ)×(v0 / v), where mopt is the optimal catalyst addition amount, m0 is the catalyst addition amount corresponding to the reference raw material grade, γ0 is the reference raw material grade, γ is the actual grade of the raw material currently used, v0 is the target reaction rate, and v is the current reaction rate.

[0012] Waste heat recovery and utilization steps: A plate heat exchanger is installed at the outlet of the cooling system of the reactor jacket. The heat exchanger transfers the waste heat carried by the cooling water to the soaking tank in the raw material pretreatment stage. The industrial control software collects the inlet and outlet water temperatures Tin and Tout of the heat exchanger and the raw material soaking tank temperature Tsoak in real time, and calculates the waste heat recovery amount Qrec, Qrec=ρcVcool×(Tout-Tin), where ρ is the density of water, c is the specific heat capacity of water, and Vcool is the cooling water flow rate. When Tsoak is lower than the pretreatment temperature Tsoak0, the industrial control software controls the opening of the bypass valve of the heat exchanger to decrease, increase the heat exchange area, and keep Tsoak at Tsoak0. The waste heat is used to preheat the raw material to reduce the energy consumption of raw material heating, and at the same time, the initial temperature of the raw material entering the reactor is closer to the optimal reaction temperature.

[0013] Furthermore, in the industrial control software modeling and analysis steps, the number of input layer nodes of the BP neural network algorithm is set to 6, the number of hidden layer nodes is set to 12, and the number of output layer nodes is set to 4 corresponding to Δn, ΔT, ΔP, and Δm. The activation function is the Sigmoid function, the network training adopts the gradient descent method, the learning rate is set to 0.01, and the number of training iterations is set to 1000. After training, the prediction error of the model on the reaction rate is controlled within 5%. After each batch of reaction is completed, the process parameters and reaction rate data of the batch are added to the training set as new samples, and the network is retrained to optimize the parameters.

[0014] Furthermore, in the temperature and pressure control steps, the optimal temperature range is set according to the reaction stage. The optimal temperature in the initial stage of the reaction is 85-90℃, the optimal temperature in the middle stage of the reaction is 90-95℃, and the optimal temperature in the later stage of the reaction is 88-92℃. The industrial control software collects T in real time through the temperature sensor.

[0015] Furthermore, it also includes: The industrial control software model iterative update steps are as follows: After each batch of reaction is completed, the industrial control software automatically stores the batch's T, P, n, Q, C, v and reaction time t data into the historical database. Each month, 1000 new sets of data added to the historical database are used to train the model. The weights and thresholds of the BP neural network are optimized using a genetic algorithm. The population size of the genetic algorithm is set to 50, the crossover probability is set to 0.8, the mutation probability is set to 0.05, and the number of iterations is set to 50. After training, the model performance is verified through small-batch trial production. The trial production batch size is 10% of the formal production batch.

[0016] Compared with existing technologies, the beneficial effects of this invention are: In terms of improving the reaction rate, a multi-layer differentiated impeller stirring device and a bottom gas distributor are used in synergy. The upper wide-blade impeller pushes the upper material downward, the middle inclined impeller strengthens lateral mixing, and the lower propeller promotes vertical circulation. Combined with the disturbance effect of inert gas, the uniformity of material mixing is greatly improved, and the mass transfer efficiency is effectively improved. At the same time, with the help of high-precision sensors and intelligent control system, the reaction temperature and pressure are adjusted in real time to keep the reaction at the optimal operating conditions, significantly accelerate the chemical reaction process, shorten the production cycle, and improve the overall production efficiency.

[0017] In terms of optimizing raw material utilization, a pretreatment process combining physical and chemical methods is used. First, magnetic impurities and large particulate impurities are removed by magnetic separation and sieving. Then, chemical impurities are transformed by soaking in a specific pretreatment solution, which greatly improves the purity and reactivity of the raw materials. The composite additives added during the reaction process can form stable complexes with harmful impurities, avoiding their interference with the main reaction and ensuring that the raw materials fully participate in the reaction. At the same time, the waste residue is washed with water and magnetically separated to recover residual manganese and valuable metals. The remaining waste residue is used to produce building materials, realizing resource recycling, greatly reducing raw material waste, lowering production costs, and alleviating environmental pressure.

[0018] In terms of intelligent production control, a sensor network covering reactors, pipelines, and pumps is constructed to collect data such as temperature, pressure, flow rate, liquid level, and product composition in real time. Through big data analysis and artificial intelligence algorithms of the central control system, the operating status of the equipment is automatically adjusted, replacing traditional manual operation and reducing operational errors. The remote monitoring and operation platform allows managers to view production data, equipment status, and alarm information anytime and anywhere. When necessary, the equipment can be remotely controlled to achieve comprehensive monitoring and precise control of the production process, reduce equipment failure rate, and ensure production continuity and product quality stability.

[0019] In terms of environmental protection and energy utilization, a multi-stage exhaust gas purification system is adopted. First, acidic gases are absorbed by alkaline spraying, then harmful gases are converted through catalytic reaction, and finally residual pollutants are removed by activated carbon adsorption to ensure that the exhaust gas meets emission standards. A heat exchanger is added to the exhaust gas treatment process to transfer the waste heat of the exhaust gas to the circulating water, which is used to preheat raw materials or as a heat source for other processes in the workshop, achieving efficient energy recovery and reducing workshop energy consumption. The application of high-efficiency catalysts further improves reaction selectivity, reduces side reactions, and reduces product impurity content, thereby improving production efficiency while ensuring product quality and achieving a dual improvement in economic and environmental benefits. Attached Figure Description

[0020] Figure 1 This is a schematic block diagram of the method for optimizing the electrolytic manganese combination reaction rate based on intelligent control proposed in this invention. Figure 2 This is a schematic diagram illustrating the impurity removal effect of each step in the raw material pretreatment process of the intelligent control-based electrolytic manganese combination reaction rate optimization method proposed in this invention. Figure 3 This diagram illustrates the removal rate of harmful gases in each stage of the tail gas treatment process using the intelligent control-based electrolytic manganese reaction rate optimization method proposed in this invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0023] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0024] Reference Figures 1 to 3 A method for optimizing the rate of electrolytic manganese combination reaction based on intelligent control, comprising: Data acquisition and preprocessing steps: A sensor network is deployed at key locations in the electrolytic manganese reaction vessel and related equipment. The sensors include temperature sensors, pressure sensors, liquid level sensors, stirring speed sensors, raw material flow sensors, and a reaction product composition analyzer. The sensors collect parameters such as temperature T, pressure P, material liquid level H, stirring speed n, raw material feed flow rate Q, and manganese ion concentration C in the product in real time. The acquisition frequency is set to 1 time / second. The acquired data is transmitted to the industrial control software via industrial Ethernet. The industrial control software filters the raw data, uses the Kalman filter algorithm to remove abnormal fluctuations caused by electromagnetic interference, and uses linear interpolation to complete missing data. The data integrity reaches more than 99%, and the measurement error is controlled within ±2%. Industrial control software modeling and analysis steps: The industrial control software constructs a reaction rate optimization model based on big data analysis and PID control algorithm. The model input is the collected T, P, n, Q, and C parameters, and the output is the stirring speed adjustment value Δn, temperature control value ΔT, pressure control value ΔP, and catalyst addition amount Δm. The industrial control software calculates the current reaction rate v in real time. The reaction rate v is evaluated by the increase in manganese ion concentration in the product per unit time, i.e., v=ΔC / Δt. At the same time, a reaction rate-process parameter correlation database is constructed by combining historical production data. The model parameters are optimized by BP neural network algorithm, and the model prediction error is controlled within 5%. Dynamic control steps of the stirring system: Based on the stirring speed adjustment value Δn output by the industrial control software, the speed of the multi-layer impeller stirring device in the reactor is controlled. The stirring device includes an upper layer wide blade impeller, a middle layer inclined blade impeller, and a lower layer propeller. The speed n1 of the upper layer wide blade impeller, the speed n2 of the middle layer inclined blade impeller, and the speed n3 of the lower layer propeller are adjusted in linkage according to a preset ratio. The ratio coefficient is dynamically corrected according to the material liquid level H. When H is higher than the preset threshold H0, the speed ratio of the lower layer propeller is increased. When H is lower than H0, the speed ratio of the upper layer wide blade impeller is increased. At the same time, the flow rate G of the inert gas introduced into the gas distributor at the bottom of the reactor is controlled. The gas flow rate G is positively correlated with the stirring speed n, and the material mixing uniformity is improved to more than 90%. Precise temperature and pressure control steps: The industrial control software drives the heating or cooling device of the outer jacket of the reactor to operate based on the ΔT and ΔP output by the reaction rate optimization model. The heating device uses an electric heating tube, and the cooling device uses circulating water cooling. Temperature control is achieved by adjusting the heating power W or the cooling water flow rate V. Pressure control is achieved through the pressure regulating valve at the top of the reactor. The valve opening K is adjusted to maintain the reaction pressure stable in the optimal range. The optimal pressure range is preset according to the raw material grade. High-grade manganese ore raw materials correspond to a higher pressure range, and low-grade raw materials correspond to a lower pressure range. Precise catalyst addition steps: Based on the catalyst addition amount Δm calculated by the industrial control software, the high-efficiency catalyst is injected into the reactor through a metering pump. The catalyst is mainly composed of transition metal oxides and is prepared by the sol-gel method. The catalyst is added in stages: 60% of the total amount is added at the initial stage of raw material feeding, 30% is added in the middle stage of the reaction (when the manganese ion concentration of the product reaches the target value of 50%), and 10% is added in the later stage of the reaction. The catalyst can fully exert its catalytic activity at each stage of the reaction. Real-time feedback and parameter iteration steps: Every 5 minutes, the industrial control software compares the current reaction rate v with the preset target reaction rate v0 and calculates the rate deviation Δv = v0 - v. If Δv > 0, that is, the current rate is lower than the target value, the industrial control software re-optimizes the model parameters and adjusts the stirring speed n, reaction temperature T, pressure P and catalyst addition amount m. If Δv ≤ 0, the current process parameters are maintained. At the same time, the industrial control software updates the reaction rate-process parameter correlation database in real time, and the model adaptability gradually improves with the accumulation of data. Anomaly warning and emergency handling steps: The industrial control software monitors in real time whether each collected parameter exceeds the safety threshold. When the temperature T exceeds the preset safety range [Tmin, Tmax], the pressure P exceeds [Pmin, Pmax], or the stirring speed n fluctuates abnormally (fluctuation amplitude > 10%), the industrial control software triggers an audible and visual warning and outputs emergency handling instructions. When the temperature exceeds the standard, the cooling water flow rate V is automatically increased. When the pressure exceeds the standard, the opening degree of the pressure regulating valve K is automatically increased. When the stirring fails, the system automatically switches to the backup stirring motor. The system ensures that the continuous operation of the reaction system is maintained at over 99.5%.

[0025] This invention also includes: Dynamic calibration steps for stirring speed: During the dynamic control of the stirring system, the industrial control software calibrates the stirring speed n in real time based on the material liquid level H and the product manganese ion concentration C. The calibrated stirring speed ncal satisfies ncal=n0+k1×(H-H0)+k2×(C0-C), where ncal is the calibrated stirring speed in r / min, n0 is the base stirring speed in r / min, k1 is the liquid level influence coefficient with a value range of 5-10 r / (min・m), H is the current liquid level in m, H0 is the optimal liquid level in m, k2 is the concentration influence coefficient with a value range of 2-5 r / (min・g / L), C0 is the target manganese ion concentration in g / L, and C is the current manganese ion concentration in g / L. Through dynamic calibration, the stirring speed is adapted to the material distribution and reaction process under different liquid levels, reducing the low mass transfer efficiency caused by the accumulation of upper material or the deposition of lower material, improving the material mixing uniformity by more than 15%, and accelerating the reaction rate by 8%-12%.

[0026] Temperature cumulative deviation compensation steps: During precise temperature control, the industrial control software calculates the temperature cumulative deviation ΔTsum over the reaction time and adjusts the control quantity of the heating or cooling device accordingly. The temperature cumulative deviation ΔTsum satisfies ΔTsum=∫(t0 to t1)(Topt-T(t))dt, where ΔTsum is the temperature cumulative deviation during the time period from t0 to t1, in °C·s, t0 is the temperature control start time in seconds, t1 is the current time in seconds, Topt is the optimal reaction temperature in °C, and T(t) is the actual temperature inside the reactor at time t in °C. When ΔTsum is positive, the actual temperature is consistently lower than the optimal temperature, and the industrial control software increases the heating power W. When ΔTsum is negative, the actual temperature is consistently higher than the optimal temperature, and the industrial control software increases the cooling water flow rate V. By compensating for the temperature cumulative deviation, the continuous impact of short-term temperature fluctuations on the reaction rate is reduced, and the reaction proceeds efficiently under a stable temperature environment.

[0027] In this invention, during the data acquisition and preprocessing steps, a raw material flow sensor is deployed at the outlet of the raw material preprocessing device. The raw material preprocessing device includes a magnetic separator, a screening device, and a preprocessing solution soaking tank. The raw material is first subjected to a magnetic separator to remove magnetic impurities. The magnetic field strength of the magnetic separator is set to 1.2T. Then, it is subjected to a screening device to remove large particles with a diameter >5mm. The screening device has a screen mesh size of 5mm. Finally, it is soaked in a dilute sulfuric acid preprocessing solution with a concentration of 10%-15%. The soaking time is dynamically adjusted according to the impurity content of the raw material. When the impurity content is >10%, the soaking time is 60 minutes, and when the impurity content is ≤10%, the soaking time is 30 minutes. The raw material flow sensor collects the actual feed flow rate Q of the raw material after preprocessing. Based on the correlation between Q and the reaction rate v, the industrial control software automatically increases the frequency of the raw material feed pump when Q is lower than the preset flow rate Qmin, so that Q is maintained in the Qmin-Qmax range, ensuring a stable raw material supply and reducing the decrease in reaction rate caused by insufficient raw materials. The reaction rate fluctuation is controlled within ±5%.

[0028] In this invention, the precise catalyst addition step involves the following preparation process for a highly efficient catalyst: Manganese salt and iron salt are dissolved in an ethanol organic solvent at a mass ratio of 3:1. The manganese salt used is manganese sulfate (MnSO4·H2O), and the iron salt is ferric sulfate (Fe2(SO4)3). Citric acid is added as a complexing agent, with the amount of complexing agent added being 20% ​​of the total mass of the metal salts. Silica (SiO2) is added as an auxiliary agent, with the amount of auxiliary agent added being 5% of the total mass of the metal salts. The mixture is stirred at 600 r / min for 30 minutes to form a uniform sol. The sol temperature is controlled at 60-70℃ and the pH value at 3.5-4.5, and the mixture is allowed to stand for 2 hours to complete gelation. The gel is dried in a 105℃ forced-air drying oven for 12 hours, then calcined in a muffle furnace at 500℃ for 2 hours. After natural cooling, the mixture is pulverized to a particle size of 100 mesh to obtain a catalyst with a specific surface area ≥100 m² / g. This catalyst improves reaction activity and shortens reaction time by 15%-20%.

[0029] This invention also includes: Catalyst addition optimization steps: The industrial control software calculates the optimal catalyst addition amount mopt based on the raw material grade γ (mass fraction of manganese in the raw material) and the current reaction rate v. mopt satisfies mopt = m0 × (γ0 / γ) × (v0 / v), where mopt is the optimal catalyst addition amount in kg, m0 is the catalyst addition amount corresponding to the baseline raw material grade in kg, γ0 is the baseline raw material grade (taken as 25%), γ is the actual grade of the raw material currently used, v0 is the target reaction rate in g / (L·h), and v is the current reaction rate in g / (L·h). Through this optimization calculation, the catalyst dosage is adjusted according to the actual raw material grade, reducing waste caused by excessive catalyst for high-grade raw materials and reducing the low reaction rate caused by insufficient catalyst for low-grade raw materials, thereby improving catalyst utilization by more than 20%.

[0030] Waste heat recovery and utilization steps: A plate heat exchanger is installed at the outlet of the cooling system of the reactor jacket. The heat exchanger transfers the waste heat carried by the cooling water to the soaking tank in the raw material pretreatment stage. The industrial control software collects the inlet and outlet water temperatures Tin and Tout of the heat exchanger and the raw material soaking tank temperature Tsoak in real time, and calculates the waste heat recovery amount Qrec, Qrec=ρcVcool×(Tout-Tin), where ρ is the density of water in kg / L, c is the specific heat capacity of water in kJ / (kg・℃), and Vcool is the cooling water flow rate in L / s. When Tsoak is lower than the pretreatment temperature Tsoak0 (Tsoak0=50℃), the industrial control software controls the opening of the bypass valve of the heat exchanger to decrease, increase the heat exchange area, improve the heat exchange efficiency, and keep Tsoak at Tsoak0. The waste heat is used to preheat the raw materials to reduce the energy consumption of raw material heating, and at the same time, the initial temperature of the raw materials entering the reactor is closer to the optimal reaction temperature, thereby increasing the rate of the reaction start-up stage.

[0031] In this invention, in the industrial control software modeling and analysis steps, the number of input layer nodes of the BP neural network algorithm is set to 6 (corresponding to T, P, n, Q, C and reaction time t), the number of hidden layer nodes is set to 12, and the number of output layer nodes is set to 4 (corresponding to Δn, ΔT, ΔP, Δm). The activation function is the Sigmoid function, the network training adopts the gradient descent method, the learning rate is set to 0.01, and the number of training iterations is set to 1000. After training, the prediction error of the model for the reaction rate is controlled within 5%. After each batch of reaction is completed, the process parameters and reaction rate data of that batch are added to the training set as new samples, and the network is retrained to optimize the parameters. The adaptability of the model to working conditions such as changes in raw material characteristics and equipment aging is improved by more than 30%.

[0032] In this invention, the optimal temperature range is set according to the reaction stage in the precise temperature and pressure control step. The optimal temperature is 85-90℃ in the initial stage of the reaction (0-1 hour), 90-95℃ in the middle stage (1-3 hours), and 88-92℃ in the later stage (3 hours to the end). The industrial control software collects T in real time through the temperature sensor. When T deviates from the optimal temperature of the current stage by ±2℃, the temperature adjustment is triggered. The adjustment range of the heating power W of the heating device is proportional to the temperature deviation ΔT=|T-Topt|. When ΔT=2℃, W is adjusted by 10%, and when ΔT=5℃, W is adjusted by 25%. The pressure control adopts a proportional control algorithm. The adjustment amount of valve opening K is ΔK=k*ΔP, where k is the pressure adjustment coefficient, which is 20% / MPa. The pressure is quickly stabilized in the optimal range. The stable temperature and pressure environment improves the reaction activity of the raw materials, thereby optimizing the reaction rate.

[0033] This invention also includes: The industrial control software model iterative update steps are as follows: After each batch of reaction is completed, the industrial control software automatically stores the T, P, n, Q, C, v and reaction time t data of that batch into the historical database. Every month, more than 1,000 sets of valid data newly added to the historical database are used to train the model. The weights and thresholds of the BP neural network are optimized using a genetic algorithm. The population size of the genetic algorithm is set to 50, the crossover probability is set to 0.8, the mutation probability is set to 0.05, and the number of iterations is set to 50. After training, the model performance is verified through small-batch trial production. The trial production batch size is 10% of the formal production batch. If the deviation between the model's predicted reaction rate and the actual reaction rate is ≤3%, the updated model is put into formal production. If the deviation is >3%, the genetic algorithm parameters are readjusted and trained until the deviation meets the requirements. Through model iterative updates, the model maintains a prediction accuracy within ±3% in the long term, providing data support for reaction rate optimization.

[0034] Specific implementation method of the method for optimizing the electrolytic manganese combination reaction rate based on intelligent control: (Example 1: Optimization of Conventional Production in the Chemical Workshop of a Medium-Sized Electrolytic Manganese Plant) This embodiment addresses the 20m³ combination reactor in the combination workshop of a medium-sized electrolytic manganese plant (daily production of 50 tons of electrolytic manganese). It is suitable for conventional operating conditions with a manganese ore grade of 20%-25% and a production batch cycle of 8 hours. The solution addresses the problems of slow reaction rate, low raw material utilization, poor temperature and pressure control accuracy, and waste of tail gas heat inherent in traditional processes. The specific implementation is as follows: I. Equipment Deployment and Parameter Setting Sensor network deployment: A PT1000 platinum resistance temperature sensor (measurement range 0-200℃, accuracy ±0.1℃) and a diffused silicon pressure sensor (measurement range 0-1.6MPa, accuracy ±0.2%FS) are installed on the top of the reactor. An ultrasonic level sensor (measurement range 0-5m, accuracy ±1mm) is installed in the middle of the reactor wall. A Hall speed sensor (measurement range 0-1500r / min, accuracy ±1r / min) is installed at the end of the stirring shaft. An electromagnetic flow sensor (measurement range 0-50m³ / h, accuracy ±0.5%FS) is installed on the raw material feed pipe. An online ion chromatograph (detecting manganese ion concentration, response time ≤30s) is installed on the product outlet pipe. All sensors are connected to the industrial Ethernet via the Modbus-RTU protocol, and the data acquisition frequency is set to 1 time / second.

[0035] Modification of the stirring and mixing device: The original single-layer straight blade was removed and replaced with a three-layer differentiated blade stirring device. The upper layer is a wide blade with a diameter of 1.2m and an angle of 15° (made of 316L stainless steel), the middle layer is an inclined blade with a diameter of 1.0m and an angle of 45°, and the lower layer is a propeller with a diameter of 0.8m and a lead of 1.0m. An annular gas distributor (2mm orifice, made of Hastelloy) is installed in the center of the bottom of the reactor, connected to a nitrogen cylinder, and equipped with a mass flow controller (control range 0-50L / min, accuracy ±1%FS).

[0036] Temperature and pressure control and catalyst preparation: The outer jacket of the reactor is equipped with electric heating tubes (total power 50kW, divided into 8 groups of independent control) and circulating water cooling pipes (equipped with electric regulating valves, diameter DN50); the catalyst is prepared according to the sol-gel method in document five: manganese sulfate (MnSO4·H2O) and ferric sulfate (Fe2(SO4)3) are dissolved in anhydrous ethanol at a mass ratio of 3:1, citric acid (20% of the total mass of metal salts) is added as a complexing agent, and silicon dioxide (5% of the total mass of metal salts) is added as an auxiliary agent. The mixture is stirred at 600r / min for 30 minutes to form a sol, and then allowed to stand at 65℃ and pH=4.0 for 2 hours to gel. After drying at 105℃ for 12 hours, the mixture is calcined at 500℃ for 2 hours and then pulverized to 100 mesh for later use.

[0037] Raw material pretreatment and tail gas treatment: Construct a raw material pretreatment production line, which will be equipped with a permanent magnet drum magnetic separator (magnetic field strength 1.2T, processing capacity 20t / h), a vibrating screen (screen mesh size 5mm), and a dilute sulfuric acid soaking tank (volume 10m³, sulfuric acid concentration 12%). The tail gas treatment system will be equipped with a Φ1.5m spray tower (filled with PP multi-faceted hollow sphere packing, layer height 2m), a Φ1.0m catalytic reactor (filled with honeycomb Pt / Al2O3 catalyst, layer height 1.5m), and a Φ0.8m activated carbon adsorption tower (filled with columnar activated carbon, layer height 2m). A plate heat exchanger (heat exchange area 10m²) will be installed between the spray tower and the catalytic reactor.

[0038] II. Full Process Implementation Steps Raw material pretreatment and feeding: The manganese ore raw material is first subjected to a magnetic separator to remove magnetic impurities (removal rate ≥95%), and then a screening machine to remove particles with a diameter >5mm (screening efficiency ≥90%). It is then sent to a soaking tank to soak in 12% dilute sulfuric acid for 45 minutes (when the impurity content is 22%). During this period, it is continuously stirred by a stirrer (speed 300r / min). After soaking, it is pumped to the reaction vessel. The electromagnetic flow sensor monitors the feed flow rate in real time, and the industrial control software controls the flow rate to be stable at 15m³ / h.

[0039] Data Acquisition and Modeling: Sensors collect real-time data on temperature T, pressure P, liquid level H, rotation speed n, flow rate Q, and manganese ion concentration C. The industrial control software uses Kalman filtering to remove electromagnetic interference (data fluctuation after filtering is ≤0.2%), and missing data is filled in using linear interpolation. Based on historical data from 100 batches (including reaction rates under different raw material grades, stirring speeds, and temperature and pressure parameters), a reaction rate optimization model is trained using a BP neural network. The input layer has 6 nodes (T, P, n, Q, C, and reaction time t), the hidden layer has 12 nodes, and the output layer has 4 nodes (Δn, ΔT, ΔP, and Δm). The model prediction error is controlled within 5%.

[0040] Stirring and Temperature / Pressure Control: The industrial control software adjusts the impeller speed ratio based on the current liquid level H (optimal liquid level set at 3m). When H=3.2m, the lower propeller speed ratio increases to 40% (total speed 600r / min), and when H=2.8m, the upper wide-blade impeller speed ratio increases to 35%. Simultaneously, the nitrogen flow rate G is positively correlated with the speed n (G=30L / min when n=600r / min), enhancing material mixing. For temperature control, the target temperature is 88℃ in the initial reaction stage (0-2 hours), adjusting the heating power when it deviates by ±2℃ (power adjusted by 10% when the deviation is 2℃). The target temperature is 92℃ in the middle reaction stage (2-6 hours) and 90℃ in the later reaction stage (6-8 hours). Pressure control is achieved through a top regulating valve, with a target pressure of 0.8MPa, adjusting the valve opening by 5% when the deviation is 0.05MPa.

[0041] Catalyst addition and reaction monitoring: The catalyst is added in stages. At the initial stage of raw material feeding, 60% of the total dosage (calculated based on 0.5% of the raw material mass) is injected through a metering pump. 30% is added in the middle stage of the reaction (when the manganese ion concentration reaches 8 g / L) and 10% is added in the later stage of the reaction (when the manganese ion concentration reaches 14 g / L). The industrial control software calculates the reaction rate v (ΔC / Δt) every 5 minutes and compares it with the target rate (2 g / (L・h)). If the rate is lower than the target, the stirring speed is increased by 5%-10% or the temperature is increased by 1-2℃.

[0042] Tail gas treatment and waste heat utilization: The tail gas generated by the chemical reaction (containing approximately 800 mg / m³ SO₂ and 200 mg / m³ NOₓ) first enters the spray tower, where it is sprayed with 20% sodium hydroxide solution (flow rate 5 m³ / h) to absorb SO₂. Then it enters the catalytic reactor (temperature 300℃) to reduce NOₓ to N₂ and oxidize CO to CO₂. Finally, it is discharged after passing through the activated carbon adsorption tower (space velocity 1000 h⁻¹). The tail gas at the outlet of the spray tower (temperature 180℃) is heated by the heat exchanger to heat the circulating water (inlet temperature 25℃, outlet temperature 85℃). The heated circulating water is used to preheat the raw materials in the soaking tank (raising the raw material temperature from 25℃ to 45℃), reducing the additional heating energy consumption of the soaking tank.

[0043] III. Data Characterization for Effectiveness Verification Comparison table of key indicators before and after optimization of conventional production: Before optimization, the traditional process suffered from uneven stirring, low temperature and pressure control precision, and a reaction cycle as long as 10 hours. Raw materials, due to numerous impurities and poor mixing, did not react fully, resulting in a utilization rate of only 75%, and product purity was only 92% affected by impurities. The exhaust gas was only treated with simple spraying, leading to high SO2 emission concentrations and ineffective waste heat recovery. This invention improves mixing uniformity through three-layer differentiated stirring and nitrogen disturbance, maintains optimal reaction conditions with high-precision temperature and pressure control, shortens the reaction cycle, and ensures full contact between raw materials and catalyst. Raw material pretreatment removes impurities, and staged catalyst addition enhances reaction activity, significantly improving raw material utilization and product purity. A multi-stage exhaust gas purification system reduces harmful gas emissions, and a heat exchanger recovers waste heat to preheat raw materials, achieving both environmental and energy-saving improvements. This fully meets the efficiency and environmental requirements of conventional production in medium-sized electrolytic manganese plants.

[0044] Example 2: High-grade raw materials and environmental protection enhancement in large-scale electrolytic manganese plants This embodiment focuses on a 40m³ chemical reaction vessel in a large-scale electrolytic manganese plant (daily production of 100 tons of electrolytic manganese). It is suitable for high-efficiency operation with a manganese ore grade of 25%-30% and a production batch cycle of 6 hours. The key features are enhanced by deep pretreatment of raw materials, intelligent remote control, zero emissions of tail gas, and full recovery of waste residue. The specific implementation is as follows: I. Equipment Deployment and Parameter Setting High-precision sensors and remote control: A fiber optic temperature sensor (measurement range 0-300℃, accuracy ±0.05℃) and a piezoelectric pressure sensor (measurement range 0-2.0MPa, accuracy ±0.1%FS) are installed on the top of the reactor. An online laser particle size analyzer (measurement range 0.1-1000μm, accuracy ±2%) is installed at different heights (1m, 2m, 3m) inside the reactor to monitor material dispersion. An X-ray fluorescence analyzer (elemental analysis accuracy ±0.01%) is installed on the raw material pretreatment line to detect the impurity content of the raw materials in real time. The central control system adopts an industrial server (2 Intel Xeon Gold 6348 processors, 128GB memory), deploys WinCC industrial control software, and builds a remote monitoring platform (supports access via mobile APP and Web, latency ≤1s), allowing managers to remotely view data and adjust parameters.

[0045] Enhanced stirring and mixing: The stirring device has been upgraded to a four-layer impeller, with the addition of a middle layer lower inclined impeller (0.9m in diameter, -30° angle), forming a three-dimensional stirring system of "upper and lower pushing - lateral shearing - bottom circulation" with the original upper layer wide impeller, middle layer upper inclined impeller, and lower layer propeller; the gas distributor adopts a double-layer annular design (upper layer aperture 1.5mm, lower layer aperture 2.5mm), connected to a nitrogen-compressed air mixed gas source (volume ratio 3:1), and the mass flow controller has a control range of 0-100L / min.

[0046] Deep pretreatment of raw materials and recycling of waste residue: After the magnetic separator (magnetic field strength 1.5T) and screening machine (screen mesh size 3mm), a high-pressure roller mill (pressure 10MPa, raw material particle size ≤1mm after treatment) is added to the raw material pretreatment line to increase the reaction specific surface area; the soaking tank adopts a two-stage series (first stage 15% dilute sulfuric acid, second stage 10% dilute sulfuric acid), each stage is equipped with an ultrasonic generator (power 500W, frequency 20kHz) to enhance impurity removal; a waste residue recycling workshop is built, and a countercurrent washing tower (5-stage water washing, water-to-solid ratio 5:1), a high gradient magnetic separator (magnetic field strength 2.0T), and a vacuum filter are deployed to recover residual manganese in the waste residue, and the tailings are sent to the building materials production line to produce lightweight partition boards.

[0047] Exhaust gas and energy system optimization: The exhaust gas treatment system is upgraded to a five-stage treatment system consisting of a spray tower, electrostatic precipitator, catalytic reactor, activated carbon adsorption, and regenerative thermal oxidizer (RTO), with an RTO thermal efficiency of ≥95%. The heat exchanger adopts a shell-and-tube type (heat exchange area of ​​20m²) to recover waste heat from the exhaust gas to heat the jacketed circulating water of the reactor (replacing part of the electric heating) and provide a heat source for the workshop heating system. Energy monitoring instruments are installed to measure the consumption of electricity, water, and nitrogen in real time, and the data is connected to the industrial control software for energy consumption analysis.

[0048] II. Full Process Implementation Steps Deep pretreatment of raw materials: High-grade manganese ore raw materials are first tested by X-ray fluorescence analyzer (impurity content 15%), then crushed into particle size ≤1mm by high-pressure roller mill, and then magnetic impurities are removed by 1.5T magnetic separator (removal rate ≥98%) and large particles are removed by 3mm sieve. Then it is sent to a two-stage soaking tank: the first stage is treated with 15% dilute sulfuric acid + ultrasonic treatment for 30 minutes (to remove carbonate impurities), and the second stage is treated with 10% dilute sulfuric acid for 20 minutes (deep impurity removal). After pretreatment, the impurity content of the raw materials is reduced to below 3%, and then pumped to the reaction vessel with the flow rate controlled at 30m³ / h.

[0049] Intelligent modeling and control: The industrial control software is based on 500 batches of historical data (including reaction data of high-grade raw materials, different stirring speeds, and temperature and pressure parameters), and uses a genetic algorithm to optimize the BP neural network model (population size 50, crossover probability 0.8, mutation probability 0.05, 50 iterations), with a model prediction error ≤3%; the remote platform displays the temperature field distribution inside the reactor in real time (generated by interpolation of data from 3 temperature sensors) and material dispersion (data from a laser particle size analyzer). When local material agglomeration (particle size > 500μm) is detected, the speed of the middle layer lower inclined blade is automatically increased by 10%-15% to enhance shear dispersion.

[0050] Reaction process optimization and catalyst addition: The target temperature is 90℃ in the initial stage (0-1.5 hours), 95℃ in the middle stage (1.5-4.5 hours), and 92℃ in the later stage (4.5-6 hours). Temperature control adopts PID-Smith prediction method with a lag compensation time of 10s and a fluctuation range of ≤±0.5℃. The target pressure is 0.9MPa, using fuzzy control algorithm with valve opening adjustment accuracy of ±1%. The catalyst addition is dynamically adjusted according to the raw material grade (0.4% when the grade is 30% and 0.6% when the grade is 25%), and the addition method is "50% in the initial stage + 35% in the middle stage + 15% in the later stage" to ensure that high-grade raw materials do not waste catalyst and low-grade raw materials react fully.

[0051] Tail gas treatment and waste residue recovery: Tail gas is treated by a spray tower (25% sodium hydroxide solution, flow rate 8m³ / h) to remove 90% of SO2, electrostatic precipitator to remove dust (emission concentration ≤10mg / m³), catalytic reactor (350℃, Pt-Rh catalyst) to reduce NOx (removal rate ≥95%), activated carbon adsorption tower to remove residual organic matter, and finally sent to RTO (800℃) to oxidize untreated harmful gases (removal rate ≥99%); the waste heat from tail gas is used to heat circulating water (inlet 25℃, outlet 90℃) through a shell-and-tube heat exchanger to provide 60% of the heat for heating the reactor; the reaction waste residue is first washed by five-stage countercurrent (the washing liquid with a manganese ion concentration ≥0.5g / L is recycled for soaking tank), and then a high-gradient magnetic separator is used to recover manganese concentrate (grade 35%). The tailings are used to produce lightweight partition boards (compressive strength ≥3.0MPa).

[0052] III. Data Characterization for Effectiveness Verification Comparison table of indicators before and after high-grade raw materials and environmentally enhanced optimization: Before optimization, traditional processes for processing high-grade raw materials suffered from incomplete pretreatment, poor stirring and dispersion, and insufficient reaction rates, resulting in a raw material utilization rate of only 80%. Incomplete tail gas treatment led to high concentrations of harmful gases and low efficiency in recovering manganese from waste heat and slag. This invention achieves deep impurity removal and refinement of raw materials through high-pressure roller milling and dual-stage ultrasonic immersion, enhancing reaction activity. Four-layer three-dimensional stirring and a mixed gas source enhance mass transfer, coupled with high-precision intelligent control, significantly increasing the reaction rate. A five-stage tail gas treatment system achieves ultra-low emissions of harmful gases, while an RTO and a shell-and-tube heat exchanger improve waste heat recovery and reduce energy consumption. Multi-stage waste slag recovery enables efficient manganese recovery and tail slag resource utilization, fully meeting the needs of large-scale electrolytic manganese plants for efficient utilization of high-grade raw materials and enhanced environmental protection. Simultaneously, remote control functionality improves production management convenience and provides stable technical support for large-scale production.

[0053] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for optimizing the reaction rate of electrolytic manganese combination based on intelligent control, characterized in that, Comprise: Data acquisition and preprocessing step: Deploy sensor network in electrolytic manganese combined reactor and associated equipment, including temperature sensor, pressure sensor, liquid level sensor, stirring speed sensor, raw material flow sensor and reaction product composition analyzer. The sensor real-time collects temperature T, pressure P, material liquid level H, stirring speed n, raw material feeding flow Q and product manganese ion concentration C parameters. The collected data is transmitted to industrial control software through Ethernet. The industrial control software filters the original data, removes abnormal fluctuations caused by electromagnetic interference using Kalman filtering algorithm, and completes the missing data using linear interpolation method; Industrial control software modeling and analysis step: The industrial control software constructs a reaction rate optimization model based on big data analysis and PID control algorithm. The model input is the collected T, P, n, Q, C parameters, and the output is the stirring speed adjustment value Δn, temperature control amount ΔT, pressure control amount ΔP and catalyst addition amount Δm. The industrial control software calculates the current reaction rate v in real time. The reaction rate v is evaluated by the increment of manganese ion concentration in the product per unit time, that is, v=ΔC / Δt. At the same time, the reaction rate-process parameter correlation database is constructed by combining historical production data, and the model parameters are optimized by BP neural network algorithm; Stirring system dynamic control step: According to the stirring speed adjustment value Δn output by the industrial control software, the speed of the paddle stirring device in the reactor is controlled. The stirring device includes upper wide-blade paddle, middle inclined-blade paddle and lower screw paddle. The upper wide-blade paddle speed n1, middle inclined-blade paddle speed n2 and lower screw paddle speed n3 are adjusted in a preset proportion. The proportion coefficient is dynamically corrected according to the material liquid level H. When H is higher than the preset threshold H0, the proportion of lower screw paddle speed is increased. When H is lower than H0, the proportion of upper wide-blade paddle speed is increased. At the same time, the flow G of inert gas introduced by the gas distributor at the bottom of the reactor is controlled. The gas flow G is positively correlated with the stirring speed n; Temperature and pressure control step: According to ΔT and ΔP output by the reaction rate optimization model, the industrial control software drives the heating device or cooling device outside the reactor to run. The heating device uses electric heating pipe, and the cooling device uses circulating water cooling. The temperature control is realized by adjusting the heating power W or the cooling water flow rate V. The pressure control is realized by the pressure regulating valve at the top of the reactor. The valve opening K is adjusted to maintain the reaction pressure stable in the optimal interval. The optimal pressure interval is preset according to the raw material grade; Catalyst addition step: According to the catalyst addition amount Δm calculated by the industrial control software, the catalyst is injected into the reactor through the metering pump. The catalyst is mainly composed of transition metal oxides and is prepared by sol-gel method. The catalyst is added in stages, that is, 60% of the total dose is added at the beginning of raw material feeding, 30% is added in the middle of the reaction, and 10% is added in the late stage of the reaction. The catalyst has catalytic activity in each stage of the reaction.

2. The method of claim 1, wherein the method is based on the intelligent control of the electrolytic manganese compound reaction rate. Also include: Real-time feedback and parameter iteration step: The industrial control software compares the current reaction rate v with the preset target reaction rate v0 every 5 minutes, calculates the rate deviation Δv = v0-v, if Δv > 0, that is, the current rate is lower than the target value, the industrial control software re-optimizes the model parameters, adjusts the stirring speed n, reaction temperature T, pressure P and catalyst addition amount m, if Δv ≤ 0, maintain the current process parameters, and the industrial control software updates the reaction rate-process parameter correlation database in real time; Abnormal early warning and emergency treatment step: The industrial control software monitors whether each collected parameter exceeds the safety threshold in real time, when the temperature T exceeds the preset safety range [Tmin, Tmax], the pressure P exceeds [Pmin, Pmax] or the stirring speed n abnormally fluctuates, the industrial control software triggers the sound and light warning, and outputs the emergency treatment instruction, automatically increases the cooling water flow rate V when the temperature exceeds the standard, automatically opens the pressure regulating valve opening K when the pressure exceeds the standard, and automatically switches to the standby stirring motor when the stirring fails; Stirring speed dynamic calibration step: The industrial control software calibrates the stirring speed n in real time according to the material liquid level H and the product manganese ion concentration C during the dynamic control of the stirring system, and the calibrated stirring speed ncal satisfies ncal = n0 + k1 × (H-H0) + k2 × (C0-C), wherein ncal is the calibrated stirring speed, n0 is the basic stirring speed, k1 is the liquid level influence coefficient, H is the current liquid level, H0 is the optimal liquid level, k2 is the concentration influence coefficient, C0 is the target manganese ion concentration, and C is the current manganese ion concentration; Through dynamic calibration, the stirring speed adapts to the material distribution state and reaction progress under different liquid levels.

3. The method of claim 1, wherein the method is based on an intelligent control of the electrolytic manganese compound reaction rate. Also includes: Temperature cumulative deviation compensation step: The industrial control software calculates the temperature cumulative deviation ΔTsum during the temperature control process, adjusts the control amount of the heating or cooling device, and the temperature cumulative deviation ΔTsum satisfies ΔTsum = ∫(t0 to t1)(Topt-T(t))dt, wherein ΔTsum is the temperature cumulative deviation in the period from t0 to t1, t0 is the temperature control start time, t1 is the current time, Topt is the optimal reaction temperature, and T(t) is the actual temperature in the reaction kettle at time t.

4. The method of claim 1, wherein the method is based on an intelligent control of the electrolytic manganese compound reaction rate. In the data acquisition and preprocessing step, the raw material flow sensor is arranged at the outlet of the raw material pretreatment device, the raw material pretreatment device includes a magnetic separator, a screening device and a pretreatment solution soaking tank, the raw material first removes magnetic impurities through the magnetic separator, the magnetic field strength of the magnetic separator is set to 1.2T, then removes the particle impurities with particle size >5mm through the screening device, and finally soaks in the dilute sulfuric acid pretreatment solution, the soaking time is dynamically adjusted according to the impurity content of the raw material.

5. The method of claim 1, wherein the method is based on the intelligent control of the electrolytic manganese compound reaction rate. In the catalyst adding step, the preparation process of the catalyst comprises: dissolving manganese salt and iron salt in ethanol organic solvent according to a mass ratio of 3:1, the manganese salt is selected from manganese sulfate, the iron salt is selected from iron sulfate, citric acid is added as a complexing agent, the amount of the complexing agent is 20% of the total mass of the metal salt, silicon dioxide is added as an additive, the amount of the additive is 5% of the total mass of the metal salt, stirring is conducted at a rotating speed of 600 r / min for 30 min to form a uniform sol, gelation is completed under the conditions of that the temperature of the sol is controlled to be 60-70 DEG C and the pH value is controlled to be 3.5-4.5 by standing for 2 h, the gel is dried in a blast drying oven at 105 DEG C for 12 h, then the gel is calcined in a muffle furnace at 500 DEG C for 2 h, and the gel is crushed to a particle size of 100 meshes after natural cooling, so that the catalyst with a specific surface area of greater than or equal to 100 m2 / g is obtained.

6. The method of claim 1, wherein the method is a method of optimizing the rate of electrolytic manganese compound reaction based on intelligent control. Further comprising: The catalyst adding amount optimization step: the industrial control software calculates the optimal catalyst adding amount mopt according to the raw material grade γ and the current reaction rate v, and the mopt satisfies mopt=m0×(γ0 / γ)×(v0 / v), wherein mopt is the optimal catalyst adding amount, m0 is the catalyst adding amount corresponding to the benchmark grade raw material, γ0 is the benchmark raw material grade, γ is the actual grade of the raw material currently used, v0 is the target reaction rate, and v is the current reaction rate.

7. The method of claim 1, wherein the method is based on an intelligent control of the electrolytic manganese compound reaction rate. Further comprising: The reaction waste heat recovery and utilization step: a plate heat exchanger is arranged at the outlet of the reaction kettle jacket cooling system, the heat exchanger transfers the reaction waste heat carried by the cooling water to the soaking tank in the raw material pretreatment stage, the industrial control software collects the water temperature Tin, Tout at the inlet and outlet of the heat exchanger and the raw material soaking tank temperature Tsoak in real time, calculates the waste heat recovery amount Qrec, and Qrec=ρcVcool×(Tout-Tin), wherein ρ is the density of water, c is the specific heat capacity of water, and Vcool is the cooling water flow rate; when Tsoak is lower than the pretreatment required temperature Tsoak0, the industrial control software controls the opening degree of the bypass valve of the heat exchanger to decrease, increases the heat exchange area, and maintains Tsoak at Tsoak0, so that the raw material is preheated by the waste heat to reduce the raw material heating energy consumption, and the initial temperature of the raw material entering the reaction kettle is closer to the optimal reaction temperature.

8. The method of claim 1, wherein the method is a method of optimizing the rate of electrolytic manganese compound reaction based on intelligent control. In the industrial control software modeling and analysis step, the number of nodes in the input layer of the BP neural network algorithm is set to 6, the number of nodes in the hidden layer is set to 12, the number of nodes in the output layer is set to 4 corresponding to Δn, ΔT, ΔP and Δm, the activation function is selected to be a Sigmoid function, the network training adopts a gradient descent method, the learning rate is set to 0.01, the training iteration number is set to 1000 times, the prediction error of the model to the reaction rate is controlled to be within 5% after the training is completed, and the industrial control software adds the batch process parameters and reaction rate data as new samples to the training set after each batch reaction is completed, and re-trains the network optimization parameters.

9. The method for optimizing the electrolytic manganese combination reaction rate based on intelligent control according to claim 1, characterized in that, In the temperature and pressure control step, the optimal temperature interval is set according to the reaction stage, the optimal temperature in the initial reaction stage is 85-90 DEG C, the optimal temperature in the middle reaction stage is 90-95 DEG C, the optimal temperature in the late reaction stage is 88-92 DEG C, and the industrial control software collects T.

10. The method of claim 1, wherein the method is a method of optimizing the rate of electrolytic manganese compound reaction based on intelligent control. Further comprising: The industrial control software model iterative updating step is: after each batch reaction is finished, the industrial control software automatically stores the batch T, P, n, Q, C, v and reaction time t data into a historical database, 1000 groups of data newly added in the historical database are subjected to model training every month, the weight and threshold of the BP neural network are optimized by using a genetic algorithm, the population size of the genetic algorithm is set to 50, the crossover probability is set to 0.8, the mutation probability is set to 0.05, the iteration number is set to 50 times, and the model performance is verified through small-batch trial production after the training is completed, and the trial production batch size is 10% of the formal production batch size.

Images