Cooperative method for optimizing industrial silicon smelting power consumption and regulating bulking agent dosage

Abstract

The invention relates to the technical field of metallurgical engineering, in particular to an industrial silicon smelting power consumption optimization and bulking agent dosage regulation and control synergistic method. Comprising the following steps: collecting dynamic parameters of the smelting process of the submerged arc furnace in real time through a multi-source sensor: constructing a collaborative optimization model of power consumption and bulking agent dosage, inputting the dynamic parameters into the model, and synchronously solving an optimal solution of an electrode current control value and a bulking agent ratio through a multi-target optimization algorithm; according to an optimization result, a composite control strategy is adopted to synchronously adjust the electrode insertion depth and the bulking agent feeding rate, so that the power consumption efficiency and the bulking agent utilization rate synchronously reach a preset threshold interval; and carrying out online iterative training on the collaborative optimization model based on historical smelting data, and dynamically correcting model parameters. According to the method, the high-precision process model is constructed through multi-source sensing data fusion, and the adaptive high-performance composite bulking agent is designed, so that the industrial difficulty that energy efficiency and furnace conditions are difficult to consider at the same time in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of metallurgical engineering technology, and in particular to a collaborative method for optimizing power consumption in industrial silicon smelting and regulating the amount of a loosening agent. Background Art

[0002] Industrial silicon smelting is a highly energy-intensive industry, with electricity consumption for its core production equipment, the submerged arc furnace, accounting for over 60% of total production costs. Traditionally, electrode current regulation and charge permeability control have been separated into separate steps: On the one hand, current distribution is optimized by adjusting electrode insertion depth to reduce power consumption; on the other hand, bulking agents (such as coke particles) are added based on empirical formulas or fixed ratios to improve charge permeability.

[0003] However, this split control mode has significant defects: first, the electrode current fluctuation is strongly coupled with the change of furnace permeability, and the optimization of a single parameter is prone to local overload or insufficient reaction; second, traditional permeability evaluation is mostly based on static furnace pressure data, without considering the dynamic change of gas flow and the real-time impact of temperature gradient on pore structure, resulting in insufficient accuracy in matching the timing and dosage of loosening agent addition. In addition, conventional loosening agents (such as ordinary coke) have low porosity (usually <50%) and small specific surface area (<300m 2 / g), which is prone to pulverization and failure in high temperature environment, further exacerbating the risk of material compaction in the furnace. Summary of the Invention

[0004] The present invention proposes a dynamic control method that can achieve coordinated optimization of power consumption efficiency and leavening agent utilization. By fusing multi-source sensor data, a high-precision process model is constructed, and an adaptive high-performance composite leavening agent is designed, which solves the industrial dilemma of balancing energy efficiency and furnace conditions in the existing technology.

[0005] The technical solution adopted by the present invention is: a collaborative method for optimizing power consumption in industrial silicon smelting and regulating the amount of bulking agent, comprising the following steps:

[0006] Step 1: Use multi-source sensors to collect dynamic parameters of the submerged arc furnace smelting process in real time, including electrode current, voltage fluctuation, furnace temperature gradient, permeability index calculated based on gas flow and furnace pressure, and exhaust CO concentration:

[0007] Step 2: Construct a collaborative optimization model for power consumption and bulking agent dosage, input the dynamic parameters of step 1 into the model, and use a multi-objective optimization algorithm to simultaneously solve the optimal solution for the electrode current control value and bulking agent ratio;

[0008] Step 3: Based on the optimization results of step 2, a composite control strategy is used to synchronously adjust the electrode insertion depth and the bulking agent feeding rate so that the power consumption efficiency and the bulking agent utilization rate simultaneously reach the preset threshold range;

[0009] Step 4: Conduct online iterative training on the collaborative optimization model in step 2 based on historical smelting data and dynamically modify the model parameters.

[0010] As a further improvement of the present invention, in step one, the charge permeability index is obtained by jointly calculating the measurement values ​​of the coupled gas flow meter and the pressure sensor array, and the calculation process includes three-dimensional parameter fusion of gas flow, pressure difference in the furnace and average furnace pressure.

[0011] As a further improvement of the present invention, the multi-objective optimization algorithm in step 2 adopts an improved non-dominated sorting genetic algorithm, and its optimization objectives include: a unit power consumption cost function with electrode current intensity and smelting time as variables; and a loosening agent utilization function with the effective action volume and actual addition amount of the loosening agent as the core.

[0012] As a further improvement of the present invention, the optimization algorithm sets a dynamic weight adjustment mechanism, which automatically increases the priority of the power consumption cost function in the optimization target when it is detected that the furnace temperature gradient exceeds a safety threshold.

[0013] As a further improvement of the present invention, the composite control strategy in step three includes: a pre-compensation control module based on the prediction of the SiO2 content of the raw material; and a dynamic correction module combined with the real-time data of the furnace thermal imaging.

[0014] As a further improvement of the present invention, the online iterative training of step four adopts a time series convolutional neural network, and the input data includes historical smelting cycle characteristics, physical properties of raw materials and environmental parameters.

[0015] As a further improvement of the present invention, the loosening agent is a composite particle of biomass carbon and nano-silicon oxide, the median particle size of which is controlled within the range of 2-5 mm and the porosity is greater than 65%.

[0016] As a further improvement of the present invention, the biochar is chemically activated, with a specific surface area of ​​not less than 800 m2 / g and an ash content of not more than 3%.

[0017] Beneficial effects of the present invention: The present invention achieves simultaneous optimization of power consumption efficiency, power utilization rate and furnace stability through the collaborative control mechanism of multi-source sensor data fusion and improved intelligent algorithm, combined with the innovation of high-performance composite bulking agent materials. While reducing power consumption per ton of silicon production by 12%-18%, it also reduces CO emissions by 30%, breaking through the bottleneck of mutual constraints between energy efficiency improvement and furnace condition deterioration in traditional processes. DETAILED DESCRIPTION

[0018] In order to make the technical problems, technical solutions and beneficial effects to be solved by this application more clearly understood, this application is further described in detail below in conjunction with the embodiments. It should be understood that the embodiments described herein are only used to explain this application and are not intended to limit this application.

[0019] The present invention provides a collaborative method for optimizing power consumption and regulating the amount of bulking agent in industrial silicon smelting, comprising the following steps:

[0020] Step 1: Use multi-source sensors to collect dynamic parameters of the submerged arc furnace smelting process in real time, including electrode current, voltage fluctuation, furnace temperature gradient, permeability index calculated based on gas flow and furnace pressure, and exhaust CO concentration:

[0021] Step 2: Construct a collaborative optimization model for power consumption and bulking agent dosage, input the dynamic parameters of step 1 into the model, and use a multi-objective optimization algorithm to simultaneously solve the optimal solution for the electrode current control value and bulking agent ratio;

[0022] Step 3: Based on the optimization results of step 2, a composite control strategy is used to synchronously adjust the electrode insertion depth and the bulking agent feeding rate so that the power consumption efficiency and the bulking agent utilization rate simultaneously reach the preset threshold range;

[0023] Step 4: Conduct online iterative training on the collaborative optimization model in step 2 based on historical smelting data and dynamically modify the model parameters.

[0024] In the present invention, in step 1, the charge permeability index is obtained by jointly calculating the measurement values ​​of the coupled gas flow meter and the pressure sensor array, and the calculation process includes the three-dimensional parameter fusion of gas flow, pressure difference in the furnace and average furnace pressure.

[0025] In the present invention, the multi-objective optimization algorithm in step 2 adopts an improved non-dominated sorting genetic algorithm, and its optimization objectives include: a unit power consumption cost function with electrode current intensity and smelting time as variables; and a loosening agent utilization function with the effective action volume and actual addition amount of the loosening agent as the core.

[0026] In the present invention, the optimization algorithm sets a dynamic weight adjustment mechanism, and when it is detected that the furnace temperature gradient exceeds the safety threshold, the priority of the power consumption cost function in the optimization target is automatically increased.

[0027] In the present invention, the composite control strategy in step three includes: a pre-compensation control module based on the prediction of the SiO2 content of the raw material; and a dynamic correction module combined with the real-time data of the furnace thermal imaging.

[0028] In the present invention, the online iterative training of step 4 adopts a temporal convolutional neural network, and the input data includes historical smelting cycle characteristics, physical properties of raw materials and environmental parameters.

[0029] In the present invention, the loosening agent is a composite particle of biomass carbon and nano-silicon oxide, the median particle size of which is controlled within the range of 2-5 mm and the porosity is greater than 65%.

[0030] In the present invention, the biomass charcoal is chemically activated, has a specific surface area of ​​not less than 800 square meters per gram, and has an ash content of not more than 3%.

[0031] Example:

[0032] An industrial silicon smelter uses a 25,000 kVA submerged arc furnace for production. The raw materials are silica (purity ≥ 99.2%) and petroleum coke (fixed carbon content ≥ 85%). The furnace diameter is 6.8 meters. The specific operations of implementing this method are as follows:

[0033] 1. Multi-source sensing system deployment

[0034] Electrode parameter monitoring: High-precision current sensors and voltage sensors are installed on the three self-baking electrodes to collect the fluctuation value of the electrode current and the secondary side voltage data in real time. The sampling frequency is 10 times per second, and the accuracy is controlled within ±0.5%.

[0035] Temperature gradient monitoring: Multiple groups of high-temperature resistant thermocouples are arranged radially along the furnace, among which a circular temperature measurement array is set up 1.2 meters away from the furnace wall, with a measuring point every 0.5 meters. The temperature data is transmitted to the control system in real time via optical fiber.

[0036] Permeability Index Calculation: A pressure sensor array and a high-temperature gas flow meter are installed below the charge surface at the furnace top to measure the pressure differential and gas flow rate in different areas. The permeability index is calculated from the dynamic data of gas flow, pressure differential within the furnace, average furnace pressure, and gas temperature, and is used to evaluate the permeability status of the charge in real time.

[0037] Exhaust gas composition analysis: A laser gas analyzer is deployed at the flue to continuously monitor the concentration of carbon monoxide (CO) in the exhaust gas, with a detection sensitivity of 0.1%.

[0038] (2) Collaborative optimization model construction

[0039] Optimization target design: (1) Power consumption efficiency optimization: With electrode current intensity and smelting time as core variables, the power consumption per unit product is reduced by dynamically adjusting the electrode insertion depth; (2) Optimization of loosening agent utilization: Based on the ratio of the effective action volume of the loosening agent to the actual addition amount, and combined with the influence of exhaust gas CO concentration on utilization, the loosening agent ratio is optimized in real time.

[0040] Intelligent optimization algorithm: An improved non-dominated sorting genetic algorithm (NSGA-II) is used to simultaneously solve the optimal combination of electrode current control value and loosening agent addition amount. When the furnace temperature gradient exceeds the safety threshold (for example, when the radial temperature difference in the furnace exceeds 120°C / meter), the algorithm automatically prioritizes reducing power consumption to ensure stable furnace conditions.

[0041] (3) Implementation of composite control strategy

[0042] Pre-compensation control: Dynamically adjusts the bulking agent addition rate based on real-time detection of the silicon dioxide (SiO2) content in the raw materials. For example, if the SiO2 content drops by 1%, the bulking agent addition rate will increase by 3% to compensate for the change in reaction efficiency.

[0043] Dynamic correction control: The temperature distribution is monitored in real time through the furnace thermal imager. If the area of ​​the local high-temperature area exceeds 15% of the total furnace area, the electrode lifting instruction is immediately triggered to reduce the insertion depth by 5-8 cm to prevent overheating; when the permeability index is lower than the set threshold, the pulse feeding mode is started, and the loosening agent feeding rate is increased by 50% within 10 seconds to quickly improve the permeability of the charge.

[0044] (4) Online model iteration and update.

[0045] A temporal convolutional neural network (TCN) is used to continuously train the collaborative optimization model: the input data includes power consumption, permeability index, raw material characteristics (such as moisture content and particle size) and environmental parameters (such as ambient temperature) in historical smelting cycles; after every five smelting furnaces, the system automatically updates the model parameters and retains the data of the most recent 1000 furnaces as a training set to ensure that the model can adapt to changes in production conditions.

[0046] (5) Preparation and use of composite loosening agent

[0047] Biochar activation treatment: Coconut shells are used as raw materials, which are crushed and then chemically activated (soaked in alkaline solution and calcined in high-temperature inert atmosphere) to obtain biochar with high specific surface area (≥800 square meters / gram) and low ash content (≤3%).

[0048] Composite particle molding: Activated biochar and nano-silicon oxide are mixed in a mass ratio of 7:3, and composite particles with a particle size of 2-5 mm and a porosity of more than 65% are prepared through a centrifugal granulator; the loosening agent maintains structural stability at high temperatures, effectively delays pulverization, and significantly improves the permeability of the charge.

[0049] Compared with the traditional process, the application of this method has achieved the following results: (1) Energy consumption is reduced: the electricity consumption per ton of silicon is reduced from 12,500 kWh to 10,600 kWh, a decrease of 15.2%; (2) Emissions are reduced: the CO concentration in the exhaust gas is reduced from 8.7% to 5.9%, with a significant emission reduction effect; (3) Efficient utilization of resources: the amount of loosening agent used in a single furnace is reduced by 18%, and the furnace stability index is increased to 0.92 (originally 0.78), effectively reducing the risk of material compaction.

[0050] Summary: This embodiment achieves a comprehensive improvement in power consumption efficiency, furnace stability, and environmental performance in industrial silicon smelting through the fusion of multi-source sensor data, dynamic optimization of intelligent algorithms, linkage of composite control strategies, and the coordinated application of high-performance loosening agents, providing a replicable technical path for the green upgrade of high-energy metallurgical processes.

[0051] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting, characterized in that: The following steps are involved: Step 1: Use multi-source sensors to collect dynamic parameters of the submerged arc furnace smelting process in real time, including electrode current, voltage fluctuation, furnace temperature gradient, permeability index calculated based on gas flow and furnace pressure, and exhaust CO concentration: Step 2: Construct a collaborative optimization model for power consumption and bulking agent dosage, input the dynamic parameters of step 1 into the model, and use a multi-objective optimization algorithm to simultaneously solve the optimal solution for the electrode current control value and bulking agent ratio; Step 3: Based on the optimization results of step 2, a composite control strategy is used to synchronously adjust the electrode insertion depth and the bulking agent feeding rate so that the power consumption efficiency and the bulking agent utilization rate simultaneously reach the preset threshold range; Step 4: Conduct online iterative training on the collaborative optimization model in step 2 based on historical smelting data and dynamically modify the model parameters.

2. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 1, characterized in that: In the step 1, the charge permeability index is obtained by jointly calculating the measured values ​​of the coupled gas flow meter and the pressure sensor array, and the calculation process includes the three-dimensional parameter fusion of gas flow, pressure difference in the furnace and average furnace pressure.

3. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 1, characterized in that: The multi-objective optimization algorithm in step 2 adopts an improved non-dominated sorting genetic algorithm, and its optimization objectives include: a unit power consumption cost function with electrode current intensity and smelting time as variables; and a loosening agent utilization function with the effective action volume and actual addition amount of the loosening agent as the core.

4. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 3, characterized in that: The optimization algorithm sets a dynamic weight adjustment mechanism, and when it is detected that the furnace temperature gradient exceeds the safety threshold, the priority of the power consumption cost function in the optimization target is automatically increased.

5. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 1, characterized in that: The composite control strategy in step three includes: a pre-compensation control module based on the prediction of the SiO2 content of the raw material; and a dynamic correction module combined with the real-time data of the furnace thermal imaging.

6. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 1, characterized in that: The online iterative training of step 4 adopts a time series convolutional neural network, and the input data includes historical smelting cycle characteristics, physical properties of raw materials and environmental parameters.

7. A collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claims 1-6, characterized in that: The loosening agent is composite particles of biomass carbon and nano-silicon oxide, the median particle size of which is controlled within the range of 2-5 mm and the porosity is greater than 65%.

8. The collaborative method for optimizing power consumption and regulating bulking agent dosage in industrial silicon smelting according to claim 7, characterized in that: The biomass charcoal is chemically activated, has a specific surface area of ​​not less than 800 square meters per gram, and has an ash content of not more than 3%.