Steam supply system and method for an alumina plant

Abstract

The application discloses an alumina plant steam supply system and method; the system comprises a power grid connection end, a high-voltage electrode steam boiler unit, a steam delivery pipe network and an intelligent control center. The method mainly comprises: collecting photovoltaic power, steam load, meteorological and power grid dispatching data in real time; using a prediction model to generate a future prediction curve of photovoltaic power generation and steam load; based on a model predictive control algorithm, taking maximizing the consumption of photovoltaic power, smoothing power grid power fluctuation and stabilizing steam supply as multiple targets, comprehensively considering the prediction curve, power grid instructions and equipment operation constraints, and rolling solving the optimal output instruction of the high-voltage electrode steam boiler unit; finally, executing the instruction to control the boiler operation and produce steam supply to the alumina plant. Through intelligent prediction and multi-target collaborative optimization control, the application effectively solves the contradiction between fluctuating green energy consumption, power grid peak shaving and stable industrial steam supply.

Description

Technical Field

[0001] This invention relates to the field of industrial steam supply and new energy consumption technology, specifically to a system and method for supplying steam to alumina plants by integrating green electricity. Background Technology

[0002] Currently, photovoltaic power generation exhibits significant intermittency and volatility, with peak output often coinciding with off-peak grid loads, leading to the problem of "curtailment." Photovoltaic power that cannot be consumed locally must be transmitted over long distances through the grid, increasing the load on critical interconnection lines and threatening the safe operation of the power grid.

[0003] To mitigate fluctuations in renewable energy sources, traditionally, deep peak shaving relies on regional thermal power generating units. Frequent and significant load adjustments by these units lead to two major drawbacks: firstly, deterioration in operational economy, with units deviating from their high-efficiency range, resulting in increased coal consumption for power generation; secondly, reduced operational safety, as main and auxiliary equipment such as boilers and turbines are subjected to severe thermal stress cycles, accelerating fatigue damage and shortening their service life.

[0004] Alumina production is a typical energy-intensive process industry, requiring a stable and continuous steam supply. Currently, its steam source mainly relies on self-owned coal-fired boilers or regional combined heat and power (CHP) units. The former has high carbon emission intensity, while the latter, when participating in deep grid peak shaving, may experience fluctuations in steam supply and parameters, affecting the stability and quality of alumina production.

[0005] Although the conversion of electrical energy into steam using electrode boilers is a known technology, existing solutions mostly treat it as a simple electrothermal conversion device, or simply start and stop it based on real-time electricity price signals. This approach fails to systematically resolve the dynamic contradictions between "fluctuating green energy consumption," "grid peak-shaving demand," and "stable industrial steam supply." Therefore, there is an urgent need for a system and method capable of intelligent coordination and optimized operation to comprehensively address the issues of green steam supply, grid ancillary services, and renewable energy consumption in alumina plants. Summary of the Invention

[0006] The technical problem this invention aims to solve is to overcome the shortcomings of existing technologies and provide a steam supply system and method for alumina plants. This system and method can intelligently absorb fluctuating green electricity, produce and supply stable steam, and simultaneously provide flexible peak-shaving capabilities for the power grid, achieving synergistic optimization of energy, economic, and environmental benefits.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A steam supply system for an alumina plant includes a power grid connection terminal, a high-pressure electrode steam boiler unit, a steam transmission pipeline network, and an intelligent control center.

[0008] The grid connection terminal is used to connect to a power network that includes photovoltaic power generation; The high-pressure electrode steam boiler unit is used to convert electrical energy into saturated or superheated steam. The steam transmission pipeline network is used to transport the steam generated by the high-pressure electrode steam boiler unit to the alumina production unit. The improvement lies in that the intelligent control center includes: The data acquisition module is used to acquire real-time weather forecast data, photovoltaic power generation data, automatic power generation control commands from the power grid, and steam load demand data from alumina plants. The prediction module, based on the data acquired by the data acquisition module, uses a trained time-series prediction model to output a photovoltaic power generation prediction curve and a steam load demand prediction curve for a specific future period. The collaborative optimization control module, whose core is the model predictive controller, has multiple objectives: maximizing photovoltaic power absorption, minimizing grid power fluctuations, and stabilizing steam supply. It establishes an optimization objective function and integrates the prediction curve, grid automatic power generation control command constraints, and the operating constraints of the high-pressure electrode steam boiler unit to perform rolling optimization and solve the problem, generating the optimal output command sequence of the high-pressure electrode steam boiler unit for a future period. The instruction issuance and execution module is used to issue the optimal output instruction sequence to the control system of the high-pressure electrode steam boiler unit and drive it to execute.

[0009] A steam supply method for an alumina plant using the above system includes the following steps: S1: Real-time acquisition of photovoltaic power, steam load, meteorological and grid dispatch data through the data acquisition module; S2: Through the prediction module, run the photovoltaic power generation prediction model and the steam load demand prediction model to obtain the future prediction curve; S3: Through the collaborative optimization control module, using the predicted curve as the feedforward input, and based on the multi-objective model predictive control algorithm, the optimal power setpoint sequence of the high-pressure electrode steam boiler unit is solved in a rolling manner under the conditions of satisfying boiler operation constraints and power grid command constraints. S4: Convert the optimal power setpoint sequence into control commands for each boiler, and control the operation of the high-pressure electrode steam boiler unit through the command issuance and execution module to produce steam and supply it to the alumina plant through the pipeline network.

[0010] The beneficial effects of this invention are: 1. Green and low-carbon: It efficiently converts previously restricted photovoltaic "waste electricity" into industrial steam, directly replacing fossil energy and significantly reducing carbon emissions in the alumina production process.

[0011] 2. Safety and Stability: Through forward-looking forecasting and multi-objective optimization, the power load of the electrode boiler is smoothed, greatly reducing the impact on the public grid and enabling rapid response to grid frequency regulation commands, thus improving the safety and stability of the power system. Simultaneously, it ensures a continuous and stable supply of steam required for alumina production.

[0012] 3. Economical and efficient: It avoids the frequent and deep peak shaving required by supporting thermal power units to adapt to new energy sources, keeping them operating within a highly efficient and economical range, thus reducing total system coal consumption and equipment wear. The electrode boiler itself has extremely high electrothermal conversion efficiency, is flexible in operation, and incurs no fuel costs.

[0013] 4. Intelligent Collaboration: Through innovative control architecture and algorithms, deep collaboration and intelligent decision-making of "source-grid-load-storage" in alumina production scenarios are realized, improving overall energy utilization efficiency. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall architecture of the steam supply system for an alumina plant according to the present invention.

[0015] Figure 2 This is a schematic diagram of the internal functional modules and data flow of the intelligent control center of the present invention.

[0016] Figure 3 This is a schematic diagram of the steam supply method for alumina plants in this invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the following embodiments are only for explaining the invention and are not intended to limit the invention.

[0018] refer to Figures 1-3 This embodiment provides a steam supply system for an alumina plant, which is specifically deployed in an alumina industrial park with a supporting thermal power plant and a centralized photovoltaic power station.

[0019] I. System Hardware Composition Power supply connection section: Two power supply lines are connected from the existing 220kV step-up substation power distribution equipment in the plant using the spare interval.

[0020] Substation and distribution section: Two split transformers (rated capacity 240 / 125-125MVA) are installed to reduce the 220kV voltage to a voltage level suitable for electrode boilers (such as 10kV or 6kV). The transformer output side is connected to a high-voltage distribution cabinet, which is equipped with vacuum circuit breakers, protection and control devices, etc.

[0021] High-pressure electrode steam boiler unit: includes 6 vertical, submerged electrode boilers, each with a rated power of 60MW, a rated evaporation capacity of 80t / h, and a designed outlet steam pressure of 2.5MPa (saturation temperature approximately 225℃). Each boiler is an independent module, equipped with pressure, temperature, water level, and conductivity monitoring instruments, as well as inlet and outlet electric valves.

[0022] Boiler feedwater system: A centralized deaerator is installed to heat the chemically treated softened water to above the boiling point and remove dissolved oxygen. Multiple high-pressure feedwater pumps (one in operation and one on standby, or multiple in operation and one on standby) are configured to pressurize the deaerated water and distribute it to each electrode boiler through a main pipe. A circulation pump is also configured to establish water circulation during the initial boiler startup.

[0023] Steam system: Steam generated by each boiler is collected into a single steam header, which is equipped with a safety valve, pressure regulating valve, flow meter, and temperature sensor. The steam header is connected to the existing steam supply network leading to the alumina plant via insulated pipes.

[0024] Intelligent Control Center: This is the "brain" of the system, consisting of high-performance industrial servers, data acquisition devices, and network switches. It communicates with the power grid dispatching system, photovoltaic power station monitoring system, the whole plant DCS, and the local PLC controllers of each electrode boiler via industrial Ethernet.

[0025] II. Software and Algorithm Implementation of the Intelligent Control Center 1. Data Acquisition Module: The dispatch data network receives short-term AGC command curves for the next 15 minutes to 4 hours and real-time ultra-short-term commands issued by the power grid.

[0026] Real-time photovoltaic power and numerical weather forecast data (irradiance, cloud cover, and temperature) for the next 72 hours are obtained from the photovoltaic power plant monitoring system via the OPC protocol.

[0027] The system reads production plans for major steam-consuming equipment such as calcining furnaces from the alumina plant's Production Management System (MES) and historical and real-time steam flow and pressure data from the steam pipeline DCS.

[0028] Read the status (running / stopping), current power, and steam parameters of each boiler from the local PLC of the electrode boiler.

[0029] 2. Prediction Module: Photovoltaic power prediction model: using a long short-term memory network (LSM) LSTMDuring the training phase, historical photovoltaic power data from the past year and corresponding meteorological observation data (total radiation, diffuse radiation, and ambient temperature) for the corresponding time period are used as the training set. In this embodiment, the LSTM network employs two hidden layers with 128 neurons per layer, using the Adam optimizer with a learning rate of 0.001. The photovoltaic power prediction target for the next 15 minutes is used, and the input features are the irradiance, temperature, and historical power time-series data from the past 6 hours. The model is trained offline and updated quarterly, with a validation set mean absolute percentage error (MAPE) ≤ 8%. During online operation, meteorological forecast data for the next 24 hours is used as input, and the output is a sequence of predicted photovoltaic power generation values ​​P_pv(k) with a time resolution of 15 minutes for the next 24 hours.

[0030] Steam load forecasting model: A seasonal autoregressive integral moving average model is adopted. Steam consumption data from the alumina plant over the past two years are analyzed to identify its daily and weekly seasonal patterns. Combined with the received production plan for the next 24 hours (such as the start-up and shutdown schedule of the calcining furnace), the model outputs a sequence of steam load demand forecasts for the next 24 hours with a time resolution of 15 minutes, Q_demand(k).

[0031] 3. Collaborative Optimization Control Module (Core): The collaborative optimization control module, with its core being a model predictive controller, establishes an optimization objective function with multiple objectives: maximizing photovoltaic power absorption, minimizing grid power fluctuations, and stabilizing steam supply. It then integrates the predicted curves, grid automatic generation control command constraints, and the operational constraints of the high-pressure electrode steam boiler unit to perform rolling optimization, generating the optimal output command sequence for the high-pressure electrode steam boiler unit over a future period. Optimization cycle: Set to rolling optimization, executed every 5 minutes.

[0032] Prediction Time Domain and Control Time Domain: During each optimization, the system behavior for the next 4 hours (i.e., k=1 to 16, time step 15 minutes) is predicted, but only the optimization result of the first time step (the next 5 minutes) is issued as the actual control command.

[0033] Optimize the objective function: The controller solves the following multi-objective optimization problem: Minimize J = [ α(k) (P_pv(k) - P_total(k)) 2 + β(k) (P_total(k) - P_total(k-1)) 2 + γ(k) (Q_demand(k) - η P_total(k)) 2 ] in: P_total(k): The total power consumption (MW) of the 6 electrode boilers at time k, which is the optimization variable.

[0034] η: The boiler's electric-to-steam conversion coefficient (t / h / MW), determined according to the design parameters (e.g., 80t / h / 60MW ≈1.333).

[0035] α(k), β(k), γ(k): Time-varying weighting coefficients. The default settings are α=0.5, β=0.3, γ=0.2. When an emergency frequency regulation command is received from the power grid, β is automatically increased to 0.6; when the photovoltaic predicted power is extremely high and curtailment is possible, α is automatically increased to 0.7.

[0036] It should be noted that the electrothermal conversion efficiency η is taken as a constant (1.333 t / h / MW) in this embodiment, which is a reasonable simplification based on the measured characteristic that the efficiency fluctuation of this type of electrode boiler is less than ±2% within the 30%~100% rated load range. For boiler types with larger efficiency fluctuations, η can be constructed as a piecewise function or polynomial fitting curve of the load rate P_total(k) / P_rated, and substituted into the objective function. This processing method still belongs to the equivalent embodiment of the present invention.

[0037] Constraints: a. Power upper and lower limits: 0 ≤ P_total(k) ≤ 360 MW (total rated power of 6 units).

[0038] b. Power ramp rate: |P_total(k) - P_total(k-1)| ≤ 60 MW / 5min (i.e., the total power change does not exceed the rated power of a single boiler).

[0039] c. Grid command constraints: AGC_floor(k) ≤ P_total(k) ≤ AGC_cap(k). This system receives the automatic generation control command value P_AGC_set(k) issued by the grid dispatch center, and constructs the constraint interval AGC_floor(k) = P_AGC_set(k) - δ and AGC_cap(k) = P_AGC_set(k) + δ according to the system frequency regulation dead zone setting parameter δ; or receives the adjustable load upper / lower limit command value directly issued by the grid. In this embodiment, the system receives the automatic generation control command value P_AGC_set(k) issued by the dispatch center, and sets δ = 5MW.

[0040] d. Steam supply safety constraints: η P_total(k) ≥ Q_demand_min, where Q_demand_min is the minimum safe steam flow rate required for the alumina production line. In this embodiment, Q_demand_min is 60t / h, which is determined by the alumina plant's process safety regulations and pre-loaded into the control system. When the production line is partially shut down, it can be manually adjusted by the operator.

[0041] Solver: A quadratic programming solver is used to efficiently solve the constrained quadratic objective function described above, obtaining the optimal total power command sequence {P_total}. (1), P_total (2), ...}.

[0042] 4. Command Issuance and Execution Module: The optimized P_total (1) (i.e., the total power setpoint for the next 5 minutes) is decomposed to each available boiler. The decomposition strategy adopts the principle of "equal incremental rate" and takes into account the minimum stable operating power of the boiler to generate the power setpoint P_i_setpoint for each boiler. The P_i_setpoint is sent to the PLC of the corresponding electrode boiler via the Modbus TCP protocol.

[0043] The boiler PLC precisely controls the input power by adjusting the electrode immersion depth or using thyristor power regulation according to the set value, thereby stably producing steam at the set pressure and flow rate.

[0044] III. Work Process After the system starts, it will run according to the following closed-loop process: 1. At time T0, the data acquisition module acquires all the latest data.

[0045] 2. The prediction module runs immediately, updating the prediction curves of P_pv(k) and Q_demand(k).

[0046] 3. The collaborative optimization control module calls the solver to solve the optimization problem for the next 4 hours based on the system state at time T0 and the latest prediction, and obtains the instruction sequence.

[0047] 4. The instruction issuance and execution module issues and executes the first instruction in the sequence (corresponding to time T0+5min).

[0048] 5. The system waits until T0+5min, then repeats steps 1-4 to achieve optimization and control of the rolling forward.

[0049] Through the detailed implementation methods described above, this system can operate intelligently under various working conditions: for example, during peak solar power generation at midday, the system automatically increases the total boiler power to near the solar power output, maximizing the absorption of green electricity; when clouds pass by causing a sudden drop in solar power, the optimization algorithm has already smoothed out the planned reduction in boiler power, avoiding the impact caused by absorbing a large amount of power from the grid; throughout the entire process, the steam supply always meets production requirements and strictly follows the grid's dispatch instructions. This fully demonstrates the integrated and synergistic advantages of this invention.

[0050] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A steam supply system for an alumina plant, characterized in that, include: The grid connection terminal is used to connect to a power network that includes photovoltaic power generation. The high-pressure electrode steam boiler unit is electrically connected to the power grid connection terminal and is used to convert electrical energy into industrial steam. A steam transmission pipeline network is connected to the steam outlet of the high-pressure electrode steam boiler unit and is used to transport steam to the alumina production unit. The intelligent control center is communicatively connected to the power grid connection terminal and the high-pressure electrode steam boiler unit; the intelligent control center includes: The data acquisition module is used to acquire real-time weather forecast data, photovoltaic power generation data, automatic power generation control commands from the power grid, and steam load demand data from alumina plants. The prediction module, connected to the data acquisition module, is used to output the photovoltaic power generation prediction curve and the steam load demand prediction curve for future periods based on the acquired data and through a time-series prediction model. The collaborative optimization control module, whose core is a model predictive controller, is configured to: establish an objective function with the optimization objectives of maximizing photovoltaic power absorption, smoothing grid power fluctuations, and stabilizing steam supply. J = Σ [ α · (P_pv(k) - P_boiler(k))² + β · (P_boiler(k) - P_boiler(k-1))² + γ · (Q_demand(k) - Q_steam(k))² ], in: k is the time index in the prediction time domain. P_pv(k) represents the predicted photovoltaic power generation at time k. P_boiler(k) is the total electrical power setting value of the high-pressure electrode steam boiler unit at time k. Q_demand(k) represents the predicted steam load at time k, and Q_steam(k) represents the predicted steam production at time k. α, β, and γ are coefficients used to weigh the weights of each optimization objective; The optimal output command sequence of the high-pressure electrode steam boiler unit is generated by combining the predicted curve, the automatic power generation control command constraints of the power grid, and the operating constraints of the high-pressure electrode steam boiler unit. The instruction issuance and execution module is used to issue the optimal output instruction sequence to the control system of the high-pressure electrode steam boiler unit.

2. The steam supply system for an alumina plant according to claim 1, characterized in that, The weighting coefficients α, β, and γ can be dynamically adjusted according to the rate and / or amplitude of change of the automatic power generation control command.

3. The steam supply system for an alumina plant according to claim 1, characterized in that, The high-pressure electrode steam boiler unit includes multiple high-pressure electrode steam boilers connected in parallel; the optimal output command sequence generated by the collaborative optimization control module includes start-up, shutdown, and power allocation commands for each high-pressure electrode steam boiler.

4. The steam supply system for an alumina plant according to any one of claims 1 to 3, characterized in that, The high-pressure electrode steam boiler unit is connected to the power grid through a spare bay of the plant's booster station.

5. A method for supplying steam to an alumina plant, characterized in that, The method, applied to the steam supply system of an alumina plant as described in any one of claims 1 to 4, comprises: Data acquisition steps: Real-time acquisition of photovoltaic power generation data, weather forecast data, automatic power generation control commands from the power grid, and steam load demand data from the alumina plant; Forecasting steps: Based on the acquired data, a forecasting model is used to generate a photovoltaic power generation forecast curve and a steam load demand forecast curve for a specific future period. Collaborative optimization control steps: With multiple objectives including maximizing photovoltaic power absorption, smoothing grid power fluctuations, and stabilizing steam supply, a model predictive control algorithm is employed, with the objective function J as follows: J = Σ [ α · (P_pv(k) - P_boiler(k))² +β · (P_boiler(k) - P_boiler(k-1))² + γ · (Q_demand(k) - Q_steam(k))² ]; Where k is the time index in the prediction time domain. P_pv(k) represents the predicted photovoltaic power generation at time k. P_boiler(k) is the total electrical power setting value of the high-pressure electrode steam boiler unit at time k. Q_demand(k) is the predicted steam load at time k. Q_steam(k) represents the predicted steam production at time k. α, β, and γ are weighting coefficients; Combining the predicted curve, the automatic power generation control command constraints of the power grid, and the operating constraints of the high-pressure electrode steam boiler unit, a rolling optimization solution is performed to generate the optimal power setpoint sequence of the high-pressure electrode steam boiler unit. Instruction execution steps: Convert the optimal power setpoint sequence into control instructions to control the operation of the high-pressure electrode steam boiler unit, produce steam, and supply it to the alumina plant through the pipeline network.

6. The steam supply method for an alumina plant according to claim 5, characterized in that, The weighting coefficients α, β, and γ are dynamically adjusted according to the rate of change and / or amplitude of the automatic power generation control command.

Images