Coal-fired unit dynamic entropy weight optimization control method and system fusing NSGA-II and PSO

Abstract

The invention discloses a coal-fired unit dynamic entropy weight optimization control method and system fusing NSGA-II and PSO. The method and system are suitable for collaborative operation optimization of a machine system, a furnace system and an electric system of a large coal-fired power generation unit. The method comprises the steps that a simulation model based on a thermodynamic system structure is established, and correction is carried out in combination with unit performance test data; constructing a multi-objective optimization function including power supply coal consumption, station service power consumption rate and waste heat utilization rate; a dynamic entropy weight distribution mechanism is introduced to carry out real-time adaptive adjustment on the target weight; iteratively solving a Pareto solution set by adopting an NSGA-II and PSO fused multi-target hybrid optimization algorithm; performing satisfaction evaluation on the candidate solutions through a fuzzy membership function, and selecting an optimal solution scheme; and finally, the optimization control parameters are issued to an execution mechanism through a DCS system, and optimization strategy closed-loop control is realized. Compared with an existing method, the method can achieve coordinated optimization among multiple targets, has high self-adaptive capacity and energy-saving effect, and is suitable for operation control requirements under complex working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of energy-saving optimization and operation control of thermal power units, and in particular to a dynamic entropy weight optimization control method and system for coal-fired units integrating NSGA-II and PSO algorithms. Background Art

[0002] Energy conservation, consumption reduction, and operational optimization of coal-fired power generation units have become important means of improving energy efficiency and reducing carbon emissions. This is especially true for large, high-parameter coal-fired power generation units, such as supercritical and double-reheat units, whose engine, furnace, and electrical systems are complex and closely coupled with each other, offering significant optimization potential and adjustment space. However, most power plants currently rely primarily on operational control strategies based on empirical rules, which suffer from significant deficiencies in terms of systematicity, multi-objective coordination, and adaptability to dynamic operating conditions.

[0003] Traditional optimization control strategies typically rely on manual experience or preset control curves for adjustment. This means that operators optimize operations based on the unit's load level or operating conditions by setting fixed control parameters such as main steam pressure, flue gas bypass damper opening, and economizer water volume. This approach relies on the operator's experience, and the adjustment method is primarily single-parameter or local optimization, making it difficult to achieve coordinated optimization control of the entire unit system. Due to the lack of system modeling and global analysis methods, the energy flow and operational impact between different subsystems are not fully considered, resulting in limited improvements in the overall operating efficiency of the unit.

[0004] In addition, the empirical control strategy has a weak ability to adapt to external disturbances and operating condition fluctuations. Under complex operating conditions such as deep peak regulation, coal quality fluctuations, and thermal load offsets, traditional rules cannot effectively adjust the control strategy, which may cause problems such as main steam pressure fluctuations and insufficient waste heat utilization. At the same time, with the increasing environmental protection constraints, unit operation optimization often faces the coexistence of multiple objectives such as reducing power supply coal consumption, reducing plant power consumption, and improving waste heat utilization. Traditional control methods lack a multi-objective coordination mechanism and usually perform synthetic optimization by manually setting the weights of the objective function. However, this method is highly subjective, and the weights are difficult to adaptively adjust according to real-time operating conditions. It has problems of poor stability and weak coordination.

[0005] In recent years, some research has attempted to incorporate multi-objective evolutionary algorithms (such as NSGA-II and PSO) into the optimization and control of coal-fired power plants, achieving some exploratory results. However, existing methods are often limited to fixed objective weights and static optimization, lacking a dynamic weight allocation mechanism linked to the system's operating status. This makes it difficult for the algorithms to maintain robustness and flexibility under complex operating conditions. Furthermore, effectively integrating the optimization results with the unit's actual control system and achieving closed-loop execution remains a bottleneck restricting the engineering application of existing optimization methods.

[0006] In summary, the existing operation optimization technology of coal-fired power generation units still has shortcomings in model accuracy, target coordination, weight distribution, adaptability and control execution closed loop. There is an urgent need for a new optimization control method that integrates system modeling capabilities, multi-objective intelligent optimization methods and dynamic weight adjustment mechanisms, and has effective integration with the unit control system to achieve global collaborative optimization and deep energy saving goals under complex working conditions. Summary of the Invention

[0007] The purpose of the present invention is to overcome the shortcomings of existing coal-fired power generation unit operation optimization methods, such as reliance on experience adjustment, poor coordination of optimization objectives, strong subjectivity in weight setting, and difficulty in closed-loop control execution. A dynamic entropy weight optimization control method for coal-fired units that integrates NSGA-II and PSO is proposed to achieve collaborative optimization control of multiple objective indicators of the unit. It is particularly suitable for large-scale high-parameter coal-fired power generation units with flue gas waste heat recovery systems.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] The dynamic entropy weight optimization control method of coal-fired units integrating NSGA-II and PSO includes the following steps:

[0010] Step S1: Establishing a thermodynamic simulation model of a large coal-fired power generation unit;

[0011] Step S2: performing parameter correction and multi-operating condition verification on the thermodynamic simulation model based on unit performance test data;

[0012] Step S3: Determine the optimization objective function of the thermodynamic simulation model, including power supply coal consumption, plant power consumption rate, and waste heat utilization rate indicators, and set key operating parameters as optimization variables;

[0013] Step S4: Solve the multi-objective function using the hybrid optimization algorithm of NSGA-II and PSO, and output the Pareto solution set;

[0014] Step S5: Introduce a dynamic entropy weight mechanism in the optimization process, dynamically calculate the weight of each objective based on the information entropy value of the objective function of the candidate solution set and update it in real time;

[0015] Step S6: Evaluate the comprehensive satisfaction of the Pareto solution set by combining the fuzzy membership function and select the optimal solution;

[0016] Step S7: The control parameters in the optimal solution are sent to the actuators through the discrete control system DCS to achieve closed-loop optimization control of the coal-fired unit.

[0017] A further improvement of the present invention is that the thermodynamic simulation model covers key units of the boiler, turbine, economizer, heater and flue gas system, and the model parameters are corrected by comparing with performance test data under typical load points, including pipeline pressure loss, turbine stage efficiency and forced draft fan and induced draft fan efficiency indicators, to ensure model accuracy.

[0018] A further improvement of the present invention is that the objective function of the power supply coal consumption adopts the following form:

[0019]

[0020] Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

[0021] A further improvement of the present invention is that the target weight calculation method of the dynamic entropy weight mechanism includes the following steps:

[0022] Step 1: Perform extreme value normalization on the objective function value of each candidate solution;

[0023] Step 2: Information entropy value of technical goal j:

[0024]

[0025] where p ij is the normalized value, n is the number of solutions involved in the evaluation, and ε is the smoothing factor;

[0026] Step 3: Calculate the coefficient of variation d j =1-E j , and find the target weight

[0027] A further improvement of the present invention is that the fuzzy membership function adopts the form of an S-type function, maps the normalized objective function to the interval [0, 1], evaluates the satisfaction of each optimization solution with respect to multiple objectives, calculates the total membership and selects the largest one as the final optimal solution.

[0028] A further improvement of the present invention is that the optimization variables include: air preheater flue gas bypass damper opening, high pressure economizer feed water flow, low pressure economizer condensate flow and air preheater circulating water flow.

[0029] A further improvement of the present invention is that the control parameters are sent down through the data interface of the unit DCS system. The control parameters include the flue gas damper opening, the flow control valve opening and the variable frequency pump speed adjustment control amount, and an adjustment rate limiting mechanism is provided.

[0030] The dynamic entropy weight optimization control system for coal-fired power units that integrates NSGA-II and PSO includes:

[0031] Thermodynamic simulation modeling module, to establish thermodynamic simulation models of large coal-fired power generation units;

[0032] An entropy weight calculation module, which performs parameter correction and multi-operating condition verification on the thermodynamic simulation model based on unit performance test data;

[0033] Optimization objective function module, which determines the optimization objective function of the thermodynamic simulation model, including power supply coal consumption, plant power consumption rate and waste heat utilization rate indicators, and sets key operating parameters as optimization variables;

[0034] The solution module uses a hybrid optimization algorithm that combines NSGA-II and PSO to solve multi-objective functions and output the Pareto solution set;

[0035] Dynamic calculation and update module, which introduces a dynamic entropy weight mechanism in the optimization process, dynamically calculates the weight of each target based on the information entropy value of the objective function of the candidate solution set and updates it in real time;

[0036] The optimal solution selection module combines the fuzzy membership function to evaluate the comprehensive satisfaction of the Pareto solution set and select the optimal solution;

[0037] The execution module sends the control parameters in the optimal solution to the actuator through the discrete control system DCS, realizing closed-loop optimization control of the coal-fired unit.

[0038] A further improvement of the present invention is that the thermodynamic simulation model in the thermodynamic simulation modeling module covers key units of the boiler, turbine, economizer, heater and flue gas system, and the model parameters are corrected by comparing with the performance test data under typical load points, including pipeline pressure loss, turbine stage efficiency and forced draft fan and induced draft fan efficiency indicators, to ensure model accuracy.

[0039] A further improvement of the present invention is that the objective function of the power supply coal consumption in the optimization objective function module adopts the following form:

[0040]

[0041] Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

[0042] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0043] The dynamic entropy weight optimization control method and system for coal-fired units that integrate NSGA-II and PSO, provided by the present invention, construct an accurate multi-system coupled thermal simulation model, and solve the problem of insufficient accuracy of "model-driven" in unit optimization; a dynamic entropy weight allocation mechanism is proposed, so that the target weights can be adjusted in real time with the operating status during the optimization process, overcoming the subjectivity and non-adaptability of fixed target weighting in traditional optimization; the NSGA-II and PSO fusion algorithm is adopted to improve the globality and local accuracy of the solution, and the optimization results are stable, which is suitable for multi-objective balance under complex constraints; the data linkage between optimization calculation and DCS control system is realized, and it has good engineering deployment capabilities and can be directly applied to the operation optimization control system of coal-fired power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the specific embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a schematic diagram of the thermal system structure of the coal-fired unit of the present invention;

[0046] Figure 2 This is a system architecture diagram of the method of the present invention;

[0047] Figure 3 To optimize the control flow graph;

[0048] Figure 4 This is the flow chart of dynamic entropy weight calculation;

[0049] Figure 5 is a schematic diagram of the fuzzy membership function;

[0050] Figure 6 This is a structural block diagram of the dynamic entropy weight optimization control system of a coal-fired unit that integrates NSGA-II and PSO in the present invention. DETAILED DESCRIPTION

[0051] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and description are to be considered as illustrative in nature and not restrictive.

[0052] In the description of the present invention, it is to be understood that when used in this specification and the appended claims, the terms “include” and “comprise” indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0053] It should also be understood that the terms used in the present specification are only for the purpose of describing particular embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0054] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0055] The accompanying drawings illustrate various schematic diagrams of structures according to embodiments disclosed herein. These figures are not drawn to scale; for clarity, some details are exaggerated and some details may be omitted. The shapes of the various regions and layers shown in the figures, as well as their relative sizes and positional relationships, are merely exemplary and may deviate in practice due to manufacturing tolerances or technical limitations. Those skilled in the art may design regions / layers with different shapes, sizes, and relative positions as needed.

[0056] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0057] Example 1

[0058] The dynamic entropy weight optimization control method for coal-fired units that integrates NSGA-II and PSO is applied to the operation optimization of the machine, furnace, and electrical systems of large coal-fired power generation units with deep utilization of flue gas waste heat. It is characterized by comprising the following steps:

[0059] Step S1: establishing a thermodynamic simulation model of the generator set;

[0060] Step S2: performing parameter correction and multi-operating condition verification on the model based on unit performance test data;

[0061] Step S3: Determine the optimization objective function, including indicators such as power supply coal consumption, plant power consumption rate, and waste heat utilization rate, and set key operating parameters as optimization variables;

[0062] Step S4: Solve the multi-objective function using the hybrid optimization algorithm of NSGA-II and PSO, and output the Pareto solution set;

[0063] Step S5: Introduce a dynamic entropy weight mechanism in the optimization process, dynamically calculate the weight of each objective based on the information entropy value of the objective function of the candidate solution set and update it in real time;

[0064] Step S6: Evaluate the comprehensive satisfaction of the Pareto solution set by combining the fuzzy membership function and select the optimal solution;

[0065] Step S7: The control parameters in the optimal solution are sent to the actuators through the discrete control system DCS to achieve closed-loop optimization control of the coal-fired unit.

[0066] In this embodiment, the thermodynamic simulation model covers key units such as boilers, turbines, economizers, heaters, and flue gas systems. The model parameters are corrected by comparing with performance test data at typical load points, including indicators such as pipeline pressure loss, turbine stage efficiency, and draft and induced draft fan efficiency, to ensure model accuracy.

[0067] In this embodiment, the objective function of the power supply coal consumption is in the following form:

[0068]

[0069] Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

[0070] In this embodiment, the target weight calculation method of the dynamic entropy weight mechanism includes the following steps:

[0071] Step 1: Perform extreme value normalization on the objective function value of each candidate solution;

[0072] Step 2: Information entropy value of technical goal j:

[0073]

[0074] where p ij is the normalized value, n is the number of solutions involved in the evaluation, and ε is the smoothing factor;

[0075] Step 3: Calculate the coefficient of variation d j =1-E j , and find the target weight

[0076] In this embodiment, the fuzzy membership function adopts the form of an S-type function, maps the normalized objective function to the interval [0, 1], evaluates the satisfaction of each optimization solution with respect to multiple objectives, calculates the total membership, and selects the largest one as the final optimal solution.

[0077] In this embodiment, the optimization variables include but are not limited to: the opening of the air preheater flue gas bypass damper (actuator type), the high-pressure economizer feed water flow (regulating valve setting value type), the low-pressure economizer condensate flow (variable frequency pump flow type), and the air preheater circulating water flow (flow regulation type).

[0078] In this embodiment, the control parameters are sent down through the data interface of the unit DCS system. The control parameters include control quantities such as the flue gas damper opening, the flow control valve opening, and the variable frequency pump speed adjustment. A regulation rate limiting mechanism is also provided to avoid system fluctuations caused by excessive regulation.

[0079] Example 2

[0080] The present invention provides a coal-fired unit dynamic entropy weight optimization control method integrating NSGA-II and PSO, including:

[0081] (1) Working condition simulation

[0082] This specific example is based on a 660MW double-reheat ultra-supercritical coal-fired power generation unit. This unit has a main steam pressure of 31MPa, a main steam temperature of 600°C, and a two-stage reheat steam temperature of approximately 610°C. The unit utilizes double reheat to improve cycle efficiency. Under a typical 50% rated load (approximately 300MW output), the unit operates in the partial load region. This region reduces boiler and turbine thermal efficiencies relative to full load, increases exhaust losses, and increases coal consumption.

[0083] This embodiment uses Ebsilon thermal simulation software to establish a thermal model of the entire unit for the 50% load steady-state operating condition, including key equipment such as boilers, steam turbines, reheaters, air preheaters, and flue gas bypass waste heat utilization systems. The model is calibrated by calling the unit design parameters (including thermal balance design values) and combining them with actual performance test data to ensure that the model can accurately reflect the characteristics of the unit under different loads. The simulation results show that the deviation between the main steam flow, coal consumption rate, exhaust gas temperature, etc. calculated by the model at the 50% load point and the test values ​​is within ±1%, which meets the accuracy requirements and provides a reliable digital test platform for optimization analysis. Figure 1 As shown in the figure, the thermal system structure of the coal-fired power generation unit involved in the present invention includes key units such as boiler, steam turbine, reheater, economizer, air preheater and flue gas waste heat recovery system. The energy flow and connection method between each device are shown in the figure, which constitute the basis of the optimization control model.

[0084] (2) Optimization variables and objective function settings

[0085] Based on the above simulation model, a multi-objective optimization control scheme integrating NSGA-II genetic algorithm and PSO particle swarm optimization is constructed. First, the optimized decision variables (optimization control parameters) and objective function are clarified as follows:

[0086] In the selection of optimization decision variables, the adjustable parameters that have a significant impact on the unit energy efficiency and waste heat recovery are selected, including:

[0087] (a) The unit electrical load instruction is given as 50% rated load as a boundary condition;

[0088] (b) The opening of the flue gas bypass damper of the air preheater controls whether part of the flue gas bypasses the air preheater and enters the low-temperature waste heat utilization circuit;

[0089] (c) High-pressure / low-pressure economizer feedwater flow distribution ratio, which adjusts the distribution of boiler feedwater between high-pressure and low-pressure economizers to affect the exhaust gas temperature and recovered heat;

[0090] (d) The cold-end circulating water flow of the air preheater (air heater), which is the water-side flow used for deep recovery of flue gas waste heat.

[0091] The above variables can all be set or influenced through DCS, and their adjustment directly affects the energy distribution and loss size of the unit's thermal system.

[0092] When selecting the objective function, the unit economy and waste heat utilization effect are considered simultaneously, and three optimization goals are set:

[0093] (a) Reduce coal consumption for power generation, that is, reduce the amount of fuel consumed per unit of net power generation;

[0094] (b) Reduce the plant power consumption rate, that is, reduce the proportion of the unit's own power consumption to the power generation, and increase the proportion of net output power;

[0095] (c) Improve the utilization rate of flue gas waste heat, that is, improve the degree of recovery and utilization of available waste heat in boiler exhaust gas (reduce exhaust gas temperature and increase the utilization of low-grade thermal energy).

[0096] These objectives are trade-offs. For example, reducing coal consumption for power generation often requires more waste heat recovery or optimizing operating parameters, but this can lead to increased power consumption for auxiliary equipment. Improving waste heat utilization (lowering exhaust gas temperatures) requires preventing corrosion caused by excessively low temperatures and potentially increasing induced draft fan loads. Therefore, this optimization problem is multi-objective, requiring a compromise between these objectives.

[0097] To quantitatively describe the multi-objective optimization problem, this embodiment constructs a comprehensive evaluation function to evaluate the pros and cons of each set of decision variables. A fuzzy membership function is used to normalize each target indicator to a satisfaction value between 0 and 1 to facilitate dimensionless comprehensive comparison. For the target to be minimized (such as power supply coal consumption and plant power consumption rate), its satisfaction membership is defined as:

[0098]

[0099] For the target to be maximized (such as waste heat utilization), define

[0100]

[0101] in, Represents the value of the j-th candidate solution on the i-th objective function; represents the function value of the j-th solution on the i-th target; f i,max , f i,min Respectively represent the minimum and maximum values ​​of the objective function in the current population. Through the above linear membership conversion, the indicators of different dimensions can be unified into an evaluation scale of "the closer to 1, the better, the closer to 0, the worse". Figure 5 As shown in the figure, the fuzzy membership function adopts the form of S-type function, and different objective function values ​​are mapped to

[0102] The interval [0, 1] is used to measure the satisfaction of each candidate solution with respect to each objective. The steepness of the function curve can be adjusted through parameters to meet the sensitivity requirements of different indicators.

[0103] Then, the information entropy weighting method is introduced to dynamically determine the weight coefficient of each target according to the distribution characteristics of the population solution. Specifically, for the i-th target, according to the membership degree of all M candidate solutions in the current population Calculate its information entropy:

[0104]

[0105] in, It represents the membership ratio of the jth solution relative to other solutions under the i-th goal, and M is the population size. Information entropy E i Between 0 and 1, it is used to represent the degree of difference of the target value in the current solution set: if E i If the value is high, it means that the differences among the solutions of the i-th target are small and the amount of information is low, and its weight should be reduced; if E i If it is low, it means that the solutions have great differences in this goal and the amount of information is high, so its weight should be increased. Based on this, the dynamic weight coefficient of the i-th goal is defined as:

[0106]

[0107] Where N is the target number (in this example, N = 3). i Satisfy∑ i w i =1, and can be adjusted dynamically with the evolution of the population, reflecting the changes in the relative importance of each goal in different optimization stages.

[0108] Finally, the comprehensive evaluation function can be expressed as the weighted sum of the satisfaction of each goal:

[0109]

[0110] Among them, F (j) F represents the comprehensive evaluation value of the jth solution. (j) The larger the value, the better the comprehensive performance of solution j in taking into account various objectives. When F reaches its maximum value and converges, the corresponding decision variable combination is considered to be the optimal operating setting under the current working conditions. Figure 2 As shown in the figure, the dynamic entropy weight optimization control method constructed by the present invention includes six major links: thermal model construction, optimization target setting, multi-objective optimization solution integrating NSGA-II and PSO, dynamic entropy weight allocation, fuzzy membership function evaluation and DCS closed-loop control. The system architecture is shown in the figure.

[0111] (3) Optimization algorithm and solution process

[0112] like Figure 3 As shown, the optimization control process of the present invention includes the steps of initial population generation, NSGA-II global search, PSO local search, dynamic entropy weight adjustment and optimal solution selection to form a closed-loop optimization control process. When solving multi-objective optimization problems, this embodiment adopts a hybrid optimization algorithm that integrates the improved NSGA-II and PSO. When the algorithm is initialized, several groups of random initial decision variables (i.e., initial populations) are generated based on methods such as Latin hypercube, and the corresponding target values ​​of each item are calculated by inputting the simulation model. Subsequently, the optimal solution is iteratively searched through the following steps:

[0113] Step 1: A global genetic algorithm search is performed, applying the selection, crossover, and mutation operations of the NSGA-II (Non-Dominated Sorting Genetic Algorithm) to the population to generate a new generation of candidate solutions. NSGA-II ranks the solution set based on Pareto goodness, using mechanisms such as crowding distance to retain non-dominated solutions for each objective and maintain diversity in the solution set. This allows exploration of various regions of the decision space, finding a series of Pareto-optimal solutions that balance different objectives.

[0114] Step 2: Particle swarm local optimization, treat the current population solution as a particle swarm, and use the PSO algorithm to further refine the search.

[0115] During the update, each particle adjusts its decision variables based on its own historical optimal solution and the global optimal solution:

[0116]

[0117] where v m is the particle velocity, x m is the particle position (i.e., a set of decision variable values), ω is the inertia weight, c1, c2 are acceleration constants, and r1, r2 are random numbers between 0 and 1; represents the optimal solution experienced by the particle itself, x best Represents the global population optimal solution. PSO uses collaborative search between particles to quickly approach the local optimal area, thereby improving convergence accuracy and speed.

[0118] Step 3: Dynamic entropy weight adjustment. In each iteration, the target weights w are updated in real time by combining the above fuzzy membership and entropy weight calculations. i , and update the comprehensive evaluation function F accordingly (j) .like Figure 4 As shown, the dynamic entropy weight calculation process includes objective function normalization, information entropy calculation, variance coefficient determination, and dynamic weight coefficient updating, ensuring that the weight assignment adapts to the solution distribution at different optimization stages. Dynamic weight adjustment means that in the early exploration phase, the algorithm focuses on global optimization, maintaining balanced weights across objectives to expand the search scope. In the later convergence phase, the algorithm adaptively increases the weights of certain key objectives to accelerate local convergence and prevent the population from stagnating at suboptimal solutions. For example, when the variance of power supply coal consumption values ​​within the population increases and information entropy decreases, the algorithm assigns a higher weight to the coal consumption objective, thereby more actively optimizing it. Conversely, if a particular objective converges among candidate solutions, with decreasing variance and increasing entropy, its weight is temporarily reduced to redirect search efforts toward other objectives that have not yet converged. This dynamic entropy weight mechanism ensures a balance between multiple objectives, enabling the algorithm to maintain both global search capability and convergence efficiency.

[0119] Step 4: Iterative optimization and termination, repeat the above genetic operation, particle swarm update and weight adjustment process to continuously evolve the population. The termination condition of the algorithm can be set as the number of iterations reaches a preset upper limit (for example, 100 generations) or the global optimal comprehensive evaluation function F is reached for several consecutive generations. best The improvement is below the threshold (convergence criterion). After the iteration is complete, the Pareto optimal solution set and the corresponding target values ​​are output. The solution with the highest overall evaluation value is selected as the recommended optimal operation plan. If necessary, a specific solution from the Pareto solution set can be selected based on the decision maker's preference (for example, a solution that focuses more on economy or energy conservation).

[0120] (4) Comparison of optimization results

[0121] The above optimization algorithm was used to simulate the 50% load condition of a 600MW unit, and the main performance indicators before and after optimization were obtained. The comparison is as follows:

[0122] Coal consumption for power generation was approximately 295g / kWh before optimization, but it dropped to approximately 292g / kWh after optimization, a reduction of approximately 3g / kWh and a 1% improvement in thermal efficiency. In other words, optimized control can save approximately 0.5 tons of coal per hour (based on a 300MW unit output) at the same load, achieving significant coal-saving results over the long term.

[0123] Before optimization, the boiler's exhaust temperature was approximately 120°C, but after optimization, it dropped to approximately 105°C. This 15°C reduction in exhaust temperature means more waste heat is recovered and utilized. Simulations show that optimized control increases waste heat absorption in the low-temperature economizer and air heater circuits, significantly reducing the proportion of waste heat emitted from the flue gas.

[0124] The flue gas waste heat utilization rate was approximately 45% before optimization and increased to approximately 53% after optimization. This 8-percentage-point increase indicates that, after optimization, a greater proportion of the low-grade heat energy at the boiler tail is effectively utilized (for preheating feedwater or combustion air), reducing energy losses. This is consistent with the reduction in exhaust gas temperature, demonstrating that the optimization scheme successfully recovered previously lost heat.

[0125] Before optimization, the auxiliary power consumption rate was approximately 5.0%. After optimization, it dropped to approximately 4.9%, a decrease of 0.1 percentage points. This slight improvement was primarily due to the optimized operating parameters reducing the load on auxiliary equipment such as the induced draft fan and feedwater pump (e.g., lower fan power consumption as exhaust gas temperature decreases, and optimized feedwater distribution reduces pump work), thereby slightly reducing the unit's own power consumption. While the reduction is not significant, it will help further increase net output power.

[0126] The above results demonstrate that the dynamic entropy weighted optimization control method, combining NSGA-II with PSO, significantly improves the unit's operating economy at 50% part load. The reduction in power generation coal consumption means less fuel consumption while maintaining the same power output, which directly reduces power generation costs and CO2 emissions. The lower exhaust temperature and improved waste heat utilization indicate that the unit is utilizing energy more efficiently, reducing the waste of low-grade thermal energy. The reduced plant power utilization rate further improves the unit's net efficiency. It is worth emphasizing that the above optimization is achieved while meeting all the unit's operating constraints. For example, the minimum exhaust temperature is limited by the sulfur content of the coal type to prevent acid dew point corrosion. In this example, the lower exhaust temperature is set at 90°C to ensure safety. During the optimization process, the main steam pressure and reheat steam temperature remained within ±0.5% of the set values, preventing excessive fluctuations in turbine thermal parameters. This demonstrates that this optimization method not only focuses on improving efficiency but also takes into account operational safety and reliability.

[0127] Example 3

[0128] like Figure 6As shown, the present invention provides a coal-fired unit dynamic entropy weight optimization control system integrating NSGA-II and PSO, including:

[0129] Thermodynamic simulation modeling module, to establish thermodynamic simulation models of large coal-fired power generation units;

[0130] An entropy weight calculation module, which performs parameter correction and multi-operating condition verification on the thermodynamic simulation model based on unit performance test data;

[0131] Optimization objective function module, which determines the optimization objective function of the thermodynamic simulation model, including power supply coal consumption, plant power consumption rate and waste heat utilization rate indicators, and sets key operating parameters as optimization variables;

[0132] The solution module uses a hybrid optimization algorithm that combines NSGA-II and PSO to solve multi-objective functions and output the Pareto solution set;

[0133] Dynamic calculation and update module, which introduces a dynamic entropy weight mechanism in the optimization process, dynamically calculates the weight of each target based on the information entropy value of the objective function of the candidate solution set and updates it in real time;

[0134] The optimal solution selection module combines the fuzzy membership function to evaluate the comprehensive satisfaction of the Pareto solution set and select the optimal solution;

[0135] The execution module sends the control parameters in the optimal solution to the actuator through the discrete control system DCS, realizing closed-loop optimization control of the coal-fired unit.

[0136] In an embodiment, the thermodynamic simulation model in the thermodynamic simulation modeling module covers key units of the boiler, turbine, economizer, heater and flue gas system. The model parameters are corrected by comparing with the performance test data under typical load points, including pipeline pressure loss, turbine stage efficiency and forced draft fan and induced draft fan efficiency indicators, to ensure model accuracy.

[0137] In an embodiment, the objective function of the power supply coal consumption in the optimization objective function module adopts the following form:

[0138]

[0139] Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

[0140] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, from all points of view, the embodiments should be regarded as illustrative and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and range of equivalents of the claims are included in the present invention. Any reference signs in the claims should not be construed as limiting the claim to which they relate.

[0141] In addition, it should be understood that although this specification describes the embodiments, not every embodiment contains only one independent technical solution. This description is for clarity only. Those skilled in the art should consider the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for the purpose of illustrating the technical concept of the present invention and cannot be used to limit the scope of protection of the present invention. Any changes made based on the technical solution in accordance with the technical concept proposed by the present invention fall within the scope of protection of the claims of the present invention.

Claims

1. A dynamic entropy weight optimization control method for coal-fired units integrating NSGA-II and PSO is characterized by: The following steps are involved: Step S1: Establishing a thermodynamic simulation model of a large coal-fired power generation unit; Step S2: performing parameter correction and multi-operating condition verification on the thermodynamic simulation model based on unit performance test data; Step S3: Determine the optimization objective function of the thermodynamic simulation model, including power supply coal consumption, plant power consumption rate, and waste heat utilization rate indicators, and set key operating parameters as optimization variables; Step S4: Solve the multi-objective function using the hybrid optimization algorithm of NSGA-II and PSO, and output the Pareto solution set; Step S5: Introduce a dynamic entropy weight mechanism in the optimization process, dynamically calculate the weight of each objective based on the information entropy value of the objective function of the candidate solution set and update it in real time; Step S6: Evaluate the comprehensive satisfaction of the Pareto solution set by combining the fuzzy membership function and select the optimal solution; Step S7: The control parameters in the optimal solution are sent to the actuators through the discrete control system DCS to achieve closed-loop optimization control of the coal-fired unit.

2. The method for optimizing the dynamic entropy weight of a coal-fired unit by integrating NSGA-II and PSO according to claim 1 is characterized in that: The thermodynamic simulation model covers key units such as boilers, turbines, economizers, air heaters, and flue gas systems. By comparing the model parameters with performance test data at typical load points, including pipeline pressure loss, turbine stage efficiency, and forced draft and induced draft fan efficiency indicators, the model accuracy is ensured.

3. The method for optimizing the dynamic entropy weight of a coal-fired power plant by integrating NSGA-II and PSO according to claim 1 is characterized in that: The objective function of power supply coal consumption is in the following form: Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

4. The method for dynamic entropy weight optimization control of coal-fired units integrating NSGA-II and PSO according to claim 1 is characterized in that: The target weight calculation method of the dynamic entropy weight mechanism includes the following steps: Step 1: Perform extreme value normalization on the objective function value of each candidate solution; Step 2: Information entropy value of technical goal j: where p ij is the normalized value, n is the number of solutions involved in the evaluation, and ε is the smoothing factor; Step 3: Calculate the coefficient of variation d j =1-E j , and find the target weight 5. The method for optimizing the dynamic entropy weight of a coal-fired unit by integrating NSGA-II and PSO according to claim 1 is characterized in that: The fuzzy membership function adopts the form of an S-type function, maps the normalized objective function to the interval [0, 1], evaluates the satisfaction of each optimization solution to multiple objectives, calculates the total membership degree and selects the largest one as the final optimal solution.

6. The method for optimizing the dynamic entropy weight of a coal-fired unit by integrating NSGA-II and PSO according to claim 1 is characterized in that: The optimization variables include: the opening of the flue gas bypass damper of the air preheater, the feed water flow of the high-pressure economizer, the condensate flow of the low-pressure economizer and the circulating water flow of the air preheater.

7. The method for optimizing the dynamic entropy weight of a coal-fired unit by integrating NSGA-II and PSO according to claim 1 is characterized in that: The control parameters are sent down through the data interface of the unit DCS system. The control parameters include the opening of the smoke damper, the opening of the flow control valve and the speed adjustment control amount of the variable frequency pump, and an adjustment rate limiting mechanism is provided.

8. A dynamic entropy weight optimization control system for coal-fired power units integrating NSGA-II and PSO, characterized by including: Thermodynamic simulation modeling module, to establish thermodynamic simulation models of large coal-fired power generation units; An entropy weight calculation module, which performs parameter correction and multi-operating condition verification on the thermodynamic simulation model based on unit performance test data; Optimization objective function module, which determines the optimization objective function of the thermodynamic simulation model, including power supply coal consumption, plant power consumption rate and waste heat utilization rate indicators, and sets key operating parameters as optimization variables; The solution module uses a hybrid optimization algorithm that combines NSGA-II and PSO to solve multi-objective functions and output the Pareto solution set; Dynamic calculation and update module, which introduces a dynamic entropy weight mechanism in the optimization process, dynamically calculates the weight of each target based on the information entropy value of the objective function of the candidate solution set and updates it in real time; The optimal solution selection module combines the fuzzy membership function to evaluate the comprehensive satisfaction of the Pareto solution set and select the optimal solution; The execution module sends the control parameters in the optimal solution to the actuator through the discrete control system DCS, realizing closed-loop optimization control of the coal-fired unit.

9. The dynamic entropy weight optimization control system of a coal-fired unit integrating NSGA-II and PSO according to claim 8 is characterized in that: The thermodynamic simulation model in the thermodynamic simulation modeling module covers key units such as boilers, turbines, economizers, air heaters, and flue gas systems. Model parameters are corrected by comparing with performance test data at typical load points, including pipeline pressure loss, turbine stage efficiency, and forced draft and induced draft fan efficiency indicators, to ensure model accuracy.

10. The coal-fired unit dynamic entropy weight optimization control system integrating NSGA-II and PSO according to claim 8 is characterized in that: The objective function of the power supply coal consumption in the optimization objective function module adopts the following form: Where B is the coal consumption for power supply, g / (kW·h); m 煤 is the hourly coal consumption of the unit, t / h; Q 煤 is the lower calorific value of coal, kJ / kg; P 净 is the net generating power of the unit, kW.

Images