Intelligent fuel management and optimized supply system for thermal power generating unit

Abstract

The invention relates to the technical field of thermal power generating unit fuel management, in particular to a thermal power generating unit fuel intelligent management and optimal supply system which performs load prediction based on a unit load prediction module, then obtains a first coal blending proportion based on a coal blending proportion calculation module, and then calculates a difference value based on a unit load monitoring module. The inventory rationality is evaluated through the size of the difference value, if the difference value is too large, a second coal blending proportion is obtained based on the coal blending proportion calculation module, then a third coal blending proportion is obtained based on the coal blending proportion optimization module, and coal blending and blending combustion are conducted according to the third coal blending proportion based on the coal blending and blending combustion control module. Compared with the prior art, the prediction deviation of the unit load prediction module is evaluated in a difference value mode, and the second coal blending proportion is corrected based on the first coal blending proportion, so that the adverse effect generated by the load prediction deviation caused by the load randomness is eliminated; and the problem of excess or shortage of fire coal inventory caused by load prediction deviation is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel management for thermal power plants, and in particular to an intelligent management and optimized supply system for fuel for thermal power plants. Background Art

[0002] Peak load regulation of thermal power units refers to adjusting the power generation of thermal power units to adapt to changes in grid load. This operation is essential to ensure the stable operation of the power system because it enables power supply to match demand, especially during peak and trough periods. Peak load regulation can not only improve the economic efficiency of the power system, but also reduce energy waste and environmental pollution.

[0003] Load forecasting and coal blending are two key issues that need to be considered in the peak load regulation process. Load forecasting predicts the power demand in the future by analyzing historical data, seasonal changes, weather conditions and other factors. This helps the power dispatching center plan the operation of thermal power units in advance to ensure sufficient power supply during peak hours. Coal blending involves adjusting the type and proportion of coal used by thermal power units according to load changes and coal quality characteristics. Reasonable coal blending can not only improve the operating efficiency and flexibility of thermal power units, but also reduce power generation costs and environmental pollution to a certain extent. The optimization and coordination of these two aspects are of great significance to achieving efficient, economical and environmentally friendly power peak load regulation.

[0004] However, in the actual operation of the power system, in addition to load forecasting and coal blending strategies, coal inventory management is also a key link that cannot be ignored. Inventory procurement strategies usually rely on forecasts of load demand, and the load changes faced by thermal power units have both predictable regularity and unpredictable randomness. For example, when renewable energy sources such as wind and hydropower fluctuate in power generation due to weather changes, thermal power units must quickly adjust to fill or reduce power supply. In this case, if inventory management is based only on forecasts, it may cause excess coal to deteriorate, or insufficient inventory to affect power supply. Therefore, people need a more accurate and flexible intelligent fuel management and optimized supply system for thermal power units to deal with the problem of excess or shortage of coal inventory caused by load forecast deviations. Summary of the invention

[0005] Therefore, the present invention provides a thermal power unit fuel intelligent management and optimization supply system to solve the problem that the existing technology cannot cope with the excess or shortage of coal inventory caused by load forecast deviation.

[0006] The present invention provides a fuel intelligent management and optimization supply system for a thermal power unit, comprising a unit load prediction module, a coal blending ratio calculation module, a unit load monitoring module, a coal blending ratio optimization module, a coal blending and combustion control module and a central control module, wherein the central control module is used for:

[0007] Based on the unit load forecasting module, forecast the unit load at the target time;

[0008] Based on the coal blending ratio calculation module, the coal blending ratio is calculated according to the predicted unit load to obtain the first coal blending ratio;

[0009] Based on the unit load monitoring module, the actual unit load at the target time is obtained, the difference between the predicted unit load and the actual unit load is calculated, and the difference is compared with the preset difference threshold;

[0010] If the difference value does not exceed the preset difference threshold, coal blending and combustion is performed at the first coal blending ratio based on the coal blending and combustion control module;

[0011] If the difference value exceeds the preset difference threshold, the coal blending ratio calculation module is used to calculate the coal blending ratio according to the actual unit load to obtain a second coal blending ratio;

[0012] Based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio;

[0013] Based on the coal blending and combustion control module, coal blending and combustion are carried out with the third coal blending ratio.

[0014] The present invention also provides a preferred solution: based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio, including:

[0015] Based on the coal blending ratio optimization module, obtain the type of coal;

[0016] According to the coal type, a solution vector is established, and each element in the solution vector corresponds to a coal type and is used to characterize the coal blending ratio of the coal type;

[0017] Calculate the difference in proportions corresponding to each type of coal in the first coal blending proportion and the second coal blending proportion, and use the difference in proportions as the value range of the corresponding element in the solution vector to establish a solution space;

[0018] Based on the solution space, taking the optimal coal inventory turnover as the optimization goal, the optimal solution vector is obtained through the preset optimization algorithm;

[0019] The coal blending ratio corresponding to the optimal solution vector is taken as the third coal blending ratio.

[0020] The present invention also provides a preferred solution: based on the solution space, taking the optimal coal inventory turnover as the optimization goal, the optimal solution vector is obtained by a preset optimization algorithm, including:

[0021] Obtain inventory data of each type of coal at the target time;

[0022] A fitness function is established based on the inventory data of each type of coal at the target time. The fitness function is used to calculate the fitness of a solution vector in the solution space. The fitness is used to characterize the quality of coal inventory turnover when each type of coal is consumed at the coal blending ratio corresponding to the solution vector starting from the target time.

[0023] Establish multiple initial solution vectors based on the solution space;

[0024] Based on the fitness function, multiple initial solution vectors are optimized through a preset optimization algorithm to obtain the optimal solution vector.

[0025] The present invention also provides a preferred solution: the inventory data of each type of coal at the target time includes the inventory value of each type of coal at the target time and the storage cost of each type of coal, and the fitness function is:

[0026]

[0027] Among them, F is the fitness, x i represents the i-th element in the solution vector, n is the total number of elements in the solution vector, I i Represents element x i The inventory value of the corresponding type of coal at the target time, p i Represents element x i Storage costs for corresponding types of coal.

[0028] The present invention also provides a preferred solution: based on the fitness function, multiple initial solution vectors are optimized by a preset optimization algorithm to obtain an optimal solution vector, including:

[0029] Obtain the purchase cycle of each type of coal, and obtain the optimized learning rate of each type of coal according to the purchase cycle, the inventory value of each type of coal at the target time, the first coal blending ratio and the second coal blending ratio;

[0030] According to the optimized learning rate, multiple initial solution vectors are optimized by a preset optimization algorithm based on the fitness function to obtain the optimal solution vector.

[0031] The present invention also provides a preferred solution: the optimization learning rate is obtained according to the following formula:

[0032]

[0033] Among them, c i is the optimal learning rate of the type of coal corresponding to the i-th element in the solution vector, f() is the preset linear function, t i A represents the purchasing cycle of the type of coal corresponding to the i-th element in the solution vector, i represents the proportion of the coal type corresponding to the i-th element in the solution vector in the first coal blending ratio, B iIt represents the coal proportion of the type of coal corresponding to the i-th element in the solution vector in the second coal proportion, min() is the minimum value function, and a is the preset unit adjustment coefficient.

[0034] The present invention also provides a preferred solution: according to the optimized learning rate, multiple initial solution vectors are optimized by a preset optimization algorithm based on the fitness function to obtain the optimal solution vector, including:

[0035] S1, obtaining multiple initial solution vectors as multiple target solution vectors;

[0036] S2, calculating the fitness of multiple target solution vectors based on the fitness function;

[0037] S3, based on fitness, obtain the global optimal solution vector and the individual optimal solution vector;

[0038] S4, determine whether the preset termination condition is met, if so, proceed to step S5, if not, based on the optimized learning rate, update the speed and position of each target solution vector according to the global optimal solution vector and the individual optimal solution vector, and re-execute step S2;

[0039] S5. Obtain the optimal solution vector according to the last updated initial solution vector.

[0040] The present invention also provides a preferred solution: the speed of the solution vector is updated by the following formula:

[0041]

[0042] in, represents the speed of the i-th element in the solution vector at the k-th iteration, pbest i Represents the i-th element in the individual optimal solution vector, gbest i represents the i-th element in the global optimal solution vector, r1 and r2 are different random numbers.

[0043] The present invention also provides a preferred solution: based on the unit load prediction module, predicting the unit load at the target time includes:

[0044] Based on the unit load forecasting module, obtain historical load data;

[0045] Based on the historical load data, the predicted unit load is obtained by using the preset LSTM neural network model.

[0046] The present invention also provides a preferred solution: based on the coal blending and combustion control module, coal blending and combustion are performed at a third coal blending ratio, including:

[0047] Based on the coal blending control module, the target supply rate of each type of coal is calculated through the third coal blending ratio;

[0048] Coal is supplied based on the target supply rate to achieve blended combustion.

[0049] The beneficial effects of adopting the above embodiment are:

[0050] The present invention provides a fuel intelligent management and optimized supply system for a thermal power unit, comprising a unit load prediction module, a coal blending ratio calculation module, a unit load monitoring module, a coal blending ratio optimization module, a coal blending and burning control module and a central control module. The central control module is used for predicting the predicted unit load at a target time based on the unit load prediction module, and then calculating the coal blending ratio according to the predicted unit load based on the coal blending ratio calculation module to obtain a first coal blending ratio. Then, based on the unit load monitoring module, the actual unit load at the target time is obtained, the difference value between the predicted unit load and the actual unit load is calculated, and the difference value is compared with a preset difference threshold value. If the difference value does not exceed the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the first coal blending ratio. If the difference value exceeds the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the first coal blending ratio. If the difference value exceeds the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the actual unit load to obtain a second coal blending ratio. Then, based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio. Finally, based on the coal blending and burning control module, the third coal blending and burning is performed. Compared with the prior art, the first coal blending ratio in the present invention is used for purchasing coal inventory, and the actual unit load is used to evaluate the deviation of the unit load prediction module in the form of difference value. If the difference value is small, it indicates that the coal inventory purchased based on the first coal blending ratio is reasonable. At this time, normal power supply based on the first coal blending ratio is sufficient. If the difference value is large, it indicates that the current coal inventory does not match the actual demand. At this time, the second coal blending ratio that meets the current demand is calculated, and it is corrected based on the first coal blending ratio to eliminate the adverse effects of the load prediction deviation caused by the randomness of the load, so that the third coal blending ratio obtained takes into account the actual coal demand and the actual inventory situation, so that the inventory can be well turned over, solving the problem of excess or shortage of coal inventory caused by load prediction deviation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A system architecture diagram of an embodiment of a system for intelligent management and optimized supply of fuel for thermal power units provided by the present invention;

[0052] Figure 2 for Figure 1 The method steps executed by the coal blending ratio optimization module are as follows:

[0053] Figure 3 for Figure 2 Specific step diagram of step S204. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Combination Figure 1 As shown, a specific embodiment of the present invention discloses a fuel intelligent management and optimization supply system for a thermal power unit, including a unit load prediction module, a coal blending ratio calculation module, a unit load monitoring module, a coal blending ratio optimization module, a coal blending and combustion control module and a central control module, wherein the central control module is used for:

[0056] Based on the unit load forecasting module, forecast the unit load at the target time;

[0057] Based on the coal blending ratio calculation module, the coal blending ratio is calculated according to the predicted unit load to obtain the first coal blending ratio;

[0058] Based on the unit load monitoring module, the actual unit load at the target time is obtained, the difference between the predicted unit load and the actual unit load is calculated, and the difference is compared with the preset difference threshold;

[0059] If the difference value does not exceed the preset difference threshold, coal blending and combustion is performed at the first coal blending ratio based on the coal blending and combustion control module;

[0060] If the difference value exceeds the preset difference threshold, the coal blending ratio calculation module is used to calculate the coal blending ratio according to the actual unit load to obtain a second coal blending ratio;

[0061] Based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio;

[0062] Based on the coal blending and combustion control module, coal blending and combustion are carried out with the third coal blending ratio.

[0063] It should be noted that in the above process, the target time refers to the time that people hope and predict, and plan and dispatch, such as the next month or a week after a few days, etc., or it can be a certain moment, and the unit load prediction module is used to predict the unit load at this time. The unit load prediction module can be set in the thermal power plant itself for operation, or it can be set in the dispatching center for operation. Similarly, the unit load monitoring module is used to monitor the real-time load of the thermal power unit from the current moment to the target time, but it should be noted that the actual unit load needs to be in the same format as the predicted unit load, which means that the actual unit load does not refer to the actual monitored load data, but to be precise, it refers to the more accurate unit load data obtained near the target time. For example, when the target time is a certain moment in the future, the predicted unit load is the current predicted load data of the unit at that moment, and the actual unit load is the actual unit load data detected at the target time. If the target time represents a future time period, the actual unit load can be considered to be the load data in the time period predicted at the beginning of the target time.

[0064] Compared with the prior art, the first coal blending ratio in the present invention is used for purchasing coal inventory, and the actual unit load is used to evaluate the deviation of the unit load prediction module in the form of difference value. If the difference value is small, it indicates that the coal inventory purchased based on the first coal blending ratio is reasonable. At this time, normal power supply based on the first coal blending ratio is sufficient. If the difference value is large, it indicates that the current coal inventory does not match the actual demand. At this time, the second coal blending ratio that meets the current demand is calculated, and it is corrected based on the first coal blending ratio to eliminate the adverse effects of the load prediction deviation caused by the randomness of the load, so that the third coal blending ratio obtained takes into account the actual coal demand and the actual inventory situation, so that the inventory can be well turned over, solving the problem of excess or shortage of coal inventory caused by load prediction deviation.

[0065] In addition, the system can be operated in real time, that is, the above process can be repeated and cycled to achieve real-time adjustment to ensure that coal blending can always take into account the specific coal needs and inventory conditions.

[0066] Further, in a preferred embodiment, the steps performed by the unit load prediction module: predicting the unit load at the target time specifically include:

[0067] Based on the unit load forecasting module, obtain historical load data;

[0068] Based on the historical load data, the predicted unit load is obtained by using the preset LSTM neural network model.

[0069] Long short-term memory network (LSTM) is a special recurrent neural network (RNN) that shows significant advantages in processing sequence data. In the above-mentioned unit load forecasting module, the superiority of LSTM is mainly reflected in the following aspects: First, LSTM can capture the long-term dependencies in time series data, which is crucial for predicting long-term load trends; second, LSTM effectively solves the common gradient vanishing or gradient explosion problems of traditional RNN in long sequence learning through its unique gating mechanism, thereby improving the stability and prediction accuracy of the model; third, the LSTM model has good generalization ability, and can learn complex patterns and laws from historical load data, and then accurately predict the unit load at the target time; finally, the flexibility and scalability of the LSTM model enable it to adapt to different forecasting tasks and data characteristics, providing strong technical support for unit load forecasting. The use of LSTM in this embodiment can effectively reduce the deviation of load forecasting, thereby ensuring the rationality of inventory procurement.

[0070] It is understandable that in practice, other methods may be used to predict load according to specific needs.

[0071] Furthermore, the steps performed by the above-mentioned unit load monitoring module are: obtaining the actual unit load at the target time, calculating the difference between the predicted unit load and the actual unit load, and comparing the difference with the preset difference threshold, wherein the difference can be flexibly defined according to the specific data structure of the predicted unit load or the actual unit load, and the difference between the two can be calculated and expressed in any existing manner. For example, when the predicted unit load and the actual unit load are the load values ​​at a certain moment, the difference between the two can be directly used as the difference value, and if the predicted unit load and the actual unit load are numerical sequences or curves, the difference value can be the difference average of the two sequences or curves or any other statistical features such as the variance of the difference.

[0072] Further, combined with Figure 2 As shown, in a preferred embodiment, the steps performed by the above-mentioned coal blending ratio optimization module: optimizing the second coal blending ratio according to the first coal blending ratio to obtain the third coal blending ratio specifically include:

[0073] S201, obtaining the type of coal based on the coal blending ratio optimization module;

[0074] S202, establishing a solution vector according to the type of coal, wherein each element in the solution vector corresponds to a type of coal and is used to characterize the coal blending ratio of the type of coal;

[0075] S203, calculating the ratio difference between the first coal blending ratio and the second coal blending ratio corresponding to each type of coal, and using the ratio difference as the value range of the corresponding element in the solution vector to establish a solution space;

[0076] S204, based on the solution space, taking the optimal coal inventory turnover as the optimization goal, obtaining the optimal solution vector through a preset optimization algorithm;

[0077] S205. The coal blending ratio corresponding to the optimal solution vector is used as the third coal blending ratio.

[0078] In the above process, the first coal blending ratio, in addition to indicating the predicted coal blending ratio, can actually be understood as the purchased inventory ratio, while the second coal blending ratio indicates the actual coal blending ratio that meets the current demand. The correction of the second coal blending ratio is to correct the coal blending ratio towards the first coal blending ratio while maintaining the proportion of the second coal blending ratio as much as possible, so that the final coal blending ratio can meet the current boiler combustion demand as much as possible, and can consume the coal inventory at the optimal speed to avoid inventory accumulation or shortage.

[0079] Therefore, the first challenge of finding the third coal-fired ratio is that it should be found between the first coal-fired ratio and the second coal-fired ratio. In addition, since the determination of the coal-fired ratio in practice needs to consider many factors, such as the stability of the thermal power unit, the carbon emission level or the cost, the second challenge of finding the third coal-fired ratio is how to ensure that it can meet multiple complex standards at the same time. Therefore, in this embodiment, the solution space is established by the difference between the two, and the third coal-fired ratio is searched by using the preset optimization algorithm. It is possible to use the optimization algorithm to achieve multi-objective optimization at the same time to solve the above two challenges and determine the third coal-fired ratio.

[0080] Specifically, combined Figure 3 As shown, in a preferred embodiment, the above step S204, based on the solution space, takes the optimal coal inventory turnover as the optimization goal, and obtains the optimal solution vector through a preset optimization algorithm, specifically including:

[0081] S301, obtaining inventory data of each type of coal at a target time;

[0082] S302, establishing a fitness function based on the inventory data of each type of coal at the target time, the fitness function is used to calculate the fitness of a solution vector in the solution space, and the fitness is used to characterize the quality of the coal inventory turnover when each type of coal is consumed at the coal blending ratio corresponding to the solution vector starting from the target time;

[0083] S303, establishing multiple initial solution vectors based on the solution space;

[0084] S304: Based on the fitness function, multiple initial solution vectors are optimized by a preset optimization algorithm to obtain an optimal solution vector.

[0085] Like the target time, the coal inventory data in the above process can also be the coal inventory data for a period of time, the coal inventory data at the beginning of a time period, or other artificially defined values ​​that can represent the inventory level. In addition, the coal inventory data can also include other inventory-related data such as the cost of each type of coal and storage conditions.

[0086] In the optimization algorithm, the fitness function is a function that evaluates the quality of candidate solutions. It maps the solution to the optimization problem to a real number or vector, and this value is usually used to indicate whether the solution is "good" or "bad". The fitness function is a core component in the optimization algorithm because it determines the direction in which the algorithm searches for the optimal solution. The goal of the present invention is to correct the second coal blending ratio to make it consistent with the inventory status, and the inventory of coal changes in real time. Therefore, in this embodiment, before performing a specific optimization process, it is necessary to first determine the fitness function based on the inventory data of each type of coal at the target time.

[0087] For example, the elements in the solution vector represent the proportion of coal blending, which actually also represent the consumption speed of coal within the target time. Then the fitness function can be: calculate the ratio of the inventory of a certain type of coal to its corresponding element in the solution vector, and then represent the estimated consumption time of the type of coal, and then find the average value of the estimated consumption time of all coals, so that the fitness of a solution vector can be obtained, and the advantages and disadvantages of different levels of fitness can be defined according to the actual situation. For example, it can be considered that in this embodiment, the larger the fitness value, the slower the coal consumption speed, the worse the coal inventory turnover capacity it reflects, and the worse the degree of fitness represented. For example, it can also be considered that in practice, due to environmental reasons, coal is easy to deteriorate. At this time, the smaller the value of the above fitness, the faster the coal consumption speed, the faster the coal turnover speed, the less storage time, and the lower the probability of deterioration, and the corresponding fitness is also better.

[0088] Further, in a preferred embodiment, the inventory data of each type of coal at the target time includes the inventory value of each type of coal at the target time and the storage cost of each type of coal, and the fitness function is:

[0089]

[0090] Among them, F is the fitness, x i represents the i-th element in the solution vector, n is the total number of elements in the solution vector, I i Represents element x i The inventory value of the corresponding type of coal at the target time, p i Represents element x i Storage costs for corresponding types of coal.

[0091] The above formula consists of two terms, the first of which is a way of calculating the fitness mentioned above, that is, calculating the rate of coal consumption. The significance of the above formula is that on this basis, the consideration of inventory cost is further added (that is, the calculation result of the second term), so that its fitness can reflect both the inventory consumption rate and inventory cost, making the third coal blending ratio more scientific and reasonable. Among them, the second term is actually to calculate the inventory cost corresponding to the solution vector in combination with the actual inventory situation and the coal consumption rate represented by the solution vector. The second term is obtained according to the integral shown below:

[0092]

[0093] Further, in a preferred embodiment, the above step S304, based on the fitness function, optimizes multiple initial solution vectors by a preset optimization algorithm to obtain the optimal solution vector, specifically includes:

[0094] Obtain the purchase cycle of each type of coal, and obtain the optimized learning rate of each type of coal according to the purchase cycle, the inventory value of each type of coal at the target time, the first coal blending ratio and the second coal blending ratio;

[0095] According to the optimized learning rate, multiple initial solution vectors are optimized by a preset optimization algorithm based on the fitness function to obtain the optimal solution vector.

[0096] The key point of the above process is to adjust the learning rate according to the procurement cycle.

[0097] In the optimization algorithm, the learning rate is a very important hyperparameter, which determines the step size of the algorithm to update the solution in each iteration. The learning rate directly affects the convergence speed of the algorithm and the quality of the final solution. In this embodiment, because the coal procurement cycle actually shows the flexibility of coal inventory turnover, if the procurement capacity of a type of coal is more flexible, then its corresponding inventory risk will also be relatively low. At this time, the optimization of this type of coal is not very meaningful, but will become a redundant calculation, affecting the overall operation speed. Therefore, in this embodiment, the learning rate of each type of coal is adjusted according to the procurement cycle, and the optimization calculation with lower benefits is eliminated to optimize the operation efficiency of the entire algorithm.

[0098] Specifically, in a preferred embodiment, the optimized learning rate is obtained according to the following formula:

[0099]

[0100] Among them, c i is the optimal learning rate of the type of coal corresponding to the i-th element in the solution vector, f() is the preset linear function, t i A represents the purchasing cycle of the type of coal corresponding to the i-th element in the solution vector, irepresents the proportion of the coal type corresponding to the i-th element in the solution vector in the first coal blending ratio, B i It represents the coal proportion of the type of coal corresponding to the i-th element in the solution vector in the second coal proportion, min() is the minimum value function, and a is the preset unit adjustment coefficient.

[0101] The meaning of the above formula is that through Calculate the minimum value of the consumption rate of the current inventory based on the first coal blending ratio and the consumption rate based on the second coal blending ratio, which represents the lower limit of the coal consumption rate that can be reflected by all solutions in the solution space. If the purchase cycle of this type of coal is greater than the lower limit, it means that there may be a situation where the coal supply is not timely. At this time, the learning rate of this type of coal should be adjusted linearly based on the purchase cycle. Obviously, its learning rate needs to be proportional to the purchase cycle. If the purchase cycle of this type of coal is less than the lower limit, it means that the replenishment time of this type of coal is less than the consumption rate. At this time, the inventory risk of this type of coal is low, and the adjustment flexibility is high. It is tolerable that the coal blending ratio of this type of coal is not optimal, so at this time, the learning rate of this type of coal is directly set to zero to greatly reduce the amount of calculation in the optimization algorithm and improve the efficiency of the algorithm.

[0102] Further, in a preferred embodiment, the above step: optimizing multiple initial solution vectors by a preset optimization algorithm based on the fitness function according to the optimized learning rate to obtain the optimal solution vector specifically includes:

[0103] S1, obtaining multiple initial solution vectors as multiple target solution vectors;

[0104] S2, calculating the fitness of multiple target solution vectors based on the fitness function;

[0105] S3, based on fitness, obtain the global optimal solution vector and the individual optimal solution vector;

[0106] S4, determine whether the preset termination condition is met, if so, proceed to step S5, if not, based on the optimized learning rate, update the speed and position of each target solution vector according to the global optimal solution vector and the individual optimal solution vector, and re-execute step S2;

[0107] S5. Obtain the optimal solution vector according to the last updated initial solution vector.

[0108] The above algorithm is actually the optimization process of the particle swarm algorithm, in which each solution vector can be regarded as a particle in the solution space, the position of the solution vector is the element content of the solution vector, the speed of the solution vector is the amplitude of the change of the solution vector position at each iteration, the global optimal solution vector is the optimal solution vector obtained at each iteration, and the individual optimal solution vector is the optimal solution vector obtained by each particle during its iteration process. It can be understood that the particle swarm algorithm is a prior art, so the above terms and some other details of the particle swarm algorithm are also understandable to those skilled in the art, so this article will not explain too much.

[0109] In addition, it can be understood that the determination of the initial solution vector and the subsequent iterative process need to meet the constraints of actual needs, such as maintaining stable boiler combustion and ensuring that carbon emissions do not exceed the standard. How to meet the above constraints is also a prior art and is not the focus of the present invention, so it will not be explained in detail.

[0110] It is worth noting that in this embodiment, the improved learning rate, i.e., the application of the optimized learning rate, is used. Specifically, in a preferred embodiment, in the above process, the speed of the solution vector is updated by the following formula:

[0111]

[0112] in, represents the speed of the i-th element in the solution vector at the k-th iteration, pbest i Represents the i-th element in the individual optimal solution vector, gbest i represents the i-th element in the global optimal solution vector, r1 and r2 are different random numbers.

[0113] The significance of the above formula is that in the prior art, the learning rates corresponding to the moving amplitude obtained according to the individual optimal solution vector and the moving amplitude obtained according to the global optimal solution vector each time the particle moves are uniformly replaced by the optimized learning rate described above, so as to reduce the frequency of optimizing the coal corresponding to the stable inventory, reduce meaningless calculations, and improve the operation efficiency of the algorithm.

[0114] It can be understood that the above process is the application of the optimized learning rate in the particle swarm algorithm. In practice, the optimized learning rate can also be expressed in different ways depending on the preset optimization algorithm. For example, in the genetic algorithm, the optimized learning rate can be used as the variation range or variation probability of the numerical value corresponding to the type of coal.

[0115] Further, in a preferred embodiment, in the above-mentioned coal blending and combustion control module, coal blending and combustion are performed at a third coal blending ratio, specifically including:

[0116] Based on the coal blending control module, the target supply rate of each type of coal is calculated through the third coal blending ratio;

[0117] Coal is supplied based on the target supply rate to achieve blended combustion.

[0118] The above process can be implemented based on any existing technology, such as adjusting the rotation speed of the bucket wheel machine in the coal bunker and the opening and closing degree of the bucket wheel to control the extraction speed of the fuel coal to achieve the target supply speed. The advantage of the above method is that it can be adjusted in real time according to the third coal blending ratio to achieve intelligent unmanned operation.

[0119] The present invention provides a fuel intelligent management and optimized supply system for a thermal power unit, comprising a unit load prediction module, a coal blending ratio calculation module, a unit load monitoring module, a coal blending ratio optimization module, a coal blending and burning control module and a central control module. The central control module is used for predicting the predicted unit load at a target time based on the unit load prediction module, and then calculating the coal blending ratio according to the predicted unit load based on the coal blending ratio calculation module to obtain a first coal blending ratio. Then, based on the unit load monitoring module, the actual unit load at the target time is obtained, the difference value between the predicted unit load and the actual unit load is calculated, and the difference value is compared with a preset difference threshold value. If the difference value does not exceed the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the first coal blending ratio. If the difference value exceeds the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the first coal blending ratio. If the difference value exceeds the preset difference threshold value, the coal blending and burning control module is used to perform coal blending and burning at the actual unit load to obtain a second coal blending ratio. Then, based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio. Finally, based on the coal blending and burning control module, the third coal blending and burning is performed. Compared with the prior art, the first coal blending ratio in the present invention is used for purchasing coal inventory, and the actual unit load is used to evaluate the deviation of the unit load prediction module in the form of difference value. If the difference value is small, it indicates that the coal inventory purchased based on the first coal blending ratio is reasonable. At this time, normal power supply based on the first coal blending ratio is sufficient. If the difference value is large, it indicates that the current coal inventory does not match the actual demand. At this time, the second coal blending ratio that meets the current demand is calculated, and it is corrected based on the first coal blending ratio to eliminate the adverse effects of the load prediction deviation caused by the randomness of the load, so that the third coal blending ratio obtained takes into account the actual coal demand and the actual inventory situation, so that the inventory can be well turned over, solving the problem of excess or shortage of coal inventory caused by load prediction deviation.

[0120] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0121] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A fuel intelligent management and optimization supply system for a thermal power unit, characterized in that: It includes unit load prediction module, coal blending ratio calculation module, unit load monitoring module, coal blending ratio optimization module, coal blending and combustion control module and central control module, among which the central control module is used for: Based on the unit load forecasting module, forecast the unit load at the target time; Based on the coal blending ratio calculation module, the coal blending ratio is calculated according to the predicted unit load to obtain the first coal blending ratio; Based on the unit load monitoring module, the actual unit load at the target time is obtained, the difference between the predicted unit load and the actual unit load is calculated, and the difference is compared with the preset difference threshold; If the difference value does not exceed the preset difference threshold, coal blending and combustion is performed at the first coal blending ratio based on the coal blending and combustion control module; If the difference value exceeds the preset difference threshold, the coal blending ratio calculation module is used to calculate the coal blending ratio according to the actual unit load to obtain a second coal blending ratio; Based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain a third coal blending ratio; Based on the coal blending and combustion control module, coal blending and combustion are carried out with the third coal blending ratio.

2. The intelligent management and optimization supply system for fuel of a thermal power unit according to claim 1 is characterized in that: Based on the coal blending ratio optimization module, the second coal blending ratio is optimized according to the first coal blending ratio to obtain the third coal blending ratio, including: Based on the coal blending ratio optimization module, obtain the type of coal; According to the coal type, a solution vector is established, and each element in the solution vector corresponds to a coal type and is used to characterize the coal blending ratio of the coal type; Calculate the difference in proportions corresponding to each type of coal in the first coal blending proportion and the second coal blending proportion, and use the difference in proportions as the value range of the corresponding element in the solution vector to establish a solution space; Based on the solution space, taking the optimal coal inventory turnover as the optimization goal, the optimal solution vector is obtained through the preset optimization algorithm; The coal blending ratio corresponding to the optimal solution vector is taken as the third coal blending ratio.

3. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 2 is characterized in that: Based on the solution space, taking the optimal coal inventory turnover as the optimization goal, the optimal solution vector is obtained through the preset optimization algorithm, including: Obtain inventory data of each type of coal at the target time; A fitness function is established based on the inventory data of each type of coal at the target time. The fitness function is used to calculate the fitness of a solution vector in the solution space. The fitness is used to characterize the quality of coal inventory turnover when each type of coal is consumed at the coal blending ratio corresponding to the solution vector starting from the target time. Establish multiple initial solution vectors based on the solution space; Based on the fitness function, multiple initial solution vectors are optimized through a preset optimization algorithm to obtain the optimal solution vector.

4. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 3 is characterized in that: The inventory data of each type of coal at the target time includes the inventory value of each type of coal at the target time and the storage cost of each type of coal. The fitness function is: Among them, F is the fitness, x i represents the i-th element in the solution vector, n is the total number of elements in the solution vector, I i Represents element x i The inventory value of the corresponding type of coal at the target time, p i Represents element x i Storage costs for corresponding types of coal.

5. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 4 is characterized in that: Based on the fitness function, multiple initial solution vectors are optimized through a preset optimization algorithm to obtain the optimal solution vector, including: Obtain the purchase cycle of each type of coal, and obtain the optimized learning rate of each type of coal according to the purchase cycle, the inventory value of each type of coal at the target time, the first coal blending ratio and the second coal blending ratio; According to the optimized learning rate, multiple initial solution vectors are optimized by a preset optimization algorithm based on the fitness function to obtain the optimal solution vector.

6. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 5 is characterized in that: The optimized learning rate is obtained according to the following formula: Among them, c i is the optimal learning rate of the type of coal corresponding to the i-th element in the solution vector, f() is the preset linear function, t i A represents the purchasing cycle of the type of coal corresponding to the i-th element in the solution vector, i represents the proportion of the coal type corresponding to the i-th element in the solution vector in the first coal blending ratio, B i It represents the coal proportion of the type of coal corresponding to the i-th element in the solution vector in the second coal proportion, min() is the minimum value function, and a is the preset unit adjustment coefficient.

7. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 6 is characterized in that: According to the optimized learning rate, multiple initial solution vectors are optimized by a preset optimization algorithm based on the fitness function to obtain the optimal solution vector, including: S1, obtaining multiple initial solution vectors as multiple target solution vectors; S2, calculating the fitness of multiple target solution vectors based on the fitness function; S3, based on fitness, obtain the global optimal solution vector and the individual optimal solution vector; S4, determine whether the preset termination condition is met, if so, proceed to step S5, if not, based on the optimized learning rate, update the speed and position of each target solution vector according to the global optimal solution vector and the individual optimal solution vector, and re-execute step S2; S5. Obtain the optimal solution vector according to the last updated initial solution vector.

8. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 7 is characterized in that: The velocity of the solution vector is updated by: in, represents the speed of the i-th element in the solution vector at the k-th iteration, pbest i Represents the i-th element in the individual optimal solution vector, gbest i represents the i-th element in the global optimal solution vector, r1 and r2 are different random numbers.

9. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 1 is characterized in that: Based on the unit load forecasting module, the unit load forecast at the target time is forecasted, including: Based on the unit load forecasting module, obtain historical load data; Based on the historical load data, the predicted unit load is obtained by using the preset LSTM neural network model.

10. The intelligent management and optimized supply system for fuel of a thermal power unit according to claim 1, characterized in that: Based on the coal blending and combustion control module, coal blending and combustion are performed at the third coal blending ratio, including: Based on the coal blending control module, the target supply rate of each type of coal is calculated through the third coal blending ratio; Coal is supplied based on the target supply rate to achieve blended combustion.

Images