Distributed control precise coal blending system based on neural network

Abstract

The present invention discloses a distributed control precision coal blending system based on a neural network, comprising a raw coal information monitoring unit, a coal blending model design unit and a dynamic coal blending scheduling unit, wherein the coal blending model design unit comprises an initial model design module, a coal supply prediction module and a model self-growth optimization module, information is collected through the raw coal information monitoring unit, and then a coal blending model is constructed through the coal blending model design unit to achieve the design of an optimal coal blending scheme, the model has a strong generalization ability, can adapt to changes in coal mine costs and changes in resource reserve conditions caused by coal mining, and can achieve dynamic adjustment configuration by continuously updating configuration schemes, which can meet user needs and has flexibility and high efficiency, and then generates an optimal coal blending prediction scheme through coal prediction and model self-growth optimization, thereby achieving early supply and scheduling management of coal resources and ensuring the continuous stability of coal blending and coal making work.

Description

Technical Field

[0001] The present invention relates to the technical field of coal blending control, and in particular to a distributed control precision coal blending system based on a neural network. Background Art

[0002] Neural networks are highly complex, including large-scale parameters and computational complexity. Through learning and training, neural network models can learn certain rules and patterns from data, which are then used for prediction and decision-making. Distributed control technology refers to distributing the computing, decision-making and execution functions of the control system to different locations in the system, and achieving control of the entire system through communication and coordination.

[0003] In the coal blending system, due to the geological conditions of the coal mining area, changes in coal resources and other reasons, the input data distribution will be offset, and the existing coal blending model may lack sufficient generalization ability when facing data changes, resulting in reduced coal blending performance and insufficient flexibility. The coal blending model design may lack integrity and comprehensiveness, making it difficult to achieve the overall optimal ratio in terms of coal combustion performance, environmental impact, coal making cost and raw coal resource reserve allocation, and it is difficult to achieve advance supply and scheduling management of coal resources, resulting in poor sustained stability of coal blending and coal making work;

[0004] In view of the above technical defects, a solution is now proposed. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the existing coal blending models have insufficient generalization ability and flexibility, lack integrity and comprehensiveness, and are difficult to achieve overall optimal blending and advance scheduling, resulting in poor sustained stability of coal blending and coal making work. Through the optimal coal blending plan, while meeting the coal making cost and combustion performance required by users, it is ensured that the environmental impact is friendly and the supply reserves are sufficient, so as to achieve the best overall coal blending effect.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A distributed control precision coal blending system based on a neural network includes a raw coal information monitoring unit, a coal blending model design unit and a dynamic coal blending scheduling unit, wherein the coal blending model design unit includes an initial model design module, a coal supply prediction module and a model self-growth optimization module;

[0008] The raw coal information monitoring unit is used to collect and obtain raw coal information: raw coal information includes coal combustion parameters and coal mine supply parameters;

[0009] The coal blending model design unit is used to analyze and process raw coal information and establish a coal blending model:

[0010] Among them, the initial model design module analyzes the combustion performance and environmental impact of raw coal through coal combustion parameters to generate the raw coal quality coefficient, and obtains the raw coal supply coefficient through coal mine supply parameter analysis, and then generates the raw coal optimization index by combining the raw coal quality coefficient with the raw coal supply coefficient, so as to construct the optimal coal blending model; the coal supply prediction module is used to conduct in-depth analysis of raw coal information, monitor and schedule the historical supply situation of different coal supply points, predict coal demand, and obtain the coal supply prediction index; the model self-growth optimization module continuously optimizes the optimal coal blending model by feeding back the coal supply prediction index to the initial model design module, thereby realizing the self-growth upgrade of the model;

[0011] The dynamic coal blending scheduling unit is used to perform distributed coal blending control operations: the control management signal is generated through the optimal coal blending model to dynamically schedule and manage the coal blending.

[0012] Furthermore, the specific process of generating the raw coal quality coefficient and the raw coal supply coefficient is as follows:

[0013] A1: First analyze the coal combustion parameters:

[0014] Coal combustion parameters include the calorific value, moisture, volatile matter, sulfur and ash of the raw coal. Any raw coal is marked as y, and the calorific value of raw coal y is marked as Qy, moisture as Wy, volatile matter as Fy, sulfur as Sy, and ash as Hy;

[0015] A1-1: First, the combustion performance evaluation coefficient Xrs of raw coal y is generated by combining the calorific value Qy, moisture Wy and volatile matter Fy of raw coal y;

[0016] A1-2: The environmental impact assessment coefficient Xhj of raw coal y is generated by combining the volatile matter Fy, sulfur Sy and ash Hy of raw coal y;

[0017] A1-3: Then, the raw coal quality coefficient ZLy of raw coal y is generated by combining the combustion performance evaluation coefficient Xrs and the environmental impact assessment coefficient Xhj of raw coal y;

[0018] A2: Reanalyze coal mine supply parameters:

[0019] Coal mine supply parameters include the cost and reserves of raw coal. The cost of raw coal y is marked as Cy and the reserves are marked as Ly.

[0020] By combining the cost Cy and the reserves Ly, the raw coal supply coefficient GYy of raw coal y is generated.

[0021] Furthermore, the specific process of constructing the optimal coal blending model is as follows:

[0022] B1: Generate the raw coal optimization index YXy by combining the raw coal quality coefficient ZLy and the raw coal supply coefficient GYy;

[0023] B1-1: There are N0 kinds of raw coal at the preset supply point, and the raw coal information of the N0 kinds of raw coal is analyzed and processed respectively, and the corresponding raw coal optimization index is obtained;

[0024] B1-2: Obtain blended coal products through coal blending operations. The blended coal products are assumed to contain n0 kinds of raw coals. Any of the raw coals is marked as i, and the raw coal preference index of raw coal i is YXi. The proportion φi of raw coal i in the blended coal products is calculated through the raw coal preference index YXi.

[0025] B1-3: Obtain the coal blending requirements of the current user, which include the minimum combustion calorific value Qz min and the maximum total coal cost Cz max , taking the user's coal blending demand as the coal blending constraint condition, and formulating the coal blending plan in combination with the raw coal optimization index;

[0026] B1-301: The coal making cycle is preset as Tz. In any coal making cycle Tz, raw coal i needs to simultaneously carry out the coal making operation process of n1 coal blending schemes. The reserve of raw coal i is marked as Li. The total amount of blended coal products produced by any coal blending scheme in one coal making cycle is preset as ZL. Combined with the proportion φi of raw coal i in the blended coal products, the coal quantity Ri required by raw coal i in any coal blending scheme is obtained, and then the total consumption Ei of raw coal i in any coal making cycle Tz is obtained;

[0027] B1-302: Then obtain the coal blending constraints and construct the coal blending index PM, and then perform data calculation on the coal blending index PM. Under the conditions of satisfying the coal blending constraints, obtain the maximum value of the coal blending index PM, and mark it as the optimal coal blending index PMmax. According to the optimal coal blending index PMmax, obtain the corresponding proportions of n0 kinds of raw coal, and generate the optimal coal blending plan and its corresponding No. 1 control management signal, so as to perform coal making operations.

[0028] Furthermore, the specific process of obtaining the coal supply forecast index is as follows:

[0029] C1: Regularly update the raw coal information of the coal supply point through data calculation, and mark the raw coal information of the coal supply point as dynamic raw coal information: the dynamic raw coal information includes the raw coal information of existing raw coal types and newly added raw coal types;

[0030] C1-1: Set the information update period Tg, obtain dynamic raw coal information regularly, and presume that there are N1 types of raw coal at the coal supply point within any information update period Tg, and mark any existing raw coal type as j;

[0031] C1-2: mark the raw coal quality coefficient of raw coal j as ZLj, and mark the cost of raw coal j as Cj and the reserve as Lj;

[0032] C1-3: With the information update cycle Tg as the horizontal coordinate and the cost Cj as the vertical coordinate, a dynamic curve graph S1 of cost Cj-information update cycle Tg is constructed; with the information update cycle Tg as the horizontal coordinate and the reserve Lj as the vertical coordinate, a dynamic curve graph S2 of reserve Lj-information update cycle Tg is constructed;

[0033] C2: Set up a curve analysis model for curve prediction analysis. The construction process of the curve analysis model is as follows:

[0034] C2-1: First extract n2 points of the curve S0, mark any point and its coordinates as p(Xp, Yp), mark the point closest to point p and its coordinates as q(Xq, Yq), obtain the slope value Kp of point p through the coordinate values ​​of point p and point q, and then obtain the slope values ​​corresponding to n2 points, and calculate the average value of the slope values ​​and mark it as the curve increase rate Ks;

[0035] C2-2: Calculate the average value PJs and standard deviation σs of the ordinates of the n2 points of the curve S0, and combine them with the curve increase rate Ks to generate the prediction index YCs of the curve S0;

[0036] C3: Substitute the dynamic curve S1 and the dynamic curve S2 into the curve analysis model in turn to obtain the corresponding prediction indexes respectively, and mark the cost prediction index of raw coal j as CBj and the reserve prediction index as CLj.

[0037] Furthermore, the specific process of optimizing the optimal coal blending model is as follows:

[0038] C4: According to the cost prediction index CBj and reserve prediction index CLj of raw coal j, the raw coal supply parameters of raw coal j are predicted and updated, and the optimal coal blending model is substituted to predict the coal blending scheme, so as to calculate and allocate coal mine resources and schedule raw coal in advance. The specific process is as follows:

[0039] C4-1: Generate the coal supply forecast coefficient GYj of raw coal j by combining the cost forecast index CBj and the reserve forecast index CLj of raw coal j;

[0040] C4-2: Feed the coal supply prediction coefficient GYj back to the initial model setting module, and then combine the raw coal quality coefficient ZLj of raw coal j with the coal supply prediction coefficient GYj to obtain the predicted raw coal preference index YXj of raw coal j and substitute it into the optimal coal blending model to generate the predicted optimal coal blending index, and then obtain the corresponding proportion of each raw coal according to the predicted optimal coal blending index, so as to generate the optimal coal blending prediction plan and its corresponding No. 2 control management signal.

[0041] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0042] The present invention collects information through a raw coal information monitoring unit, and then constructs a coal blending model through a coal blending model design unit to achieve the design of an optimal coal blending scheme. The model has a strong generalization ability and can adapt to changes in coal mine costs and changes in resource reserve conditions caused by coal mining. It can dynamically adjust the configuration by continuously updating the configuration scheme, meet user needs, and has flexibility and high efficiency. Then, through coal combustion prediction and model self-growth optimization, an optimal coal blending prediction scheme is generated, and the raw coal reserves of each supply point are scheduled in advance, thereby realizing the early supply and scheduling management of coal resources and ensuring the continuous stability of coal blending and coal making work.

[0043] Among them, the coal blending model first evaluates the quality of raw coal by monitoring the combustion performance and environmental impact of coal, and generates a raw coal optimization index based on the coal mine supply cost and storage situation. At the same time, it forms coal blending constraints based on user coal blending needs and actual coal resource reserves. Then, the raw coal ratio is set by combining the raw coal optimization index with the coal blending constraints to obtain the optimal coal blending plan. On the premise of meeting the user's required coal making cost and combustion performance, it ensures environmental friendliness and sufficient supply reserves, thereby achieving the best overall coal blending effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 The overall module schematic diagram of the present invention is shown;

[0045] Figure 2 The overall process schematic diagram of the present invention is shown;

[0046] Figure 3 A schematic flow chart of the coal blending model of the present invention is shown. DETAILED DESCRIPTION

[0047] 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.

[0048] Embodiment 1:

[0049] like Figure 1-3As shown, the distributed control precision coal blending system based on neural network includes a raw coal information monitoring unit, a coal blending model design unit and a dynamic coal blending scheduling unit, wherein the coal blending model design unit includes an initial model design module, a coal supply prediction module and a model self-growth optimization module;

[0050] Among them, the signal connection between the raw coal information monitoring unit, the coal blending model design unit and the dynamic coal blending scheduling unit, and the signal connection between the initial model design module, the coal supply prediction module and the model self-growth optimization module;

[0051] The working steps are as follows:

[0052] S1: The raw coal information monitoring unit collects and obtains raw coal information: the raw coal information includes coal combustion parameters and coal mine supply parameters;

[0053] S2: The coal blending model design unit analyzes and processes the raw coal information and establishes a coal blending model;

[0054] A neural network is a computational model that simulates the interconnections between neurons in the human brain. It is usually highly complex, including large-scale parameters and computational complexity. The neural network model can learn certain rules and patterns from data through learning and training, and then use them for prediction and decision-making. Therefore, the present invention establishes a coal blending model that is actually applied to coal blending work scenarios through a neural network model. The specific process is as follows:

[0055] S2-1: The initial model design module analyzes the combustion performance and environmental impact of raw coal through coal combustion parameters to generate the raw coal quality coefficient, and obtains the raw coal supply coefficient through coal mine supply parameter analysis. Then, the raw coal quality coefficient and the raw coal supply coefficient are combined to generate the raw coal optimization index, thereby constructing the optimal coal blending model;

[0056] S2-101: The specific process of generating the raw coal quality coefficient and the raw coal supply coefficient is:

[0057] A1: First analyze the coal combustion parameters:

[0058] Coal combustion parameters include the calorific value, moisture, volatile matter, sulfur and ash of the raw coal. Any raw coal is marked as y, and the calorific value of raw coal y is marked as Qy, moisture as Wy, volatile matter as Fy, sulfur as Sy, and ash as Hy;

[0059] A1-1: First, the combustion performance evaluation coefficient Xrs of raw coal y is generated by combining the calorific value Qy, moisture Wy and volatile matter Fy of raw coal y:

[0060] Among them, α1, α2 and α3 are weight factor coefficients of calorific value Qy, moisture Wy and volatile matter Fy, respectively, and α1, α2 and α3 are all greater than 0. The weight factor coefficients are obtained by calculating a large amount of experimental data. When the calorific value Qy and volatile matter Fy are higher and the moisture Wy is lower, the combustion performance evaluation coefficient Xrs is higher, indicating that the combustion performance of raw coal y is better.

[0061] A1-2: The environmental impact assessment coefficient Xhj of raw coal y is generated by combining the volatile matter Fy, sulfur Sy and ash Hy of raw coal y: Xhj = β1*Fy+β2*Sy+β3*Hy

[0062] Among them, β1, β2 and β3 are weight factor coefficients of volatile matter Fy, sulfur content Sy and ash content Hy, respectively, and β1, β2 and β3 are all greater than 0; when the volatile matter Fy, sulfur content Sy and ash content Hy are higher, the environmental impact assessment coefficient Xhj is higher, indicating that the impact of raw coal y on the environment is more serious;

[0063] A1-3: Then, by combining the combustion performance evaluation coefficient Xrs and the environmental impact evaluation coefficient Xhj of the raw coal y, the raw coal quality coefficient ZLy of the raw coal y is generated:

[0064] Among them, ω1 and ω2 are weight coefficients of the combustion performance evaluation coefficient Xrs and the environmental impact assessment coefficient Xhj, respectively, and ω1 and ω2 are both greater than 0; when the combustion performance evaluation coefficient Xrs is higher and the environmental impact assessment coefficient Xhj is lower, the raw coal quality coefficient ZLy is higher, which means that when the combustion performance of raw coal y is better and the impact on the environment is smaller, the quality effect of raw coal y is better, therefore, this type of raw coal is preferred for coal blending operation;

[0065] A2: Reanalyze coal mine supply parameters:

[0066] Coal mine supply parameters include the cost and reserves of raw coal. The cost of raw coal y is marked as Cy and the reserves are marked as Ly.

[0067] By combining the cost Cy and the reserves Ly, the raw coal supply coefficient GYy of raw coal y is generated:

[0068]

[0069] Among them, ε1 and ε2 are the weight coefficients of cost Cy and reserve Ly respectively, and both ε1 and ε2 are greater than 0; when the cost Cy is lower and the reserve Ly is higher, the raw coal supply coefficient GYy is higher, which means that when the cost of raw coal is lower and the reserve is higher, the supply effect of this type of raw coal is better, therefore, this type of raw coal is preferred for coal blending operation;

[0070] S2-102: The specific process of constructing the optimal coal blending model is:

[0071] B1: The raw coal quality coefficient ZLy and the raw coal supply coefficient GYy are combined to generate the raw coal optimization index YXy: YXy=μ1*ZLy+μ2*GYy

[0072] Among them, μ1 and μ2 are the weight coefficients of the raw coal quality coefficient ZLy and the raw coal supply coefficient GYy, respectively, and both μ1 and μ2 are greater than 0; when the raw coal quality coefficient ZLy and the raw coal supply coefficient GYy of raw coal y are higher, the corresponding raw coal preference index YXy is higher, which means that the better the combustion quality of raw coal y is, the more friendly the environmental impact is, and the better the supply effect is, the higher the priority of raw coal y is;

[0073] B1-1: N0 kinds of raw coal are preset, and the raw coal information of N0 kinds of raw coal is analyzed and processed through neural network and distributed computing control technology to obtain the corresponding raw coal optimization index;

[0074] B1-2: A blended coal product is obtained by blending several kinds of raw coals. The blended coal product is assumed to contain n0 kinds of raw coals. Any one of the raw coals is marked as i, and the raw coal preference index of raw coal i is YXi;

[0075] The proportion of raw coal i in the blended coal product is assumed to be φi: Among them, λi is the adjustment factor, which is obtained by simulating and calculating a large amount of data; when the raw coal optimization index YXi of raw coal i is higher, its proportion φi in the blended coal product is higher, and the adjustment factor is set to adjust the slight error of the raw coal proportion, so as to achieve the overall optimal effect of coal blending;

[0076] B1-3: Obtain the coal blending requirements of the current user, which include the minimum combustion calorific value Qz min and the maximum total coal cost Cz max , taking the user's coal blending demand as the coal blending constraint condition, and formulating the coal blending plan in combination with the raw coal optimization index;

[0077] B1-301: The coal making cycle is preset as Tz. In any coal making cycle Tz, raw coal i needs to simultaneously carry out the coal making operation process of n1 coal blending schemes. The reserve of raw coal i is marked as Li. The total amount of blended coal products produced by any coal blending scheme in one coal making cycle is preset as ZL. Combined with the proportion of raw coal i in the blended coal products φi, the coal quantity Ri required by raw coal i in any coal blending scheme is obtained, and then the total consumption Ei of raw coal i in any coal making cycle Tz is obtained:

[0078] B1-302: Then the coal blending constraint conditions are obtained as follows:

[0079] Constructing coal blending index PM:

[0080] When the calorific value Qi of raw coal i is higher and the cost Ci is lower, the effect of the coal making scheme is better. When the coal blending index PM of the coal making plant is higher, it means that the overall coal blending effect presented by the coal making plant in the process of simultaneous coal making of n1 coal blending schemes is better.

[0081] Then, the coal blending index PM is calculated, and the maximum value of the coal blending index PM is obtained under the coal blending constraint condition, and it is marked as the optimal coal blending index PMmax. According to the optimal coal blending index PMmax, the corresponding proportions of n0 kinds of raw coal are obtained, that is, the specific value of the proportion φi of raw coal i is obtained, and the optimal coal blending plan and control management signal are generated, so as to carry out coal making operation;

[0082] Among them, according to the actual reserves of raw coal and the actual coal production needs of each user, the overall coal mine resource allocation and configuration are carried out. Under the premise of low cost, high heat and sufficient reserves of the coal blending scheme, the overall coal blending and coal production effect is optimized. In addition, the model has strong generalization ability and can adapt to changes in coal mine conditions, including changes in coal mine costs and reserves caused by resource mining. By continuously updating the configuration plan, dynamic adjustment configuration is achieved, which can meet user needs and is flexible and efficient.

[0083] S2-2: The coal supply forecasting module is used to conduct in-depth analysis of raw coal information, monitor and dispatch the historical supply conditions of different coal supply points, forecast coal demand, and thus obtain the coal supply forecasting index;

[0084] Distributed control technology refers to the distribution of the calculation, decision-making and execution functions of the control system to different locations of the system, and the control of the entire system is achieved through communication and coordination. Data processing equipment is set up at different coal supply points as different port locations of the system to synchronously collect corresponding raw coal information and monitor fluctuations, as well as subsequent data processing, and realize the overall coal distribution scheduling control of the entire system through network communication technology;

[0085] C1: Data is calculated through distributed computing technology to regularly update the raw coal information of coal supply points, and the raw coal information of coal supply points is marked as dynamic raw coal information: dynamic raw coal information includes raw coal information of existing raw coal types and newly added raw coal types. Through dynamic raw coal information, the newly added raw coal types and their corresponding raw coal information of coal supply points, as well as the changes in coal mine supply parameters corresponding to existing raw coal types, are monitored;

[0086] C1-1: Set the information update period Tg, obtain dynamic raw coal information regularly, and presume that there are N1 types of raw coal at the coal supply point within any information update period Tg. Mark any existing raw coal type as j, then the coal combustion parameters of raw coal j remain unchanged, and the coal mine supply parameters are in a dynamic change state;

[0087] C1-2: mark the raw coal quality coefficient of raw coal j as ZLj, and mark the cost of raw coal j as Cj and the reserve as Lj;

[0088] C1-3: With the information update cycle Tg as the horizontal coordinate and the cost Cj as the vertical coordinate, a dynamic curve graph S1 of cost Cj-information update cycle Tg is constructed; with the information update cycle Tg as the horizontal coordinate and the reserve Lj as the vertical coordinate, a dynamic curve graph S2 of reserve Lj-information update cycle Tg is constructed;

[0089] C2: Set up a curve analysis model for curve prediction analysis. The construction process of the curve analysis model is as follows:

[0090] C2-1: First extract n2 points of curve S0, mark any point and its coordinates as p(Xp, Yp), mark the point closest to point p and its coordinates as q(Xq, Yq), and obtain the slope value Kp of point p through the coordinate values ​​of point p and point q: Then, the slope values ​​corresponding to the n2 points are obtained, and the average value of the slope values ​​is calculated and marked as the curve increase rate Ks:

[0091] C2-2: Calculate the average value PJs and standard deviation σs of the ordinates of the n2 points of the curve S0, and combine them with the curve increase rate Ks to generate the prediction index YCs of the curve S0: YCs = PJs*(1+Ks) σs ;

[0092] When the ordinate mean value PJs, ordinate standard deviation σs and curve increase rate Ks of curve S0 are higher, the prediction index YCs of curve S0 is higher;

[0093] C3: Substitute the dynamic curve S1 and the dynamic curve S2 into the curve analysis model in turn to obtain the corresponding prediction indexes respectively, and mark the cost prediction index of raw coal j as CBj and the reserve prediction index as CLj;

[0094] When the cost prediction index CBj of raw coal j is higher, it means that the cost prediction of raw coal j is higher due to the complex changes in coal mine conditions; when the reserve prediction index CLj of raw coal j is higher, it means that more reserves of raw coal j need to be supplied and replenished;

[0095] S2-3: The model self-growth optimization module continuously optimizes the optimal coal blending model by feeding back the coal supply forecast index to the initial model design module, thereby achieving model self-growth upgrade;

[0096] C4: According to the cost prediction index CBj and reserve prediction index CLj of raw coal j, the raw coal supply parameters of raw coal j are predicted and updated, and the optimal coal blending model is substituted to predict the coal blending scheme, so as to calculate and allocate coal mine resources and schedule raw coal in advance. The specific process is as follows:

[0097] C4-1: Similar to the formula of raw coal supply coefficient GYy, the coal supply forecast coefficient GYj of raw coal j is generated by combining the cost forecast index CBj and the reserve forecast index CLj of raw coal j:

[0098]

[0099] C4-2: Feed the coal supply prediction coefficient GYj back to the initial model setting module, and then obtain the raw coal prediction optimization index YXj of raw coal j by combining the raw coal quality coefficient ZLj of raw coal j with the coal supply prediction coefficient GYj, and build the optimal coal blending prediction model. Similar to the optimal coal blending model, generate the optimal coal blending prediction index and its corresponding optimal coal blending prediction scheme;

[0100] According to the optimal coal blending prediction plan, the raw coal reserves of each supply point are scheduled in advance, so as to realize the advance supply and scheduling management of coal resources and ensure the continuous stability of coal blending and coal making;

[0101] S3: The dynamic coal blending scheduling unit performs distributed coal blending control operations: generates control management signals through the optimal coal blending model to dynamically schedule and manage coal blending;

[0102] The specific process of distributed coal blending control operation is as follows:

[0103] Distributed remote control technology is used to control the sending and receiving of management signals, so as to remotely operate and control each coal supply point to achieve real-time adjustment of coal supply scheduling;

[0104] Among them, by receiving control management signals, text editing and visualization are performed in the background center of the coal supply point. The text content is the proportion of each raw coal in the current optimal coal blending plan and its demand. The raw coal is dispatched between coal supply points according to the raw coal demand, and the raw coal resources stored in the coal supply point are extracted for coal blending and coal making operations.

[0105] In summary, the present invention collects information through the raw coal information monitoring unit, and then constructs the coal blending model through the coal blending model design unit to achieve the design of the optimal coal blending scheme. The model has strong generalization ability and can adapt to changes in coal mine conditions, including changes in coal mine costs and changes in resource reserves caused by coal mining. It realizes dynamic adjustment configuration by continuously updating the configuration scheme, which can meet user needs and has flexibility and high efficiency. Then, through coal combustion prediction and model self-growth optimization, the optimal coal blending prediction scheme is generated to perform mutual advance scheduling of the raw coal reserves of each supply point, thereby realizing the advance supply and scheduling management of coal resources and ensuring the continuous stability of coal blending and coal making work.

[0106] Among them, the coal blending model first evaluates the quality of raw coal by monitoring the combustion performance and environmental impact of coal, and generates a raw coal optimization index based on the coal mine supply cost and storage situation. At the same time, it forms coal blending constraints based on user coal blending needs and actual coal resource reserves. Then, the raw coal ratio is set by combining the raw coal optimization index with the coal blending constraints to obtain the optimal coal blending plan. On the premise of meeting the user's required coal making cost and combustion performance, it ensures environmental friendliness and sufficient supply reserves, thereby achieving the best overall coal blending effect.

[0107] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technical personnel in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0108] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0109] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. Distributed control precision coal blending system based on neural network, characterized by: It includes a raw coal information monitoring unit, a coal blending model design unit and a dynamic coal blending scheduling unit, wherein the coal blending model design unit includes an initial model design module, a coal supply prediction module and a model self-growth optimization module; The raw coal information monitoring unit is used to collect and obtain raw coal information: raw coal information includes coal combustion parameters and coal mine supply parameters; The coal blending model design unit is used to analyze and process raw coal information and establish a coal blending model: Among them, the initial model design module analyzes the combustion performance and environmental impact of raw coal through coal combustion parameters to generate the raw coal quality coefficient, and obtains the raw coal supply coefficient through coal mine supply parameter analysis, and then generates the raw coal optimization index by combining the raw coal quality coefficient with the raw coal supply coefficient, so as to construct the optimal coal blending model; the coal supply prediction module is used to conduct in-depth analysis of raw coal information, monitor and schedule the historical supply situation of different coal supply points, predict coal demand, and obtain the coal supply prediction index; the model self-growth optimization module continuously optimizes the optimal coal blending model by feeding back the coal supply prediction index to the initial model design module, thereby realizing the self-growth upgrade of the model; S2-101: The specific process of generating the raw coal quality coefficient and the raw coal supply coefficient is: A1: First analyze the coal combustion parameters: Coal combustion parameters include the calorific value, moisture, volatile matter, sulfur and ash of the raw coal. Any raw coal is marked as y, and the calorific value of raw coal y is marked as Qy, moisture as Wy, volatile matter as Fy, sulfur as Sy, and ash as Hy; A1-1: First, the combustion performance evaluation coefficient Xrs of raw coal y is generated by combining the calorific value Qy, moisture Wy and volatile matter Fy of raw coal y: ; Among them, α1, α2 and α3 are weight factor coefficients of calorific value Qy, water content Wy and volatile matter Fy respectively, and α1, α2 and α3 are all greater than 0. The weight factor coefficients are obtained by calculating a large amount of experimental data in advance. A1-2: The environmental impact assessment coefficient Xhj of raw coal y is generated by combining the volatile matter Fy, sulfur Sy and ash Hy of raw coal y: ; Among them, β1, β2 and β3 are the weight factor coefficients of volatile matter Fy, sulfur content Sy and ash content Hy, and β1, β2 and β3 are all greater than 0; A1-3: Then, by combining the combustion performance evaluation coefficient Xrs and the environmental impact evaluation coefficient Xhj of the raw coal y, the raw coal quality coefficient ZLy of the raw coal y is generated: ; Among them, ω1 and ω2 are weight coefficients of the combustion performance evaluation coefficient Xrs and the environmental impact assessment coefficient Xhj, respectively, and ω1 and ω2 are both greater than 0; A2: Reanalyze coal mine supply parameters: Coal mine supply parameters include the cost and reserves of raw coal. The cost of raw coal y is marked as Cy and the reserves are marked as Ly. By combining the cost Cy and the reserves Ly, the raw coal supply coefficient GYy of raw coal y is generated: ； Among them, ε1 and ε2 are the weight coefficients of cost Cy and reserves Ly respectively, and both ε1 and ε2 are greater than 0; The specific process of constructing the optimal coal blending model is as follows: B1: By combining the raw coal quality coefficient ZLy with the raw coal supply coefficient GYy, the raw coal optimization index YXy is generated: ; Among them, μ1 and μ2 are the weight coefficients of the raw coal quality coefficient ZLy and the raw coal supply coefficient GYy, respectively, and both μ1 and μ2 are greater than 0; B1-1: N0 kinds of raw coal are preset, and the raw coal information of N0 kinds of raw coal is analyzed and processed through neural network and distributed computing control technology to obtain the corresponding raw coal optimization index; B1-2: A blended coal product is obtained by blending several kinds of raw coals. The blended coal product is assumed to contain n0 kinds of raw coals. Any one of the raw coals is marked as i, and the raw coal preference index of raw coal i is YXi; The proportion of raw coal i in the blended coal product is assumed to be φi: , where λi is the adjustment factor, which is derived by simulating and calculating a large amount of data; B1-3: Obtain the coal blending requirements of the current user, which include the minimum combustion calorific value Qz min and the maximum total coal cost Cz max , taking the user's coal blending demand as the coal blending constraint condition, and formulating the coal blending plan in combination with the raw coal optimization index; B1-301: The coal making cycle is preset as Tz. In any coal making cycle Tz, raw coal i needs to simultaneously carry out the coal making operation process of n1 coal blending schemes. The reserve of raw coal i is marked as Li. The total amount of blended coal products produced by any coal blending scheme in one coal making cycle is preset as ZL. Combined with the proportion of raw coal i in the blended coal products φi, the coal quantity Ri required by raw coal i in any coal blending scheme is obtained, and then the total consumption Ei of raw coal i in any coal making cycle Tz is obtained: ; B1-302: Then the coal blending constraint conditions are obtained as follows: ; Constructing coal blending index PM: , where Qi is the calorific value of raw coal i, Ci is the cost of raw coal i; Then, the coal blending index PM is calculated, and the maximum value of the coal blending index PM is obtained under the coal blending constraint condition, and it is marked as the optimal coal blending index PMmax. According to the optimal coal blending index PMmax, the corresponding proportions of n0 kinds of raw coal are obtained, that is, the specific value of the proportion φi of raw coal i is obtained, and the optimal coal blending plan and control management signal are generated; The dynamic coal blending scheduling unit is used to perform distributed coal blending control operations: the control management signal is generated through the optimal coal blending model to dynamically schedule and manage the coal blending.

2. The distributed control accurate coal blending system based on neural network according to claim 1 is characterized by: The specific process of obtaining the coal supply forecast index is as follows: C1: Data is calculated through distributed computing technology, and the raw coal information of the coal supply point is marked as dynamic raw coal information: the dynamic raw coal information includes the raw coal information of existing raw coal types and newly added raw coal types. The new raw coal types and their corresponding raw coal information of the coal supply point, as well as the changes in the coal mine supply parameters corresponding to the existing raw coal types, are monitored through the dynamic raw coal information; C1-1: Set the information update period Tg, obtain dynamic raw coal information regularly, and presume that there are N1 types of raw coal at the coal supply point within any information update period Tg. Mark any existing raw coal type as j, then the coal combustion parameters of raw coal j remain unchanged, and the coal mine supply parameters are in a dynamic change state; C1-2: mark the raw coal quality coefficient of raw coal j as ZLj, and mark the cost of raw coal j as Cj and the reserve as Lj; C1-3: With the information update cycle Tg as the horizontal coordinate and the cost Cj as the vertical coordinate, a dynamic curve graph S1 of cost Cj-information update cycle Tg is constructed; with the information update cycle Tg as the horizontal coordinate and the reserve Lj as the vertical coordinate, a dynamic curve graph S2 of reserve Lj-information update cycle Tg is constructed; C2: Set up a curve analysis model for curve prediction analysis. The construction process of the curve analysis model is as follows: C2-1: First extract n2 points of curve S0, mark any point and its coordinates as p (Xp, Yp), mark the point closest to point p and its coordinates as q (Xq, Yq), and obtain the slope value Kp of point p through the coordinate values ​​of point p and point q: , and then obtain the slope values ​​corresponding to n2 points, and calculate the average value of the slope values ​​and mark it as the curve increase rate Ks: ; C2-2: Calculate the average value PJs and standard deviation σs of the ordinates of the n2 points of the curve S0, and combine them with the curve increase rate Ks to generate the prediction index YCs of the curve S0: ; When the ordinate mean value PJs, ordinate standard deviation σs and curve increase rate Ks of curve S0 are higher, the prediction index YCs of curve S0 is higher; C3: Substitute the dynamic curve S1 and the dynamic curve S2 into the curve analysis model in turn to obtain the corresponding prediction indexes respectively, and mark the cost prediction index of raw coal j as CBj and the reserve prediction index as CLj.

3. The distributed control accurate coal blending system based on neural network according to claim 2 is characterized by: The specific process of optimizing the optimal coal blending model is as follows: C4: According to the cost prediction index CBj and reserve prediction index CLj of raw coal j, the raw coal supply parameters of raw coal j are predicted and updated, and the optimal coal blending model is substituted to predict the coal blending scheme, so as to calculate and allocate coal mine resources and schedule raw coal in advance. The specific process is as follows: C4-1: The coal supply forecast coefficient GYj of raw coal j is generated by combining the cost forecast index CBj and the reserve forecast index CLj of raw coal j: ； C4-2: Feed the coal supply prediction coefficient GYj back to the initial model setting module, and then combine the raw coal quality coefficient ZLj of raw coal j with the coal supply prediction coefficient GYj to obtain the predicted raw coal preference index YXj of raw coal j and substitute it into the optimal coal blending model to generate the predicted optimal coal blending index. Then, according to the predicted optimal coal blending index, the corresponding proportion of each raw coal is obtained, thereby generating the optimal coal blending prediction plan and its corresponding No. 2 control management signal, so as to carry out advance scheduling operations for coal making.

Images