Method and device for real-time regulation and control of matte quality in complex copper resource smelting process

Through the optimization of PID single-neuron network algorithm and learning differential sparrow search algorithm, the precise control of the quality of ice copper during copper smelting is achieved, and the problems of inaccurate quality control and untimely impurity regulation in traditional smelting methods are solved, and the smelting is improved The stability and efficiency of the process reduce the production of harmful substances.

CN118938643APending Publication Date: 2024-11-12YIMEN COPPER CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410816991.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-24
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

During the traditional copper smelting process, the quality control of ice copper relies on manual experience and cannot achieve precise control. The impurity elements are not regulated in time, which cannot meet the sustainable development requirements of modern industry.

Method used

The PID single-neuron network algorithm is used to combine fluorescence analysis and real-time monitoring to automatically adjust the speed of copper concentrate transmission belt, control the content of main elements and impurity elements through the inner and outer rings, and optimize the weight value of the neural network by using the learning differential sparrow search algorithm to achieve Precise control of copper quality.

Benefits of technology

The stability and consistency of copper quality has been improved, manual intervention has been reduced, production efficiency and stability has been improved, production costs have been reduced, and the production of harmful substances has been reduced, which has promoted the virtuous cycle of the ecological environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118938643A_ABST
    Figure CN118938643A_ABST
Patent Text Reader

Abstract

The invention discloses a method and device for real-time regulation and control of matte quality in a complex copper resource smelting process, and the method comprises the following steps: setting a target matte quality fluctuation range according to the influence of different in-furnace copper concentrate type element contents on the matte quality; according to the target quality range and the actual quality, the PID single-neuron network algorithm is used for automatically adjusting the speed of a conveying belt to which all the in-furnace copper concentrates belong, and the quality of the smelting product matte is adjusted and controlled; wherein the quality of the matte is monitored in real time according to fluorescence analysis, and data of a controlled object is fed back to the PID control system; and calculating the corresponding control quantity according to the deviation between the current quality and the target quality. By the adoption of the technical scheme, accurate control over elements such as Cu, Fe, S, As, Sb and Pb can be achieved, and therefore the quality of the matte is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of metallurgy, and in particular relates to a method and device for real-time control of matte quality in a complex copper resource smelting process. Background Art

[0002] Copper is an important industrial metal, widely used in the fields of electricity, electronics, and machinery manufacturing. At present, copper smelting mainly adopts pyrometallurgy. Pyrometallurgy mainly includes two processes: smelting and blowing. Smelting is the process of converting copper concentrate into matte. The traditional smelting method relies solely on manual experience to control the quality of matte, and it is impossible to accurately control the quality of matte. At the same time, there are also problems such as untimely control of impurity elements, which does not meet the sustainable development requirements of modern industry.

[0003] In recent years, under the background of "dual carbon" and "green development", copper smelting enterprises have ushered in the development direction of "green development" and "digital transformation". Some enterprises in the domestic copper smelting industry have carried out exploration of smart smelting, but online control is the initial stage of smart smelting. Traditional PID control cannot coordinate and control the copper smelting process with multiple inputs. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and device for real-time regulation of matte quality in a complex copper resource smelting process, which can achieve precise control of elements such as Cu, Fe, S, As, Sb, Pb, etc., thereby improving the quality of matte.

[0005] To achieve the above object, the present invention adopts the following technical solution:

[0006] A method for real-time control of matte quality in a complex copper resource smelting process comprises the following steps:

[0007] According to the influence of element content of different types of copper concentrate on the quality of matte, the target matte quality fluctuation range is set;

[0008] According to the target quality range and actual quality, the PID single neuron network algorithm is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace to achieve regulation and control of the quality of the smelting product, the matte. Among them, the quality of the matte is monitored in real time based on fluorescence analysis, and the data of the controlled object is fed back to the PID control system. According to the deviation between the current quality and the target quality, the corresponding control amount is calculated.

[0009] Preferably, based on the deviation between the control amount and the target amount analyzed by the fluorescence analyzer, the inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

[0010] Preferably, the outer loop system uses PID control, and the parameters of KP, KI, and KD of the PID control system are adjusted according to the feedback quantity to monitor the quality of the matte in real time; for the inner loop system, the controlled object is controlled by multiple inputs, and a single neuron PID is used for each input quantity.

[0011] Preferably, the single neuron PID control uses a learning differential sparrow search algorithm to perform multi-dimensional optimization on the initial weight values ​​of the neural network.

[0012] The present invention also provides a device for real-time control of matte quality in a complex copper resource smelting process, comprising:

[0013] The first processing module is used to set the target matte quality fluctuation range according to the influence of the element content of different types of copper concentrates entering the furnace on the quality of matte;

[0014] The second processing module is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace according to the target quality range and the actual quality, using the PID single neuron network algorithm to achieve regulation and control of the quality of the smelting product matte; wherein, the quality of the matte is monitored in real time according to fluorescence analysis, and the data of the controlled object is fed back to the PID control system; and the corresponding control amount is calculated according to the deviation between the current quality and the target quality.

[0015] Preferably, based on the deviation between the control amount and the target amount analyzed by the fluorescence analyzer, the inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

[0016] Preferably, the outer loop system uses PID control, and the parameters of KP, KI, and KD of the PID control system are adjusted according to the feedback quantity to monitor the quality of the matte in real time; for the inner loop system, the controlled object is controlled by multiple inputs, and a single neuron PID is used for each input quantity.

[0017] Preferably, the single neuron PID control uses a learning differential sparrow search algorithm to perform multi-dimensional optimization on the initial weight values ​​of the neural network.

[0018] The beneficial effects of the present invention are:

[0019] Through PID adjustment, relevant parameters can be accurately adjusted according to real-time matte quality data to improve the stability and consistency of matte quality. The PID control system can achieve automatic adjustment, reduce manual intervention, and improve production efficiency and stability. By accurately controlling the relevant parameters of matte production, production costs can be reduced, while reducing the generation of harmful substances As, Sb, and Pb, thereby reducing impurity removal operations, and promoting a virtuous cycle of the ecological environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0021] Figure 1 This is the structure diagram of the matte quality PID control system;

[0022] Figure 2 This is the structure diagram of the single neuron PID control system with multiple input variables;

[0023] Figure 3 This is the simulation diagram of PID adjustment for matte quality;

[0024] Figure 4 This is the simulation diagram of Cu proportion PID regulation;

[0025] Figure 5 This is the S proportion PID adjustment simulation diagram;

[0026] Figure 6 This is the simulation diagram of As proportion PID regulation;

[0027] Figure 7 This is the Pb proportion PID adjustment simulation diagram;

[0028] Figure 8 This is the simulation diagram of PID adjustment for matte quality;

[0029] Fig. 9 This is the simulation diagram of Cu proportion PID regulation;

[0030] Fig.10 This is the S proportion PID adjustment simulation diagram;

[0031] Fig.11 This is the simulation diagram of As proportion PID regulation;

[0032] Fig.12 This is the Pb proportion PID adjustment simulation diagram. DETAILED DESCRIPTION

[0033] 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.

[0034] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] Embodiment 1:

[0036] like Figure 1 , 2 As shown, the present invention provides a real-time control method for the quality of copper matte in the complex copper resource smelting process, which adopts PID control, and realizes precise control of the quality of copper matte by adjusting the content of main elements Cu, Fe, S, impurity elements As, Sb, Pb and oxygen in the smelting process. The specific implementation steps are as follows:

[0037] According to the influence of element content of different types of copper concentrate on the quality of matte, the target matte quality fluctuation range is set;

[0038] According to the target quality range and actual quality, the PID single neuron network algorithm is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace to achieve regulation and control of the quality of the smelting product, the matte. Among them, the quality of the matte is monitored in real time based on fluorescence analysis, and the data of the controlled object is fed back to the PID control system. According to the deviation between the current quality and the target quality, the corresponding control amount is calculated.

[0039] As an implementation method of an embodiment of the present invention, according to the deviation between the control amount and the target amount analyzed by the fluorescence analyzer, the inner loop system adjusts elements such as Cu, Fe, and S, and the outer loop system adjusts the quality of the matte.

[0040] Based on the continuous adjustment of the inner and outer loop PID, the quality of the matte gradually approaches the target value while maintaining a stable output.

[0041] The entire system consists of two closed loops. The inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

[0042] The outer loop system uses traditional PID control to adjust the relevant parameters of the PID control system - KP, KI, and KD - according to the feedback amount, monitor the quality of the matte in real time, achieve the target matte quality and make it stable.

[0043] For the inner loop system, the controlled object is controlled by multiple inputs, and each input needs to be controlled. In this case, a single neuron PID is used for each input quantity. The single neuron PID can optimize the network connection weights and key parameters in real time. The error signal between the feedback quantity and the input quantity is converted into a neuron network input variable after passing through the state converter. The connection weight of the neuron is approximately equal to the coefficients of the proportional, integral, and differential terms of the PID controller. By adjusting the connection weight, the PID control parameters can be adjusted online.

[0044] Single neuron PID control also needs to use the learning-based differential sparrow search algorithm (LDSSA) to optimize the initial weight value of the neural network in multiple dimensions to prevent the stability of achieving the target matte quality from being delayed due to weight fluctuations.

[0045] The weights are optimized using the learning differential sparrow search algorithm (LDSSA). When the individual sparrow is not threatened and does not exceed the set maximum number of stagnant steps, the optimization is still performed according to the standard sparrow search algorithm. If the individual is threatened, a learning mutation operation is performed to avoid the "aggregation" phenomenon. When the individual is not threatened, the update position is set:

[0046]

[0047] in, is the updated position, Indicates the position before updating. represents the global optimal position; f i and f w They represent the global best fitness value and the worst fitness value respectively; ε represents the minimum constant, the purpose of which is to prevent the denominator from being 0; K is a random number between positive and negative 1;

[0048] When an individual is threatened, a learning mutation operation is performed:

[0049] W new =W best ±F×(w i -w j )

[0050] Where ± indicates upward or downward difference; W new is the position after mutation, W best represents the optimal position before mutation, w i 、w j Indicates two suboptimal positions.

[0051]

[0052] Where F is the scaling factor, which is adaptively changed according to the two individuals of the generated difference vector.

[0053] W new Substitute the values ​​into the following formula:

[0054]

[0055] This is the updated position.

[0056] Through the timely feedback of the inner and outer loop systems analyzed by the fluorescence analyzer, it can be known that the inner loop controls the furnace temperature, oxygen concentration, additive input, the input ratio of the main elements Cu, Fe, S, and the impurity elements As, Sb, Pb, etc., and is adjusted to obtain the proportion parameters that meet the target matte quality; the outer loop controls the matte quality, and according to the deviation between the fluorescence analyzer and the target matte quality, it is adjusted to obtain matte that meets the target matte quality.

[0057] The embodiment of the present invention first determines the influence of Cu, Fe, S, As, Sb, and Pb elements on the quality of matte. Secondly, a single neuron PID control algorithm is designed, and the PID is adjusted using the sparrow optimization algorithm. According to the target matte quality range and the current quality, the input proportion of elements such as Cu, Fe, S, As, Sb, and Pb and the oxygen flow rate are automatically adjusted; the composition of matte is monitored in real time by a fluorescence analyzer, and the data is fed back to the inner and outer loop PID control systems respectively. Again, the control system feeds back the controlled object to the oxygen and flow control devices of the element control devices such as Cu, Fe, S, As, Sb, and Pb to achieve precise adjustment of the above parameters; after continuous PID adjustment, the quality of matte is kept floating at the target value.

[0058] Embodiment 2:

[0059] like Figure 1 , 2 As shown, an embodiment of the present invention provides a method for real-time control of matte quality in a complex copper resource smelting process, comprising the following steps:

[0060] A quality sensor and a fluorescence analyzer are installed on the ore feeding conveyor belt to collect the addition amount information of each ore; the content of Cu, Fe, S, As, Sb, and Pb elements in each ore is determined by the fluorescence analyzer;

[0061] Collect matte data information and transmit it to the PID controller;

[0062] Design a single neuron PID control based on the sparrow optimization algorithm and obtain the optimization parameters;

[0063] The deviation between the smelting data analysis and the target element content ratio is fed back to each single neuron to form PID feedback regulation, and the ore type regulation and management is carried out according to the feedback data.

[0064] As an implementation method of an embodiment of the present invention, according to the chemical composition of copper ore in the copper smelting process, the influence of As, Sb, and Pb impurity elements on the quality of matte is analyzed, and the proportion of impurity element content is controlled.

[0065] As an implementation method of the embodiment of the present invention, the element analysis of copper ore is carried out by using a fluorescence analyzer in the conveyor belt of each ore type, and the element composition data of the ore type entering the furnace is collected in real time.

[0066] As an implementation method of the present invention, a single neuron PID is used for one-by-one control, and a learning-based differential sparrow search algorithm (LDSSA) is used to optimize the initial weight value of the neural network in multiple dimensions. The single neuron PID is used to accurately control the multi-input controlled object, and the specific operation is as follows: the error signal is transformed into a neural network input variable through a state converter, and the connection weight of the neuron is approximately equal to the coefficient of the proportional, integral, and differential terms of the PID controller; by adjusting the connection weight, the PID control parameters are adjusted online.

[0067] Furthermore, the learning differential sparrow search algorithm (LDSSA) adopts a differential mutation strategy, that is, to avoid clustering when searching for the optimal position of an individual, the optimal solution in the current iteration and two random suboptimal positions are selected in the algorithm to perform a learning mutation operation. The position parameters before and after the mutation (W represents the position before and after the mutation, and F is the scaling factor) are set as follows:

[0068] W new =W best ±F×(w i -w j )

[0069] Where ± indicates upward or downward difference; W new is the position after mutation, W best represents the optimal position before mutation, w i 、w j Indicates two suboptimal positions.

[0070]

[0071] Where F is the scaling factor, which is adaptively changed according to the two individuals of the generated differential vector; F u , F1 is the adaptive change value.

[0072] As an implementation mode of an embodiment of the present invention, feedback adjustment is performed based on the error between a preset impurity element content ratio and a current impurity element content ratio. If the content of impurity elements such as As, Sb, Pb, etc. is higher than the preset content ratio, the amount of the corresponding ore type fed into the furnace is reduced, that is, a command is issued through the PID controller to adjust and reduce the rotation speed of the conveyor belt belonging to the corresponding ore type; if the content of impurity elements such as As, Sb, Pb, etc. is lower than the preset content ratio, the amount of the corresponding ore type fed into the furnace is increased, that is, a command is issued through the PID controller to adjust and increase the rotation speed of the conveyor belt.

[0073] As an implementation method of an embodiment of the present invention, the amount of ore entering the furnace is reduced, and a learning differential sparrow search algorithm (LDSSA) is used to update the parameters of the weights of the neural network. The weight parameters updated by each neuron are adjusted through PID. The rotation speed of the conveyor belt corresponding to the ore with an impurity element content higher than a preset level is reduced, so that the amount of ore entering the furnace tends to be stable as soon as possible and reaches a preset proportion of the corresponding impurity element content; the same applies to the impurity element content lower than the preset level.

[0074] The embodiments of the present invention can be applied to the matte production process in the copper smelting industry, and are suitable for smelting equipment and process flows of various scales.

[0075] Experiment 1:

[0076] In order to verify the present invention, a model was established to carry out simulation experiments.

[0077] First, according to the chemical composition of copper concentrate, when the target matte grade is 75%, the proportions of main elements and impurity elements are shown in Table 1 below:

[0078] Table 1

[0079]

[0080] Establish a matte quality PID control system and input the target matte;

[0081] According to the chemical composition and thermodynamic properties of the copper ore, the proportions of the main elements Cu, Fe, S, impurity elements As, Sb, Pb and other elements corresponding to the single neuron PID network are set, and then the furnace temperature, oxygen flow rate and additive input are determined. The neural network and LDSSA optimization algorithm are initialized, and the initial weights in the network are randomly generated by the software, the learning rate is selected as 0.4, the neuron proportional coefficient is 0.3, and the sampling period is 1s.

[0082] The accused was smelted;

[0083] After detection by fluorescence analyzer, the quality of matte obtained after smelting meets the target matte quality, and the main elements and impurity elements meet the target proportions. The main element Cu accounts for 75.1%, S accounts for 18.55%, both meet the target proportions and tend to be stable, and the impurity elements As and Pb account for 0.25% and 1.08% respectively, which also meet the target proportions and tend to be stable.

[0084] By connecting the smelting results to an oscilloscope, we can clearly observe the rise time and overshoot before the matte grade and element ratio stabilize. Figure 3It can be seen that the actual matte grade obtained deviates from the target matte grade in the first 30 seconds, and the results are fed back to the inner loop and outer loop respectively. After the outer loop controlled system is adjusted, the matte grade approaches 75% and reaches stability; the input of the inner loop controlled system is provided by the output of the neural network, and the output of the controlled object is fed back to the input of the neural network. The content of elements such as Cu, Fe, S, As, Sb, and Pb is observed through an oscilloscope. The initial weight values ​​of the input layer and hidden layer that need to be optimized are fed back to the single neural network, and the improved sparrow search algorithm is used for optimization. The corresponding deviation is calculated to obtain a simulation diagram that tends to be stable. This example takes the main element Cu and the impurity elements As and Pb as examples to observe the PID control system based on the LDSSA algorithm.

[0085] Figure 3 This is the simulation diagram of PID adjustment of matte grade. To prove the effectiveness of the method proposed in this invention, it is compared with the traditional PID control system. Figure 3 It can be seen that the PID control system based on the LDSSA algorithm has approached the target value of the matte grade at 20 seconds, and the neural network using the optimization algorithm has greatly improved the speed and accuracy of tracking the given value compared with the neural network with fixed initial values, and the target value of the controlled object is more accurate and stable. The traditional PID control system reaches stability after 20 seconds, and its stability is significantly lower than that of the method proposed in the present invention.

[0086] Figure 4 This is the simulation diagram of Cu proportion PID adjustment. Compared with the traditional PID control system. Figure 4 It can be seen that in the PID control system based on the LDSSA algorithm, Cu reaches stability before 10 seconds, and the optimized single neuron PID reduces the fluctuation of Cu proportion, and approaches the target value more stably and faster. However, the traditional PID control system needs continuous adjustment in overshoot and adjustment time before it becomes stable, and the adjustment time exceeds 10 seconds, and its stability is lower than the method proposed in the present invention.

[0087] Figure 5 This is the S proportion PID adjustment simulation diagram. As the Cu proportion gradually tends to be stable, the S element also tends to be stable after regulation, and it is more stable than the traditional PID control towards the target proportion.

[0088] Figure 6 , Figure 7 The following is a simulation diagram of the PID adjustment of the proportion of impurity elements As and Pb, compared with the traditional PID control system. Figure 6It can be seen that in the PID control system based on the LDSSA algorithm, the impurity element As reaches stability before 8s, and the optimized single neuron PID makes the As proportion fluctuate less. However, the traditional PID control system has overshoot and a long adjustment time. It tends to be stable after continuous adjustment, and the adjustment time is close to 10s. The stability is lower than the method proposed in the present invention. It can also be proved that the impurity element Pb reaches the target proportion first and tends to be stable.

[0089] Experiment 2:

[0090] Based on Experiment 1, this example uses the traditional PID system as an example to compare the system's rapidity.

[0091] First, a model was established for simulation experiments. According to the chemical composition of copper concentrate, when the target matte grade is 73%, the proportions of main elements and impurity elements are shown in Table 2 below:

[0092] Table 2

[0093]

[0094] Establish a matte quality PID control system and input the target matte;

[0095] According to the chemical composition and thermodynamic properties of copper ore, the proportions of main elements Cu, Fe, S, impurity elements As, Sb, Pb and other elements corresponding to the single neuron PID network are set, and then the furnace temperature, oxygen flow rate and additive input are determined. The neural network is initialized, and the initial weights in the network are randomly generated by the software. The learning rate is selected as 0.4, the neuron proportional coefficient is 0.3, and the sampling period is 1s.

[0096] The accused was smelted;

[0097] After detection by fluorescence analyzer, the quality of matte obtained after smelting meets the target matte quality, and the main elements and impurity elements meet the target proportions. The main element Cu accounts for 74.08%, which meets the target Cu proportion and tends to be stable. The impurity elements As, Sb, and Pb also reach the target proportion and tend to be stable. This example also takes the main element Cu and the impurity elements As and Pb as examples to observe the PID control system.

[0098] By connecting the smelting results to an oscilloscope, we can clearly observe the rise time and overshoot before the matte quality and element ratio stabilize. Figure 5It can be seen that the actual matte quality obtained deviates from the target matte quality in the first 30 seconds, and then the results are fed back to the inner and outer loops respectively. After the outer loop controlled system is adjusted, the matte grade approaches 72.89% and reaches stability; the input of the inner loop controlled system is provided by the output of the neural network, and the output of the controlled object is fed back to the input of the neural network. The content of elements such as Cu, Fe, S, As, Sb, and Pb is observed through an oscilloscope, and the initial weight values ​​of the input layer and hidden layer that need to be optimized are fed back to the single neural network, and the corresponding deviation is calculated to obtain the simulation diagram.

[0099] Figure 8 This is the simulation diagram of PID adjustment for matte quality. Figure 8 It can be seen that the system tends to be stable at 30 seconds, and the quality of the matte approaches 73%, reaching the target high quality of matte, but the adjustment time is after 30 seconds. The method provided by the present invention achieves the target matte quality before 30 seconds. Therefore, the rapidity of the traditional PID control system is lower than that of the method provided by the present invention.

[0100] Fig. 9 That is the PID adjustment simulation diagram of the main element Cu proportion. Figure 7 It can be seen that the Cu content tends to be stable at 20 seconds, and the method provided by the present invention achieves it before 10 seconds. Compared with the traditional PID control system, the method provided by the present invention is more effective in terms of rapidity.

[0101] Fig.10 The PID adjustment simulation diagram of the main element S ratio is Fig.10 It can be seen that element S tends to be stable after 5 seconds, and its rapidity is obviously lower than that of the method provided by the present invention.

[0102] Fig.11 , Fig.12 They are the PID adjustment simulation diagrams of the proportion of impurity elements As and Pb respectively. Fig.10 It can be seen that the As content tends to be stable at 10s, and the method provided by the present invention reaches it in 8s. Compared with the traditional PID control system, the method provided by the present invention is more effective in terms of rapidity. However, the content of impurity element Pb in the smelting process is very small, and its rapidity is significantly reduced.

[0103] Experiment 1 and Experiment 2 were compared in terms of stability and rapidity. The method for real-time control of matte quality in a complex copper resource smelting process proposed by the present invention is more accurate, rapid and stable than the traditional PID system based on neural network and optimization algorithm, which verifies the effectiveness of the present invention.

[0104] Embodiment 3:

[0105] The embodiment of the present invention further provides a device for real-time control of matte quality in a complex copper resource smelting process, comprising:

[0106] The first processing module is used to set the target matte quality fluctuation range according to the influence of the element content of different types of copper concentrates entering the furnace on the quality of matte;

[0107] The second processing module is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace according to the target quality range and the actual quality, using the PID single neuron network algorithm to achieve regulation and control of the quality of the smelting product matte; wherein, the quality of the matte is monitored in real time according to fluorescence analysis, and the data of the controlled object is fed back to the PID control system; and the corresponding control amount is calculated according to the deviation between the current quality and the target quality.

[0108] As an implementation method of an embodiment of the present invention, according to the deviation between the control amount and the target amount analyzed by the fluorescence analyzer, the inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

[0109] As an implementation method of an embodiment of the present invention, the outer loop system uses PID control, and the parameters of KP, KI, and KD of the PID control system are adjusted according to the feedback quantity to monitor the quality of matte in real time; for the inner loop system, the controlled object is controlled by multiple inputs, and a single neuron PID is used for each input quantity.

[0110] As an implementation method of an embodiment of the present invention, the single neuron PID control uses a learning differential sparrow search algorithm to perform multi-dimensional optimization on the initial weight value of the neural network.

[0111] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for real-time control of matte quality in a complex copper resource smelting process, characterized in that: The following steps are involved: According to the influence of the element content of different types of copper concentrate on the quality of matte, the target matte quality fluctuation range is set; According to the target quality range and actual quality, the PID single neuron network algorithm is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace to achieve regulation and control of the quality of the smelting product, the matte. Among them, the quality of the matte is monitored in real time based on fluorescence analysis, and the data of the controlled object is fed back to the PID control system. According to the deviation between the current quality and the target quality, the corresponding control amount is calculated.

2. The method for real-time control of matte quality in a complex copper resource smelting process according to claim 1, characterized in that: According to the deviation between the controlled amount and the target amount analyzed by the fluorescence analyzer, the inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

3. The method for real-time control of matte quality in a complex copper resource smelting process according to claim 2, characterized in that: The outer loop system uses PID control, and the parameters of KP, KI, and KD of the PID control system are adjusted according to the feedback quantity to monitor the quality of matte in real time. For the inner loop system, the controlled object is controlled by multiple inputs, and a single neuron PID is used for each input quantity.

4. The method for real-time control of matte quality in a complex copper resource smelting process according to claim 3, characterized in that: The single neuron PID control uses a learning differential sparrow search algorithm to optimize the initial weight value of the neural network in multiple dimensions.

5. A device for real-time control of matte quality in a complex copper resource smelting process, characterized in that: include: The first processing module is used to set the target matte quality fluctuation range according to the influence of the element content of different types of copper concentrates entering the furnace on the quality of matte; The second processing module is used to automatically adjust the speed of the conveyor belts of each copper concentrate entering the furnace according to the target quality range and the actual quality, using the PID single neuron network algorithm to achieve regulation and control of the quality of the smelting product matte; wherein, the quality of the matte is monitored in real time according to fluorescence analysis, and the data of the controlled object is fed back to the PID control system; and the corresponding control amount is calculated according to the deviation between the current quality and the target quality.

6. The device for real-time control of matte quality in a complex copper resource smelting process according to claim 5, characterized in that: According to the deviation between the controlled amount and the target amount analyzed by the fluorescence analyzer, the inner loop controls the main elements Cu, Fe, S, impurity elements As, Pb, Bi, Sb, temperature, oxygen concentration, and the proportion of additives and auxiliary materials, and the outer loop controls the target matte quality.

7. The device for real-time control of matte quality in a complex copper resource smelting process according to claim 6, characterized in that: The outer loop system uses PID control, and the parameters of KP, KI, and KD of the PID control system are adjusted according to the feedback quantity to monitor the quality of matte in real time. For the inner loop system, the controlled object is controlled by multiple inputs, and a single neuron PID is used for each input quantity.

8. The device for real-time control of matte quality in a complex copper resource smelting process according to claim 7, characterized in that: The single neuron PID control uses a learning differential sparrow search algorithm to optimize the initial weight value of the neural network in multiple dimensions.