CN117920446B - Semi-autogenous mill running state optimization method based on digital twin - Google Patents
Semi-autogenous mill running state optimization method based on digital twin Download PDFInfo
- Publication number
- CN117920446B CN117920446B CN202410326898.5A CN202410326898A CN117920446B CN 117920446 B CN117920446 B CN 117920446B CN 202410326898 A CN202410326898 A CN 202410326898A CN 117920446 B CN117920446 B CN 117920446B
- Authority
- CN
- China
- Prior art keywords
- semi
- autogenous mill
- model
- follows
- rate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000005457 optimization Methods 0.000 title claims description 13
- 239000000463 material Substances 0.000 claims abstract description 18
- 229910052500 inorganic mineral Inorganic materials 0.000 claims abstract description 9
- 239000011707 mineral Substances 0.000 claims abstract description 9
- 229910000831 Steel Inorganic materials 0.000 claims description 31
- 239000010959 steel Substances 0.000 claims description 31
- 230000006870 function Effects 0.000 claims description 22
- 238000013016 damping Methods 0.000 claims description 20
- 238000011156 evaluation Methods 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000000227 grinding Methods 0.000 claims description 14
- 230000008569 process Effects 0.000 claims description 14
- 239000002245 particle Substances 0.000 claims description 13
- 230000004913 activation Effects 0.000 claims description 12
- 210000002569 neuron Anatomy 0.000 claims description 12
- 230000014509 gene expression Effects 0.000 claims description 11
- 230000003068 static effect Effects 0.000 claims description 11
- 230000007246 mechanism Effects 0.000 claims description 10
- 238000011084 recovery Methods 0.000 claims description 7
- 230000035772 mutation Effects 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 230000002068 genetic effect Effects 0.000 claims description 4
- 108090000623 proteins and genes Proteins 0.000 claims description 4
- 206010064571 Gene mutation Diseases 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 4
- 238000004088 simulation Methods 0.000 abstract description 3
- 238000012545 processing Methods 0.000 abstract description 2
- 238000004458 analytical method Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 4
- 238000000926 separation method Methods 0.000 description 3
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011439 discrete element method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011112 process operation Methods 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Crushing And Grinding (AREA)
Abstract
The invention relates to the technical field of mineral processing equipment, and discloses a method for optimizing the running state of a semi-autogenous mill based on digital twin. According to the method for optimizing the running state of the semi-autogenous mill based on digital twin, the influence of the ball-to-ball ratio on the power is further considered, the parameters are continuously optimized through real-time simulation on the power, the power is stabilized within a certain range, the production efficiency is improved, the safety risk is reduced, the running states of mineral materials and media in the semi-autogenous mill can be better displayed through observing a speed cloud picture, and meanwhile, the time lag is greatly reduced, and the real-time state of equipment can be reflected.
Description
Technical Field
The invention relates to the technical field of mineral processing equipment, in particular to a semi-autogenous mill running state optimization method based on digital twin.
Background
Ore grinding is an industrial process, and accounts for a large proportion in the ore dressing industry, and a semi-autogenous mill is main mining equipment in the link. The main working device of the semi-autogenous mill is a roller which is horizontally arranged and rotatable, a lifter is arranged on the inner wall of the roller, and the inner wall of the roller comprises a roller surface and two side end surfaces. When the semi-autogenous mill operates, ore entering the cylinder from one end of the cylinder is impacted, ground and crushed between the steel balls and the ore and under the impact of the cylinder, and then is discharged from the other end of the cylinder.
At present, aiming at parameter optimization of a semi-autogenous mill, the grinding performance of the semi-autogenous mill is researched by a discrete element method, the influence rule of each working parameter on the grinding performance of the mill is analyzed, and the working parameters are optimized and analyzed based on a uniform design method and a hierarchical analysis method, but the real-time performance is insufficient, and the factor consideration is incomplete; meanwhile, aiming at the problems that sensitivity of multi-source (multi-channel) mechanical signals collected in a ball mill system under different operation conditions has variability, redundancy, complementarity and the like exist in the information of the load parameters of the included mill, a multi-operation condition multi-channel mechanical signal analysis, evaluation and optimization combination method for MLP prediction is provided, but mechanism support is lacked, the adaptability of a conclusion on an industrial mill is to be verified, and further research also needs to be conducted in depth by combining the multi-operation condition and a multi-component signal self-adaptive decomposition algorithm.
Disclosure of Invention
(One) solving the technical problems
Aiming at the defects of the prior art, the invention provides a semi-autogenous mill running state optimization method based on digital twinning, which has the advantages of continuously optimizing parameters by simulating power in real time, ensuring stable power and the like, and solves the technical problems.
(II) technical scheme
In order to achieve the above purpose, the present invention provides the following technical solutions: a semi-autogenous mill running state optimization method based on digital twin comprises the following steps:
S1, building a three-dimensional model according to a semi-autogenous mill, building a mechanism model of mineral aggregate and steel balls according to a grinding process, and combining the three-dimensional model and the mechanism model to build a digital twin model;
s2, setting boundary conditions required by the ore grinding process, and setting filling rate, rotation rate and ball ratio;
s3, transmitting the filling rate, the rotation rate and the ball ratio to a receiving end of the semi-autogenous mill through a digital twin model, starting to operate according to initial operation parameters, and collecting actual operation parameters in real time;
S4, predicting the power value by combining the real-time operation parameters in the step S3 to obtain a predicted power value;
s5, evaluating a prediction result according to a preset evaluation index;
s6, judging whether the parameters need to be optimized, if the evaluation index is met, not changing the current operation parameters, executing S7, if the evaluation index is not met, optimizing the current operation parameters, correcting the digital twin model, inputting the optimized parameters into a semi-autogenous mill, adjusting the ore feeding amount in real time, and executing S7;
and S7, judging whether the ore grinding process is finished, if so, finishing the operation, and if not, circulating the steps S4 to S6.
As a preferred embodiment of the present invention, the mechanism model in the step S1 includes a contact model for obtaining a contact force between ore particles and a motion modelThe specific calculation formula is as follows:
wherein, Is the normal spring rate and,Is the normal damping coefficient of the device,Is the normal overlap between the ore particles,Is the relative normal velocity;
wherein, Indicating the tangential contact force, the contact force is,Representing the smallest set to be taken for the internal data,Is the tangential spring rate and the spring force,Is the tangential damping coefficient of the material,Is the tangential overlap between the ore particles,Is the relative tangential velocity of the light beam,Is the coefficient of static friction.
As a preferred embodiment of the present invention, the normal spring rateAnd normal damping coefficientThe calculation formula of (2) is as follows:
wherein, Is an effective young's modulus, and is excellent in the mechanical strength,Is the effective radius of the lens and is,Is the effective mass of the product, and the product has the advantages of high quality,Is the normal overlap between the ore particles,Representing the evolution of the internal data,Represents the damping coefficient of the damping device,Represents normal stiffness;
The tangential spring rate And tangential damping coefficientThe calculation formula of (2) is as follows:
wherein, Is the effective shear modulus.
As a preferable mode of the invention, the damping coefficientAnd normal stiffnessThe calculation formula of (2) is as follows:
wherein, Is the coefficient of restitution,* Representing a logarithmic function.
As a preferable technical scheme of the invention, the motion model is used for analyzing the motion state of the steel ball in the semi-autogenous mill, and the steel ball in the semi-autogenous mill is separated from the cylinder body of the semi-autogenous millRate of rotation atFilling rateCalculating and simultaneously setting the inner steel ball at the falling pointAnalysis of the movement of the part, wherein the rotational speed isFilling rateThe calculation formula of (2) is as follows:
wherein, Is the angular speed of the operation of the semi-autogenous mill cylinder,For the critical angular velocity to be the one,Is the radius of the inner layer of the steel ball,Is the radius of the semi-autogenous mill cylinder.
As a preferred embodiment of the present invention, the boundary conditions in the step S2 include material characteristic parameters including poisson' S ratio, material density and shear modulus, and physical mechanical characteristics including a coefficient of restitution, static friction and sliding friction between materials.
As a preferable technical scheme of the invention, a prediction model is established for the prediction process in the step S4, and the ore feeding amount, the rotation rate, the ball ratio and the filling rate are used as input layersSimultaneously build the hidden layer asIn which there is presentIndividual nodes, corresponding output layersWherein the hidden layerAnd an output layerThe output expression of (2) is as follows:
wherein, Input layer of representationThe number of nodes in the network is,Indicating hidden layer numberThe outputs of the nodes of the respective pairs of nodes,Representing the corresponding hidden layerNode output of individualThe value of the one of the values,Representing an input layerHidden layerThe weight of the corresponding node in between,Representing hidden layersAnd an output layerThe weight of the corresponding node in between,AndThe hidden layer neuron activation function and the output layer neuron activation function are represented respectively,Representing input toThe individual data are summed up and,Representing hidden layersThe data are summed and，。
As a preferred embodiment of the present invention, the error function of the prediction modelThe expression is as follows:
wherein, The number of input samples represented is the number of samples,The output value is represented by a value of the output,The desired value is indicated to be the desired value,Representing summing the internal data;
Weighting by chain law AndAdjusting to obtain the adjusted corresponding weightAndThe specific expression is as follows:
wherein, Representing the calculation of the partial derivatives,The learning rate is indicated as being indicative of the learning rate,AndRepresenting the derivative of the hidden layer neuron activation function and the output layer neuron activation function, respectively.
As a preferred embodiment of the present invention, the evaluation index in step S5 isThe expression of (2) is as follows:
wherein, The predicted value is represented by a value of the prediction,The true value is represented by a value that is true,Representing summing the internal N data.
As a preferred embodiment of the present invention, the judging process in the step S6 is as follows:
s6.1, pair evaluation index Judging the value of (1), if the evaluation index isDirectly ending without changing the current ore feeding amount, if the evaluation index isS6.2 is performed;
S6.2、 The current ore feeding amount is optimized through a genetic algorithm, gene mutation is carried out according to random mutation point positions through mutation operation, namely, original gene values are inverted, the fitness of each individual is calculated again, and parameters with the highest fitness are input into a semi-autogenous mill as optimization parameters, so that the ore feeding amount is adjusted in real time.
Compared with the prior art, the invention provides a semi-autogenous mill running state optimization method based on digital twinning, which has the following beneficial effects:
According to the invention, the influence of ball comparison on power is further considered by a digital twin-based semi-autogenous mill operation parameter optimization method, a digital twin-based model of the semi-autogenous mill capable of dynamically interacting is established by a digital twin technology, parameters are continuously optimized by real-time simulation on power, so that power is stabilized in a certain range, the production efficiency is improved, the safety risk is reduced, the movement states of mineral materials and media in the semi-autogenous mill can be better displayed by observing a speed cloud picture, meanwhile, the simulation power is output on line by the grinding process operation parameters of the semi-autogenous mill, the system can immediately carry out decision adjustment, the time lag is greatly reduced, and the real-time state of equipment can be reflected to a certain extent.
Drawings
FIG. 1 is a schematic flow chart of the present invention;
FIG. 2 is a schematic diagram of the operation flow of the semi-autogenous mill of the invention;
FIG. 3 is a schematic view of a contact model of the present invention;
FIG. 4 is a schematic diagram of the motion profile of the medium of the present invention;
FIG. 5 is a schematic diagram of simulated speeds of a semi-autogenous mill according to the present invention;
FIG. 6 is a schematic diagram of the power variation of the semi-autogenous mill of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1-6, a method for optimizing the running state of a semi-autogenous mill based on digital twin comprises the following steps:
S1, building a three-dimensional model according to a semi-autogenous mill, building a mechanism model of mineral aggregate and steel balls according to a grinding process, and combining the three-dimensional model and the mechanism model to build a digital twin model, wherein the mechanism model comprises a contact model and a motion model;
the contact model is used for obtaining the contact force between ore particles The specific calculation formula is as follows:
wherein, Is the normal spring rate and,Is the normal damping coefficient of the device,Is the normal overlap between the ore particles,Is the relative normal velocity;
wherein, Indicating the tangential contact force, the contact force is,Representing the smallest set to be taken for the internal data,Is the tangential spring rate and the spring force,Is the tangential damping coefficient of the material,Is the tangential overlap between the ore particles,Is the relative tangential velocity of the light beam,Is the static coefficient of friction, the normal spring rateAnd normal damping coefficientThe calculation formula of (2) is as follows:
wherein, Is an effective young's modulus, and is excellent in the mechanical strength,Is the effective radius of the lens and is,Is the effective mass of the product, and the product has the advantages of high quality,Is the normal overlap between the ore particles,Representing the evolution of the internal data,Represents the damping coefficient of the damping device,Expressed as normal stiffness, tangential spring stiffnessAnd tangential damping coefficientThe calculation formula of (2) is as follows:
wherein, Damping coefficient for effective shear modulusAnd normal stiffnessThe calculation formula of (2) is as follows:
wherein, Is the coefficient of restitution,* Representing a logarithmic function;
The motion model is specifically as follows, and mutual interference between the steel ball and mineral aggregate is ignored, and the steel ball and the mineral aggregate are regarded as particles. Assuming that the medium at the separation point A is a steel ball, when the medium leaves the cylinder, the separation angle formed by the medium and the Y direction is Finally, the liquid falls to a falling point B on the right side of the cylinder body, and forms a falling angle with the X direction; angular velocity of. The following analysis was performed for the disengagement point a and the drop-back point B, respectively:
the stress relation of the steel ball at the point A can be known:
In the X-Y coordinate system, it is available according to the geometrical relationship: And (2) and ；
The steel ball motion trail equation of the point A can be obtained:；
At the point A where the ball is separated from the ball At this time, its component speed in x and y
The motion analysis of the point A is mainly aimed at the steel ball at the outermost layer of the cylinder, and when the steel ball at the inner layer of the cylinder reaches a certain height to separate, the motion analysis can be obtained:
At the same time, the rotation speed rate can be obtained Filling rateThe calculation formula of (2) is as follows:
wherein, Is the angular speed of the operation of the semi-autogenous mill cylinder,For the critical angular velocity to be the one,The radius of the inner layer of the steel ball movement is as follows;
the fact that the movement of the steel balls on the inner layer and the outer layer is inconsistent is indicated, when the steel balls on the outer layer are thrown out, the disengaging angle of the steel balls on the outer layer is related to the rotating speed of the cylinder, and the inner layer is influenced by the filling rate;
the steel ball moves in a parabolic manner after leaving the point A, and the track equation is as follows:
Meanwhile, when the cylinder moves clockwise, the track equation of the steel ball is as follows:
Further, the position of the B point is known as ，And then get
From geometrical relationships，；
The height of the steel ball rising away from the point a (the vertical distance between the highest point reached and the disengagement point) is:
thus when falling from the highest point to point B, the vertical total height is:
At the point B, the separation speed of the steel ball in the x and y directions is as follows
The B-point velocity is known as:，
further, the normal and tangential speeds of the point B are known to be:
S2, setting boundary conditions required by the ore grinding process, setting filling rate, rotation rate and ball ratio, setting boundary conditions required by the ore grinding process according to production conditions, wherein the boundary conditions comprise material characteristic parameters and physical mechanical characteristics, the material characteristic parameters comprise Poisson 'S ratio, material density and shear modulus, the physical mechanical characteristics comprise collision recovery coefficient, static friction and sliding friction between materials, the effective diameter of ore granularity is 260mm, 200mm, 180mm, 80mm and 60mm in sequence, the medium size is 125mm, the shear modulus of the ore is 1x108Pa, the material density is 2600kg/m3, the Poisson' S ratio is 0.3, the shear modulus of the steel ball is 7x1010Pa, the material density is 7800kg/m3, the Poisson 'S ratio is 0.3, the shear modulus of the semi-autogenous mill cylinder liner is 7x1010Pa, the material density is 7800kg/m3, the Poisson' S ratio is 0.3, the recovery coefficient between ores is 0.3, the static friction coefficient is 0.65, and the sliding friction coefficient is 0.05; the recovery coefficient between the ore and the steel ball is 0.5, the static friction coefficient is 0.4, and the sliding friction coefficient is 0.05; the recovery coefficient between the ore and the lining plate is 0.5, the static friction coefficient is 0.4, and the sliding friction coefficient is 0.1; the recovery coefficient between the steel balls is 0.75, the static friction coefficient is 0.35, and the sliding friction coefficient is 0.1; the recovery coefficient between the steel ball and the lining plate is 0.75, the static friction coefficient is 0.5, and the sliding friction coefficient is 0.2; and the initial filling rate is set to be 35%, the rotation rate is set to be 75%, and the ball ratio is set to be 1:1, a step of;
S3, transmitting parameters such as filling rate, rotation rate, material ball ratio, ore feeding amount and the like to the entity equipment through a digital twin model, starting operation of the entity equipment according to the initial operation parameters, and receiving actual operation parameters such as temperature, oil pressure, power and the like acquired by the entity equipment in real time;
S4, predicting the power value by combining the real-time operation parameters in the step S3 to obtain a predicted power value, and establishing a prediction model by taking the ore feeding amount, the rotation rate, the ball ratio and the filling rate as input layers Simultaneously build the hidden layer asIn which there is presentIndividual nodes, corresponding output layersWherein the hidden layerAnd an output layerThe output expression of (2) is as follows:
wherein, Input layer of representationThe number of nodes in the network is,Indicating hidden layer numberThe outputs of the nodes of the respective pairs of nodes,Representing the corresponding hidden layerNode output of individualThe value of the one of the values,Representing an input layerHidden layerThe weight of the corresponding node in between,Representing hidden layersAnd an output layerThe weight of the corresponding node in between,AndThe hidden layer neuron activation function and the output layer neuron activation function are represented respectively,Representing input toThe individual data are summed up and,Representing hidden layersThe data are summed and，;
Error function of predictive modelThe expression is as follows:
wherein, The number of input samples represented is the number of samples,The output value is represented by a value of the output,The desired value is indicated to be the desired value,Representing summing the internal data; initializing the weight range from-1 to 1, setting the bias value to 1, selecting a sample immediately after 100 times of training, calculating the output of a hidden layer and an output layer, updating the weight through an error function to obtain a predicted power predicted value, and updating the weight through a chain ruleAndAdjusting to obtain the adjusted corresponding weightAndThe specific expression is as follows:
wherein, Representing the calculation of the partial derivatives,The learning rate is indicated as being indicative of the learning rate,AndDerived functions representing hidden layer neuron activation functions and output layer neuron activation functions, respectively
S5, evaluating the prediction result according to a preset evaluation indexThe expression of (2) is as follows:
wherein, The predicted value is represented by a value of the prediction,The true value is represented by a value that is true,Representing summing the internal N data;
S6, if If the ore feeding amount is less than or equal to 2, the current ore feeding amount is not changed, ifIf the ratio is greater than 2, optimizing the current ore feeding amount through a genetic algorithm, firstly carrying out individual coding on the basis of binary system such as a transformation rate, a ball ratio, a filling rate, the ore feeding amount and the like, setting a population scale to be 4, calculating the fitness of different individuals, determining the genetic probability of the individuals through the ratio of the fitness of the individuals to the fitness of the sum after calculating the fitness sum, carrying out random pairing through cross operation, setting random cross points, mutually exchanging part of genes of paired individuals, carrying out gene mutation according to the random mutation point position through mutation operation, namely, taking the original gene value as a reverse, calculating the fitness of each individual again, inputting the parameter with the highest fitness into a semi-autogenous mill as an optimization parameter, and regulating the ore feeding amount in real time;
S7, judging whether to finish the ore grinding process, if yes, finishing the operation, and if no, circulating the steps S4 to S6
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (6)
1. A semi-autogenous mill running state optimization method based on digital twin is characterized in that: the method comprises the following steps:
S1, building a three-dimensional model according to a semi-autogenous mill, building a mechanism model of mineral aggregate and steel balls according to a grinding process, and combining the three-dimensional model and the mechanism model to build a digital twin model;
The mechanism model in the step S1 includes a contact model and a motion model, wherein the contact model is used for obtaining a contact force F n between ore particles, and a specific calculation formula is as follows:
Fn＝-Knδn+Cnvn
Wherein K n is the normal spring rate, C n is the normal damping coefficient, δ n is the normal overlap between ore particles, v n is the relative normal velocity;
Ft＝-min{μFn,Ktδt+Ctvt}
Where F t represents tangential contact force, min { } represents the smallest set of internal data, K t is tangential spring rate, C t is tangential damping coefficient, delta t is tangential overlap between ore particles, v t is relative tangential velocity, μ is static friction coefficient;
the normal spring rate K n and the normal damping coefficient C n are calculated as follows:
Wherein E * is the effective Young's modulus, R * is the effective radius, m * is the effective mass, delta n is the normal overlap between ore particles, The method is characterized in that the method comprises the steps of performing square operation on internal data, wherein beta represents a damping coefficient, and S n represents normal stiffness;
The tangential spring rate K t and tangential damping coefficient C t are calculated as follows:
Wherein G * is the effective shear modulus;
The damping coefficient beta and the normal stiffness S n are calculated according to the following formula:
wherein ε is the recovery coefficient and ln represents the logarithmic function;
The motion model is used for analyzing the motion state of the steel ball in the semi-autogenous mill, and the motion condition of the steel ball in the falling point B is analyzed by calculating the rotation speed rate N and the filling rate xi k of the steel ball in the barrel body of the semi-autogenous mill at the falling point A, wherein the calculation formula of the rotation speed rate N and the filling rate xi k is as follows:
Wherein ω is the angular velocity falling to the falling back point B, ω cri is the critical angular velocity, R k is the inner layer radius of the steel ball movement, and R is the semi-autogenous mill barrel radius;
s2, setting boundary conditions required by the ore grinding process, and setting filling rate, rotation rate and ball ratio;
s3, transmitting the filling rate, the rotation rate and the ball ratio to a receiving end of the semi-autogenous mill through a digital twin model, starting to operate according to initial operation parameters, and collecting actual operation parameters in real time;
S4, predicting the power value by combining the real-time operation parameters in the step S3 to obtain a predicted power value;
s5, evaluating a prediction result according to a preset evaluation index;
s6, judging whether the parameters need to be optimized, if the evaluation index is met, not changing the current operation parameters, executing S7, if the evaluation index is not met, optimizing the current operation parameters, correcting the digital twin model, inputting the optimized parameters into a semi-autogenous mill, adjusting the ore feeding amount in real time, and executing S7;
and S7, judging whether the ore grinding process is finished, if so, finishing the operation, and if not, circulating the steps S4 to S6.
2. The method for optimizing the running state of a semi-autogenous mill based on digital twinning according to claim 1, wherein the method comprises the following steps: the boundary conditions in the step S2 include material characteristic parameters including poisson' S ratio, material density, and shear modulus, and physical mechanical characteristics including a coefficient of restitution, static friction, and sliding friction between materials.
3. The method for optimizing the running state of a semi-autogenous mill based on digital twinning according to claim 1, wherein the method comprises the following steps: establishing a prediction model for the prediction process in the step S4, taking the ore feeding quantity, the rotation rate, the ball ratio and the filling rate as an input layer X, simultaneously establishing a hidden layer Z as an output layer Y with q nodes, wherein the output expressions of the hidden layer Z and the output layer Y are as follows:
Wherein X i represents the ith node of the input layer, Z k represents the output corresponding to the kth node of the hidden layer, Y j represents the jth value corresponding to the kth node output of the hidden layer, V ki represents the weight of the corresponding node between the input layer X and the hidden layer Z, W jk represents the weight of the corresponding node between the hidden layer Z and the output layer Y, f 1 and f 2 represent the hidden layer neuron activation function and the output layer neuron activation function, respectively, Representing the summation of the n data entered,/>The q data representing the hidden layer are summed and k= {1,2, …, q }, j= {1,2, …, m }.
4. A method for optimizing the operating state of a digital twin based semi-autogenous mill as recited in claim 3, wherein: the step S4 further includes an error function E p, where the error function E p is expressed as follows:
Where p represents the number of input samples, y jp represents the output value, y' jp represents the desired value, Representing summing the internal data;
The weights W jk and V ki are adjusted through a chain rule to obtain adjusted corresponding weights delta W jk and delta V ki, and the specific expression is as follows:
wherein, Representing partial derivative calculations, η represents learning rate, and f '1 and f' 2 represent derivative functions of hidden layer neuron activation function and output layer neuron activation function, respectively.
5. The method for optimizing the running state of a semi-autogenous mill based on digital twinning according to claim 1, wherein the method comprises the following steps: the expression of the evaluation index RMSE in the step S5 is as follows:
wherein, Represents a predicted value, X i represents a true value,/>Representing summing the internal N data.
6. The method for optimizing the running state of the semi-autogenous mill based on digital twinning according to claim 5, wherein the method comprises the following steps of: the judging process in the step S6 is as follows:
S6.1, judging the value of the evaluation index RMSE, if the evaluation index RMSE is less than or equal to 2, directly ending, not changing the current ore feeding amount, and if the evaluation index RMSE is more than 2, executing S6.2;
S6.2, optimizing the current ore feeding amount through a genetic algorithm, carrying out gene mutation according to random mutation point positions through mutation operation, namely reversing original gene values, calculating the fitness of each individual again, inputting the parameter with the highest fitness into a semi-autogenous mill as an optimization parameter, and adjusting the ore feeding amount in real time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410326898.5A CN117920446B (en) | 2024-03-21 | 2024-03-21 | Semi-autogenous mill running state optimization method based on digital twin |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410326898.5A CN117920446B (en) | 2024-03-21 | 2024-03-21 | Semi-autogenous mill running state optimization method based on digital twin |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117920446A CN117920446A (en) | 2024-04-26 |
| CN117920446B true CN117920446B (en) | 2024-05-31 |
Family
ID=90754209
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410326898.5A Active CN117920446B (en) | 2024-03-21 | 2024-03-21 | Semi-autogenous mill running state optimization method based on digital twin |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117920446B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118122480A (en) * | 2024-05-08 | 2024-06-04 | 昆明理工大学 | Semi-autogenous mill state monitoring method and system based on digital twin |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11253828A (en) * | 1998-03-09 | 1999-09-21 | Sekisui Chem Co Ltd | Working power estimating method of inner and outer cylinder interrotating type ball mill |
| CN112115649A (en) * | 2020-09-29 | 2020-12-22 | 郑州轻工业大学 | Vertical mill multi-field coupling system process parameter optimization method based on digital twin |
| CN112347689A (en) * | 2020-10-30 | 2021-02-09 | 昆明理工大学 | Discrete element method-based semi-autogenous mill lining plate shape optimization method |
| EP4033321A1 (en) * | 2021-01-22 | 2022-07-27 | Tata Consultancy Services Limited | System and method for performance and health monitoring to optimize operation of a pulverizer mill |
| WO2022236064A2 (en) * | 2021-05-06 | 2022-11-10 | Strong Force Iot Portfolio 2016, Llc | Quantum, biological, computer vision, and neural network systems for industrial internet of things |
| CN115845991A (en) * | 2022-12-06 | 2023-03-28 | 昆明理工大学 | Method for determining steel ball grading by predicting semi-autogenous grinding hard stone crushing effect based on Tavares crushing model |
| CN117313435A (en) * | 2023-11-29 | 2023-12-29 | 昆明理工大学 | Speed characteristic comparison analysis method for different medium proportioning schemes based on discrete element method |
| CN117299335A (en) * | 2023-10-08 | 2023-12-29 | 太原理工大学 | Ball mill ore discharge granularity control method and system based on digital twin technology |
| CN117399153A (en) * | 2023-09-25 | 2024-01-16 | 武汉科大电控设备有限公司 | Self-adaptive control method and system for mill |
| CN117541717A (en) * | 2023-10-27 | 2024-02-09 | 鞍钢集团矿业有限公司 | Modeling method of digital twin body of ball mill |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2012216845B2 (en) * | 2011-09-20 | 2017-12-07 | Russell Mineral Equipment Pty Ltd | Method And System For Simulating A Mill Reline |
| CN109499697B (en) * | 2018-12-17 | 2019-10-25 | 东北大学 | A kind of parameter determination method of three mill tubes high-frequency vibration grinding machine |
| CN114330168B (en) * | 2021-12-30 | 2022-06-21 | 中国科学院力学研究所 | Method for dynamically evaluating slope safety |
-
2024
- 2024-03-21 CN CN202410326898.5A patent/CN117920446B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11253828A (en) * | 1998-03-09 | 1999-09-21 | Sekisui Chem Co Ltd | Working power estimating method of inner and outer cylinder interrotating type ball mill |
| CN112115649A (en) * | 2020-09-29 | 2020-12-22 | 郑州轻工业大学 | Vertical mill multi-field coupling system process parameter optimization method based on digital twin |
| CN112347689A (en) * | 2020-10-30 | 2021-02-09 | 昆明理工大学 | Discrete element method-based semi-autogenous mill lining plate shape optimization method |
| EP4033321A1 (en) * | 2021-01-22 | 2022-07-27 | Tata Consultancy Services Limited | System and method for performance and health monitoring to optimize operation of a pulverizer mill |
| WO2022236064A2 (en) * | 2021-05-06 | 2022-11-10 | Strong Force Iot Portfolio 2016, Llc | Quantum, biological, computer vision, and neural network systems for industrial internet of things |
| CN115845991A (en) * | 2022-12-06 | 2023-03-28 | 昆明理工大学 | Method for determining steel ball grading by predicting semi-autogenous grinding hard stone crushing effect based on Tavares crushing model |
| CN117399153A (en) * | 2023-09-25 | 2024-01-16 | 武汉科大电控设备有限公司 | Self-adaptive control method and system for mill |
| CN117299335A (en) * | 2023-10-08 | 2023-12-29 | 太原理工大学 | Ball mill ore discharge granularity control method and system based on digital twin technology |
| CN117541717A (en) * | 2023-10-27 | 2024-02-09 | 鞍钢集团矿业有限公司 | Modeling method of digital twin body of ball mill |
| CN117313435A (en) * | 2023-11-29 | 2023-12-29 | 昆明理工大学 | Speed characteristic comparison analysis method for different medium proportioning schemes based on discrete element method |
Non-Patent Citations (6)
| Title |
|---|
| 半自磨机优化控制技术研究;王庆凯;杨佳伟;邹国斌;余刚;;有色冶金设计与研究;20160625(03);4-9 * |
| 半自磨机梯形衬板的磨损预测及其优化;张学东;董为民;汪滋润;孙珊珊;周海燕;;矿冶;20180601(03);61-65 * |
| 基于制粉数字孪生系统的智能磨煤机运行组合控制算法;李珂;山东电力高等专科学校学报;20231031;第26卷(第5期);48-53 * |
| 基于离散元法的半自磨机磨矿过程的仿真研究;蔡改贫;林龙飞;熊洋;王俊;;矿山机械;20141210(12);89-92 * |
| 融合数字孪生模型的滚动轴承性能退化趋势预测方法;陈靖宇等;北京航空航天大学学报;20231225;第1-14页 * |
| 高能球磨机的工艺参数及搅拌叶片尺寸对粉体粒度影响的研究与模拟;乔颖博;中国优秀硕士学位论文全文数据库工程科技I辑;20200415;B020-290 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117920446A (en) | 2024-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117920446B (en) | Semi-autogenous mill running state optimization method based on digital twin | |
| US20060293817A1 (en) | Intelligent electronically-controlled suspension system based on soft computing optimizer | |
| CN109446236B (en) | Cement particle size distribution prediction method based on random distribution | |
| CN111079906A (en) | Cement product specific surface area prediction method and system based on long-time and short-time memory network | |
| CN104375478B (en) | A kind of method and device of Rolling production process product quality on-line prediction and optimization | |
| CN113128124B (en) | Multi-grade C-Mn steel mechanical property prediction method based on improved neural network | |
| Yang et al. | Development of blast furnace burden distribution process modeling and control | |
| CN103942422B (en) | Granular-computation-based long-term prediction method for converter gas holder positions in metallurgy industry | |
| US20230080871A1 (en) | Method for estimating throat temperature of blast furnace based on multilayer ore-to-coke ratio distribution model | |
| CN111882869A (en) | Deep learning traffic flow prediction method considering adverse weather | |
| CN113591374B (en) | Cement finished product specific surface area prediction method based on convolution simple circulation unit network | |
| CN111444942A (en) | Intelligent forecasting method and system for silicon content of blast furnace molten iron | |
| CN113761787A (en) | Blast furnace molten iron silicon content online prediction method and system based on deep migration network | |
| CN113591240A (en) | Gear grinding machine thermal error model modeling method based on bidirectional LSTM network | |
| CN113517037A (en) | Method and system for predicting sintering ore FeO by fusing data and knowledge | |
| CN117299335A (en) | Ball mill ore discharge granularity control method and system based on digital twin technology | |
| Ganesan et al. | Multiobjective optimization of green sand mould system using chaotic differential evolution | |
| Wang et al. | RBF neural network soft-sensor modeling of rotary kiln pellet quality indices optimized by biogeography-based optimization algorithm | |
| CN112036090B (en) | Linear hydrodynamic polishing waviness prediction optimization method | |
| CN115774931A (en) | Method for predicting and classifying particle flow in rotary drum based on data driving model | |
| CN113406957B (en) | Mobile robot autonomous navigation method based on immune deep reinforcement learning | |
| CN114290335A (en) | Robot track planning method | |
| Yang et al. | Process optimization and quality prediction of laser aided additive manufacturing SS 420 based on RSM and WOA-Bi-LSTM | |
| CN114813347A (en) | Method for predicting mechanical property of ultrathin niobium strip | |
| Sun et al. | Reconstruction and classification of 3D burden surfaces based on two model drived data fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |