Semi-autogenous mill load state prediction method and system based on fuzzy neural network

Abstract

The invention discloses a semi-autogenous mill load state prediction method and system based on a fuzzy neural network, and the method comprises the steps: obtaining first data corresponding to a to-be-predicted load state semi-autogenous mill, and preprocessing the first data; second data are generated and obtained, a corresponding fuzzy logic rule base is established based on the second data, and meanwhile the mapping relation between the monitoring data of the semi-autogenous mill and the load state is established; according to the fuzzy logic rule base and based on a neural network algorithm, creating a corresponding fuzzy neural network model; based on the fuzzy neural network model and in combination with the first data, third data corresponding to the semi-autogenous mill is generated; wherein the third data is the prediction data of the load state of the semi-autogenous mill, and according to the corresponding system, accurate identification of the load of the semi-autogenous mill is achieved through a fuzzy neural network algorithm, so that early warning is conducted in time, operation parameters are adjusted in time, and on the premise that safe and stable operation of equipment is guaranteed, the ore grinding efficiency can be improved.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent monitoring and control of mining machinery and equipment, and in particular relates to a method and system for predicting the load state of a semi-autogenous grinding mill based on a fuzzy neural network. Background Art

[0002] At present, SAG mills are widely used in mining and metallurgical industries. They are the core equipment in the mineral processing production process. Affected by different ore properties and complex working conditions, the load state of SAG mills will change constantly. If the load is too high, it will lead to increased equipment loss, increased energy consumption, and even production accidents; if the load is too low and the mill filling rate is too low, it will lead to low grinding efficiency and reduced production capacity. Therefore, accurate identification of load status has an important impact on equipment operation efficiency and mineral processing production capacity.

[0003] Traditional SAG mill load monitoring methods mostly rely on physical sensors and simple signal processing algorithms, which are difficult to cope with complex operating conditions and nonlinear characteristics. Fuzzy neural network combines the advantages of fuzzy logic and neural network, accurately describes the fuzzy relationship between SAG mill monitoring data and load status through fuzzy logic, and extracts and infers hidden knowledge information through the powerful learning and knowledge acquisition ability of neural network. It has unique advantages in identifying SAG mill load status. Existing applications based on fuzzy neural network are mainly concentrated in the fields of image recognition and pattern classification, and have not yet been effectively applied in SAG mill load warning.

[0004] Therefore, in view of the above technical problems and defects, it is urgent to design and develop a method and system for predicting the load state of a semi-autogenous mill based on a fuzzy neural network. Summary of the invention

[0005] In order to overcome the shortcomings and difficulties of the above-mentioned prior art, the purpose of the present invention is to provide a method and system for predicting the load state of a semi-autogenous mill based on a fuzzy neural network, which utilizes a fuzzy neural network algorithm to accurately identify the load of the semi-autogenous mill, thereby providing timely warning and adjusting operating parameters, while ensuring the safe and stable operation of the equipment, and also improving the grinding efficiency.

[0006] The first object of the present invention is to provide a method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network; the second object of the present invention is to provide a system for predicting the load state of a semi-autogenous mill based on a fuzzy neural network.

[0007] The first object of the present invention is achieved in that the method comprises the following steps: Acquire first data corresponding to a semi-autogenous mill of a load state to be predicted in real time, and pre-process the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; Generate and obtain second data, and establish a corresponding fuzzy logic rule base based on the second data, and at the same time construct a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; According to the fuzzy logic rule base and based on the neural network algorithm, a corresponding fuzzy neural network model is created; Based on the fuzzy neural network model and in combination with the first data, third data corresponding to the semi-autogenous mill is generated; wherein the third data is the load state prediction data of the semi-autogenous mill.

[0008] Furthermore, the real-time acquisition of first data corresponding to the SAG mill with a load state to be predicted and preprocessing the first data further includes: In combination with a Kalman filter algorithm, the first data is denoised, and the first data is normalized based on a Z-score normalization method; Generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: (1) In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

[0009] Further, the generating and acquiring of the second data, establishing a corresponding fuzzy logic rule base based on the second data, and constructing a mapping relationship between the semi-autogenous mill monitoring data and the load state also includes: Based on the fuzzy logic rule base, the first data is divided into three corresponding thresholds: medium and low; According to the second data, a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill is constructed; wherein the calculation formula is as follows: (2) In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule.

[0010] Furthermore, the creating a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm also includes: Based on the neural network algorithm, the fuzzy rules in the fuzzy logic rule base are fused and processed, and a corresponding fuzzy neural network model is created; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer.

[0011] Furthermore, the creating a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm also includes: The input quantity in the fuzzy neural network model is fuzzified using the Gaussian membership function, and the input quantity is mapped to the corresponding fuzzy set interval; wherein the calculation formula is as follows: (3) In the formula, C i and are the center point and width of the membership function respectively, and the output is defined as ; According to the fuzzy logic rule base, node output data corresponding to the rule layer is generated.

[0012] Further, the generating of third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data also includes: According to the third data and based on the rule control threshold data, fourth data corresponding to the semi-autogenous mill is generated; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; Based on the fourth data, and in combination with the programmable processing logic controller, correct and adjust the operating parameter data corresponding to the semi-autogenous mill in real time; Generate and obtain fifth data corresponding to the semi-autogenous mill load state to be predicted, and generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

[0013] The second object of the present invention is achieved in that the system is applied to the semi-autogenous mill load state prediction method based on fuzzy neural network, and the system comprises: A data acquisition processing unit, used for acquiring first data corresponding to the semi-autogenous mill of the load state to be predicted in real time, and preprocessing the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; A data generation and construction unit is used to generate and obtain second data, and establish a corresponding fuzzy logic rule base based on the second data, and at the same time construct a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; A model creation unit, used to create a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm; A prediction data generating unit is used to generate third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data; wherein the third data is the prediction data of the load state of the semi-autogenous mill.

[0014] Furthermore, the data acquisition processing unit further includes: A first processing module, used to reduce noise of the first data in combination with a Kalman filter algorithm, and normalize the first data based on a Z-score normalization method; The first generating module is used to generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: (1) In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

[0015] Furthermore, the data generation construction unit further includes: A second generating module is used to classify the first data into three corresponding thresholds, namely, high, medium and low, based on the fuzzy logic rule base; The first construction module is used to construct a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill according to the second data; wherein the calculation formula is as follows: (2) In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule; And / or, the model creation unit further includes: A second processing module is used to fuse and process the fuzzy rules in the fuzzy logic rule base based on the neural network algorithm, and create a corresponding fuzzy neural network model; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer; The data mapping module is used to fuzzify the input quantity in the fuzzy neural network model using Gaussian membership function and map the input quantity to the corresponding fuzzy set interval; wherein the calculation formula is as follows: (3) In the formula, C i and are the center point and width of the membership function respectively, and the output is defined as ; The third generating module is used to generate node output data corresponding to the rule layer according to the fuzzy logic rule base.

[0016] Furthermore, the prediction data generating unit further includes: a fourth generating module, configured to generate fourth data corresponding to the semi-autogenous mill according to the third data and based on the rule-controlled threshold data; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; A correction and adjustment module, for correcting and adjusting the operating parameter data corresponding to the semi-autogenous mill in real time based on the fourth data and in combination with a programmable processing logic controller; The fifth generating module is used to generate and obtain fifth data corresponding to the semi-autogenous mill whose load state is to be predicted, and to generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

[0017] The present invention uses a method to obtain first data corresponding to a semi-autogenous mill with a load state to be predicted in real time, and pre-process the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; second data is generated and obtained, and a corresponding fuzzy logic rule base is established based on the second data, and a mapping relationship between the semi-autogenous mill monitoring data and the load state is constructed at the same time; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; according to the fuzzy logic rule base and based on a neural network algorithm, a corresponding fuzzy neural network model is created; based on the fuzzy neural network model and in combination with the first data, third data corresponding to the semi-autogenous mill is generated; wherein the third data is the semi-autogenous mill load state prediction data, and a system corresponding to the method uses a fuzzy neural network algorithm to realize accurate identification of the semi-autogenous mill load, thereby timely warning and adjusting operating parameters, and on the premise of ensuring safe and stable operation of the equipment, the grinding efficiency can also be improved.

[0018] That is to say, the scheme of the present invention integrates the advantages of fuzzy logic and neural network, extracts the fuzzy relationship between the monitoring data of the semi-autogenous mill and the load state, improves the recognition accuracy of the load state of the semi-autogenous mill, overcomes the problem of weak adaptability of the original method to the working conditions, effectively improves the operating stability and operating efficiency of the mill, and provides reliable technical support for the intelligent control of the mine beneficiation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 It is a schematic diagram of the steps of a semi-autogenous mill load early warning method according to an embodiment of a method for predicting a semi-autogenous mill load state based on a fuzzy neural network; Figure 2 A schematic diagram of a waveform of a basic data set of a semi-autogenous mill of a method for predicting a load state of a semi-autogenous mill based on a fuzzy neural network according to the present invention; Figure 3 A schematic diagram of a fuzzy neural network structure for identifying a load state of a semi-autogenous mill of a fuzzy neural network-based semi-autogenous mill load state prediction method of the present invention; Figure 4 A schematic diagram of a Gaussian membership function of a method for predicting a load state of a semi-autogenous mill based on a fuzzy neural network according to the present invention; Figure 5 It is a schematic diagram of a control signal transmission framework of a method for predicting a load state of a semi-autogenous mill based on a fuzzy neural network according to the present invention; Figure 6 A schematic diagram showing the accuracy of semi-autogenous mill load state recognition and early warning results of a semi-autogenous mill load state prediction method based on a fuzzy neural network according to the present invention; Figure 7 A schematic flow chart of a method for predicting a load state of a semi-autogenous mill based on a fuzzy neural network according to the present invention; Figure 8 The present invention is a schematic diagram of a system architecture for predicting the load state of a semi-autogenous mill based on a fuzzy neural network. DETAILED DESCRIPTION

[0021] In order to better understand the purpose, technical solutions and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0022] The present invention may also be implemented or applied through other different specific examples, and the details in this specification may also be modified and changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0023] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0024] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. Secondly, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0025] The present invention will be further described in detail below in conjunction with the accompanying drawings. Figure 1-Figure 7 As shown, the present invention provides a method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network, wherein the method comprises the following steps: S01, acquiring first data corresponding to a semi-autogenous mill with a load state to be predicted in real time, and preprocessing the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; S02, generating and acquiring second data, and establishing a corresponding fuzzy logic rule base based on the second data, and constructing a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; S03. Creating a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm; S04. Based on the fuzzy neural network model and in combination with the first data, generate third data corresponding to the semi-autogenous mill; wherein the third data is the semi-autogenous mill load state prediction data.

[0026] The method of acquiring first data corresponding to the semi-autogenous mill with a load state to be predicted in real time and preprocessing the first data further includes: S011. De-noise the first data by using a Kalman filter algorithm, and normalize the first data based on a Z-score normalization method; S012. Generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: (1) In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

[0027] The generating and acquiring of the second data, establishing a corresponding fuzzy logic rule base based on the second data, and constructing a mapping relationship between the semi-autogenous mill monitoring data and the load state, further includes: S021. Based on the fuzzy logic rule base, the first data is divided into three corresponding thresholds: medium and low; S022. According to the second data, a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill is constructed; wherein the calculation formula is as follows: (2) In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule.

[0028] The method of creating a corresponding fuzzy neural network model based on the fuzzy logic rule base and the neural network algorithm also includes: S031. Based on the neural network algorithm, the fuzzy rules in the fuzzy logic rule base are integrated and processed, and a corresponding fuzzy neural network model is created; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer.

[0029] The method of creating a corresponding fuzzy neural network model based on the fuzzy logic rule base and the neural network algorithm also includes: S032, using Gaussian membership function to fuzzify the input in the fuzzy neural network model, and mapping the input to the corresponding fuzzy set interval; wherein the calculation formula is as follows: (3) In the formula, Ci and are the center point and width of the membership function respectively, and the output is defined as ; S033. Generate node output data corresponding to the rule layer according to the fuzzy logic rule base.

[0030] The method of generating third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data further includes: S041, generating fourth data corresponding to the semi-autogenous mill according to the third data and based on the rule control threshold data; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; S042, based on the fourth data and in combination with the programmable processing logic controller, correct and adjust the operating parameter data corresponding to the semi-autogenous mill in real time; S043. Generate and obtain fifth data corresponding to the semi-autogenous mill load state to be predicted, and generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

[0031] Specifically, in an embodiment of the present invention, the present invention provides a semi-autogenous mill load warning method based on fuzzy neural network, which integrates the respective advantages of fuzzy logic and neural network, extracts the fuzzy relationship between the semi-autogenous mill monitoring data and the load status, automatically identifies the semi-autogenous mill load status, and solves the above-mentioned technical defects.

[0032] The method comprises the following steps: S1: pre-processing the detection data of the semi-autogenous mill sensor based on the Kalman filter algorithm, completing data noise reduction and normalization, and forming a basic data set; S2: establishing a fuzzy logic rule base in combination with expert experience and knowledge, and constructing a mapping relationship between the semi-autogenous mill monitoring data and the load state; S3: integrating the fuzzy rules and the neural network algorithm to establish a fuzzy neural network model, and identifying the load state of the semi-autogenous mill; S4: combining a programmable logic controller and a human-computer interaction interface to realize the load early warning and operation parameter optimization of the semi-autogenous mill.

[0033] The preprocessing of the SAG mill sensor detection data in step S1 also includes the steps of: the SAG mill sensor detection data includes grinding sound, oil pressure and power, using the Kalman filter algorithm to reduce data noise, and completing data normalization based on the Z-score normalization method to eliminate data dimension differences, and obtaining a basic data set, as shown in the following formula: (1) In the formula, i=1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

[0034] The fuzzy logic rule base of the semi-autogenous mill in step S2 divides the grinding sound, oil pressure and power into three thresholds of high, medium and low. According to the experience and knowledge of on-site mineral processing experts, a fuzzy logic rule base R is constructed for the three operating states of the semi-autogenous mill, namely underload, normal load and overload. j , described by if-then rules, as shown below: (2) In the formula, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule.

[0035] The fuzzy neural network model for identifying the load state of the semi-autogenous mill in step S3 includes a four-layer network structure of an input layer, a fuzzification layer, a rule layer and an output layer: S1.1: The first layer is the input layer, which selects the preprocessed sensor data As the neural network input, the number of nodes is the number of input variables, and the output is ; S1.2: The second layer is the fuzzification layer, which uses Gaussian membership function for fuzzification processing and maps the input quantity to the fuzzy set interval, as shown in the following formula: (3) In the formula, C i and are the center point and width of the membership function respectively, and the output is defined as .

[0036] S1.3: The third layer is the rule layer, which is used to implement fuzzy reasoning. The output of the node represents the matching degree between the input sample and the rule. Combined with the fuzzy logic rule base, the node output of this layer is obtained as shown in the following formula: (4) S1.4: The fourth layer is the output layer, which is used to complete the center average defuzzification operation and obtain numerical accurate output, as shown in the following formula: (5) In the formula, ω j is the link weight between the rule layer and the output layer, and m is the number of rules in the output layer.

[0037] The output result of the fuzzy neural network model for identifying the load state of the semi-autogenous mill is combined with the fuzzy logic rule base to determine that if it is a normal state, the process ends; if it is an underload or overload state, the process enters the early warning and decision-making stage.

[0038] The SAG mill load early warning and operation parameter optimization in step S4 further include the following steps: S1.1: Obtain the current abnormal state (underload or overload) of the SAG mill through the human-computer interaction interface; S1.2: Generate corresponding control signals according to pre-set expert control rules; S1.3: The control signal is passed through the programmable logic controller to adjust the grinding operation parameters such as feed rate and water supply.

[0039] S1.4: Re-collect the sensor data of the SAG mill after parameter adjustment to determine the load status.

[0040] The fuzzy neural network adopts the error back propagation algorithm to correct the weights, and its error correction formula is: (6) In the formula, z i is the expected output value.

[0041] The network link weight ω j Correction in the negative direction, the expression is as follows: (7) In the formula, η is the learning factor. Through continuous training, the network parameters are optimized and the model accuracy is improved.

[0042] That is to say, Figure 1 As shown, the present invention provides a load early warning method for a semi-autogenous mill based on a fuzzy neural network, comprising the following steps: S1: Preprocess the SAG mill sensor detection data based on the Kalman filter algorithm to complete data noise reduction and normalization to form a basic data set; S2: Establish a fuzzy logic rule base based on expert experience and knowledge to construct a mapping relationship between the SAG mill monitoring data and the load state; S3: Establish a fuzzy neural network model by integrating fuzzy rules and neural network algorithms to identify the SAG mill load state; S4: Combine the programmable logic controller and the human-computer interaction interface to realize the SAG mill load warning and operation parameter optimization.

[0043] The preprocessing of the SAG mill sensor detection data in step S1 further includes the steps of: the SAG mill sensor detection data includes grinding sound , oil pressure and power , the Kalman filter algorithm is used to reduce data noise, and the data is normalized based on the Z-score normalization method to eliminate the data dimension difference and obtain the basic data set, such as Figure 2 As shown, the specific expression is: (1) The fuzzy logic rule base of the semi-autogenous mill in step S2 divides grinding noise, oil pressure and power into three thresholds of high, medium and low. The relationship of signal status bits is shown in Table 1.

[0044] Table 1 Signal status bit relationship

[0045] According to the experience and knowledge of on-site mineral processing experts, a fuzzy logic rule base R is constructed for the three operating states of the semi-autogenous mill: underload, normal load and overload. j , described by if-then rules, the fuzzy rules are as follows:

[0046] The fuzzy neural network model for identifying the load state of the semi-autogenous mill in step S3 has a network structure as follows: Figure 3 As shown, the network structure includes 4 layers: input layer, fuzzification layer, rule layer and output layer: S1.1: The first layer is the input layer, which selects the preprocessed sensor data As the neural network input, the number of nodes is the number of input variables, and the output is ; S1.2: The second layer is the fuzzification layer, which uses Gaussian membership function for fuzzification processing, such as Figure 4 As shown, the input Mapped to the fuzzy set interval, as shown below: (3) Take C i =0.5,σ i = 0.8, the output is defined as .

[0047] S1.3: The third layer is the rule layer, which is used to realize fuzzy reasoning. The output of the node represents the matching degree between the input sample and the rule. Combined with the fuzzy logic rule base, the three operating states of the semi-autogenous mill, underload, normal load and overload, are represented by numbers 1, 2 and 3, and the relationship between the input and output signals is obtained. The fuzzy rule base of the load warning is shown in Table 2: Table 2 SAG mill load warning fuzzy rule base

[0048] S1.4: The fourth layer is the output layer, which is used to complete the center average defuzzification operation and obtain numerical accurate output, as shown in the following formula: (5) In the formula, ω j is the link weight between the rule layer and the output layer, and m=27 is the number of rules in the output layer.

[0049] The output result of the fuzzy neural network model for identifying the load state of the semi-autogenous mill is combined with the fuzzy logic rule base to determine that if it is a normal state, the process ends; if it is an underload or overload state, the process enters the early warning and decision-making stage.

[0050] The control signal transmission framework of the SAG mill load warning and operation parameter optimization in step S4 is as follows: Figure 5 As shown, the steps include: S1.1: Obtain the current abnormal state (underload or overload) of the SAG mill through the human-computer interaction interface; S1.2: Generate corresponding data control signals according to pre-set expert control rules; S1.3: The data control signal is passed through the programmable logic controller to adjust the grinding operation parameters such as feed rate and water supply.

[0051] S1.4: Re-collect the sensor data of the SAG mill after parameter adjustment to determine the load status.

[0052] The fuzzy neural network adopts the error back propagation algorithm to correct the weights, and its error correction formula is: (6) In the formula, z i is the expected output value.

[0053] The network link weight ω j Correction in the negative direction, the expression is as follows: (7) In the formula, η =0.95 is the learning factor. Through continuous training, the network parameters are optimized and the model accuracy is improved.

[0054] The fuzzy neural network training was completed by selecting the historical operation data of the semi-autogenous mill of a mineral processing plant for three months and deploying it in the industrial site. The statistical analysis results of the load identification and early warning accuracy of the semi-autogenous mill are as follows: Figure 6It can be seen that the SAG mill load early warning method based on fuzzy neural network can realize the SAG mill load early warning and ensure the safe, efficient and stable operation of the equipment.

[0055] To achieve the above object, the present invention also provides a semi-autogenous mill load state prediction system based on fuzzy neural network, such as Figure 8 As shown, the system is applied to the SAG mill load state prediction method based on fuzzy neural network, and the system includes: A data acquisition processing unit, used for acquiring first data corresponding to the semi-autogenous mill of the load state to be predicted in real time, and preprocessing the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; A data generation and construction unit is used to generate and obtain second data, and establish a corresponding fuzzy logic rule base based on the second data, and at the same time construct a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; A model creation unit, used to create a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm; A prediction data generating unit is used to generate third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data; wherein the third data is the prediction data of the load state of the semi-autogenous mill.

[0056] The data acquisition processing unit further includes: A first processing module, used to reduce noise of the first data in combination with a Kalman filter algorithm, and normalize the first data based on a Z-score normalization method; The first generating module is used to generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: (1) In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

[0057] The data generation construction unit further includes: A second generating module is used to classify the first data into three corresponding thresholds, namely, high, medium and low, based on the fuzzy logic rule base; The first construction module is used to construct a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill according to the second data; wherein the calculation formula is as follows: (2) In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule; And / or, the model creation unit further includes: A second processing module is used to fuse and process the fuzzy rules in the fuzzy logic rule base based on the neural network algorithm, and create a corresponding fuzzy neural network model; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer; The data mapping module is used to fuzzify the input quantity in the fuzzy neural network model using Gaussian membership function and map the input quantity to the corresponding fuzzy set interval; wherein the calculation formula is as follows: (3) In the formula, C i and σ i are the center point and width of the membership function respectively, and the output is defined as ; The third generating module is used to generate node output data corresponding to the rule layer according to the fuzzy logic rule base.

[0058] The prediction data generating unit further includes: a fourth generating module, configured to generate fourth data corresponding to the semi-autogenous mill according to the third data and based on the rule-controlled threshold data; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; A correction and adjustment module, for correcting and adjusting the operating parameter data corresponding to the semi-autogenous mill in real time based on the fourth data and in combination with a programmable processing logic controller; The fifth generating module is used to generate and obtain fifth data corresponding to the semi-autogenous mill whose load state is to be predicted, and to generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

[0059] In the system solution embodiment of the present invention, the method steps involved in the prediction of the load state of a semi-autogenous mill based on a fuzzy neural network have been described above in detail. That is to say, the functional modules in the system are used to implement the steps or sub-steps in the above method embodiment, which will not be repeated here.

[0060] The present invention uses a method to obtain first data corresponding to a semi-autogenous mill with a load state to be predicted in real time, and pre-process the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; second data is generated and obtained, and a corresponding fuzzy logic rule base is established based on the second data, and a mapping relationship between the semi-autogenous mill monitoring data and the load state is constructed at the same time; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; according to the fuzzy logic rule base and based on a neural network algorithm, a corresponding fuzzy neural network model is created; based on the fuzzy neural network model and in combination with the first data, third data corresponding to the semi-autogenous mill is generated; wherein the third data is the semi-autogenous mill load state prediction data, and a system corresponding to the method uses a fuzzy neural network algorithm to realize accurate identification of the semi-autogenous mill load, thereby timely warning and adjusting operating parameters, and on the premise of ensuring safe and stable operation of the equipment, the grinding efficiency can also be improved.

[0061] That is to say, the scheme of the present invention integrates the advantages of fuzzy logic and neural network, extracts the fuzzy relationship between the monitoring data of the semi-autogenous mill and the load state, improves the recognition accuracy of the load state of the semi-autogenous mill, overcomes the problem of weak adaptability of the original method to the working conditions, effectively improves the operating stability and operating efficiency of the mill, and provides reliable technical support for the intelligent control of the mine beneficiation process.

[0062] In other words, the present invention provides a method for early warning of semi-autogenous mill load based on fuzzy neural network. First, the sensor detection data of the semi-autogenous mill is preprocessed based on the Kalman filter algorithm to form a basic data set; then, a fuzzy logic rule base is established in combination with expert experience and knowledge to construct a mapping relationship between the monitoring data of the semi-autogenous mill and the load state; secondly, a fuzzy neural network model is established by integrating fuzzy rules and neural network algorithms to identify the load state of the semi-autogenous mill; finally, a programmable logic controller and a human-computer interaction interface are combined to realize early warning of semi-autogenous mill load and optimization of operating parameters. The present invention utilizes a fuzzy neural network algorithm to extract the fuzzy relationship between the monitoring data of the semi-autogenous mill and the load state, realize accurate identification and timely early warning of the semi-autogenous mill load, effectively avoid low load or overload operation of the equipment, provide decision support for the optimization control of the semi-autogenous mill, and improve the intelligent level of mill operation and maintenance in the mining and metallurgical industries.

[0063] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network, characterized in that: The method comprises the steps of: Acquire first data corresponding to a semi-autogenous mill of a load state to be predicted in real time, and pre-process the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; Generate and obtain second data, and establish a corresponding fuzzy logic rule base based on the second data, and at the same time construct a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; According to the fuzzy logic rule base and based on the neural network algorithm, a corresponding fuzzy neural network model is created; Based on the fuzzy neural network model and in combination with the first data, third data corresponding to the semi-autogenous mill is generated; wherein the third data is the load state prediction data of the semi-autogenous mill.

2. A method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network according to claim 1, characterized in that: The method of acquiring first data corresponding to the semi-autogenous mill with a load state to be predicted in real time and preprocessing the first data further includes: In combination with a Kalman filter algorithm, the first data is denoised, and the first data is normalized based on a Z-score normalization method; Generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: （1） In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

3. The method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network according to claim 1, characterized in that: The generating and acquiring of the second data, establishing a corresponding fuzzy logic rule base based on the second data, and constructing a mapping relationship between the semi-autogenous mill monitoring data and the load state, further includes: Based on the fuzzy logic rule base, the first data is divided into three corresponding thresholds: medium and low; According to the second data, a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill is constructed; wherein the calculation formula is as follows: （2） In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule.

4. The method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network according to claim 1, characterized in that: The method of creating a corresponding fuzzy neural network model based on the fuzzy logic rule base and the neural network algorithm also includes: Based on the neural network algorithm, the fuzzy rules in the fuzzy logic rule base are fused and processed, and a corresponding fuzzy neural network model is created; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer.

5. A method for predicting a semi-autogenous mill load state based on a fuzzy neural network according to claim 1 or 4, characterized in that: The method of creating a corresponding fuzzy neural network model based on the fuzzy logic rule base and the neural network algorithm also includes: The input quantity in the fuzzy neural network model is fuzzified using the Gaussian membership function, and the input quantity is mapped to the corresponding fuzzy set interval; wherein the calculation formula is as follows: （3） In the formula, C i and are the center point and width of the membership function respectively, and the output is defined as ; According to the fuzzy logic rule base, node output data corresponding to the rule layer is generated.

6. The method for predicting the load state of a semi-autogenous mill based on a fuzzy neural network according to claim 1, characterized in that: The method of generating third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data further includes: According to the third data and based on the rule control threshold data, fourth data corresponding to the semi-autogenous mill is generated; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; Based on the fourth data, and in combination with the programmable processing logic controller, correct and adjust the operating parameter data corresponding to the semi-autogenous mill in real time; Generate and obtain fifth data corresponding to the semi-autogenous mill load state to be predicted, and generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

7. A semi-autogenous mill load state prediction system based on fuzzy neural network, characterized in that: The system is applied to the semi-autogenous mill load state prediction method based on fuzzy neural network as claimed in any one of claims 1 to 6, and the system comprises: A data acquisition processing unit, used for acquiring first data corresponding to the semi-autogenous mill of the load state to be predicted in real time, and preprocessing the first data; wherein the first data is sensor detection data corresponding to the semi-autogenous mill; A data generation and construction unit is used to generate and obtain second data, and establish a corresponding fuzzy logic rule base based on the second data, and at the same time construct a mapping relationship between the semi-autogenous mill monitoring data and the load state; wherein the second data is the existing basic empirical data corresponding to the semi-autogenous mill; A model creation unit, used to create a corresponding fuzzy neural network model according to the fuzzy logic rule base and based on a neural network algorithm; A prediction data generating unit is used to generate third data corresponding to the semi-autogenous mill based on the fuzzy neural network model and in combination with the first data; wherein the third data is the prediction data of the load state of the semi-autogenous mill.

8. A fuzzy neural network based semi-autogenous mill load state prediction system according to claim 7, characterized in that: The data acquisition processing unit further includes: A first processing module, used to reduce noise of the first data in combination with a Kalman filter algorithm, and normalize the first data based on a Z-score normalization method; The first generating module is used to generate a basic data set corresponding to the first data; wherein the calculation formula is as follows: （1） In the formula, i =1,2,3 are the number of sensors, N represents the sampling time, express i Sensor in t The sampling value at the time.

9. A fuzzy neural network based semi-autogenous mill load state prediction system according to claim 7, characterized in that: The data generation construction unit further includes: A second generating module is used to classify the first data into three corresponding thresholds, namely, high, medium and low, based on the fuzzy logic rule base; The first construction module is used to construct a fuzzy logic rule base of three operating states of underload, normal load and overload corresponding to the second data and the semi-autogenous mill according to the second data; wherein the calculation formula is as follows: （2） In the formula, R j is the fuzzy logic rule base, For the Rule No. The fuzzy set of inputs, β j is the fuzzy single value of the conclusion of the jth rule; And / or, the model creation unit further includes: A second processing module is used to fuse and process the fuzzy rules in the fuzzy logic rule base based on the neural network algorithm, and create a corresponding fuzzy neural network model; wherein the fuzzy neural network model includes an input layer, a fuzzification layer, a rule layer and an output layer; The data mapping module is used to fuzzify the input quantity in the fuzzy neural network model using Gaussian membership function and map the input quantity to the corresponding fuzzy set interval; wherein the calculation formula is as follows: （3） In the formula, C i and are the center point and width of the membership function respectively, and the output is defined as ; The third generating module is used to generate node output data corresponding to the rule layer according to the fuzzy logic rule base.

10. A fuzzy neural network based semi-autogenous mill load state prediction system according to claim 6 or 7, characterized in that: The prediction data generating unit further includes: a fourth generating module, configured to generate fourth data corresponding to the semi-autogenous mill according to the third data and based on the rule-controlled threshold data; wherein the fourth data is control signal data corresponding to the semi-autogenous mill; A correction and adjustment module, for correcting and adjusting the operating parameter data corresponding to the semi-autogenous mill in real time based on the fourth data and in combination with a programmable processing logic controller; The fifth generating module is used to generate and obtain fifth data corresponding to the semi-autogenous mill whose load state is to be predicted, and to generate corresponding sixth data based on the fifth data; wherein the fifth data is the semi-autogenous mill sensor detection data after re-collection of parameter adjustment; and the sixth data is the corresponding semi-autogenous mill load state prediction data generated based on the fifth data.

Images