Fuzzy logic-based smelting furnace burden proportioning optimization method and system

Abstract

The invention relates to a smelting furnace burden proportioning optimization method and system based on fuzzy logic. The method comprises the following steps: inputting obtained furnace temperature and furnace pressure into a fuzzy logic system for fuzzification processing and one-heat coding; then, feature extraction is carried out through a time sequence correlation feature extractor based on a deep neural network model, and Hash mapping coding is carried out to obtain a Hash embedding temperature time sequence correlation feature vector and a Hash embedding pressure time sequence correlation feature vector; and carrying out interactive fusion on the Hash embedded temperature time sequence correlation feature vector and the Hash embedded pressure time sequence correlation feature vector to obtain a temperature-pressure time sequence interaction correlation feature so as to determine a recommended furnace charge ratio. In this way, rapid adaptation can be achieved when raw material components and furnace conditions change, the furnace burden proportion is optimized, and therefore the stability and continuity of the production process are kept.

Description

Technical Field

[0001] The present disclosure relates to the technical field of smelting furnace charge, and in particular, to a smelting furnace charge ratio optimization method and system based on fuzzy logic. Background Art

[0002] In the smelting industry, the optimization of the smelting charge ratio is crucial to improving smelting efficiency, reducing production costs and improving the quality of the final product. The ratio of raw material components directly affects the chemical reaction process of the charge and the quality characteristics of the final product. Therefore, by optimizing the ratio of raw material components, a more efficient smelting process can be achieved, production efficiency can be improved, energy consumption and production costs can be reduced, and the quality of the final product can be improved.

[0003] Fuzzy logic can effectively handle the uncertainty and ambiguity in the smelting process, enabling the system to make reasonable decisions in complex environments. In other words, the fuzzy logic system has strong robustness and can maintain stable performance in the face of noise interference or data fluctuations. Therefore, the smelting charge ratio optimization method based on fuzzy logic has strong adaptability. It can use the principles of fuzzy mathematics to handle the uncertainty and ambiguity in the ingredients by simulating the human decision-making process to achieve a better charge ratio.

[0004] However, the traditional fuzzy logic-based optimization method for smelting charge ratio requires the definition of a set of fuzzy rules. Specifically, it is necessary to rely on expert knowledge and experience to establish a rule base to map fuzzy input to fuzzy output. For example, "If the raw material composition is high and the temperature is low, increase the ratio of a certain raw material." However, the rule base is often static and difficult to adapt to the real-time changes in raw material composition and furnace conditions, resulting in poor optimization of smelting charge ratio. Moreover, the process of establishing a rule base needs to rely on experiments and experience, which is not only very time-consuming, but also easily affected by subjective factors, resulting in the possibility that the establishment of the rule base may be incomplete or inaccurate, limiting its scope of application in complex smelting processes.

[0005] Therefore, an optimized fuzzy logic-based smelting charge ratio optimization scheme is desired. Summary of the invention

[0006] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0007] In a first aspect, the present disclosure provides a method for optimizing smelting charge ratio based on fuzzy logic, the method comprising:

[0008] Acquire the time series of the temperature and pressure in the furnace collected by the temperature sensor and the pressure sensor;

[0009] Inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzification processing to obtain a fuzzified temperature series and a fuzzified pressure series;

[0010] One-hot encoding the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0011] The sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector are respectively subjected to feature extraction by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0012] Performing hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector;

[0013] Interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0014] Based on the temperature-pressure time series interactive correlation characteristics, the recommended charge ratio is determined.

[0015] Optionally, the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional extended convolutional layer.

[0016] Optionally, feature extraction is performed on the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector, including: processing the sequence of the temperature time series fuzzy input vector by the time series correlation feature extractor based on the one-dimensional extended convolution layer with the following feature extraction formula to obtain the temperature time series correlation feature vector; wherein the feature extraction formula is:

[0017] X 1:n =X1⊕X2⊕…⊕X n

[0018] C i =f(ω·X i:i+l-1 +b)

[0019] C=[C1,C2,...,C n-l+1]

[0020] Among them, X1,X2,...,X n represents each temperature time series fuzzy input vector in the sequence of the temperature time series fuzzy input vector, X 1:n represents the concatenated vector of the sequence of the temperature time series fuzzification input vector, ⊕ represents the connection operation, X i:i+l-1 X represents the sequence of the temperature time series fuzzy input vector i ,X i+1 ,...,X i+l-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i represents each temperature local time series association feature vector in the temperature time series association feature vector, [·,·] represents vector concatenation, and C is the temperature time series association feature vector.

[0021] Optionally, performing hash mapping encoding on the temperature timing association feature vector and the pressure timing association feature vector to obtain a hash embedded temperature timing association feature vector and a hash embedded pressure timing association feature vector, comprising: performing hash mapping encoding on the temperature timing association feature vector and the pressure timing association feature vector according to the following hash formula to obtain the hash embedded temperature timing association feature vector and the hash embedded pressure timing association feature vector; wherein the hash formula is:

[0022]

[0023] Among them, v i is one of the temperature time series associated feature vectors and the pressure time series associated feature vectors, f θ (·) represents the fully connected encoding of the feature vector, sign(·) represents the sign function, It is the corresponding feature vector in the hash embedded temperature timing associated feature vector and the hash embedded pressure timing associated feature vector.

[0024] Optionally, interactively fusing the hash-embedded temperature timing association feature vector and the hash-embedded pressure timing association feature vector to obtain a temperature-pressure timing interaction association feature, including: interactively fusing the hash-embedded temperature timing association feature vector and the hash-embedded pressure timing association feature vector using the following interactive formula to obtain a temperature-pressure timing interaction association feature vector as the temperature-pressure timing interaction association feature; wherein the interactive formula is:

[0025]

[0026] in, is the hash embedding temperature time series associated feature vector, is the hash embedding pressure time series associated feature vector, is the square of the L2 norm of the vector, L2 is a hyperparameter, v c is the temperature-pressure time series interaction correlation feature vector.

[0027] Optionally, based on the temperature-pressure time series interaction correlation feature, determining a recommended charge ratio includes: passing the temperature-pressure time series interaction correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

[0028] Optionally, passing the temperature-pressure time series interactive correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio, comprising: optimizing the temperature-pressure time series interactive correlation feature vector to obtain an optimized temperature-pressure time series interactive correlation feature vector; passing the optimized temperature-pressure time series interactive correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

[0029] In a second aspect, the present disclosure provides a smelting charge ratio optimization system based on fuzzy logic, the system comprising:

[0030] A furnace temperature and pressure acquisition module is used to acquire the time series of the furnace temperature and the furnace pressure collected by the temperature sensor and the pressure sensor;

[0031] A fuzzy processing module, used for inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzy processing to obtain a fuzzy temperature sequence and a fuzzy pressure sequence;

[0032] A one-hot encoding module, used for performing one-hot encoding on the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0033] A feature extraction module, used for performing feature extraction on the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively through a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0034] A hash mapping encoding module, used for performing hash mapping encoding on the temperature time series associated feature vector and the pressure time series associated feature vector to obtain a hash embedded temperature time series associated feature vector and a hash embedded pressure time series associated feature vector;

[0035] An interactive fusion module, used for interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0036] The recommended charge ratio determination module is used to determine the recommended charge ratio based on the temperature-pressure time series interaction correlation characteristics.

[0037] Optionally, the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional extended convolutional layer.

[0038] The above technical solution is adopted to input the obtained furnace temperature and pressure into the fuzzy logic system for fuzzy processing and unique hot encoding; then, the time series correlation feature extractor based on the deep neural network model is used for feature extraction and hash mapping encoding to obtain the hash embedded temperature time series correlation feature vector and the hash embedded pressure time series correlation feature vector; the hash embedded temperature time series correlation feature vector and the hash embedded pressure time series correlation feature vector are interactively fused to obtain the temperature-pressure time series interactive correlation feature to determine the recommended charge ratio. In this way, it is possible to quickly adapt to changes in raw material composition and furnace conditions, optimize the charge ratio, and thus maintain the stability and continuity of the production process.

[0039] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0041] Figure 1 It is a flow chart of a method for optimizing smelting charge ratio based on fuzzy logic according to an exemplary embodiment.

[0042] Figure 2 It is a block diagram of a smelting charge ratio optimization system based on fuzzy logic according to an exemplary embodiment.

[0043] Figure 3 It is a block diagram of an electronic device according to an exemplary embodiment.

[0044] Figure 4 It is an application scenario diagram of a smelting charge ratio optimization method based on fuzzy logic according to an exemplary embodiment. DETAILED DESCRIPTION

[0045] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0046] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0047] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0048] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0049] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0050] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0051] First, in response to the above technical problems, in the technical solution of this application, a smelting charge ratio optimization method based on fuzzy logic is proposed, which can monitor and collect the temperature and pressure data in the furnace in real time through temperature sensors and pressure sensors, and use the fuzzy logic system to convert these time series data into fuzzy data, and use artificial intelligence-based data processing and analysis algorithms at the back end to perform time series collaborative correlation analysis of temperature fuzzy data and pressure fuzzy data, so as to use the time series correlation pattern and change between the temperature and pressure in the furnace to more accurately recommend and adjust the charge ratio. In this way, it can quickly adapt to changes in raw material composition and furnace conditions, optimize the charge ratio, and maintain the stability and continuity of the production process.

[0052] In order to solve the above problems, the present disclosure provides a smelting furnace charge ratio optimization method and system based on fuzzy logic, which inputs the obtained furnace temperature and furnace pressure into the fuzzy logic system for fuzzification processing and unique hot encoding; then, a time series correlation feature extractor based on a deep neural network model is used for feature extraction and hash mapping encoding to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector; the hash embedded temperature time series correlation feature vector and the hash embedded pressure time series correlation feature vector are interactively fused to obtain the temperature-pressure time series interactive correlation feature to determine the recommended furnace charge ratio. In this way, it is possible to quickly adapt to changes in raw material composition and furnace conditions, optimize the furnace charge ratio, and thus maintain the stability and continuity of the production process.

[0053] The specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0054] Figure 1 is a flow chart of a smelting charge ratio optimization method based on fuzzy logic according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0055] Step 101, obtaining the time series of the temperature and pressure in the furnace collected by the temperature sensor and the pressure sensor;

[0056] Step 102: Input the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzification processing to obtain a fuzzified temperature sequence and a fuzzified pressure sequence;

[0057] Step 103: One-hot encode the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0058] Step 104: extracting features from the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively by using a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0059] Step 105: Perform hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector;

[0060] Step 106: interactively fuse the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0061] Step 107: Determine a recommended furnace charge ratio based on the temperature-pressure time series interaction correlation characteristics.

[0062] Among them, the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional extended convolutional layer.

[0063] Specifically, in the technical solution of the present application, first, the time series of the furnace temperature and the furnace pressure collected by the temperature sensor and the pressure sensor are obtained. Then, considering that the parameters such as temperature and pressure in the smelting process usually change continuously, in order to be able to optimize the smelting charge ratio better based on fuzzy logic, it is necessary to input the time series of the furnace temperature and the furnace pressure into the fuzzy logic system for fuzzification processing to obtain the sequence of fuzzified temperature and the sequence of fuzzified pressure. It should be understood that the fuzzy logic system can handle the fuzziness and uncertainty of the data. By inputting the time series data of the furnace temperature and the furnace pressure into the fuzzy logic system for fuzzification processing, the precise furnace temperature and pressure data at different time points can be converted into fuzzy language variables (such as "low", "medium", "high"), so that the system can better handle the fuzziness and uncertainty in actual situations.

[0064] Then, in order to avoid the defects and problems brought by the traditional fuzzy logic using the rule base to optimize the smelting charge ratio, in the technical solution of the present application, artificial intelligence technology and deep learning algorithm are used to perform time series correlation analysis of fuzzified charge temperature data and pressure data. Therefore, in order to convert the fuzzy language variables into a numerical form suitable for deep learning model processing, in the technical solution of the present application, the sequence of the fuzzified temperature and the sequence of the fuzzified pressure are further uniquely encoded to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors. It should be understood that uniquely encoding is a coding method that converts categorical variables into binary vectors, each category corresponds to a binary vector, in which only one element is 1 and the rest are 0. Specifically, for the fuzzified temperature and pressure sequences, each fuzzy language variable (such as "low", "medium", "high") corresponds to a binary vector, and these binary vectors are arranged in chronological order to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors.

[0065] Further, a one-dimensional convolution layer is used to analyze and extract the time series correlation features of the sequence of the temperature time series fuzzification input vector and the sequence of the pressure time series fuzzification input vector, which can help the system understand the time series variation law of the temperature and pressure in the furnace. In particular, considering that it is difficult to capture the time series correlation features of data under different step lengths due to the limitations of one-dimensional convolution, in order to extract higher-level, more abstract and more comprehensive features from the fuzzified time series data of the furnace temperature and furnace pressure, so as to better understand and utilize the time series information of the furnace temperature and furnace pressure to optimize the furnace charge ratio, in the technical solution of the present application, the sequence of the temperature time series fuzzification input vector and the sequence of the pressure time series fuzzification input vector are passed through a time series correlation feature extractor based on a one-dimensional extended convolution layer to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector. It should be understood that compared with the one-dimensional convolution layer, the one-dimensional extended convolution layer introduces a hyperparameter called the expansion rate to control the number of null values ​​0 in the convolution kernel. The expansion rate defines the spacing between each time point when the convolution kernel processes the blurred time series data of furnace pressure and furnace temperature. After the expanded convolution operation, the time series correlation characteristics of furnace parameter data at different step lengths and at a long distance can be captured, which is crucial for the optimization of furnace charge ratio.

[0066] In one embodiment of the present disclosure, a time series correlation feature extractor based on a deep neural network model is used to extract features from the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector, including: processing the sequence of the temperature time series fuzzy input vector through the time series correlation feature extractor based on the one-dimensional extended convolution layer with the following feature extraction formula to obtain the temperature time series correlation feature vector; wherein the feature extraction formula is:

[0067] X 1:n =X1⊕X2⊕…⊕X n

[0068] C i =f(ω·X i:i+l-1 +b)

[0069] C=[C1,C2,...,C n-l+1 ]

[0070] Among them, X1,X2,...,X n represents each temperature time series fuzzy input vector in the sequence of the temperature time series fuzzy input vector, X 1:n represents the concatenated vector of the sequence of the temperature time series fuzzification input vector, ⊕ represents the connection operation, X i:i+l-1X represents the sequence of the temperature time series fuzzy input vector i ,X i+1 ,...,X i+l-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i represents each temperature local time series association feature vector in the temperature time series association feature vector, [·,·] represents vector concatenation, and C is the temperature time series association feature vector.

[0071] It should be understood that since the temperature time series correlation feature vector and the pressure time series correlation feature vector respectively contain the time series correlation features of the furnace temperature and furnace pressure at different step sizes after fuzzification, and since the temperature data and the pressure data have different data distribution ranges and units, this will produce feature differences, thereby affecting the accuracy of the time series correlation analysis of the furnace temperature and pressure. Based on this, in order to establish a connection between the high-dimensional time series correlation feature vectors of temperature and pressure, so as to capture the correlation and mutual influence between the temperature and pressure in the furnace, in the technical solution of the present application, it is necessary to perform hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector. Through hash mapping encoding, the temperature time series correlation feature vector and the pressure time series correlation feature vector can be mapped to a common low-dimensional hash embedding space, while retaining the correlation features between the time series features of the furnace temperature and the furnace pressure data, providing a compact and information-rich feature representation for the optimization of the charge ratio.

[0072] In one embodiment of the present disclosure, the temperature timing association feature vector and the pressure timing association feature vector are hash-mapped to obtain a hash-embedded temperature timing association feature vector and a hash-embedded pressure timing association feature vector, including: performing hash-map encoding on the temperature timing association feature vector and the pressure timing association feature vector using the following hash formula to obtain the hash-embedded temperature timing association feature vector and the hash-embedded pressure timing association feature vector; wherein the hash formula is:

[0073]

[0074] Among them, v i is one of the temperature time series associated feature vectors and the pressure time series associated feature vectors, f θ (·) represents the fully connected encoding of the feature vector, sign(·) represents the sign function, It is the corresponding feature vector in the hash embedded temperature timing associated feature vector and the hash embedded pressure timing associated feature vector.

[0075] During the smelting process, there is a complex interactive relationship between temperature and pressure. For example, when the temperature rises, the pressure may also rise, which indicates that the raw material ratio needs to be adjusted to control the furnace conditions. The hash embedded temperature time series association feature vector and the hash embedded pressure time series association feature vector after hash space embedding mapping respectively reflect the time series pattern and change trend of the furnace temperature and furnace pressure after fuzzification. There is interactive correlation feature information between the furnace temperature and the furnace pressure, which is crucial for the optimization of the charge ratio. Based on this, in the technical solution of the present application, the hash embedded temperature time series association feature vector and the hash embedded pressure time series association feature vector are further interactively fused to obtain the temperature-pressure time series interactive correlation feature vector. By interactively fusion, the hash embedded temperature time series association feature vector and the hash embedded pressure time series association feature vector can capture the complex correlation relationship and time series characteristics between the furnace temperature and pressure, thereby improving the accuracy of the charge ratio optimization.

[0076] In one embodiment of the present disclosure, the Hash embedded temperature timing correlation feature vector and the Hash embedded pressure timing correlation feature vector are interactively fused to obtain a temperature-pressure timing interaction correlation feature, including: interactively fusion of the Hash embedded temperature timing correlation feature vector and the Hash embedded pressure timing correlation feature vector using the following interactive formula to obtain a temperature-pressure timing interaction correlation feature vector as the temperature-pressure timing interaction correlation feature; wherein the interactive formula is:

[0077]

[0078] in, is the hash embedding temperature time series associated feature vector, is the hash embedding pressure time series associated feature vector, is the square of the L2 norm of the vector, L2 is a hyperparameter, v c is the temperature-pressure time series interaction correlation feature vector.

[0079] Then, the temperature-pressure time series interactive correlation feature vector is passed through a decoder-based smelting furnace charge ratio recommender to obtain a recommended furnace charge ratio. That is to say, the time series interactive correlation feature between the furnace temperature and the furnace pressure is used to perform decoding regression of the smelting furnace charge ratio, so as to perform recommendation control of the furnace charge ratio. It should be understood that in the optimization of the smelting furnace charge ratio, the raw material composition ratio is an important variable, and the content and composition of different raw materials will directly affect the quality and performance of the charge. By optimizing the ratio of raw material components, the smelting efficiency can be improved, the production cost can be reduced, and the quality of the final product can be improved. For example, in a smelting furnace, a variety of raw materials need to be input for smelting, and these raw materials include metal ores, fuels, limestone, etc., and each raw material has different chemical compositions, such as metal content, ash content, sulfur content, etc. By optimizing the ratio of raw material components, the input ratio of different raw materials can be adjusted according to the requirements and production goals of the smelting furnace to achieve the goals of improving smelting efficiency, reducing production costs and improving product quality.

[0080] In one embodiment of the present disclosure, based on the temperature-pressure time series interaction correlation feature, a recommended charge ratio is determined, including: passing the temperature-pressure time series interaction correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

[0081] In one embodiment of the present disclosure, the temperature-pressure timing interaction correlation feature vector is passed through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio, including: optimizing the temperature-pressure timing interaction correlation feature vector to obtain an optimized temperature-pressure timing interaction correlation feature vector; passing the optimized temperature-pressure timing interaction correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

[0082] In the above technical scheme, when the time series of the temperature and pressure in the furnace are input into the fuzzy logic system for fuzzification processing to obtain the sequence of fuzzified temperature and the sequence of fuzzified pressure, considering that the fuzzy logic system strengthens the main trend of the time series distribution of the temperature and pressure in the furnace by fuzzifying the local time series distribution, it will cause the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector to have a certain source time series distribution difference. Therefore, the temperature time series correlation feature vector and the pressure time series correlation feature vector obtained by passing the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector through the time series correlation feature extractor based on the one-dimensional extended convolution layer will also have a time series correlation feature distribution difference, so that the temperature-pressure time series interactive correlation feature vector obtained by interactive fusion has an eigenvalue granularity distribution inconsistency, that is, the temperature-pressure time series interactive correlation feature vector has a distribution discreteness based on the eigenvalue numerical distribution, thereby affecting the temperature-pressure time series interactive correlation feature vector The convergence effect of the class regression iteration of the classifier.

[0083] Therefore, the applicant of the present application hopes to improve the overall semantically consistent aggregate expression effect of the temperature-pressure time series interactive correlation feature vector, thereby optimizing it, including:

[0084] Perform feature clustering on all eigenvalues ​​of the temperature-pressure time series interaction correlation feature vector to obtain a temperature-pressure time series interaction correlation intra-class feature set Out-of-class feature set interactively associated with temperature-pressure time series in, represents the intra-class feature set of temperature-pressure time series interaction correlation, f 1i represents the ith eigenvalue of the feature set within the temperature-pressure time series interaction correlation class, f 2j Represents the jth eigenvalue of the out-of-class feature set of temperature-pressure time series interaction correlation;

[0085] Calculate the ratio λ=k / n of the number of eigenvalues ​​in the feature set within the temperature-pressure time series interaction association class to the number of eigenvalues ​​of all features of the temperature-pressure time series interaction association feature vector, where k represents the number of eigenvalues ​​in the feature set within the temperature-pressure time series interaction association class, n represents the number of eigenvalues ​​of all features of the temperature-pressure time series interaction association feature vector, and λ represents the ratio;

[0086] Calculating the ratio of the λ-th power of the sum of the absolute values ​​of all eigenvalues ​​in the eigenvector of the temperature-pressure time series interaction association class to the λ-th power of the sum of the absolute values ​​of all eigenvalues ​​in the temperature-pressure time series interaction association eigenvector to obtain a first temperature-pressure time series interaction association weight;

[0087] Calculate the ratio of the λ / 2 power of the sum of squares of all eigenvalues ​​in the eigenvector of the temperature-pressure time series interaction association class to the λ / 2 power of the sum of squares of all eigenvalues ​​in the temperature-pressure time series interaction association eigenvector to obtain a second temperature-pressure time series interaction association weight;

[0088] That is, the first temperature-pressure time series interaction correlation weight w1 and the second temperature-pressure time series interaction correlation weight w2 are expressed as:

[0089]

[0090] Among them, f p represents each eigenvalue of the temperature-pressure time series interactive correlation eigenvector, w1 represents a first temperature-pressure time series interactive correlation weight, and w2 represents a second temperature-pressure time series interactive correlation weight;

[0091] Then, for each feature value in the feature set within the temperature-pressure time series interaction association class, the product of the feature value and the second temperature-pressure time series interaction association weight is calculated, and then the product of the feature value and the first temperature-pressure time series interaction association weight is added to obtain the optimized feature value f' 1i =w2×f 1i +w1, where f' 1i represents the i-th optimized feature value of the feature set within the temperature-pressure time series interaction association class;

[0092] For each eigenvalue in the temperature-pressure time series interaction correlation out-of-class feature set, calculate the product of the eigenvalue and the first temperature-pressure time series interaction correlation weight to obtain the optimized eigenvalue f 2j '=w1×f 2j , where f 2j ' represents the jth optimized feature value of the out-of-class feature set based on the temperature-pressure time series interaction correlation;

[0093] Finally, the optimized feature values ​​based on the temperature-pressure time series interaction correlation class internal feature set and the temperature-pressure time series interaction correlation class external feature set are combined to form an optimized temperature-pressure time series interaction correlation feature vector.

[0094] That is, while the temperature-pressure time series interaction correlation feature vector is subjected to feature clustering, the interaction description of the key feature information of the temperature-pressure time series interaction correlation feature vector in the clustering process is carried out, and the geometric equivariant topology of the feature is constructed by low-rank harmonic modulation of the clustering features of the temperature-pressure time series interaction correlation feature vector and the equivariance of the feature as a whole, so as to obtain the graphical distribution translation and rotation symmetry of the clustering features of the temperature-pressure time series interaction correlation feature vector relative to the feature as a whole, thereby introducing geometric message passing in the feature expression of the temperature-pressure time series interaction correlation feature vector, and realizing the clustering mapping symmetry of the temperature-pressure time series interaction correlation feature vector through manipulation of irreducible low-rank order coefficients, thereby improving the consistency of the clustering-based feature representation of the temperature-pressure time series interaction correlation feature vector, and thus improving the accuracy of the classification results.

[0095] In this way, artificial intelligence technology can be used to depict the time series correlation pattern between the fuzzy data of temperature and pressure in the furnace, so as to make more accurate recommendations on the adjustment of the charge ratio. In this way, it is possible to quickly adapt and optimize the charge ratio when the raw material composition and furnace conditions change, thereby maintaining the stability and continuity of the production process.

[0096] In summary, the above scheme uses temperature sensors and pressure sensors to monitor and collect temperature and pressure data in the furnace in real time, and uses the fuzzy logic system to convert these time series data into fuzzy data, and uses artificial intelligence-based data processing and analysis algorithms at the back end to perform time series collaborative correlation analysis of temperature fuzzy data and pressure fuzzy data, so as to use the time series correlation pattern and changes between the temperature and pressure in the furnace to more accurately recommend adjustments to the charge ratio. In this way, it is possible to quickly adapt to changes in raw material composition and furnace conditions, optimize the charge ratio, and maintain the stability and continuity of the production process.

[0097] Figure 2 FIG. 1 is a block diagram of a smelting charge ratio optimization system based on fuzzy logic according to an exemplary embodiment. Figure 2 As shown, the system 200 includes:

[0098] A furnace temperature and pressure acquisition module 201 is used to acquire the time series of the furnace temperature and the furnace pressure collected by the temperature sensor and the pressure sensor;

[0099] A fuzzy processing module 202 is used to input the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzy processing to obtain a fuzzy temperature sequence and a fuzzy pressure sequence;

[0100] A one-hot encoding module 203, used for performing one-hot encoding on the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0101] A feature extraction module 204 is used to extract features from the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively through a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0102] A hash mapping encoding module 205 is used to perform hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector;

[0103] An interactive fusion module 206, configured to interactively fuse the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0104] The recommended charge ratio determination module 207 is used to determine the recommended charge ratio based on the temperature-pressure time series interaction correlation feature.

[0105] Among them, the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional extended convolutional layer.

[0106] Reference below Figure 3 , which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0107] like Figure 3As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0108] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0109] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0110] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0111] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0112] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0113] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0114] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0115] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, a test parameter acquisition module may also be described as a "module for acquiring device test parameters corresponding to a target device".

[0116] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0117] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0118] According to one or more embodiments of the present disclosure, Example 1 provides a smelting charge ratio optimization method based on fuzzy logic, the method comprising:

[0119] Acquire the time series of the temperature and pressure in the furnace collected by the temperature sensor and the pressure sensor;

[0120] Inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzification processing to obtain a fuzzified temperature series and a fuzzified pressure series;

[0121] One-hot encoding the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0122] The sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector are respectively subjected to feature extraction by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0123] Performing hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector;

[0124] Interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0125] Based on the temperature-pressure time series interactive correlation characteristics, the recommended charge ratio is determined.

[0126] According to one or more embodiments of the present disclosure, Example 2 provides the method of Example 1, wherein the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional extended convolutional layer.

[0127] According to one or more embodiments of the present disclosure, Example 3 provides the method of Example 2, wherein the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector are respectively subjected to feature extraction by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector, including: processing the sequence of the temperature time series fuzzy input vector by the time series correlation feature extractor based on a one-dimensional extended convolution layer with the following feature extraction formula to obtain the temperature time series correlation feature vector;

[0128] Among them, the feature extraction formula is:

[0129] X 1:n =X1⊕X2⊕…⊕X n

[0130] C i =f(ω·X i:i+j-l +b)

[0131] C=[C1,C2,...,C n-l+1 ]

[0132] Among them, X1,X2,...,X n represents each temperature time series fuzzy input vector in the sequence of the temperature time series fuzzy input vector, X 1:n represents the concatenated vector of the sequence of the temperature time series fuzzification input vector, ⊕ represents the connection operation, X i:i+j-l X represents the sequence of the temperature time series fuzzy input vector i ,X i+1 ,...,X i+j-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i represents each temperature local time series association feature vector in the temperature time series association feature vector, [·,·] represents vector concatenation, and C is the temperature time series association feature vector.

[0133] According to one or more embodiments of the present disclosure, Example 4 provides the method of Example 3, performing hash mapping encoding on the temperature timing association feature vector and the pressure timing association feature vector to obtain a hash embedded temperature timing association feature vector and a hash embedded pressure timing association feature vector, including: performing hash mapping encoding on the temperature timing association feature vector and the pressure timing association feature vector according to the following hash formula to obtain the hash embedded temperature timing association feature vector and the hash embedded pressure timing association feature vector;

[0134] The hash formula is:

[0135]

[0136] Among them, v i is one of the temperature time series associated feature vectors and the pressure time series associated feature vectors, f θ (·) represents the fully connected encoding of the feature vector, sign(·) represents the sign function, It is the corresponding feature vector in the hash embedded temperature timing associated feature vector and the hash embedded pressure timing associated feature vector.

[0137] According to one or more embodiments of the present disclosure, Example 5 provides the method of Example 4, interactively fusing the hash embedded temperature timing association feature vector and the hash embedded pressure timing association feature vector to obtain a temperature-pressure timing interaction association feature, including: interactively fusing the hash embedded temperature timing association feature vector and the hash embedded pressure timing association feature vector using the following interactive formula to obtain a temperature-pressure timing interaction association feature vector as the temperature-pressure timing interaction association feature;

[0138] Wherein, the interaction formula is:

[0139]

[0140] in, is the hash embedding temperature time series associated feature vector, is the hash embedding pressure time series associated feature vector, is the square of the L2 norm of the vector, L2 is a hyperparameter, v c is the temperature-pressure time series interaction correlation feature vector.

[0141] According to one or more embodiments of the present disclosure, Example 6 provides the method of Example 5, which determines the recommended charge ratio based on the temperature-pressure timing interaction correlation feature, including: passing the temperature-pressure timing interaction correlation feature vector through a decoder-based smelting charge ratio recommender to obtain the recommended charge ratio.

[0142] According to one or more embodiments of the present disclosure, Example 7 provides the method of Example 6, further comprising a training step: for training the temporal correlation feature extractor based on the one-dimensional expanded convolutional layer and the decoder-based smelting charge ratio recommender.

[0143] According to one or more embodiments of the present disclosure, Example 8 provides the method of Example 7, wherein the training step includes:

[0144] Acquire the time series of the temperature and pressure in the training furnace collected by the temperature sensor and the pressure sensor;

[0145] Inputting the time series of the training furnace temperature and the training furnace pressure into the fuzzy logic system for fuzzification processing to obtain a training fuzzified temperature sequence and a training fuzzified pressure sequence;

[0146] One-hot encoding the training fuzzified temperature sequence and the training fuzzified pressure sequence to obtain a training temperature time series fuzzified input vector sequence and a training pressure time series fuzzified input vector sequence;

[0147] The time series correlation feature extractor based on the deep neural network model performs feature extraction on the sequence of the training temperature time series fuzzy input vector and the sequence of the training pressure time series fuzzy input vector respectively to obtain a training temperature time series correlation feature vector and a training pressure time series correlation feature vector;

[0148] Performing hash mapping encoding on the training temperature time series associated feature vector and the training pressure time series associated feature vector to obtain a training hash embedded temperature time series associated feature vector and a training hash embedded pressure time series associated feature vector;

[0149] Interactively fusing the training hash-embedded temperature time series correlation feature vector and the training hash-embedded pressure time series correlation feature vector to obtain a training temperature-pressure time series interactive correlation feature vector;

[0150] Passing the training temperature-pressure time series interactive correlation feature vector through the decoder-based smelting furnace charge ratio recommender to obtain a decoding loss function value;

[0151] The one-dimensional extended convolutional layer-based temporal correlation feature extractor and the decoder-based smelting charge ratio recommender are trained based on the decoding loss function value, and the training temperature-pressure temporal interaction correlation feature vector is iteratively optimized in each iteration.

[0152] According to one or more embodiments of the present disclosure, Example 9 provides a smelting charge ratio optimization system based on fuzzy logic, the system comprising:

[0153] A furnace temperature and pressure acquisition module is used to acquire the time series of the furnace temperature and the furnace pressure collected by the temperature sensor and the pressure sensor;

[0154] A fuzzy processing module, used for inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzy processing to obtain a fuzzy temperature sequence and a fuzzy pressure sequence;

[0155] A one-hot encoding module, used for performing one-hot encoding on the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors;

[0156] A feature extraction module, used for performing feature extraction on the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively through a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector;

[0157] A hash mapping encoding module, used for performing hash mapping encoding on the temperature time series associated feature vector and the pressure time series associated feature vector to obtain a hash embedded temperature time series associated feature vector and a hash embedded pressure time series associated feature vector;

[0158] An interactive fusion module, used for interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature;

[0159] The recommended charge ratio determination module is used to determine the recommended charge ratio based on the temperature-pressure time series interaction correlation characteristics.

[0160] According to one or more embodiments of the present disclosure, Example 10 provides the method of Example 9, wherein the temporal correlation feature extractor based on the deep neural network model is a temporal correlation feature extractor based on a one-dimensional expanded convolutional layer.

[0161] Figure 4 is an application scenario diagram of a smelting charge ratio optimization method based on fuzzy logic according to an exemplary embodiment. Figure 4 As shown, in this application scenario, first, the furnace temperature collected by the temperature sensor and the pressure sensor is obtained (for example, Figure 4 C1) and the time series of the furnace pressure (e.g. Figure 4Then, the obtained furnace temperature and furnace pressure are input to a server (for example, Figure 4 In S) shown in , the server is capable of processing the furnace temperature and the furnace pressure based on a fuzzy logic smelting charge ratio optimization algorithm to determine a recommended charge ratio.

[0162] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0163] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0164] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A smelting charge ratio optimization method based on fuzzy logic, characterized in that: include: Acquire the time series of the temperature and pressure in the furnace collected by the temperature sensor and the pressure sensor; Inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzification processing to obtain a fuzzified temperature series and a fuzzified pressure series; One-hot encoding the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors; The sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector are respectively subjected to feature extraction by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector; Performing hash mapping encoding on the temperature time series correlation feature vector and the pressure time series correlation feature vector to obtain a hash embedded temperature time series correlation feature vector and a hash embedded pressure time series correlation feature vector; Interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature; Based on the temperature-pressure time series interactive correlation characteristics, the recommended charge ratio is determined.

2. The fuzzy logic-based smelting charge ratio optimization method according to claim 1 is characterized in that: The temporal association feature extractor based on the deep neural network model is a temporal association feature extractor based on a one-dimensional extended convolutional layer.

3. The fuzzy logic-based smelting charge ratio optimization method according to claim 2, characterized in that: The sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector are respectively subjected to feature extraction by a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector, including: the sequence of the temperature time series fuzzy input vector is processed by the time series correlation feature extractor based on a one-dimensional extended convolution layer according to the following feature extraction formula to obtain the temperature time series correlation feature vector; Among them, the feature extraction formula is: X 1:n =X1⊕X2⊕...⊕X n C i =f(ω·X i:i+l-1 +b) C=[C1,C2,...,Cn -l+1 ] Among them, X1,X2,...,X n represent the respective temperature time series fuzzy input vectors in the sequence of the temperature time series fuzzy input vectors, X 1:n represents the concatenated vector of the sequence of the temperature time series fuzzification input vector, ⊕ represents the connection operation, X i:i+l-1 X represents the sequence of the temperature time series fuzzy input vector i ,X i+1 ,...,X i+l-1 , ω and b represent the weight matrix and the offset vector respectively, f(·) represents the convolution operation, Ci represents each temperature local timing association feature vector in the temperature timing association feature vector, [·,·] represents vector concatenation, and c is the temperature timing association feature vector.

4. The fuzzy logic-based smelting charge ratio optimization method according to claim 3 is characterized in that: The temperature timing correlation feature vector and the pressure timing correlation feature vector are hash-mapped to obtain a hash-embedded temperature timing correlation feature vector and a hash-embedded pressure timing correlation feature vector, including: hash-mapped to obtain the hash-embedded temperature timing correlation feature vector and the hash-embedded pressure timing correlation feature vector by using the following hash formula; The hash formula is: Among them, v i is one of the temperature time series associated feature vectors and the pressure time series associated feature vectors, f θ (·) represents the fully connected encoding of the feature vector, sign(·) represents the sign function, It is the corresponding feature vector in the hash embedded temperature timing associated feature vector and the hash embedded pressure timing associated feature vector.

5. The fuzzy logic-based smelting charge ratio optimization method according to claim 4 is characterized in that: Interactively fusing the hash-embedded temperature timing correlation feature vector and the hash-embedded pressure timing correlation feature vector to obtain a temperature-pressure timing interaction correlation feature, including: interactively fusing the hash-embedded temperature timing correlation feature vector and the hash-embedded pressure timing correlation feature vector using the following interactive formula to obtain a temperature-pressure timing interaction correlation feature vector as the temperature-pressure timing interaction correlation feature; Wherein, the interaction formula is: in, is the hash embedding temperature time series associated feature vector, is the hash embedding pressure time series associated feature vector, is the square of the L2 norm of the vector, L2 is a hyperparameter, v c is the temperature-pressure time series interaction correlation feature vector.

6. The fuzzy logic-based smelting charge ratio optimization method according to claim 5, characterized in that: Based on the temperature-pressure time series interaction correlation feature, a recommended charge ratio is determined, including: passing the temperature-pressure time series interaction correlation feature vector through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

7. The fuzzy logic-based smelting charge ratio optimization method according to claim 6, characterized in that: The temperature-pressure time series interactive correlation feature vector is passed through a smelting furnace charge ratio recommender based on a decoder to obtain a recommended furnace charge ratio, including: Optimizing the temperature-pressure time series interaction correlation feature vector to obtain an optimized temperature-pressure time series interaction correlation feature vector; The optimized temperature-pressure time series interactive correlation feature vector is passed through a decoder-based smelting charge ratio recommender to obtain a recommended charge ratio.

8. A smelting charge ratio optimization system based on fuzzy logic, characterized in that: include: A furnace temperature and pressure acquisition module is used to acquire the time series of the furnace temperature and the furnace pressure collected by the temperature sensor and the pressure sensor; A fuzzy processing module, used for inputting the time series of the temperature and pressure in the furnace into a fuzzy logic system for fuzzy processing to obtain a fuzzy temperature sequence and a fuzzy pressure sequence; A one-hot encoding module, used for performing one-hot encoding on the fuzzified temperature sequence and the fuzzified pressure sequence to obtain a sequence of temperature time series fuzzified input vectors and a sequence of pressure time series fuzzified input vectors; A feature extraction module, used for performing feature extraction on the sequence of the temperature time series fuzzy input vector and the sequence of the pressure time series fuzzy input vector respectively through a time series correlation feature extractor based on a deep neural network model to obtain a temperature time series correlation feature vector and a pressure time series correlation feature vector; A hash mapping encoding module, used for performing hash mapping encoding on the temperature time series associated feature vector and the pressure time series associated feature vector to obtain a hash embedded temperature time series associated feature vector and a hash embedded pressure time series associated feature vector; An interactive fusion module, used for interactively fusing the hash-embedded temperature time series correlation feature vector and the hash-embedded pressure time series correlation feature vector to obtain a temperature-pressure time series interactive correlation feature; The recommended charge ratio determination module is used to determine the recommended charge ratio based on the temperature-pressure time series interaction correlation characteristics.

9. The fuzzy logic-based smelting charge ratio optimization system according to claim 8, characterized in that: The temporal association feature extractor based on the deep neural network model is a temporal association feature extractor based on a one-dimensional extended convolutional layer.

Images