Steady state detection method and system in aluminum oxide production evaporation process

Abstract

The invention relates to the technical field of steady-state detection of an aluminum oxide production evaporation process, in particular to a steady-state detection method and system of the aluminum oxide production evaporation process. Comprising the steps that firstly, a steady-state variable correlation analysis method is provided, variable correlation characteristics in an initial data matrix are obtained through a discrete equation, and significance sorting is carried out on the correlation characteristics to analyze correlation coefficients of steady-state variables; the method can be used for accurately analyzing the action relationship of various variables in the aluminum oxide production and evaporation process; secondly, a steady-state detection method is provided, according to the method, abnormal variable data are detected by establishing a steady-state detection function, accurate evaluation is conducted through evaluation indexes at different time points, and the change situation of evaporation variables at specific time is accurately known; finally, a self-adaptive evaporation variable coordination method is provided, the optimal evaporation variable is solved through multiple times of iterative optimization so as to meet the control target, and the quality of aluminum oxide production is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of steady-state detection of an evaporation process in the production of aluminum oxide, and in particular to a steady-state detection method and system for an evaporation process in the production of aluminum oxide. Background Art

[0002] Alumina is an important inorganic compound with a white crystal structure. Alumina has the following physical properties: high hardness, with a Mohs hardness of 9, second only to diamond and silicon carbide; high melting point, about 2072°C; good corrosion resistance to most acids and alkalis, especially at high temperatures.

[0003] Due to the above excellent physical and chemical properties, alumina has been widely used in the fields of industry, science and medicine. For example, in the industrial field, alumina is the main component of high-temperature resistant materials such as refractory ceramics, refractory bricks and refractory coatings, and is used for lining and protection of various industrial furnaces and metal smelting equipment; in the scientific field, alumina is an ideal carrier for many catalysts, used to support and disperse active components, and improve the efficiency and stability of catalysts; in the medical field, alumina is used as an important component of artificial joints and dental filling materials.

[0004] The production process of alumina usually includes the following main steps: extraction and pretreatment of alumina raw materials, preparation, refining and processing of aluminum hydroxide; among them, the refining process of alumina is very important in its production. The purpose of refining is to remove impurities and further purify so that the alumina can meet specific quality standards and application requirements. The first task in the refining process is evaporation and concentration. The alumina solution is often heated and evaporated by an evaporator to remove water and concentrate the solution to the required concentration.

[0005] The evaporation process of alumina solution is a critical step, and its steady-state has an important impact on the quality of the product. The steady-state of the evaporation process in alumina production can not only maintain the solution concentration and ensure the quality of alumina, but also improve production efficiency and increase cost-benefit. Therefore, it is very necessary to conduct steady-state detection in the evaporation process of alumina production.

[0006] There are many excellent methods in the prior art for steady-state detection of the evaporation process of alumina production. For example, the Chinese patent application publication number CN108664000A discloses a steady-state detection method and system for the evaporation process of alumina production, including: selecting multiple steady-state detection variables from the process variables according to the correlation between the process variables and the concentration of the four-flash discharge in the evaporation process; using the improved K-means algorithm to perform outlier detection on the first sample data set corresponding to each steady-state detection variable to obtain the second sample data set corresponding to each steady-state detection variable; obtaining the change speed and change acceleration of each steady-state detection variable at any time in the time period to be detected according to the second sample data set corresponding to each steady-state detection variable; obtaining the multivariate steady-state evaluation index at any time according to the change speed and change acceleration of each steady-state detection variable at any time, and judging whether the evaporation process is in a steady state at any time according to the multivariate steady-state evaluation index. However, this steady-state detection method only relies on the collected data to analyze the correlation of the process variables, but the accuracy of the detection cannot be guaranteed when the data is wrong. In order to further improve the accuracy of steady-state detection of the evaporation process of alumina production and ensure the production quality of the product.

[0007] To this end, the present invention proposes a steady-state detection method and system for the evaporation process of aluminum oxide production. Summary of the invention

[0008] The purpose of the present invention is to provide a steady-state detection method and system for the evaporation process of aluminum oxide production, which can realize the steady-state detection of evaporation variables and the adaptive adjustment of evaporation variables in the evaporation process of aluminum oxide production; the method and system mainly include: first, collecting relevant historical data of aluminum oxide in multiple production evaporation processes according to timestamps, and using an evaporation variable analysis method to perform correlation analysis on the collected historical data; the analysis method obtains variable correlation characteristics in the initial data matrix through discrete equations, and performs significance sorting according to the variable correlation characteristics to obtain the relevant differences between the variables, and selects steady-state variables in combination with the initial matrix; according to the selection results, a steady-state correlation coefficient matrix is ​​obtained; the method helps to optimize the selection process of evaporation variables, ensuring that the selected variables are conducive to improving the subsequent stability The accuracy of steady-state detection; secondly, the present invention proposes a steady-state detection method, which establishes a steady-state detection function, and the function performs steady-state detection on evaporation variables at different time points through a steady-state anomaly detection algorithm, and establishes a multivariate steady-state evaluation index in the function through historical data, and the index can accurately evaluate the evaporation variables; this method can not only provide an accurate evaluation of the steady state of the evaporation variables, but also accurately understand the changes in the evaporation variables at a specific time; finally, the present invention proposes a method for adaptively coordinating evaporation variables in an evaporation process, and adaptively adjusts the evaporation variables; this method achieves the control target under the condition that the evaporation variables are adjusted without losing stability, and obtains the optimal evaporation variables through multiple iterative optimizations to meet the set control targets, thereby ensuring the quality of alumina production.

[0009] To achieve the above object, the present invention provides the following technical solutions:

[0010] A steady-state detection method for an evaporation process in aluminum oxide production, comprising:

[0011] Obtain relevant historical data of the alumina during multiple production evaporation processes according to the timestamp; wherein the relevant historical data include: historical raw liquid parameters, historical evaporation variables, historical evaporation efficiency and historical product concentration at the outlet of the final flash evaporator;

[0012] Furthermore, the relevant historical data is processed to obtain standardized relevant historical data; wherein the data processing includes: processing abnormal values ​​and missing values ​​in the relevant historical data to obtain first relevant historical data; normalizing the first relevant historical data to obtain second relevant historical data; and standardizing the second relevant historical data to obtain the standardized relevant historical data.

[0013] Furthermore, the historical evaporation variables are marked; wherein the marking content includes: 1 represents a stable state; 0 represents an unstable state;

[0014] Furthermore, the standardized relevant historical data is subjected to evaporation variable analysis to obtain an evaporation variable correlation matrix; wherein the calculation process of the evaporation variable correlation matrix includes: establishing an initial matrix according to the standardized relevant historical data; the initial matrix is ​​expressed as: in, Expressed as the standardized historical stock solution parameter at the mth time point; SV mn It is expressed as the nth standardized historical evaporation variable at the mth time point; Expressed as the normalized historical product concentration at the mth time point;

[0015] Substitute each row of data in the initial matrix into the discrete equation; wherein the discrete equation formula is:

[0016]

[0017] in, Expressed as the normalized historical product concentration of the i-th row in the initial matrix; Represented as the standardized historical stock solution parameter of the i-th row in the initial matrix; SV i It is represented as the normalized historical evaporation variable vector of the ith row in the initial matrix; f() is represented as a nonlinear analytical function; E() is represented as the environmental function of the aluminum oxide in the production evaporation process;

[0018] Solving the correlation coefficient vector of each data in the initial matrix according to the discrete equation, and establishing the evaporation variable correlation matrix;

[0019] Furthermore, the evaporation variable correlation matrix is ​​used to calculate a steady-state correlation coefficient matrix; wherein the calculation process of the steady-state correlation coefficient matrix includes: ranking the evaporation variable correlation matrix by significance to obtain a significance ranking result of the evaporation variables; selecting steady-state variables based on the significance ranking result of the evaporation variables combined with the marking content of the initial matrix; and obtaining the steady-state correlation coefficient matrix using a correlation analysis method based on the result of the steady-state variable selection.

[0020] Further, the steady-state detection is performed on the current alumina production evaporation process in combination with the steady-state correlation coefficient matrix to obtain a steady-state detection result; wherein the steady-state detection process includes: obtaining a multivariate steady-state evaluation index corresponding to the timestamp according to the marked standardized related historical data; wherein the multivariate steady-state evaluation index is expressed as:

[0021]

[0022] Among them, SI(t) represents the steady-state index at time t; n represents the total number of evaporation variables; X i(t) represents the parameter value of the i-th evaporation variable at time t; σ i (t) represents the mean value of the ith evaporation variable at the tth time; μ i (t) represents the variance of the i-th evaporation variable at time t;

[0023] A steady-state detection function is constructed according to the multivariate steady-state evaluation index; wherein the steady-state detection function is expressed as:

[0024]

[0025] Wherein, SS() represents a steady-state detection function; X(t) represents a vector of evaporation variables of the aluminum oxide during the production evaporation process; AnoD() represents a steady-state anomaly detection algorithm; the steady-state anomaly detection algorithm includes:

[0026] Combining the steady-state correlation coefficient matrix with the standardized historical evaporation variables to construct an extended variable set;

[0027] The extended variable set is used to train a random forest classifier; wherein the output layer of the random forest classifier uses the multivariate steady-state evaluation index as a judgment criterion.

[0028] Furthermore, adaptive evaporation variable coordination is performed in the evaporation process according to the steady-state detection result; wherein the process of adaptive evaporation variable coordination in the evaporation process includes: obtaining the stock solution parameters and evaporation variables at the current moment; setting the control target of the evaporation process; wherein the control target includes: maximum product concentration and maximum evaporation efficiency; using an evaporation variable optimization algorithm to obtain optimized evaporation variables; wherein the steps of the evaporation variable optimization algorithm include:

[0029] Step 1: taking the evaporation variable as the initial optimal evaporation variable;

[0030] Step 2: evaluating the correlation of the initial optimal evaporation variables through the steady-state correlation coefficient matrix;

[0031] Step 3: Taking the highest value of the correlation as the parent generation, performing crossover and mutation operations to generate new evaporation variables;

[0032] Step 4: updating the initial optimal evaporation variable according to the new evaporation variable;

[0033] Step 5: Repeat step 2 to step 4 until the stop condition is met to obtain the optimized evaporation variable.

[0034] The optimized evaporation variable is subjected to steady-state detection; when the optimized evaporation variable is in steady state and can achieve the control target, the optimal evaporation variable is obtained; otherwise, the evaporation variable optimization algorithm is repeated.

[0035] A steady-state detection system for an evaporation process of aluminum oxide production, the system comprising: a monitoring unit, a data acquisition unit, a data processing unit, a steady-state variable analysis unit, a steady-state detection unit, a variable control unit and a feedback unit;

[0036] Wherein, the monitoring unit is used for monitoring the aluminum oxide during the production evaporation process;

[0037] The data acquisition unit is used to collect various data of the alumina production evaporation process;

[0038] The data processing unit is used to process the data collected by the data collection unit; wherein the data processing unit includes: processing abnormal values ​​and missing values ​​in the relevant historical data to obtain first relevant historical data; normalizing the first relevant historical data to obtain second relevant historical data; and standardizing the second relevant historical data to obtain standardized relevant historical data.

[0039] The steady-state variable analysis unit is used to perform correlation analysis on the variables of the aluminum oxide during the production evaporation process; wherein the steady-state variable analysis unit includes: establishing an initial matrix based on standardized relevant historical data; the initial matrix is ​​expressed as: in, Expressed as the standardized historical stock solution parameter at the mth time point; SV mn It is expressed as the nth standardized historical evaporation variable at the mth time point; Expressed as the normalized historical product concentration at the mth time point;

[0040] Furthermore, each row of data in the initial matrix is ​​substituted into the discrete equation; wherein the discrete equation formula is:

[0041]

[0042] in, Expressed as the normalized historical product concentration of the i-th row in the initial matrix; Represented as the standardized historical stock solution parameter of the i-th row in the initial matrix; SV i It is represented as the normalized historical evaporation variable vector of the ith row in the initial matrix; f() is represented as a nonlinear analytical function; E() is represented as the environmental function of the aluminum oxide in the production evaporation process;

[0043] Further, solving the correlation coefficient vector of each data in the initial matrix according to the discrete equation, and establishing the evaporation variable correlation matrix;

[0044] Further, the evaporation variable correlation matrix is ​​sorted by significance to obtain a significance sorting result of the evaporation variable;

[0045] Further, steady-state variable selection is performed according to the significance ranking result of the evaporation variables combined with the marking content of the initial matrix;

[0046] Furthermore, the steady-state correlation coefficient matrix is ​​obtained by using a correlation analysis method according to the result of the steady-state variable selection.

[0047] The steady-state detection unit is used to perform steady-state detection on various variables in the evaporation process of aluminum oxide production; wherein the steady-state detection unit includes:

[0048] The multivariate steady-state evaluation index of the corresponding timestamp is obtained according to the marked standardized related historical data; wherein the multivariate steady-state evaluation index is expressed as:

[0049]

[0050] Among them, SI(t) represents the steady-state index at time t; n represents the total number of evaporation variables; X i (t) represents the parameter value of the i-th evaporation variable at time t; σ i (t) represents the mean value of the ith evaporation variable at the tth time; μ i (t) represents the variance of the i-th evaporation variable at time t;

[0051] A steady-state detection function is constructed according to the multivariate steady-state evaluation index; wherein the steady-state detection function is expressed as:

[0052]

[0053] Wherein, SS() represents a steady-state detection function; X(t) represents a vector of evaporation variables of the aluminum oxide during the production evaporation process; AnoD() represents a steady-state anomaly detection algorithm; the steady-state anomaly detection algorithm includes:

[0054] Combining the steady-state correlation coefficient matrix with the standardized historical evaporation variables to construct an extended variable set;

[0055] The extended variable set is used to train a random forest classifier; wherein the output layer of the random forest classifier uses the multivariate steady-state evaluation index as a judgment criterion.

[0056] The variable control unit is used to perform steady-state adjustment on the variable;

[0057] The feedback unit is used to feed back parameter changes and evaporation conditions during the aluminum oxide production evaporation process to relevant staff.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1. The present invention proposes a method for analyzing the correlation of steady-state variables in the evaporation process of aluminum oxide production, which is used to analyze the collected historical data; the method obtains the variable correlation characteristics in the initial data matrix through discrete equations, and performs significance sorting according to the variable correlation characteristics to obtain the correlation differences between the variables, and selects steady-state variables in combination with the initial matrix; the selected steady-state variables are subjected to correlation analysis to obtain the final steady-state variable correlation coefficient matrix; the method can accurately analyze the functional relationship between the variables in the evaporation process of aluminum oxide production, ensure that the selected variables are conducive to improving the accuracy of subsequent steady-state detection, and ensure the quality of aluminum oxide production.

[0060] 2. The present invention proposes a steady-state detection function; the function performs steady-state detection on evaporation variables at different time points through a steady-state anomaly detection algorithm, and establishes a multivariate steady-state evaluation index through historical data in the function, which accurately evaluates the evaporation variables; this method can not only provide an accurate evaluation of the steady state of the evaporation variables, but also accurately understand the changes in the evaporation variables at a specific time, thereby ensuring the quality of alumina production.

[0061] 3. The present invention proposes a method for adaptively coordinating evaporation variables in an evaporation process, which can adaptively adjust the evaporation variables according to steady-state detection results; this method achieves the control target without losing stability by adjusting the evaporation variables, and obtains the optimal evaporation variables through multiple iterative optimizations to meet the set control targets, thereby ensuring the quality of alumina production. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A flow chart of a steady-state detection method for an aluminum oxide production evaporation process provided by an embodiment of the present invention;

[0063] Figure 2 A structural diagram of a steady-state detection system for an aluminum oxide production evaporation process provided by an embodiment of the present invention;

[0064] Figure 3 A schematic diagram of a four-effect evaporation process provided by an embodiment of the present invention;

[0065] Figure 4 A schematic diagram of variable data of the aluminum hydride production evaporation process provided by an embodiment of the present invention;

[0066] Figure 5 A schematic diagram of a single variable steady-state detection result provided by an embodiment of the present invention;

[0067] Figure 6 A schematic diagram of steady-state detection results under the action of multiple variables provided by an embodiment of the present invention;

[0068] Figure 7 A flow chart of an adaptive evaporation variable coordination method provided by an embodiment of the present invention;

[0069] Figure 8 A schematic diagram of a steady-state detection result after coordination of evaporation variables provided in an embodiment of the present invention.

[0070] In the figure: 101, the first condensate water tank; 102, the second condensate water tank; 103, the third condensate water tank; 104, the fourth condensate water tank; 201, the first evaporator; 202, the second evaporator; 203, the third evaporator; 204, the fourth evaporator; 301, the first preheater; 302, the second preheater; 303, the third preheater; 401, the first flash evaporator; 402, the second flash evaporator 2; 403, the third flash evaporator; 501, the condenser; 601, the raw steam; 701, the raw liquid. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] Alumina, as an important inorganic compound, has a white crystal structure and has high hardness and melting point characteristics.

[0073] Due to its excellent physical and chemical properties, alumina has been widely used in industry, science, and medicine. For example, in the industrial field, alumina is the main component for manufacturing high-temperature resistant materials such as refractory ceramics, refractory bricks, and refractory coatings, and is used for lining and protection of various industrial furnaces and metal smelting equipment. In the scientific field, alumina is an ideal carrier for many catalysts, used to support and disperse active components and improve the efficiency and stability of catalysts. In the medical field, alumina is used as an important component of artificial joints and dental filling materials because of its biocompatibility and corrosion resistance.

[0074] The production process of alumina usually includes the following main steps: extraction and pretreatment of alumina raw materials, preparation, refining and processing of aluminum hydroxide; among them, the refining process of alumina is very important in its production. The purpose of refining is to remove impurities and further purify so that the alumina can meet specific quality standards and application requirements. The first task in the refining process is evaporation and concentration. The alumina solution is often heated and evaporated by an evaporator to remove water and concentrate the solution to the required concentration.

[0075] In the process of aluminum oxide production evaporation, the outlet material concentration detection lags a lot, which makes it impossible to detect in time. Therefore, the current technology detects the evaporation variables of the aluminum oxide production evaporation process in a steady state to reflect the evaporation of aluminum oxide; however, there are some defects in the detection process. Therefore, the present invention proposes a steady-state detection method and system for the aluminum oxide production evaporation process; the method and system proposed by the present invention will be described in detail from two embodiments below.

[0076] Embodiment 1:

[0077] In the embodiment of the present application, the proposed method and system are combined to achieve the purpose of the present invention; first, Figure 1 The specific steps of the method proposed in the present invention are given in the following, including: S10. Obtain relevant historical data of alumina during multiple production evaporation processes according to timestamps; S20. Preprocess relevant historical data to obtain standardized relevant historical data; S30. Mark evaporation variables in standardized relevant historical data; S40. Analyze evaporation variables and calculate steady-state correlation coefficients; S50. Perform steady-state detection; S60. Adaptively adjust evaporation variables. Correspondingly, in Figure 2 The structure of the steady-state detection system for the evaporation process of aluminum oxide production proposed by the present invention is given, which includes: a monitoring unit, a data acquisition unit, a data processing unit, a steady-state variable analysis unit, a steady-state detection unit, a variable control unit and a feedback unit. Figure 1 The method steps can be Figure 2 The units in implement the corresponding functions. Finally, the process of combining the two will be explained from the following content.

[0078] Specifically, the evaporation process of alumina production can be monitored in real time by the monitoring unit of the system; wherein the monitoring unit mainly monitors data through multiple types of sensors, including: temperature sensors, concentration sensors, flow meters and pressure sensors, etc.; wherein the temperature sensor is used to monitor the temperature inside the evaporation equipment (evaporator, heater and cooler, etc.), the process temperature of the solution or mixture entering or leaving the evaporation equipment, and the temperature of the heat transfer medium (such as hot water and steam, etc.); the production of alumina usually involves dissolving alumina or other substances in a solvent to form a solution, and the concentration sensor is used to monitor the concentration of alumina in these solutions; in the production of alumina, the solution is one of the liquids that often need to flow, and the flow meter is used to measure the flow rate of the solution in the pipeline or equipment; in addition to the solution, the heat transfer medium such as water or steam also plays an important role in the production of alumina, and also needs to be measured by a flow meter; when heating or cooling the evaporation equipment, hot water, steam or other liquids are usually used as heat transfer media. The pressure sensor is used to monitor the pressure changes of these media to ensure their normal flow and use status in the equipment.

[0079] Specifically, the data acquisition unit collects data during the evaporation process of aluminum oxide production through the sensor of the monitoring unit and saves the data for subsequent analysis; the function of this unit corresponds to the above-mentioned step S10;

[0080] The data collected by the data collection unit has a time series characteristic; each piece of data is timestamped according to the set collection time so that relevant staff can query it;

[0081] Specifically, the data processing unit of the system processes the collected relevant historical data, corresponding to the above-mentioned step S20; wherein the relevant historical data include: historical raw solution parameters (temperature, concentration and flow rate, etc.), historical evaporation variables (evaporator inlet solution flow rate, evaporator outlet concentrated solution flow rate, evaporator inlet solution temperature, evaporator outlet concentrated solution temperature, evaporator internal steam pressure, evaporator external ambient temperature, evaporator inlet solution concentration and evaporator outlet concentrated solution concentration, etc.), historical evaporation efficiency and historical product concentration at the outlet of the final flash evaporator;

[0082] Among them, the data processing process in the data processing unit includes: processing the abnormal values ​​and missing values ​​in the relevant historical data to obtain the first relevant historical data; wherein, the processing of abnormal values ​​or missing values ​​can adopt the deletion method and interpolation method to correct the data; normalizing the first relevant historical data to obtain the second relevant historical data; standardizing the second relevant historical data to obtain the standardized relevant historical data.

[0083] Refer to Table 1, which lists some of the processed data;

[0084] Table 1. Some historical data related to standardization

[0085]

[0086] Table 1 lists some relevant historical data after preprocessing at different time nodes, mainly including eight variable parameters, namely: first evaporator discharge, evaporation mother liquor, first evaporator flow, first evaporator pressure, causticizing stock solution, evaporation stock solution tank temperature, heat exchanger outlet water temperature and causticizing tank temperature; among which, the first evaporator is the material quantity parameter output by the discharge pipe of the first evaporator in the production evaporation process.

[0087] The data processing unit of the system described in the embodiment of the present application processes the collected relevant historical data; wherein, the data processing includes: outliers, missing value processing can help to remove errors or incomplete parts in the data, thereby improving the quality and accuracy of the data, thereby avoiding large errors in subsequent steady-state detection; normalization eliminates the complexity of data interpretation and analysis introduced by unit differences. This makes the data easier to compare and analyze, ensuring the consistency and comparability of the data; standardization can eliminate the effects introduced by dimensional differences between different variables; in the evaporation process of alumina production, the parameters involved may include temperature, pressure, and concentration, etc., and the dimensions of these parameters may be different. Standardization scales each variable to the same scale to ensure that they have similar importance in the analysis, and avoids certain variables from having too much influence on the test results due to their values.

[0088] Specifically, according to Figure 1 Step S30 of marking the variables in the processed data; wherein the marking content includes: 1 represents a stable state; 0 represents an unstable state; the marking process is pre-marked by relevant staff through observation of data;

[0089] Specifically, the steady-state correlation analysis of the marked data is implemented by the steady-state variable analysis unit of the system, corresponding to the above-mentioned step S40; wherein the specific analysis process includes:

[0090] An initial matrix is ​​established based on standardized relevant historical data; the initial matrix is ​​expressed as: in, Expressed as the standardized historical stock solution parameter at the mth time point; SV mn It is expressed as the nth standardized historical evaporation variable at the mth time point; Expressed as the normalized historical product concentration at the mth time point;

[0091] Substitute each row of data in the initial matrix into the discrete equation; wherein the discrete equation formula is:

[0092]

[0093] in, Expressed as the normalized historical product concentration of the i-th row in the initial matrix; Represented as the standardized historical stock solution parameter of the i-th row in the initial matrix; SV i It is represented as the normalized historical evaporation variable vector of the ith row in the initial matrix; f() is represented as a nonlinear analytical function; E() is represented as the environmental function of the aluminum oxide in the production evaporation process;

[0094] Among them, an environmental function is added to the above discrete equation, and the function is used to simulate the environmental conditions of the aluminum oxide production evaporation model; in the embodiment of the present application, the aluminum oxide production evaporation model can be composed of 4 evaporators, 3 flash evaporators, 3 preheaters and 4 condensation water tanks to realize a four-effect evaporation process; refer to Figure 3 , the figure shows a schematic diagram of the process of four-effect evaporation; among them, Figure 3 The process includes:

[0095] (1) Liquid flow

[0096] The raw liquid 701 enters the third evaporator 203 and the fourth evaporator 204 from the raw material tank through the feed pump, and the outlet liquid of the fourth evaporator 204 is pumped into the third preheater 303 by the feed pump. After the liquid temperature is heated to a temperature close to the boiling point of the third evaporator 203 in the third preheater 303, it enters the third evaporator 203, and then enters the first evaporator 201 and the second evaporator 202 in the same way. The liquid flows continuously in each evaporator through the circulation pump. The outlet liquid of the first evaporator 201 enters the first flash evaporator 401, the second flash evaporator 402 and the third flash evaporator 403 in turn by the pressure difference.

[0097] (2) Steam process

[0098] The raw steam 601 enters the first evaporator 201, and the secondary steam generated by the first evaporator 201 enters the second evaporator 201 and the first preheater 301 respectively; similarly, the secondary steam of the second evaporator 202 enters the third evaporator 203 and the second preheater 302 respectively, and the secondary steam of the third evaporator 203 enters the fourth evaporator 204 and the third preheater 303 respectively, and the secondary steam of the fourth evaporator 204 enters the vacuum pump through the condenser 501. The steam generated by evaporation in each condensate tank enters the corresponding evaporator for reuse, among which part of the steam generated by the first condensate tank 101 enters the third condensate tank 103. The secondary steam of the first flash evaporator 401, the second flash evaporator 402 and the third flash evaporator 403 enter the first preheater 301, the second preheater 302 and the third preheater 303 respectively.

[0099] (3) Condensate process

[0100] The condensed water generated by the first evaporator 201 is pumped to the condensed water monitoring station through a qualified water pump after self-evaporation, the condensed water in the second condensed water tank 102 enters the third condensed water tank 103, recovers heat from self-evaporation to the third evaporator 203, and is pumped to the condensed water detection station by an unqualified condensed water pump, and the condensed water generated by the fourth evaporator 204 is directly pumped to the condensed water monitoring station by a qualified condensed water pump.

[0101] Specifically, the environmental function includes four modules: an evaporator module, a flash evaporator module, a preheater module and a condensed water tank module;

[0102] The evaporator module is represented as: Among them, T evap,i Expressed as the temperature of evaporator i; Q in,i Expressed as the feed heat of evaporator i; Q steam,i Expressed as the steam heat output of evaporator i; Q out,i Expressed as other heat loss or output of evaporator i; Cp i Expressed as the specific heat capacity of evaporator i material; m i Expressed as the mass of the material in evaporator i;

[0103] The flash evaporator module is represented as: Among them, M liquid,j Expressed as the mass of the liquid phase of flash evaporator j; M vapor,j Expressed as the mass of the gas phase in flash evaporator j; Expressed as the steam mass flow rate of flash evaporator j;

[0104] The preheater module is represented by: Q preheat,k =U k *A k *(T steam -T feed,k )(k=1,2,3); where Q preheat,k Expressed as the heat transfer rate of the preheater k; U k Expressed as the heat transfer coefficient of the preheater k; A k Expressed as the heat transfer area of ​​the preheater k; T steam Expressed as the temperature of the steam; T feed,k Expressed as the feed temperature of preheater k;

[0105] The condensate tank module is represented by: M condensed,l =f condensation_rat (T steam,l )(l=1,2,3,4); where M condensed,l Expressed as the condensate production rate of the condensate tank l; f condensation_rat () is expressed as the condensate production rate function; T steam,l Expressed as the flow rate of the condensate tank l;

[0106] By combining the modules of the above components, an integrated environmental model can be established to describe the dynamic evolution of the entire alumina production evaporation process and solved using the Runge–Kutta method as f().

[0107] Specifically, the correlation coefficient vector of each data in the initial matrix is ​​solved according to the discrete equation, and the evaporation variable correlation matrix is ​​established; refer to Table 2, which gives a table of the evaporation variable correlation matrix in this application example;

[0108] Table 2. Values ​​of correlation matrix of evaporation variables

[0109] Evaporator temperature Feed temperature Steam production Feed mass flow rate Evaporator temperature 1.00 0.85 0.60 0.72 Feed temperature 0.85 1.00 0.55 0.68 Steam production 0.60 0.55 1.00 0.45 Feed mass flow rate 0.72 0.68 0.45 1.00

[0110] Specifically, the evaporation variable correlation matrix is ​​ranked by significance to obtain the evaporation variable significance ranking result; the significance calculation process according to the data in Table 2 is:

[0111] Evaporator temperature: 1.00+0.85+0.60+0.72=3.17;

[0112] Feed temperature: 0.85+1.00+0.55+0.68=3.08;

[0113] Steam volume: 0.72+0.68+0.45+1.00=2.85;

[0114] Feed mass flow rate: 0.60+0.55+0.100+0.45=2.60;

[0115] The results of the above significance ranking are given in Table 3;

[0116] Table 3. Significance ranking results

[0117] Ranking Evaporation variable Sum of correlation coefficients 1 Evaporator temperature 3.17 2 Feed temperature 3.08 3 Steam production 2.85 4 Feed mass flow rate 2.60

[0118] Table 3 shows the four evaporation variables, namely, evaporator temperature, feed temperature, steam production, and feed mass flow rate. The sum of the correlation coefficients of evaporator temperature is 3.17, the highest, and therefore ranks first in the significance ranking. The sum of the correlation coefficients of feed temperature is 3.08, closely following, ranking second. The sum of the correlation coefficients of steam production is 2.85, ranking third. The sum of the correlation coefficients of feed mass flow rate is 2.60, ranking fourth.

[0119] Specifically, steady-state variables are selected according to the significance ranking results of the evaporation variables combined with the marking content of the initial matrix; wherein the selection of the steady-state variables can be screened by a threshold method;

[0120] Specifically, the steady-state correlation coefficient matrix is ​​obtained by using a correlation analysis method according to the result of the steady-state variable selection.

[0121] In the embodiment of the present application, a method for analyzing the correlation of steady-state variables in the evaporation process of aluminum oxide production is adopted to analyze the collected historical data; the method obtains the variable correlation characteristics in the initial data matrix through discrete equations, and performs significance sorting according to the variable correlation characteristics to obtain the correlation differences between the variables, and selects steady-state variables in combination with the initial matrix; correlation analysis is performed on the selected steady-state variables to obtain the final steady-state variable correlation coefficient matrix; the method can accurately analyze the functional relationship between the various variables in the evaporation process of aluminum oxide production, and ensure that the selected variables are conducive to improving the accuracy of subsequent steady-state detection.

[0122] Specifically, according to the correlation analysis of the steady-state variables in the historical data, the influence state between the variables in the evaporation process of alumina production can be obtained; next, steady-state detection is performed according to the analysis results, and this function is implemented by the steady-state detection unit of the system. Figure 1 In step S50, the process is as follows:

[0123] Specifically, a multivariate steady-state evaluation index corresponding to the timestamp is obtained according to the marked standardized related historical data; wherein the multivariate steady-state evaluation index is expressed as:

[0124]

[0125] Among them, SI(t) represents the steady-state index at time t; n represents the total number of evaporation variables; X i (t) represents the parameter value of the i-th evaporation variable at time t; σ i (t) represents the mean value of the ith evaporation variable at the tth time; μ i (t) represents the variance of the i-th evaporation variable at time t;

[0126] In order to achieve steady-state detection of various variables in the evaporation process of aluminum oxide production, the embodiment of the present application adopts a multivariable steady-state evaluation index for evaluation; the multivariable steady-state evaluation index is calculated according to different time nodes in the historical data, so as to achieve more accurate variable monitoring and steady-state detection.

[0127] Specifically, a steady-state detection function is constructed according to the multivariate steady-state evaluation index; wherein the steady-state detection function is expressed as:

[0128]

[0129] Wherein, SS() represents a steady-state detection function; X(t) represents a vector of evaporation variables of the aluminum oxide during the production evaporation process; AnoD() represents a steady-state anomaly detection algorithm;

[0130] Combining the steady-state correlation coefficient matrix with the standardized historical evaporation variables to construct an extended variable set;

[0131] Using the extended variable set to train a random forest classifier; wherein the output layer of the random forest classifier uses the multivariate steady-state evaluation index as a judgment criterion;

[0132] In the embodiment of the present application, three main operating parameters, namely steam flow, pressure and raw liquid flow, are selected as the basis for process steady-state judgment. The actual evaporation process is sampled once every 5 minutes, 240 sets of data are collected within 20 consecutive hours, and the steady-state detection unit is used to perform steady-state judgment; wherein, referring to Figure 4 ,exist Figure 4 The actual values ​​of the new steam flow, pressure and raw liquid flow during the actual measurement are given in Figure 5 The steady-state test results of the above three variables are given in Figure 5 A non-zero ordinate indicates that the moment is in a steady state (in order to distinguish the three detection results, the two monostable detection results are respectively increased by 0.2 and 0.3), and 0 indicates that the corresponding moment is in a non-steady state.

[0133] In order to further illustrate the steady-state detection results of the evaporation process of alumina production under the interaction of multiple variables, Figure 6 The test results within 20 hours are given in.

[0134] In the embodiment of the present application, a steady-state detection method is adopted by establishing a steady-state detection function; the function performs steady-state detection on evaporation variables at different time points through a steady-state anomaly detection algorithm, and establishes a multivariate steady-state evaluation index through historical data in the function, and the index accurately evaluates the evaporation variables; the method can not only provide an accurate evaluation of the steady state of the evaporation variable, but also accurately understand the changes in the evaporation variables at a specific time.

[0135] Specifically, the evaporation variable is regulated by the variable control unit of the system, corresponding to the above-mentioned step S60; see Figure 7 The specific control process is as follows:

[0136] Obtain the original liquid parameters and evaporation variables at the current moment; set the control target of the evaporation process;

[0137] Wherein, the control objectives include: maximum product concentration and maximum evaporation efficiency; using evaporation variable optimization algorithm to obtain optimized evaporation variables;

[0138] The steps of the evaporation variable optimization algorithm include:

[0139] Step 1: taking the evaporation variable as the initial optimal evaporation variable;

[0140] Step 2: evaluating the correlation of the initial optimal evaporation variables through the steady-state correlation coefficient matrix;

[0141] The evaluation process of step 2 includes:

[0142] According to the strength and direction of each relationship in the steady-state correlation coefficient matrix; the value range of the correlation coefficient is from -1 to +1, close to +1 indicates a strong positive correlation, close to -1 indicates a strong negative correlation; close to 0 indicates no correlation;

[0143] The values ​​of the initial optimal evaporation variables are combined and calculated according to the median of the steady-state correlation coefficient matrix.

[0144] Step 3: Taking the highest value of the correlation as the parent generation, performing crossover and mutation operations to generate new evaporation variables;

[0145] Step 4: updating the initial optimal evaporation variable according to the new evaporation variable;

[0146] Step 5: Repeat step 2 to step 4 until the stop condition is met to obtain the optimized evaporation variable.

[0147] Performing steady-state detection on the optimized evaporation variable; when the optimized evaporation variable is in a steady state and can achieve the control target, obtaining the optimal evaporation variable; otherwise repeating the evaporation variable optimization algorithm;

[0148] In this embodiment, according to the above-mentioned four-effect evaporation process, it takes about 90 minutes for the current data, so 60 consecutive groups of evaporation variables are selected for coordination in this evaporation process; Table 4 shows the comparison between the coordination results of the variable control unit and the measured results.

[0149] Table 4. Comparison of evaporation variable coordination results and measured values

[0150]

[0151] The table shows the comparison between the coordination results of the evaporation variables and the actual measured values; the coordinated evaporation variables listed in the table are representative process control variables in the entire four-effect evaporation process; it can be seen from the table that the error between the method proposed by the present invention and the actual detection value is small; for example, the actual measured value of the steam flow rate of evaporator 3 in the table is 20.050, and the value after coordination of the present invention is 19.537, with an error of 0.513; the actual measured value of the preheater pressure is 0.520, and the value after coordination of the present invention is 0.517, with an error of 0.003.

[0152] In addition, according to the embodiments of the present application Figure 6 Steady-state detection of the evaporation process of alumina production in the middle to make variable adaptive adjustments; Figure 8 The results show that after coordination, the detection of the evaporation process in the continuous time period is in a stable state.

[0153] In the embodiment of the present application, a method for adaptively coordinating evaporation variables in an evaporation process is adopted to adaptively adjust the evaporation variables; this method achieves the control target without losing stability when adjusting the evaporation variables, and obtains the optimal evaporation variables through multiple iterative optimizations to meet the set control targets, thereby ensuring the quality of alumina production.

[0154] Specifically, the feedback unit of the system feeds back the parameter changes and evaporation conditions during the alumina production evaporation process to relevant staff.

[0155] In the embodiment of the present application, the method of the present invention is combined with the system to realize the steady-state detection of aluminum oxide in the four-effect evaporation process and the coordination of variables; the present invention mainly starts from the following main aspects; first, the relevant historical data of aluminum oxide in multiple production evaporation processes are collected according to the timestamp, and an evaporation variable analysis method is used to perform correlation analysis on the collected historical data; the analysis method obtains the variable correlation characteristics in the initial data matrix through discrete equations, and performs significance sorting according to the variable correlation characteristics to obtain the relevant differences between the variables, and selects the steady-state variables in combination with the initial matrix; according to the selection results, a steady-state correlation coefficient matrix is ​​obtained; this method helps to optimize the selection process of evaporation variables, ensuring that the selected variables are conducive to improving the subsequent steady-state detection Precision; secondly, the present invention proposes a steady-state detection method, which establishes a steady-state detection function, and the function performs steady-state detection on evaporation variables at different time points through a steady-state anomaly detection algorithm, and establishes a multivariate steady-state evaluation index in the function through historical data, and the index can accurately evaluate the evaporation variables; this method can not only provide an accurate evaluation of the steady state of the evaporation variables, but also accurately understand the changes in the evaporation variables at a specific time; finally, a method for adaptively coordinating evaporation variables in the evaporation process is adopted to adaptively adjust the evaporation variables; this method achieves the control target under the condition that the evaporation variables are adjusted without losing stability, and obtains the optimal evaporation variables through multiple iterative optimizations to meet the set control targets, thereby ensuring the quality of alumina production.

[0156] Embodiment 2:

[0157] In Example 1, the method of the present invention is combined with the system to realize the steady-state detection of aluminum oxide in the four-effect evaporation process; in order to further illustrate the feasibility of the method proposed by the present invention, a technical description will be further provided in the embodiment of the present application; the specific detection process is basically the same as that of Example 1, and mainly includes:

[0158] Acquire relevant historical data of the alumina during multiple production evaporation processes according to the timestamp;

[0159] Further, data processing is performed on the relevant historical data to obtain standardized relevant historical data;

[0160] Further, according to the observations of relevant staff, steady-state marking was performed;

[0161] Furthermore, the standardized related historical data is subjected to evaporation variable analysis to obtain an evaporation variable correlation matrix; wherein the calculation process of the evaporation variable correlation matrix includes:

[0162] An initial matrix is ​​established based on the standardized related historical data; the initial matrix is ​​expressed as: in, Expressed as the standardized historical stock solution parameter at the mth time point; SV mn It is expressed as the nth standardized historical evaporation variable at the mth time point; Expressed as the normalized historical product concentration at the mth time point;

[0163] Substitute each row of data in the initial matrix into the discrete equation; wherein the discrete equation formula is:

[0164]

[0165] in, Expressed as the normalized historical product concentration of the i-th row in the initial matrix; Represented as the standardized historical stock solution parameter of the i-th row in the initial matrix; SV i It is represented as the normalized historical evaporation variable vector of the ith row in the initial matrix; f() is represented as a nonlinear analytical function; E() is represented as the environmental function of the aluminum oxide in the production evaporation process;

[0166] Solving the correlation coefficient vector of each data in the initial matrix according to the discrete equation, and establishing the evaporation variable correlation matrix;

[0167] Further, the steady-state correlation coefficient matrix is ​​calculated using the evaporation variable correlation matrix;

[0168] Furthermore, steady-state detection is performed; the steady-state detection process includes: obtaining a multivariate steady-state evaluation index corresponding to the timestamp based on the marked standardized related historical data;

[0169] Wherein, the multivariate steady-state evaluation index is expressed as:

[0170]

[0171] Among them, SI(t) represents the steady-state index at time t; n represents the total number of evaporation variables; X i (t) represents the parameter value of the i-th evaporation variable at time t; σ i (t) represents the mean value of the ith evaporation variable at the tth time; μ i (t) represents the variance of the i-th evaporation variable at time t;

[0172] A steady-state detection function is constructed according to the multivariate steady-state evaluation index; wherein the steady-state detection function is expressed as:

[0173]

[0174] Wherein, SS() represents a steady-state detection function; X(t) represents a vector of evaporation variables of the aluminum oxide during the production evaporation process; AnoD() represents a steady-state anomaly detection algorithm;

[0175] Furthermore, adaptive evaporation variable coordination is performed in the evaporation process according to the detection results; wherein the process of adaptive evaporation variable coordination in the evaporation process includes: obtaining the stock solution parameters and evaporation variables at the current moment; setting the control target of the evaporation process; wherein the control target includes: maximum product concentration and maximum evaporation efficiency; using an evaporation variable optimization algorithm to obtain optimized evaporation variables; wherein the steps of the evaporation variable optimization algorithm include:

[0176] Step 1: taking the evaporation variable as the initial optimal evaporation variable;

[0177] Step 2: evaluating the correlation of the initial optimal evaporation variables through the steady-state correlation coefficient matrix;

[0178] Step 3: Taking the highest value of the correlation as the parent generation, performing crossover and mutation operations to generate new evaporation variables;

[0179] Step 4: updating the initial optimal evaporation variable according to the new evaporation variable;

[0180] Step 5: Repeat step 2 to step 4 until the stop condition is met to obtain the optimized evaporation variable.

[0181] The optimized evaporation variable is subjected to steady-state detection; when the optimized evaporation variable is in steady state and can achieve the control target, the optimal evaporation variable is obtained; otherwise, the evaporation variable optimization algorithm is repeated.

[0182] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A steady-state detection method for the evaporation process of aluminum oxide production, characterized in that: include: The relevant historical data of the alumina in multiple production evaporation processes are obtained according to the timestamp; wherein the relevant historical data include: historical raw liquid parameters, historical evaporation variables, historical evaporation efficiency and historical product concentration at the outlet of the final flash evaporator; data processing is performed on the relevant historical data to obtain standardized relevant historical data; the historical evaporation variables are marked; wherein the content of the mark includes: 1 represents a stable state; 0 represents an unsteady state; evaporation variable analysis is performed on the standardized relevant historical data to obtain an evaporation variable correlation matrix; a steady-state correlation coefficient matrix is ​​calculated; and the current A steady-state detection is performed on the evaporation process of the previous alumina production to obtain a steady-state detection result; an adaptive evaporation variable coordination is performed on the evaporation process according to the steady-state detection result; wherein the process of adaptive evaporation variable coordination of the evaporation process includes: obtaining the raw liquid parameters and evaporation variables at the current moment; setting the control target of the evaporation process; wherein the control target includes: maximum product concentration and maximum evaporation efficiency; using an evaporation variable optimization algorithm to obtain an optimized evaporation variable; a steady-state detection is performed on the optimized evaporation variable; when the optimized evaporation variable is in a steady state and can achieve the control target, the optimal evaporation variable is obtained; otherwise, the evaporation variable optimization algorithm is repeated.

2. The steady-state detection method for the evaporation process of aluminum oxide production according to claim 1, characterized in that: The data processing includes: Processing abnormal values ​​and missing values ​​in the relevant historical data to obtain first relevant historical data; Normalizing the first relevant historical data to obtain second relevant historical data; The second relevant historical data is standardized to obtain the standardized relevant historical data.

3. The steady-state detection method for the evaporation process of aluminum oxide production according to claim 1, characterized in that: The process of calculating the steady-state correlation coefficient matrix includes: Establish an initial matrix based on the standardized relevant historical data; substitute each row of data in the initial matrix into the discrete equation; Solving the correlation coefficient vector of each data in the initial matrix according to the discrete equation, and establishing the evaporation variable correlation matrix; Performing significance sorting on the evaporation variable correlation matrix to obtain a significance sorting result of the evaporation variable; Selecting steady-state variables according to the significance ranking results of the evaporation variables combined with the marking content of the initial matrix; The steady-state correlation coefficient matrix is ​​obtained by using a correlation analysis method according to the result of the steady-state variable selection.

4. The method for steady-state detection of an aluminum oxide production evaporation process according to claim 1, characterized in that: The steady-state detection process includes: obtaining a multivariate steady-state evaluation index corresponding to the timestamp based on the marked standardized related historical data; constructing a steady-state detection function based on the multivariate steady-state evaluation index; and using the steady-state detection function to perform a steady-state detection on the alumina production evaporation process to obtain the steady-state detection result.

5. The steady-state detection method for the evaporation process of aluminum oxide production according to claim 1, characterized in that: The steps of the evaporation variable optimization algorithm include: Step 1: taking the evaporation variable as the initial optimal evaporation variable; Step 2: evaluating the correlation of the initial optimal evaporation variables through the steady-state correlation coefficient matrix; Step 3: Taking the highest value of the correlation as the parent generation, performing crossover and mutation operations to generate new evaporation variables; Step 4: updating the initial optimal evaporation variable according to the new evaporation variable; Step 5: Repeat step 2 to step 4 until the stop condition is met to obtain the optimized evaporation variable.

6. A steady-state detection system for the evaporation process of aluminum oxide production, characterized in that: The system includes: a monitoring unit, a data acquisition unit, a data processing unit, a steady-state variable analysis unit, a steady-state detection unit, a variable control unit and a feedback unit; wherein the monitoring unit is used to monitor the aluminum oxide during the production evaporation process; the data acquisition unit is used to collect various data of the aluminum oxide production evaporation process; the data processing unit is used to process the data collected by the data acquisition unit; the steady-state variable analysis unit is used to perform correlation analysis on the variables of the aluminum oxide during the production evaporation process; the steady-state detection unit is used to perform steady-state detection on various variables in the aluminum oxide production evaporation process; the variable control unit is used to perform steady-state adjustment on the variables; and the feedback unit is used to feed back the parameter changes and evaporation conditions in the aluminum oxide production evaporation process to relevant staff.

7. A steady-state detection system for an aluminum oxide production evaporation process according to claim 6, characterized in that: The data processing unit comprises: Processing abnormal values ​​and missing values ​​in the relevant historical data to obtain first relevant historical data; Normalizing the first relevant historical data to obtain second relevant historical data; The second relevant historical data is standardized to obtain standardized relevant historical data.

8. The steady-state detection system for the evaporation process of aluminum oxide production according to claim 6, characterized in that: The steady-state variable analysis unit includes: establishing an initial matrix based on standardized relevant historical data; Solving the correlation coefficient vector of each data in the initial matrix according to the discrete equation, and establishing the evaporation variable correlation matrix; A steady-state correlation coefficient matrix is ​​calculated based on the evaporation variable correlation matrix.

9. A steady-state detection system for an aluminum oxide production evaporation process according to claim 8, characterized in that: The steady-state correlation coefficient matrix includes: performing significance sorting on the evaporation variable correlation matrix to obtain a significance sorting result of the evaporation variable; performing steady-state variable selection according to the significance sorting result of the evaporation variable combined with the marking content of the initial matrix; and obtaining the steady-state correlation coefficient matrix by using a correlation analysis method according to the result of the steady-state variable selection.

10. The steady-state detection system for the evaporation process of aluminum oxide production according to claim 6, characterized in that: The steady-state detection unit includes: obtaining a multivariate steady-state evaluation index corresponding to a timestamp according to the marked standardized related historical data; constructing a steady-state detection function according to the multivariate steady-state evaluation index; and performing a steady-state detection on the alumina production evaporation process using the steady-state detection function to obtain a steady-state detection result.

Images