A by-product magnesium sulfate refining method and system based on recrystallization

Abstract

The present application relates to the field of chemical purification technology, and proposes a by-product magnesium sulfate refining method and system based on recrystallization, which comprises: performing quantitative analysis and statistics on raw material component detection data to obtain impurity distribution data; accordingly, low-temperature hot melting process parameter search is carried out to determine the hot melting temperature interval and generate hot melting control instructions; the calcium ion concentration data after the execution of the instructions is analyzed to identify abnormal fluctuations, and calcium removal correction parameters are obtained through seed induction logic judgment; the parameters are fused and fitted with the preset crystallization data to generate a two-stage gradient cooling curve, dynamically compensate the crystallization temperature deviation, and obtain the crystallization process control data; the corresponding crystal particle size distribution is evaluated to form the crystal growth quality characterization; relying on the results, the mother liquor reflux material balance is verified to generate feedback data, which is used to update the impurity distribution data to realize full-process closed-loop regulation and control; the present application can improve the efficiency of a by-product magnesium sulfate refining method based on recrystallization.

Description

Technical Field

[0001] This invention relates to the field of chemical purification technology, and in particular to a method and system for refining magnesium sulfate by-products based on recrystallization. Background Technology

[0002] In the recrystallization and refining of by-product magnesium sulfate, existing processes are unable to perform precise analysis of the components of raw material impurities, cannot dynamically match low-temperature thermal dissolution parameters according to the distribution of impurities, lack real-time monitoring and dynamic correction methods for the removal of key impurities such as calcium ions, and often adopt a single cooling mode during the crystallization stage, which is prone to problems such as temperature deviation, uneven crystal particle size, and secondary nucleation, making it difficult to guarantee the overall refining quality and process stability.

[0003] Traditional refining processes lack a complete data feedback and closed-loop control system. Material deviations during the mother liquor recycling process cannot be corrected in a timely manner, and process parameters cannot be adaptively adjusted according to the state of raw materials and production conditions. This results in low refining efficiency and insufficient purity of the by-product magnesium sulfate. Therefore, improving the overall performance and product quality of the recrystallization refining of by-product magnesium sulfate has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for refining magnesium sulfate by-products based on recrystallization, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for refining magnesium sulfate by-products based on recrystallization, comprising:

[0006] D1. Quantitatively analyze and statistically analyze the composition detection data of the raw materials to obtain the impurity distribution data of the raw materials;

[0007] D2. Based on the impurity distribution data, perform a parameter space search on the low-temperature hot melting process of the raw material to obtain the hot melting temperature range of the raw material, and generate a hot melting control command for the raw material according to the hot melting temperature range.

[0008] D3. Perform trend analysis on the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material.

[0009] D4. The calcium removal correction parameters and preset crystallization data are fused together, and the fused data is curve fitted to obtain two gradient cooling curves of the raw material. Based on the two gradient cooling curves, the real-time temperature deviation of the raw material is dynamically compensated to obtain the crystallization process control data of the raw material.

[0010] D5. The crystal grain size distribution data under the control of the crystallization process is graded and evaluated to obtain the crystal growth quality assessment characterization of the raw material.

[0011] D6. Based on the crystal growth quality assessment characterization, perform material balance verification on the mother liquor circulation reflux data of the raw material to obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

[0012] In a preferred embodiment, the step of quantitatively analyzing and statistically analyzing the component detection data of the raw material to obtain the impurity distribution data of the raw material includes:

[0013] The composition detection data of the raw materials are collected, and multi-dimensional feature extraction is performed on the composition detection data to obtain the impurity feature vector of the raw materials;

[0014] Principal component screening is performed on the impurity feature vector to obtain the core impurity component data of the raw material;

[0015] The core impurity component data is discretized and segmented to obtain the impurity concentration range distribution sequence of the raw material;

[0016] Frequency statistics were performed on the distribution sequence of impurity concentration ranges to obtain the impurity frequency distribution histogram of the raw material;

[0017] Based on the impurity frequency distribution histogram, cluster analysis is performed on the impurity distribution pattern of the raw material to obtain impurity aggregation characteristic data of the raw material.

[0018] By fusing the impurity aggregation characteristic data with the core impurity component data, the impurity distribution data of the raw material is obtained.

[0019] In a preferred embodiment, the step of performing a parameter space search on the low-temperature hot melting process of the raw material based on the impurity distribution data to obtain the hot melting temperature range of the raw material, and generating a hot melting control command for the raw material according to the hot melting temperature range, includes:

[0020] The impurity distribution data is matched and filtered with a preset process parameter library to obtain the candidate parameter space for the raw material;

[0021] An orthogonal ergodic analysis is performed on the candidate parameter space to obtain the parameter combination evaluation data of the raw materials;

[0022] Based on the evaluation data of the parameter combination, the candidate parameter space is pruned and optimized to obtain the optimal parameter subspace of the raw material;

[0023] Based on the optimal parameter subspace, the tolerance boundary of the impurity distribution data is determined to obtain the thermal dissolution constraint conditions of the raw material.

[0024] The hot melt constraint conditions are mapped to interval boundary data, and the interval boundary data is used as the hot melt temperature range of the raw material.

[0025] Based on the hot melting temperature range, the heating rate data in the optimal parameter subspace are arranged in time sequence to obtain the hot melting time sequence parameter sequence of the raw material.

[0026] The hot melting timing parameter sequence is converted into an instruction encoding format to obtain the hot melting control instruction for the raw material.

[0027] In a preferred embodiment, the step of performing trend analysis on the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and performing seed-induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material, includes:

[0028] The calcium ion concentration monitoring data after the execution of the heat dissolution control command is divided into time-series sliding window segments to obtain local concentration change segments of the calcium ion concentration monitoring data. The slope consistency of the local concentration change segments is checked to obtain the concentration change trend sequence of the calcium ion concentration monitoring data.

[0029] Based on the concentration change trend sequence, identify abrupt breaks in the calcium ion concentration monitoring data;

[0030] The mutation fragments were screened by both duration and amplitude to obtain abnormal fluctuation data of the calcium ion concentration monitoring data.

[0031] Based on the fluctuation start time and fluctuation decay rate characteristics of the abnormal fluctuation data, the seed-induced resistance characteristics of the raw material are obtained, and the seed-induced resistance characteristics are judged by logic threshold to obtain the seed-induced failure identifier of the raw material.

[0032] Based on the seed-induced failure indicator, the energy accumulation of the abnormal fluctuation data is analyzed in reverse to obtain the calcium ion supersaturation deviation of the raw material.

[0033] The deviation of calcium ion supersaturation is converted into a concentration control step size, which is then used as the calcium removal correction parameter for the raw material.

[0034] In a preferred embodiment, the process involves fusing the calcium removal correction parameters with preset crystallization data, performing curve fitting on the fused data to obtain two gradient cooling curves for the raw material, and dynamically compensating for the real-time temperature deviation of the raw material's crystallization based on these two gradient cooling curves to obtain crystallization process control data for the raw material, including:

[0035] The correlation between the calcium removal correction parameter and the calcium ion-related impurity concentration in the impurity distribution data is calibrated to obtain the temperature threshold of the calcium removal section of the raw material.

[0036] Based on the temperature threshold of the calcium removal section, the cooling range of the raw material is divided into a high-temperature calcium removal section and a low-temperature crystallization section, thus obtaining the section division identifier sequence of the raw material.

[0037] Based on the segment division identifier sequence, the driving characteristics of the calcium removal correction parameter in the high-temperature segment and the driving characteristics of the historical cooling data of the crystallization process control data in the low-temperature segment are extracted respectively to obtain the high-temperature segment driving characteristics and low-temperature segment driving characteristics of the raw material.

[0038] Piecewise curve fitting is performed on the high-temperature driving characteristics and the low-temperature driving characteristics to obtain two gradient cooling curves of the raw material, which include the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment.

[0039] Based on the cooling slope of the high-temperature section and the cooling slope of the low-temperature section, dynamic compensation within the section and smoothing processing between sections are performed on the real-time temperature deviation of the crystallization to obtain the crystallization process control data of the raw material.

[0040] In a preferred embodiment, the step of performing intra-segment dynamic compensation and inter-segment switching smoothing processing on the real-time temperature deviation of crystallization based on the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment, respectively, to obtain the crystallization process control data of the raw material, includes:

[0041] Based on the real-time temperature deviation of crystallization and the segment division identifier sequence, extract the target cooling slope corresponding to the two gradient cooling curves;

[0042] Numerical analysis was performed on the real-time temperature deviation of the crystallization to obtain the deviation amplitude sequence of the raw material;

[0043] Using the impurity concentration gradient in the impurity distribution data as a constraint factor for the intra-segment compensation intensity, the compensation weight coefficient of the raw material is obtained.

[0044] The corrected cooling slope of the raw material is calculated based on the target cooling slope, the deviation amplitude sequence, and the compensation weighting coefficient.

[0045] Based on the temperature threshold of the calcium removal section, the modified cooling slope of the adjacent sections is transferred by a moving average to obtain the crystallization process control data of the raw material.

[0046] In a preferred embodiment, the formula for calculating the corrected cooling slope is as follows:

[0047] ;

[0048] in, The first of the corrected cooling slope The slope at each moment The first of the target cooling slopes The slope at each moment The compensation weight coefficient is... The first of the real-time temperature deviations of the crystallization Amplitude at each moment, The maximum value in the deviation amplitude sequence. It is a very small constant.

[0049] In a preferred embodiment, the step of classifying and evaluating the crystal grain size distribution data under the influence of the crystallization process control data to obtain a characterization of the crystal growth quality of the raw material includes:

[0050] The crystal particle size distribution data under the control of the crystallization process is divided into particle size grades according to the particle size range, and the number of crystal particles in different particle size grades is accumulated and statistically analyzed to obtain the particle size distribution histogram of the raw material.

[0051] Extract the quantity decay gradient between adjacent particle size grades in the particle size distribution histogram to obtain the crystal growth rate consistency coefficient of the raw material;

[0052] Based on the crystal growth rate consistency coefficient, outlier detection is performed on the abnormal particle size peaks in the particle size distribution histogram to obtain the secondary nucleation intensity identifier of the raw material.

[0053] By fusing the crystal growth rate consistency coefficient with the secondary nucleation intensity identifier and mapping the fused data to a quality level, a quality assessment characterization of the raw material's crystal growth is obtained.

[0054] In a preferred embodiment, the step of performing material balance verification on the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization to obtain feedback data of the raw material, and updating the impurity distribution data of the raw material based on the feedback data, includes:

[0055] The mother liquor reflux data of the raw material is time-series sliced ​​to obtain the mother liquor reflux feature sequence of the raw material;

[0056] Based on the crystal growth quality assessment characterization, the mother liquor reflux characteristic sequence is screened by benchmarking to obtain the impurity residue deviation vector of the raw material;

[0057] Based on the impurity residue deviation vector, the mother liquor reflux characteristic sequence is corrected for deviation to obtain the feedback data of the raw material;

[0058] The feedback data and the impurity distribution data of the raw material are concatenated to obtain the updated impurity distribution data of the raw material.

[0059] To address the aforementioned problems, the present invention also provides a recrystallization-based system for refining magnesium sulfate byproducts, the system comprising:

[0060] The impurity distribution analysis module is used to perform quantitative analysis and statistical analysis on the component detection data of the raw material to obtain the impurity distribution data of the raw material.

[0061] The hot melting parameter optimization module is used to perform parameter space search on the low-temperature hot melting process of the raw material based on the impurity distribution data, obtain the hot melting temperature range of the raw material, and generate hot melting control instructions for the raw material according to the hot melting temperature range.

[0062] The calcium ion anomaly identification module is used to perform trend analysis on the calcium ion concentration monitoring data after the execution of the hot melting control command, obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material.

[0063] The two-segment crystallization curve generation module is used to fuse the calcium removal correction parameters and preset crystallization data, and to perform curve fitting on the fused data to obtain two-segment gradient cooling curves of the raw material. Based on the two-segment gradient cooling curves, the real-time temperature deviation of the raw material crystallization is dynamically compensated to obtain the crystallization process control data of the raw material.

[0064] The crystal growth grading and evaluation module is used to grade and evaluate the crystal grain size distribution data under the control of the crystallization process, and obtain the crystal growth quality evaluation characterization of the raw material.

[0065] The material balance feedback update module is used to perform material balance verification on the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization, obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

[0066] Compared with the prior art, the present invention has the following beneficial effects:

[0067] 1. This invention, through full-process data analysis and closed-loop control mechanism, can accurately analyze the distribution of impurities in raw materials, and determine suitable low-temperature hot melting process parameters by combining parameter space search. It can effectively stabilize the operation of the hot melting process, accurately identify abnormal fluctuations in calcium ion concentration and output calcium removal correction parameters, and significantly improve the impurity removal effect of by-product magnesium sulfate and the purity of raw material refining.

[0068] 2. This invention employs a crystallization control method that combines two-stage gradient cooling curves with dynamic temperature deviation compensation, which ensures uniform crystal growth. Simultaneously, it achieves iterative data updates through crystal quality assessment and mother liquor reflux material balance verification, continuously optimizing the operating parameters of each process, improving the overall production stability and the quality of magnesium sulfate crystal products, and strengthening the continuous operation capability of the entire refining process. Attached Figure Description

[0069] Figure 1 This is a schematic flow diagram of a method for refining magnesium sulfate by-product based on recrystallization, provided in an embodiment of the present invention.

[0070] Figure 2 A functional block diagram of a recrystallization-based by-product magnesium sulfate refining system provided in an embodiment of the present invention;

[0071] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0072] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0073] This application provides a method for refining by-product magnesium sulfate based on recrystallization. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for refining by-product magnesium sulfate based on recrystallization can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0074] Reference Figure 1The diagram shown is a flow chart of a method for refining magnesium sulfate by-product based on recrystallization according to an embodiment of the present invention. In this embodiment, the method for refining magnesium sulfate by-product based on recrystallization includes:

[0075] D1. Quantitatively analyze and statistically analyze the composition detection data of the raw materials to obtain the impurity distribution data of the raw materials;

[0076] In this embodiment of the invention, the step of quantitatively analyzing and statistically analyzing the component detection data of the raw material to obtain the impurity distribution data of the raw material includes:

[0077] The composition detection data of the raw materials are collected, and multi-dimensional feature extraction is performed on the composition detection data to obtain the impurity feature vector of the raw materials;

[0078] Principal component screening is performed on the impurity feature vector to obtain the core impurity component data of the raw material;

[0079] The core impurity component data is discretized and segmented to obtain the impurity concentration range distribution sequence of the raw material;

[0080] Frequency statistics were performed on the distribution sequence of impurity concentration ranges to obtain the impurity frequency distribution histogram of the raw material;

[0081] Based on the impurity frequency distribution histogram, cluster analysis is performed on the impurity distribution pattern of the raw material to obtain impurity aggregation characteristic data of the raw material.

[0082] By fusing the impurity aggregation characteristic data with the core impurity component data, the impurity distribution data of the raw material is obtained.

[0083] Using industrial online component detection equipment, the entire batch of raw material components was tested at the magnesium sulfate feed point. The raw material was magnesium sulfate, a by-product of cellulose acetate production, obtained by pretreatment of wastewater from the esterification and neutralization processes of cellulose acetate. Its main component was magnesium sulfate, containing impurities such as calcium ions, trace organic matter, and sodium salts. It needed to be refined and purified through recrystallization. Complete component detection data was continuously acquired at a fixed detection frequency. Multidimensional feature extraction was performed on the collected component detection data to identify the specific types of impurities in the raw material, the actual concentration of each impurity, the mass percentage of each impurity in the overall material, and the physical form of the impurities in the magnesium sulfate raw material system. All the identified feature information was then orderly combined according to a unified and regular arrangement rule to form a structurally complete and clearly defined impurity feature vector.

[0084] Principal component analysis (PCA) was performed on the generated impurity feature vectors. The PCA algorithm was used to compare the content of each feature in the impurity feature vector with its impact on the process. Substances that directly affect the subsequent low-temperature hot melting, recrystallization, and impurity removal processes were identified as core impurities. Substances with extremely low content that do not interfere with the entire refining process were identified as trace impurities. The feature information corresponding to all trace impurities was removed, and all relevant information corresponding to the core impurities was retained. After screening, accurate core impurity component data was obtained.

[0085] Discretization and segmentation processing were carried out on the core impurity component data after it was sorted out. A unified concentration range division standard was set in combination with the actual control requirements of the by-product magnesium sulfate refining process for various impurity concentrations. According to the established division standard, the continuously changing impurity concentration values ​​were divided into multiple independent concentration segments with clear ranges. The concentration value of each core impurity was assigned to the corresponding concentration segment. All information was sorted and arranged in sequence according to the impurity type and the corresponding concentration segment. After sorting, a clear impurity concentration range distribution sequence was formed.

[0086] Frequency statistics were performed on the generated impurity concentration range distribution sequence. The total number of times the corresponding impurity data appeared in each concentration range was counted one by one. During the statistical process, the impurities were strictly classified and distinguished according to their types to ensure that the statistical data of different impurities were independent and not confused. The frequency information obtained from the statistics was matched with the corresponding concentration range one by one. The data was arranged in a way that the horizontal axis represents the concentration range and the vertical axis represents the frequency of occurrence. After the arrangement was completed, an impurity frequency distribution histogram was generated to intuitively present the data distribution pattern.

[0087] Cluster analysis was conducted based on the completed impurity frequency distribution histogram. Different distribution morphology categories were divided according to the overall aggregation state of impurities and the density of the same impurity in different concentration ranges. Impurity data with highly similar distribution patterns and consistent aggregation states were grouped into the same category. The aggregation forms and distribution characteristics of impurities corresponding to different categories were distinguished, and the relevant features of each category were fully recorded. After the recording was completed, impurity aggregation feature data that can reflect the impurity aggregation pattern were obtained.

[0088] The obtained impurity aggregation characteristic data is fully integrated with the previously sorted core impurity component data. The information reflecting the impurity aggregation state and the information reflecting the basic components and concentrations of the impurities are completely spliced ​​and integrated to ensure that the correspondence between the two types of data is accurate. The integrated information fully covers all the contents of the core impurities inside the raw material, such as composition, concentration, distribution range, frequency of occurrence and aggregation form. After the integration process is completed, the final impurity distribution data that meets the process requirements is obtained.

[0089] The beneficial effects are that this entire data analysis process can comprehensively and meticulously dissect the relevant information of various core impurities inside the by-product magnesium sulfate raw material, completely restore the composition, concentration distribution and aggregation state of the impurities, and provide real and reliable data support for the subsequent determination of low-temperature hot dissolution process parameters, removal of calcium ion impurities and control of recrystallization process. It ensures the rationality of the matching of by-product magnesium sulfate refining process parameters from the source, effectively avoids the abnormal process operation caused by incomplete understanding of raw material impurities, improves the overall stability of subsequent impurity removal and crystallization operations, and meets the actual application needs of magnesium sulfate salt refining production.

[0090] D2. Based on the impurity distribution data, perform a parameter space search on the low-temperature hot melting process of the raw material to obtain the hot melting temperature range of the raw material, and generate a hot melting control command for the raw material according to the hot melting temperature range.

[0091] In this embodiment of the invention, the step of performing a parameter space search on the low-temperature hot melting process of the raw material based on the impurity distribution data to obtain the hot melting temperature range of the raw material, and generating a hot melting control command for the raw material according to the hot melting temperature range, includes:

[0092] The impurity distribution data is matched and filtered with a preset process parameter library to obtain the candidate parameter space for the raw material;

[0093] An orthogonal ergodic analysis is performed on the candidate parameter space to obtain the parameter combination evaluation data of the raw materials;

[0094] Based on the evaluation data of the parameter combination, the candidate parameter space is pruned and optimized to obtain the optimal parameter subspace of the raw material;

[0095] Based on the optimal parameter subspace, the tolerance boundary of the impurity distribution data is determined to obtain the thermal dissolution constraint conditions of the raw material.

[0096] The hot melt constraint conditions are mapped to interval boundary data, and the interval boundary data is used as the hot melt temperature range of the raw material.

[0097] Based on the hot melting temperature range, the heating rate data in the optimal parameter subspace are arranged in time sequence to obtain the hot melting time sequence parameter sequence of the raw material.

[0098] The hot melting timing parameter sequence is converted into an instruction encoding format to obtain the hot melting control instruction for the raw material.

[0099] The pre-built low-temperature hot melting preset process parameter library is retrieved. This database fully includes all types of process parameters such as hot melting temperature, heating rate, holding time, and liquid-solid ratio under different impurity conditions, as well as the legal value range of each parameter. The acquired impurity distribution data is used as the matching basis. The degree of fit between each set of parameters in the parameter library and the current raw material impurity state is compared one by one. Parameters that do not match the raw material impurity state at all are eliminated. All parameters and parameter combinations that meet the basic process requirements are retained. In this way, the candidate parameter space that is suitable for the current raw material conditions is defined.

[0100] For the defined candidate parameter space, L9(3) is used. 4 The orthogonal array was used to perform traversal operations. The hot melting temperature, heating rate, and holding time were selected as the core analysis dimensions. Each parameter combination in the candidate parameter space was put into the process simulation. The raw material dissolution effect, the preliminary removal effect of impurities, and the energy consumption of each parameter combination were recorded throughout the process. The weighted proportions of the evaluation indicators were 0.5 for dissolution rate, 0.3 for preliminary removal rate of calcium ions, and 0.2 for unit energy consumption. Each parameter combination was comprehensively evaluated according to a unified process evaluation standard. All evaluation results and corresponding parameter information were fully summarized and compiled to form detailed parameter combination evaluation data.

[0101] The candidate parameter space is pruned and optimized based on the generated parameter combination evaluation data. Fixed process qualification criteria are set, and all parameter combinations that do not meet the qualification criteria are eliminated. At the same time, parameters that conflict with each other or cannot work together in actual equipment are removed. Only parameter combinations that have excellent overall performance, are compatible with each other, and can be stably used are retained. After layer-by-layer screening and simplification, the optimal parameter subspace with reasonable parameter matching and suitable for the current raw materials is formed.

[0102] Based on the optimal parameter subspace obtained by screening, and combined with the information such as the concentration and aggregation morphology of various impurities recorded in the impurity distribution data, the impurity tolerance boundary is determined. Combined with the established standard for the impurity tolerance capacity of the low-temperature hot melting process, the limit state under different parameter conditions where impurities will not precipitate and will not affect the hot melting effect is determined. The various restrictions that the raw materials must comply with during the hot melting process are clarified. After sorting out and summarizing all the restrictions, the hot melting constraint conditions that standardize the operation of the entire hot melting process are obtained.

[0103] The completed hot-melting constraints are transformed into interval boundary data that can be directly used for equipment control. This process defines the overall reference temperature range for low-temperature hot-melting as 30~70℃, uses pure water as solvent, has a solid-liquid mass ratio of 1:4~1:8, a stirring speed of 120~180r / min, and a single batch holding time of 20~40min. Combining the strict requirements for the upper and lower temperature limits in the constraints, the lowest and highest temperature values ​​that can be used in the hot-melting process are determined. These two sets of values ​​together constitute a complete temperature range, which is the hot-melting temperature range that is suitable for the current raw material state.

[0104] Based on the established hot-melting temperature range, all parameters related to the heating rate are extracted from the optimal parameter subspace. Following the complete operational sequence of the low-temperature hot-melting process from starting heating, continuous temperature rise to constant temperature maintenance, the heating rates corresponding to different stages are arranged and combined sequentially. This ensures that the rhythm of heating rate change and the operation stage of the temperature range are completely synchronized. After the orderly arrangement is completed, a coherent hot-melting time sequence parameter sequence that conforms to the process flow is formed.

[0105] Following the unified instruction coding rules used by industrial temperature control equipment, with 10 seconds as the minimum time unit, the standard Modbus protocol is used to encode the prepared hot melt timing parameter sequence. The format is converted so that each timing parameter, parameter execution duration, and switching node in the sequence is converted into coded content that the equipment can recognize. This ensures that the coded content corresponds one-to-one with the actual process parameters and that the transmission logic is clear. After all the format conversions are completed, hot melt control instructions that can be directly issued to the equipment on site are generated.

[0106] The beneficial effects are that the process parameter optimization and instruction generation process can accurately screen suitable low-temperature hot melting process parameters based on the actual impurity state of the raw materials, eliminate inferior parameter combinations layer by layer, define a scientific and reasonable hot melting temperature range and heating rhythm, and finally generate standardized and usable equipment control instructions. This allows the low-temperature hot melting process to operate in perfect match with the characteristics of the raw materials, ensuring that the magnesium sulfate raw materials are fully dissolved and impurities are initially removed, while reducing process energy consumption and operating condition fluctuations, improving the operating efficiency and process stability of the low-temperature hot melting process, and laying a good foundation for subsequent calcium removal and crystallization processes.

[0107] D3. Perform trend analysis on the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material.

[0108] In this embodiment of the invention, the step of performing trend analysis on the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and performing seed-induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material, includes:

[0109] The calcium ion concentration monitoring data after the execution of the heat dissolution control command is divided into time-series sliding window segments to obtain local concentration change segments of the calcium ion concentration monitoring data. The slope consistency of the local concentration change segments is checked to obtain the concentration change trend sequence of the calcium ion concentration monitoring data.

[0110] Based on the concentration change trend sequence, identify abrupt breaks in the calcium ion concentration monitoring data;

[0111] The mutation fragments were screened by both duration and amplitude to obtain abnormal fluctuation data of the calcium ion concentration monitoring data.

[0112] Based on the fluctuation start time and fluctuation decay rate characteristics of the abnormal fluctuation data, the seed-induced resistance characteristics of the raw material are obtained, and the seed-induced resistance characteristics are judged by logic threshold to obtain the seed-induced failure identifier of the raw material.

[0113] Based on the seed-induced failure indicator, the energy accumulation of the abnormal fluctuation data is analyzed in reverse to obtain the calcium ion supersaturation deviation of the raw material.

[0114] The deviation of calcium ion supersaturation is converted into a concentration control step size, which is then used as the calcium removal correction parameter for the raw material.

[0115] Continuous calcium ion concentration monitoring data was collected after the thermal dissolution control command was executed. The sliding window size was 60 seconds and the sliding step was 10 seconds. The entire monitoring data was divided into time-series sliding window segments according to a fixed window range and fixed window movement interval. Each independent window captured the concentration information within the corresponding time period. After splitting, multiple interconnected but independent local concentration change segments were obtained. The overall trend of concentration change over time within each local concentration change segment was examined one by one. The rate of concentration change at different time points within the same segment was compared to determine whether the rhythm of concentration change within a single segment remained consistent. The overall trend of change corresponding to all segments was sorted and arranged in sequence, and after continuous summarization, a complete and coherent concentration change trend sequence was formed.

[0116] Following the concentration change trend sequence, the concentration change status is checked segment by segment. The threshold for identifying abrupt fragments is a concentration change rate ≥ 0.02 g / (L・min) and a duration ≥ 15 s. By comparing with the overall trend of stable calcium ion concentration changes under normal operating conditions, segments in the sequence where the concentration change rhythm suddenly deviates from the normal pattern and the change trend shows a clear turning point are identified. These segments that are significantly different from the normal change state are abrupt fragments in the calcium ion concentration monitoring data.

[0117] For all the mutant fragments identified, a dual screening process was conducted based on duration and amplitude. The complete duration from the onset of change to the return to a stable state for each mutant fragment was recorded. At the same time, the overall deviation of the calcium ion concentration within the fragment compared to the normal value was recorded. The data was checked against the standard duration and standard deviation amplitude requirements set by the on-site process. Fragments whose duration and deviation amplitude both exceeded the standard range were retained, while fragments whose indicators all met the normal range were removed. Finally, the abnormal fluctuation data of calcium ion concentration monitoring data were identified.

[0118] The abnormal fluctuation data was analyzed, starting from the specific time point when the fluctuation began, and the overall speed at which the fluctuation gradually weakened until it disappeared was recorded. Combining these two types of information, the actual situation of the seed crystal's induction effect being hindered during the calcium ion precipitation process was judged. After comprehensively summarizing the relevant information, the seed crystal induction resistance characteristics of the raw material were obtained. Magnesium sulfate single crystals with a particle size of 80~120μm were used as seed crystals, added at a ratio of 0.8%~1.5% of the liquid volume. The addition time was after the low-temperature hot melting was completed and before the cooling process was started. According to the pre-set judgment criteria of the process, the level boundaries corresponding to different resistances were divided. The obtained seed crystal induction resistance characteristics were compared with the predetermined level boundaries one by one. Based on the comparison results, it was determined whether the seed crystal induction effect was normal, thereby generating the corresponding seed crystal induction failure indicator.

[0119] By combining the identified seed-induced failure indicators, the actual working state of the seed crystal under the current operating conditions is determined. The overall change in calcium ion concentration deviating from the standard state during the entire process of abnormal fluctuation data is statistically analyzed. The cause of the entire change is deduced by tracing the formation process of concentration fluctuation. The difference between the actual saturation state and the ideal saturation state of calcium ions in the current system is identified. After a complete reverse analysis process, the deviation of calcium ion supersaturation in the raw material is obtained.

[0120] Based on the control rules of calcium removal operations in the refining process, the calculated calcium ion supersaturation deviation is converted into the concentration adjustment range that can be executed in actual production according to the corresponding matching relationship. The change scale to be executed for each adjustment is determined according to the process operation logic. The determined concentration control step size is directly used as the raw material calcium removal correction parameter for this process adjustment.

[0121] The beneficial effects are that this data processing and parameter generation process can accurately capture subtle changes and abnormal fluctuations in calcium ion concentration after the hot melting process, accurately determine the operating status of the seed induction process, derive the actual calcium ion supersaturation deviation based on the actual fluctuation, and convert it into calcium removal correction parameters that can be implemented. This allows for timely process adjustments to address calcium ion exceedances, strengthens the dynamic impurity removal capability of the entire process, ensures stable removal of calcium ion impurities from magnesium sulfate raw materials, reduces the adverse effects of calcium ions on subsequent crystallization processes, and improves the purity of finished products and the overall stability of the process operation.

[0122] D4. The calcium removal correction parameters and preset crystallization data are fused together, and the fused data is curve fitted to obtain two gradient cooling curves of the raw material. Based on the two gradient cooling curves, the real-time temperature deviation of the raw material is dynamically compensated to obtain the crystallization process control data of the raw material.

[0123] In this embodiment of the invention, the process of fusing the calcium removal correction parameters with preset crystallization data, and performing curve fitting on the fused data to obtain two gradient cooling curves for the raw material, and then dynamically compensating for the real-time temperature deviation of the raw material's crystallization based on the two gradient cooling curves, yields crystallization process control data for the raw material, including:

[0124] The correlation between the calcium removal correction parameter and the calcium ion-related impurity concentration in the impurity distribution data is calibrated to obtain the temperature threshold of the calcium removal section of the raw material.

[0125] Based on the temperature threshold of the calcium removal section, the cooling range of the raw material is divided into a high-temperature calcium removal section and a low-temperature crystallization section, thus obtaining the section division identifier sequence of the raw material.

[0126] Based on the segment division identifier sequence, the driving characteristics of the calcium removal correction parameter in the high-temperature segment and the driving characteristics of the historical cooling data of the crystallization process control data in the low-temperature segment are extracted respectively to obtain the high-temperature segment driving characteristics and low-temperature segment driving characteristics of the raw material.

[0127] Piecewise curve fitting is performed on the high-temperature driving characteristics and the low-temperature driving characteristics to obtain two gradient cooling curves of the raw material, which include the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment.

[0128] Based on the cooling slope of the high-temperature section and the cooling slope of the low-temperature section, dynamic compensation within the section and smoothing processing between sections are performed on the real-time temperature deviation of the crystallization to obtain the crystallization process control data of the raw material.

[0129] The method involves performing intra-segment dynamic compensation and inter-segment switching smoothing processing on the real-time temperature deviation of the crystallization based on the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment, respectively, to obtain the crystallization process control data of the raw material, including:

[0130] Based on the real-time temperature deviation of crystallization and the segment division identifier sequence, extract the target cooling slope corresponding to the two gradient cooling curves;

[0131] Numerical analysis was performed on the real-time temperature deviation of the crystallization to obtain the deviation amplitude sequence of the raw material;

[0132] Using the impurity concentration gradient in the impurity distribution data as a constraint factor for the intra-segment compensation intensity, the compensation weight coefficient of the raw material is obtained.

[0133] The corrected cooling slope of the raw material is calculated based on the target cooling slope, the deviation amplitude sequence, and the compensation weighting coefficient.

[0134] Based on the temperature threshold of the calcium removal section, the modified cooling slope of the adjacent sections is transferred by a moving average to obtain the crystallization process control data of the raw material.

[0135] The formula for calculating the corrected cooling slope is as follows:

[0136] ;

[0137] in, The first of the corrected cooling slope The slope at each moment The first of the target cooling slopes The slope at each moment The compensation weight coefficient is... The first of the real-time temperature deviations of the crystallization Amplitude at each moment, The maximum value in the deviation amplitude sequence. It is a very small constant.

[0138] The calcium ion-related impurity concentrations recorded in the calcium removal correction parameters and impurity distribution data were correlated and analyzed. The crystallization process was carried out under normal pressure, with a magnesium sulfate mass concentration of 25%~32%, a system pH of 5.5~7.0, a crystallizer stirring speed of 80~140 r / min, and a total crystallization time of 4~8 h per batch. Considering the overall temperature requirements of the calcium removal process, the boundary temperature of the calcium removal section under normal operating conditions is 45~55℃. The critical temperature value that can effectively remove calcium ion impurities was determined, and the temperature threshold of the calcium removal section of the raw material was obtained after the correlation calibration was completed.

[0139] The overall cooling range is divided into segments based on the established temperature threshold for the calcium removal segment. The temperature value corresponding to the threshold is used as the dividing point between the two segments. The segment above the dividing point is designated as the high-temperature calcium removal segment, and the segment below the dividing point is designated as the low-temperature crystallization segment. The segment category of each operating node is marked in the order of temperature change. After continuously organizing all the marking information, a segment division identifier sequence for the raw materials is formed.

[0140] By referring to the complete segment division identifier sequence to distinguish different operating intervals, within the interval corresponding to the high-temperature decalcification interval, we sort out the various manifestations and change patterns of the influence of the decalcification correction parameters on temperature control, thereby extracting the driving characteristics of the high-temperature interval. Within the interval corresponding to the low-temperature crystallization interval, we retrieve historical cooling data of the crystallization process retained from past production, sort out the various manifestations and change patterns of this data on temperature changes in the low-temperature interval, thereby extracting the driving characteristics of the low-temperature interval.

[0141] Piecewise curve fitting was performed on the extracted high-temperature and low-temperature driving features respectively. Corresponding change curves were plotted according to the actual temperature change over time in each interval. The two curves were connected to form a complete two-stage gradient cooling curve of the raw material. The entire curve can intuitively show the overall temperature drop rate in the high-temperature segment, i.e., the temperature drop slope in the high-temperature segment, and at the same time, it can also intuitively show the overall temperature drop rate in the low-temperature segment, i.e., the temperature drop slope in the low-temperature segment.

[0142] Based on the cooling slopes of the high-temperature and low-temperature sections corresponding to the two gradient cooling curves, and combined with the real-time temperature deviation of crystallization under the current production conditions, temperature compensation operations are carried out within their respective ranges. At the same time, transition processing is performed at the junction of the high-temperature decalcification section and the low-temperature crystallization section to ensure a smooth transition of temperature change. After a complete compensation and smoothing process, the crystallization process control data of the raw materials is obtained.

[0143] By combining the real-time temperature deviation of the crystallization and the segment division identifier sequence, the current cooling range of the equipment is determined. The cooling rate standard that matches the current range is selected from the two gradient cooling curves. After extracting the corresponding content, the target cooling slope that needs to be used is determined.

[0144] The real-time temperature deviations generated throughout the crystallization process were numerically analyzed one by one. The actual magnitude of the temperature deviation from the standard value at each operating moment was recorded. All deviation values ​​were arranged and integrated in chronological order. After the analysis was completed, the deviation amplitude sequence of the raw materials was obtained.

[0145] The gradient changes of various impurity concentrations in the material system are retrieved from the impurity distribution data. These changes are used as a reference to limit the magnitude of compensation. The compensation level corresponding to different impurity concentration gradients is determined in combination with the matching rules preset in the process, thereby determining the compensation weight coefficient that is suitable for the current operating conditions.

[0146] By combining the established target cooling slope, the complete deviation amplitude sequence, and the corresponding compensation weight coefficient, the temperature change amplitude that needs to be adjusted at each operating moment is comprehensively determined according to the established process control logic. The adjustment values ​​corresponding to each time node are determined one by one, and the corrected slope of the raw material is obtained through calculation.

[0147] The first of the target cooling slope The slope at each moment is derived from a pre-set sequence of cooling process curve data. The compensation weighting coefficients are fixed parameters determined during the process debugging phase. The first moment of the real-time temperature deviation during crystallization... The amplitude at each moment is derived from the absolute value of the difference between the current crystallization temperature and the target temperature. The maximum value in the deviation amplitude sequence is derived from the maximum value of the temperature deviation amplitude at all moments during the crystallization process. The minimum constant is a fixed, small positive number set to prevent the denominator from being zero.

[0148] The corrected cooling slope of the first The slope at time step 1 is used to dynamically compensate for the target cooling slope in real time based on the current temperature deviation. The corrected cooling slope at time step 2 is then used to... The slope at time step 1 is adjusted adaptively by introducing a normalized proportion of the deviation amplitude to achieve the cooling rate adapting to the magnitude of the deviation. The corrected cooling slope at time step 2... The slope at time step 1 is used to enhance cooling control when the temperature deviation is large, and to maintain control stability when the deviation is small. The corrected cooling slope at time step 2 is... The slope at each moment ensures the smoothness of the cooling process and the synchronization of the deviation response.

[0149] When the real-time temperature deviation during crystallization is the When the amplitude increases at a given moment, the corrected cooling slope is at its first... The absolute value of the slope at time point increases accordingly. When the real-time temperature deviation during crystallization... When the amplitude decreases at time n, the corrected cooling slope is at time n. The absolute value of the slope at time point 1 decreases accordingly. When the real-time temperature deviation of crystallization... When the amplitude is zero at the nth moment, the corrected cooling slope is... At the t moment, the slope is equal to the target cooling slope. The slope at time step 1. As the maximum value in the deviation amplitude sequence increases, the slope at the 1st moment of the same crystallization real-time temperature deviation... The corrected cooling slope corresponding to the amplitude at time n. The magnitude of the slope change at each time point decreases accordingly.

[0150] The temperature threshold of the decalcification section will be used as the dividing point between the two intervals. The sliding average transition processing will be carried out on the modified cooling slope of the adjacent intervals on both sides of the dividing point to gradually weaken the rate change at the junction of the two intervals, so that the temperature drop rhythm of the two intervals will transition naturally. After smoothing, the crystallization process control data that is officially issued for use will be obtained.

[0151] The beneficial effects are that the process can scientifically divide the cooling range and build a gradient cooling curve based on the calcium removal requirements. It can perform precise dynamic compensation and smooth transition between segments for real-time temperature deviations. This allows the temperature changes during the crystallization stage to continuously remove residual calcium ion impurities while maintaining a stable and uniform cooling rhythm. It effectively avoids problems such as uneven crystal particle size and deterioration of crystallization effect caused by temperature fluctuations and sudden rate changes. It further optimizes the growth environment of magnesium sulfate crystals and improves the overall quality of the crystallized product. At the same time, relying on the impurity state constraint compensation, the temperature control action is made to fit the actual working conditions of the raw materials, enhancing the adaptive control capability and operational stability of the entire crystallization process.

[0152] D5. The crystal grain size distribution data under the control of the crystallization process is graded and evaluated to obtain the crystal growth quality assessment characterization of the raw material.

[0153] In this embodiment of the invention, the step of classifying and evaluating the crystal grain size distribution data under the control of the crystallization process to obtain a characterization of the crystal growth quality of the raw material includes:

[0154] The crystal particle size distribution data under the control of the crystallization process is divided into particle size grades according to the particle size range, and the number of crystal particles in different particle size grades is accumulated and statistically analyzed to obtain the particle size distribution histogram of the raw material.

[0155] Extract the quantity decay gradient between adjacent particle size grades in the particle size distribution histogram to obtain the crystal growth rate consistency coefficient of the raw material;

[0156] Based on the crystal growth rate consistency coefficient, outlier detection is performed on the abnormal particle size peaks in the particle size distribution histogram to obtain the secondary nucleation intensity identifier of the raw material.

[0157] By fusing the crystal growth rate consistency coefficient with the secondary nucleation intensity identifier and mapping the fused data to a quality level, a quality assessment characterization of the raw material's crystal growth is obtained.

[0158] According to the particle size classification standards set for industrial crystallization production, the crystal particle size distribution data generated under the control data of the crystallization process are classified one by one, and crystal particles of different sizes are assigned to corresponding particle size intervals to complete the overall particle size classification. Then, according to the classified particle size grades, the total number of crystal particles contained in the corresponding intervals is counted one by one. The counted particle number is correlated with the corresponding particle size grade, and all information is integrated in a standardized arrangement to finally generate a particle size distribution histogram that can intuitively show the particle distribution.

[0159] Observe the changes in the number of crystal particles corresponding to two adjacent particle size levels in the particle size distribution histogram, and statistically analyze the overall trend of the gradual decrease in the number of particles when transitioning from one particle size level to the next. Completely analyze the overall pattern corresponding to this trend, and determine whether the crystal growth rate remains uniform based on the analyzed number decay pattern. Calculate the crystal growth rate consistency coefficient by linearly fitting the slope of the particle number in adjacent intervals, with a value ranging from 0 to 1. The closer the value is to 1, the better the uniformity. After sorting out the relevant judgment information, the crystal growth rate consistency coefficient of the raw material is obtained.

[0160] By combining the obtained crystal growth rate consistency coefficient, the distribution law of normal crystal growth is determined. Based on this law, the positions of abrupt increases in the number of particles in the particle size distribution histogram are comprehensively investigated. The 3σ outlier detection algorithm is used to identify abnormal particle size peaks that deviate from the normal distribution state. After completing the screening and judgment of abnormal points, the strength level of secondary nucleation phenomenon is judged according to the prominence of abnormal peaks. The secondary nucleation intensity is divided into three levels: weak, medium, and strong according to the deviation. The corresponding state categories are classified according to the judgment results, and finally the secondary nucleation intensity label of the raw material is obtained.

[0161] The information contained in the crystal growth rate consistency coefficient and the secondary nucleation intensity indicator is integrated and summarized. The two types of content characterizing the crystal state are combined into a unified whole data. Then, the integrated data is classified into grades according to the preset quality evaluation criteria. The weighting coefficients of the two indicators are 0.7 and 0.3, respectively. The final quality is divided into four grades: excellent, good, qualified, and unqualified. Different data combinations are assigned to the corresponding quality levels. After completing the grade mapping, a raw material crystal growth quality evaluation characterization that can comprehensively evaluate the crystal state is obtained.

[0162] The beneficial effects are that this evaluation process can complete the crystal quality detection and evaluation from multiple dimensions such as particle distribution, growth uniformity, and secondary nucleation, accurately identify abnormal problems in the crystallization process, objectively reflect the actual growth state of magnesium sulfate crystals, provide a real and effective basis for subsequent mother liquor reflux verification and process parameter updates, facilitate staff to grasp the crystallization conditions in a timely manner, continuously optimize production conditions, ensure uniform crystal particle size and stable phase of the produced crystals, and improve the overall quality of the by-product magnesium sulfate.

[0163] D6. Based on the crystal growth quality assessment characterization, perform material balance verification on the mother liquor circulation reflux data of the raw material to obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

[0164] In this embodiment of the invention, the step of performing material balance verification on the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization to obtain feedback data of the raw material, and updating the impurity distribution data of the raw material based on the feedback data, includes:

[0165] The mother liquor reflux data of the raw material is time-series sliced ​​to obtain the mother liquor reflux feature sequence of the raw material;

[0166] Based on the crystal growth quality assessment characterization, the mother liquor reflux characteristic sequence is screened by benchmarking to obtain the impurity residue deviation vector of the raw material;

[0167] Based on the impurity residue deviation vector, the mother liquor reflux characteristic sequence is corrected for deviation to obtain the feedback data of the raw material;

[0168] The feedback data and the impurity distribution data of the raw material are concatenated to obtain the updated impurity distribution data of the raw material.

[0169] The complete mother liquor circulation reflux data is processed by time-series slicing according to continuous time nodes in the mother liquor circulation reflux process. The continuous data is split into segments with fixed time spans. The relevant information such as mother liquor composition, impurity content, and material flow rate in each independent time period are formed into separate unit contents. All the split unit contents are arranged and combined in chronological order to form a time-coherent and information-complete mother liquor reflux feature sequence.

[0170] Combining the overall production conditions and impurity residue status reflected by the generated crystal growth quality assessment characterization, the data of each item in the mother liquor reflux characteristic sequence and the standard mother liquor component data set by the process are compared and screened one by one. The difference between the actual residual amount and the standard residual amount of each type of impurity in the mother liquor is statistically analyzed. The difference information corresponding to each type of impurity is sorted and combined in a fixed order to form an impurity residue deviation vector that can intuitively reflect the deviation state of impurity residue.

[0171] Material balance is based on mass, with a statistical cycle of 1 minute. Deviation correction uses a linear compensation algorithm, relying on the deviation values ​​marked by the obtained impurity residual deviation vector. It adjusts and corrects each data item with deviation in the mother liquor reflux characteristic sequence, supplementing missing content of material and impurity related data or removing abnormal content that exceeds the standard range according to the actual deviation size. The corrected data is made to fit the actual state of the mother liquor and the process standard requirements. After all correction work is completed, feedback data for backend data updates is obtained.

[0172] The feedback data, including information on impurity residues and material states, is compiled and then systematically integrated with the raw material impurity distribution data retained from previous processes. This process preserves the basic content of the original impurity distribution data while overlaying the latest operating conditions information from the feedback data. The integrated data allows for a synchronized reflection of the initial impurity state of the raw materials and the changes in impurities brought in by the mother liquor reflux, ultimately resulting in comprehensive and up-to-date impurity distribution data that is consistent with real-time operating conditions.

[0173] The beneficial effects are that this process can fully verify and correct the mother liquor reflux data based on the crystal growth quality results, accurately capture the residual deviation of impurities in the mother liquor circulation process, realize the dynamic update of the front-end impurity distribution data through data splicing, open up the data closed loop of the entire refining process, and allow subsequent batch production to adjust process parameters according to the latest impurity status, continuously correct production deviations, ensure the material balance and stability of the mother liquor circulation system, and maintain the long-term operation quality and product qualification rate of the by-product magnesium sulfate refining process.

[0174] like Figure 2 The diagram shown is a functional block diagram of a magnesium sulfate refining system based on recrystallization according to an embodiment of the present invention.

[0175] The recrystallization-based by-product magnesium sulfate refining system 100 described in this invention can be installed in an electronic device. Depending on the functions implemented, the recrystallization-based by-product magnesium sulfate refining system 100 may include an impurity distribution analysis module 101, a hot-melt parameter optimization module 102, a calcium ion anomaly identification module 103, a two-segment crystallization curve generation module 104, a crystal growth grading evaluation module 105, and a material balance feedback update module 106. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by an electronic device processor and perform a fixed function, stored in the electronic device's memory.

[0176] In this embodiment, the functions of each module / unit are as follows:

[0177] The impurity distribution analysis module 101 is used to perform quantitative analysis and statistical analysis on the component detection data of the raw material to obtain the impurity distribution data of the raw material.

[0178] The hot melting parameter optimization module 102 is used to perform parameter space search on the low-temperature hot melting process of the raw material based on the impurity distribution data, obtain the hot melting temperature range of the raw material, and generate hot melting control instructions for the raw material according to the hot melting temperature range.

[0179] The calcium ion anomaly identification module 103 is used to perform trend analysis on the calcium ion concentration monitoring data after executing the hot melting control command, obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material.

[0180] The two-segment crystallization curve generation module 104 is used to fuse the calcium removal correction parameters and preset crystallization data, and to perform curve fitting on the fused data to obtain two-segment gradient cooling curves of the raw material. Based on the two-segment gradient cooling curves, the real-time temperature deviation of the raw material is dynamically compensated to obtain the crystallization process control data of the raw material.

[0181] The crystal growth grading evaluation module 105 is used to grade and evaluate the crystal grain size distribution data under the control of the crystallization process, so as to obtain the crystal growth quality evaluation characterization of the raw material.

[0182] The material balance feedback update module 106 is used to perform material balance verification on the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization, obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

[0183] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0184] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0185] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0186] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0187] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0188] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for refining magnesium sulfate byproducts based on recrystallization, characterized in that, The method includes: D1. Quantitatively analyze and statistically analyze the composition detection data of the raw materials to obtain the impurity distribution data of the raw materials; D2. Based on the impurity distribution data, perform a parameter space search on the low-temperature hot melting process of the raw material to obtain the hot melting temperature range of the raw material, and generate a hot melting control command for the raw material according to the hot melting temperature range. D3. Perform trend analysis on the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material. D4. The calcium removal correction parameters and preset crystallization data are fused together, and the fused data is curve fitted to obtain two gradient cooling curves of the raw material. Based on the two gradient cooling curves, the real-time temperature deviation of the raw material is dynamically compensated to obtain the crystallization process control data of the raw material. D5. The crystal grain size distribution data under the control of the crystallization process is graded and evaluated to obtain the crystal growth quality assessment characterization of the raw material. D6. Based on the crystal growth quality assessment characterization, perform material balance verification on the mother liquor circulation reflux data of the raw material to obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

2. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, The quantitative analysis and statistical analysis of the component detection data of the raw materials to obtain the impurity distribution data of the raw materials includes: The composition detection data of the raw materials are collected, and multi-dimensional feature extraction is performed on the composition detection data to obtain the impurity feature vector of the raw materials; Principal component screening is performed on the impurity feature vector to obtain the core impurity component data of the raw material; The core impurity component data is discretized and segmented to obtain the impurity concentration range distribution sequence of the raw material; Frequency statistics were performed on the distribution sequence of impurity concentration ranges to obtain the impurity frequency distribution histogram of the raw material; Based on the impurity frequency distribution histogram, cluster analysis is performed on the impurity distribution pattern of the raw material to obtain impurity aggregation characteristic data of the raw material. By fusing the impurity aggregation characteristic data with the core impurity component data, the impurity distribution data of the raw material is obtained.

3. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, Based on the impurity distribution data, a parameter space search is performed on the low-temperature hot melting process of the raw material to obtain the hot melting temperature range of the raw material, and a hot melting control command for the raw material is generated according to the hot melting temperature range, including: The impurity distribution data is matched and filtered with a preset process parameter library to obtain the candidate parameter space for the raw material; An orthogonal ergodic analysis is performed on the candidate parameter space to obtain the parameter combination evaluation data of the raw materials; Based on the evaluation data of the parameter combination, the candidate parameter space is pruned and optimized to obtain the optimal parameter subspace of the raw material; Based on the optimal parameter subspace, the tolerance boundary of the impurity distribution data is determined to obtain the thermal dissolution constraint conditions of the raw material. The hot melt constraint conditions are mapped to interval boundary data, and the interval boundary data is used as the hot melt temperature range of the raw material. Based on the hot melting temperature range, the heating rate data in the optimal parameter subspace are arranged in time sequence to obtain the hot melting time sequence parameter sequence of the raw material. The hot melting timing parameter sequence is converted into an instruction encoding format to obtain the hot melting control instruction for the raw material.

4. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, The process involves trend analysis of the calcium ion concentration monitoring data after executing the hot melt control command to obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and seed-induced logic determination of the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material, including: The calcium ion concentration monitoring data after the execution of the heat dissolution control command is divided into time-series sliding window segments to obtain local concentration change segments of the calcium ion concentration monitoring data. The slope consistency of the local concentration change segments is checked to obtain the concentration change trend sequence of the calcium ion concentration monitoring data. Based on the concentration change trend sequence, identify abrupt breaks in the calcium ion concentration monitoring data; The mutation fragments were screened by both duration and amplitude to obtain abnormal fluctuation data of the calcium ion concentration monitoring data. Based on the fluctuation start time and fluctuation decay rate characteristics of the abnormal fluctuation data, the seed-induced resistance characteristics of the raw material are obtained, and the seed-induced resistance characteristics are judged by logic threshold to obtain the seed-induced failure identifier of the raw material. Based on the seed-induced failure indicator, the energy accumulation of the abnormal fluctuation data is analyzed in reverse to obtain the calcium ion supersaturation deviation of the raw material. The deviation of calcium ion supersaturation is converted into a concentration control step size, which is then used as the calcium removal correction parameter for the raw material.

5. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, The process involves fusing the calcium removal correction parameters with preset crystallization data, performing curve fitting on the fused data to obtain two gradient cooling curves for the raw material, and dynamically compensating for the real-time temperature deviation of the raw material's crystallization based on these two gradient cooling curves to obtain crystallization process control data for the raw material, including: The correlation between the calcium removal correction parameter and the calcium ion-related impurity concentration in the impurity distribution data is calibrated to obtain the temperature threshold of the calcium removal section of the raw material. Based on the temperature threshold of the calcium removal section, the cooling range of the raw material is divided into a high-temperature calcium removal section and a low-temperature crystallization section, thus obtaining the section division identifier sequence of the raw material. Based on the segment division identifier sequence, the driving characteristics of the calcium removal correction parameter in the high-temperature segment and the driving characteristics of the historical cooling data of the crystallization process control data in the low-temperature segment are extracted respectively to obtain the high-temperature segment driving characteristics and low-temperature segment driving characteristics of the raw material. Piecewise curve fitting is performed on the high-temperature driving characteristics and the low-temperature driving characteristics to obtain two gradient cooling curves of the raw material, which include the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment. Based on the cooling slope of the high-temperature section and the cooling slope of the low-temperature section, dynamic compensation within the section and smoothing processing between sections are performed on the real-time temperature deviation of the crystallization to obtain the crystallization process control data of the raw material.

6. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 5, characterized in that, The method involves performing intra-segment dynamic compensation and inter-segment switching smoothing processing on the real-time temperature deviation of the crystallization based on the cooling slope of the high-temperature segment and the cooling slope of the low-temperature segment, respectively, to obtain the crystallization process control data of the raw material, including: Based on the real-time temperature deviation of crystallization and the segment division identifier sequence, extract the target cooling slope corresponding to the two gradient cooling curves; Numerical analysis was performed on the real-time temperature deviation of the crystallization to obtain the deviation amplitude sequence of the raw material; Using the impurity concentration gradient in the impurity distribution data as a constraint factor for the intra-segment compensation intensity, the compensation weight coefficient of the raw material is obtained. The corrected cooling slope of the raw material is calculated based on the target cooling slope, the deviation amplitude sequence, and the compensation weighting coefficient. Based on the temperature threshold of the calcium removal section, the modified cooling slope of the adjacent sections is transferred by a moving average to obtain the crystallization process control data of the raw material.

7. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 6, characterized in that, The formula for calculating the corrected cooling slope is as follows: ； in, The first of the corrected cooling slope The slope at each moment The first of the target cooling slopes The slope at each moment The compensation weight coefficient is... The first of the real-time temperature deviations of the crystallization Amplitude at each moment, The maximum value in the deviation amplitude sequence. It is a very small constant.

8. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, The step of classifying and evaluating the crystal grain size distribution data under the control of the crystallization process to obtain a characterization of the crystal growth quality of the raw material includes: The crystal particle size distribution data under the control of the crystallization process is divided into particle size grades according to the particle size range, and the number of crystal particles in different particle size grades is accumulated and statistically analyzed to obtain the particle size distribution histogram of the raw material. Extract the quantity decay gradient between adjacent particle size grades in the particle size distribution histogram to obtain the crystal growth rate consistency coefficient of the raw material; Based on the crystal growth rate consistency coefficient, outlier detection is performed on the abnormal particle size peaks in the particle size distribution histogram to obtain the secondary nucleation intensity identifier of the raw material. By fusing the crystal growth rate consistency coefficient with the secondary nucleation intensity identifier and mapping the fused data to a quality level, a quality assessment characterization of the raw material's crystal growth is obtained.

9. The method for refining magnesium sulfate by-product based on recrystallization as described in claim 1, characterized in that, The process of verifying the material balance of the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization, obtaining feedback data of the raw material, and updating the impurity distribution data of the raw material based on the feedback data includes: The mother liquor reflux data of the raw material is time-series sliced ​​to obtain the mother liquor reflux feature sequence of the raw material; Based on the crystal growth quality assessment characterization, the mother liquor reflux characteristic sequence is screened by benchmarking to obtain the impurity residue deviation vector of the raw material; Based on the impurity residue deviation vector, the mother liquor reflux characteristic sequence is corrected for deviation to obtain the feedback data of the raw material; The feedback data and the impurity distribution data of the raw material are concatenated to obtain the updated impurity distribution data of the raw material.

10. A system for refining magnesium sulfate byproducts based on recrystallization, characterized in that, For implementing the recrystallization-based method for refining byproduct magnesium sulfate as described in claim 1, the system comprises: The impurity distribution analysis module is used to perform quantitative analysis and statistical analysis on the component detection data of the raw material to obtain the impurity distribution data of the raw material. The hot melting parameter optimization module is used to perform parameter space search on the low-temperature hot melting process of the raw material based on the impurity distribution data, obtain the hot melting temperature range of the raw material, and generate hot melting control instructions for the raw material according to the hot melting temperature range. The calcium ion anomaly identification module is used to perform trend analysis on the calcium ion concentration monitoring data after the execution of the hot melting control command, obtain abnormal fluctuation data of the calcium ion concentration monitoring data, and perform seed induction logic judgment on the abnormal fluctuation data to obtain the calcium removal correction parameters of the raw material. The two-segment crystallization curve generation module is used to fuse the calcium removal correction parameters and preset crystallization data, and to perform curve fitting on the fused data to obtain two-segment gradient cooling curves of the raw material. Based on the two-segment gradient cooling curves, the real-time temperature deviation of the raw material crystallization is dynamically compensated to obtain the crystallization process control data of the raw material. The crystal growth grading and evaluation module is used to grade and evaluate the crystal grain size distribution data under the control of the crystallization process, and obtain the crystal growth quality evaluation characterization of the raw material. The material balance feedback update module is used to perform material balance verification on the mother liquor circulation reflux data of the raw material based on the crystal growth quality assessment characterization, obtain the feedback data of the raw material, and update the impurity distribution data of the raw material based on the feedback data.

Images