Boric acid production material scheduling method and system based on internet of things

Abstract

A boron acid production material scheduling method and system based on the Internet of Things, the method comprising: acquiring multi-source data from a boron acid production line, and preprocessing the multi-source data to map the preprocessed multi-source data to a digital twin model. The scheduling requirements of the current workshop are extracted from the initial parameters of the digital twin model and the preprocessed multi-source data vector, and a multi-objective scheduling model is constructed when the scheduling requirements meet the workshop production requirements. The turbidity, crystalline particle characteristics and color change of the solution on the production line are monitored in real time through an optical system to obtain optical monitoring data, and the internal crystallization state and impurity content of the solution are determined in combination with the multi-source data to perform scheduling when the crystallization state or impurity content deviates from the normal range. The optical monitoring data and scheduling execution are fed back to the digital twin model based on the abnormal indication to correct the scheduling strategy and process parameters of the boron acid production line, thereby improving the intelligentization and production efficiency of boron acid production.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of Internet of Things and production scheduling control, and more particularly, relates to a borax production material scheduling method and system based on Internet of Things. BACKGROUND

[0002] Material scheduling is an important technology in modern production and logistics, which involves the management and optimization of material demand, inventory, transportation and distribution. With the expansion of production scale and the intensification of market competition, the importance of material scheduling is increasingly prominent.

[0003] Currently, the material scheduling on the borax production line still relies heavily on manual work, and traditional borax production line data measurement mostly uses single instruments. However, single instrument data coverage is insufficient, and the collection cycle is relatively long, resulting in low material scheduling efficiency and high error rate. In addition, due to the particularity of borax production, the requirements for the environment are often more stringent. For example, the temperature, flow rate, and turbidity, crystallization and other parameters of the liquid inside the reaction kettle have important influence on the production. However, the existing technology often ignores the monitoring of the state of the internal liquid of borax production, and ignoring the state of the reaction liquid may lead to production according to the original feeding and scheduling mode in abnormal state, which affects the yield and production quality of borax to some extent. SUMMARY

[0004] To solve the problems in the prior art, the present application aims to solve the above-mentioned defects, and further provides a borax production material scheduling method and system based on Internet of Things.

[0005] The present application adopts the following technical solutions.

[0006] The present application discloses a borax production material scheduling method based on Internet of Things, which comprises:

[0007] Obtaining multi-source data from the borax production line, and preprocessing the multi-source data to map the preprocessed multi-source data to a digital twin model;

[0008] Extracting the scheduling demand of the current workshop from the initial parameters of the digital twin model and the preprocessed multi-source data vector, and constructing a multi-objective scheduling model when the scheduling demand meets the production demand of the workshop;

[0009] Real-time monitoring of the turbidity, crystallization particle characteristics and color change of the solution on the production line through an optical system to obtain optical monitoring data, and determining the internal crystallization state and impurity content of the solution in combination with the multi-source data to perform scheduling when the crystallization state or impurity content deviates from the normal range;

[0010] feed back the optical monitoring data and scheduling execution feedback to the digital twin model based on the abnormal indication to correct the scheduling strategy and process parameters of the boric acid production line;

[0011] The multi-source data includes sensor data of temperature, flow, pH and pressure installed at multiple key nodes of the boric acid production line, the preprocessing includes missing value filling and denoising filtering, and the multi-objective scheduling model is used to output scheduling decisions according to scheduling requirements and resource lists.

[0012] Further, the multi-source data from the boric acid production line is obtained, and the multi-source data is preprocessed to map the preprocessed multi-source data to the digital twin model, including:

[0013] Raw data vectors are generated based on instantaneous readings from different sensors of the boric acid production line, and the raw data vectors are filled and denoised to obtain preprocessed data vectors;

[0014] Based on the preprocessed data vectors, determine the key process variables, and map the key process variables to the digital twin structure to build the digital twin model, and obtain the initial parameter vector of the digital twin model;

[0015] The key process variables include temperature, boron ore impurity content and flow, and the initial parameter vector includes initial modeling values of reaction temperature, flow and pH of the boric acid production line.

[0016] Further, the multi-source data from the boric acid production line is obtained, and the multi-source data is preprocessed to map the preprocessed multi-source data to the digital twin model, including:

[0017] Compare the established digital twin model with the actual monitoring data to obtain the error between the model output and the actual monitoring data, and determine whether the error exceeds a first threshold;

[0018] When the error exceeds the first threshold, the initial parameters of the digital twin model are preprocessed or the bias of the sensor is corrected to obtain the verification result of the digital twin model and the corrected model parameters.

[0019] Further, the scheduling requirements of the current workshop are extracted from the initial parameters of the digital twin model and the preprocessed multi-source data vectors, and when the scheduling requirements meet the production requirements of the workshop, a multi-objective scheduling model is constructed, including:

[0020] According to the inventory data, flow, temperature and pH of the borate production line in the digital twin model, key parameters of total feeding amount, feeding frequency and expected yield are determined, and a scheduling demand vector is generated by combining the equipment capacity and personnel configuration of the workshop;

[0021] A multi-objective function for defining a decision variable vector is set, and a scheduling decision variable vector is obtained based on the multi-objective function to construct the multi-objective scheduling model;

[0022] The multi-objective function includes minimizing total transportation time, minimizing energy consumption and maximizing capacity utilization, and the decision variable vector includes route selection of all batch feeding times or transportation vehicles.

[0023] Further, the scheduling demand of the current workshop is extracted from the initial parameters of the digital twin model and the preprocessed multi-source data vector, and when the scheduling demand meets the production demand of the workshop, a multi-objective scheduling model is constructed, which further comprises:

[0024] When the optical sensor monitors the abnormality of solution turbidity or crystalline particle size, a first feedback signal is triggered and transmitted through the data interaction interface between the multi-objective scheduling model and the optical system, and when the multi-objective scheduling model captures the first feedback signal, the corresponding feeding time or batch order is fine-tuned;

[0025] When the crystallization rate of the current feeding batch exceeds a second threshold, a second feedback signal is triggered and transmitted, and when the multi-objective scheduling model captures the second feedback signal, the subsequent material delivery time is delayed or the single feeding amount is reduced;

[0026] An adaptive function is constructed based on the first feedback signal and the second feedback signal to obtain a revised scheduling decision based on the adaptive function on the basis of the original adjustment strategy, and the revised scheduling decision is used to generate scheduling execution instructions.

[0027] Further, the optical system is used to monitor the turbidity, crystalline particle characteristics and color change of the solution in real time to obtain optical monitoring data, and the solution internal crystallization state and impurity content are determined in combination with the multi-source data to perform scheduling execution when the crystallization state or impurity content deviates from the normal range, which comprises:

[0028] According to the set sampling frequency, the optical sensor installed at the discharge port of the reaction kettle or the inlet of the crystallization workshop is used to monitor the turbidity, crystalline particle characteristics and color change of the solution in real time to obtain an optical data vector, and the optical data vector is fused with the multi-source data to construct a fusion vector;

[0029] Anomaly detection is performed on the fusion vector based on setting a normal working range threshold or a machine learning algorithm, and when the fusion vector is abnormal, a control system is triggered to generate an abnormal index quantity for early warning and automatic adjustment.

[0030] A control strategy function is obtained based on the abnormal index quantity and a correction quantity of the original adjustment strategy, and when the abnormal index quantity is a first value, a production scheduling variable is revised by the digital twin model to obtain a revised scheduling scheme.

[0031] Further, the optical monitoring data and scheduling execution are fed back to the digital twin model based on the abnormal indication to correct the scheduling strategy and process parameters of the boric acid production line, including:

[0032] Execution instructions and execution data logs are obtained based on the revised scheduling scheme, and a historical data matrix is constructed in combination with the abnormal record and the optical monitoring data, so as to predict the crystallization state or energy consumption at different times based on the historical data matrix by using a machine learning model;

[0033] The scheduling strategy and the optical threshold are optimized based on the prediction result of the machine learning model to construct an optimization objective function, obtain updated new model parameters, and map the new model parameters to the digital twin model.

[0034] The second aspect of the present application discloses a boric acid production material scheduling system based on the Internet of Things, the system comprising:

[0035] A data preprocessing module is configured to obtain multi-source data from a boric acid production line and preprocess the multi-source data to map the preprocessed multi-source data to a digital twin model.

[0036] A model construction module is configured to extract scheduling requirements of a current workshop from initial parameters of the digital twin model and a preprocessed multi-source data vector, and construct a multi-objective scheduling model when the scheduling requirements meet production requirements of the workshop.

[0037] A scheduling execution module is configured to obtain optical monitoring data by real-time monitoring of turbidity, crystallization particle characteristics and color change of a solution on a production line by an optical system, and determine an internal crystallization state and impurity content of the solution in combination with the multi-source data to perform scheduling execution when the crystallization state or the impurity content deviates from a normal range.

[0038] A scheduling strategy adjustment module is configured to feed back the optical monitoring data and scheduling execution to the digital twin model based on an abnormal indication to correct the scheduling strategy and process parameters of the boric acid production line.

[0039] The multi-source data includes sensor data on temperature, flow rate, pH, and pressure installed at multiple key nodes on the boric acid production line. The preprocessing includes missing value imputation and noise reduction filtering. The multi-objective scheduling model is used to output scheduling decisions based on scheduling requirements and resource lists.

[0040] A third aspect of the present invention discloses a terminal, including a processor and a storage medium; characterized in that:

[0041] The storage medium is used to store instructions;

[0042] The processor is configured to operate according to the instructions to perform the steps of the method described in the first aspect.

[0043] A fourth aspect of the present invention discloses a computer-readable storage medium having a computer program stored thereon, characterized in that the program, when executed by a processor, implements the steps of the method described in the first aspect.

[0044] The beneficial effects of the present invention are as follows: Compared with the prior art, the present invention has the following advantages:

[0045] By acquiring multi-source sensor data from the boric acid production line and preprocessing the data, a digital twin model is mapped onto the preprocessed data. The current scheduling requirements of the workshop are then extracted from the initial parameters of the digital twin model and the preprocessed multi-source data vectors. A multi-objective scheduling model is constructed when the scheduling requirements meet the workshop's production needs. Subsequently, optical monitoring systems are used to monitor the turbidity, crystal particle characteristics, and color changes of the solution on the production line in real time, obtaining optical monitoring data. This data, combined with multi-source data, determines the internal crystallization state and impurity content of the solution, allowing for scheduling execution when the crystallization state or impurity content deviates from the normal range. Finally, based on anomaly indicators, the optical monitoring data and scheduling execution status are fed back to the digital twin model to correct the scheduling strategy and process parameters of the boric acid production line. This method, by fusing optical monitoring and multi-source sensor data to correct the scheduling strategy of the boric acid production line, utilizes a large sensor coverage and real-time data acquisition, effectively avoiding losses caused by adhering to the original scheduling strategy when anomalies occur in boric acid production. The assistance of the digital twin model also improves the intelligence and efficiency of boric acid production to a certain extent, ensuring both the output and quality of boric acid. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the structure of a boric acid production material scheduling method based on the Internet of Things provided by the present invention. Detailed Implementation

[0047] The application will be further described below with reference to the drawings. The following examples are only used to more clearly illustrate the technical solutions of the application, and cannot be used to limit the protection scope of the application.

[0048] As shown in Figure 1 In one embodiment, a boron acid production material scheduling method based on the Internet of Things includes the following steps:

[0049] Step S110, acquiring multi-source data from the boron acid production line, and preprocessing the multi-source data to map the preprocessed multi-source data to a digital twin model.

[0050] Wherein, the multi-source data includes sensor data of temperature, flow, pH and pressure installed at multiple key nodes on the boron acid production line, the preprocessing includes missing value filling and denoising filtering, and the key nodes include raw material warehouse, conveying pipeline and reaction kettle.

[0051] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the application specifically includes the following steps:

[0052] Step S111, generating an original data vector based on the instantaneous readings from different sensors of the boron acid production line, and performing missing value filling and denoising filtering on the original data vector to obtain a preprocessed data vector.

[0053] Step S112, determining key process variables based on the preprocessed data vector, and mapping the key process variables to a digital twin structure to build a digital twin model, and obtaining an initial parameter vector of the digital twin model.

[0054] Wherein, the key process variables include temperature, boron ore impurity content and flow, and the initial parameter vector includes initial modeling values of reaction temperature, flow and pH of the boron acid production line.

[0055] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the application specifically further includes the following steps:

[0056] Step S113, comparing the established digital twin model with the actual monitoring data to obtain the error between the model output and the actual monitoring data, and determining whether the error exceeds a first threshold.

[0057] Step S114, when the error exceeds the first threshold, reprocessing the initial parameters of the digital twin model or correcting the bias of the sensor to obtain a verification result of the digital twin model and a corrected model parameter.

[0058] Step S120, extract the scheduling demand of the current workshop from the initial parameters of the digital twin model and the pre-processed multi-source data vector, and construct a multi-objective scheduling model when the scheduling demand meets the production demand of the workshop.

[0059] The multi-objective scheduling model is configured to output scheduling decisions according to the scheduling demand and the resource list.

[0060] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the present application specifically comprises the following steps:

[0061] Step S121, according to the inventory data, flow, temperature and pH of the boron acid production line in the digital twin model, determine the key parameters of the total amount of raw materials, the frequency of raw materials and the expected yield, and combine the workshop equipment capacity and personnel configuration to generate a scheduling demand vector.

[0062] Step S122, set a multi-objective function for defining a decision variable vector, and obtain a scheduling decision variable vector based on the multi-objective function to construct a multi-objective scheduling model.

[0063] The multi-objective function includes minimizing total transportation time, minimizing energy consumption, and maximizing capacity utilization, and the variable vector for decision includes route selection of all batches of raw materials or transportation vehicles.

[0064] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the present application specifically further comprises the following steps:

[0065] Step S123, when the optical sensor monitors the abnormality of solution turbidity or crystalline particle size, trigger and transmit a first feedback signal through a data interaction interface between the multi-objective scheduling model and the optical system, and fine-tune the corresponding feeding time or batch order when the multi-objective scheduling model captures the first feedback signal.

[0066] Step S124, when the crystallization rate of the current feeding batch exceeds a second threshold, trigger and transmit a second feedback signal, and delay the subsequent material delivery time or reduce the amount of single feeding when the multi-objective scheduling model captures the second feedback signal.

[0067] Step S125, construct an adaptive function based on the first feedback signal and the second feedback signal, and obtain a revised scheduling decision based on the adaptive function on the basis of the original adjustment strategy, and the revised scheduling decision is used to generate a scheduling execution instruction.

[0068] Step S130, the turbidity, crystallization particle characteristics and color change of the solution on the production line are monitored in real time through the optical system to obtain optical monitoring data, and the internal crystallization state and impurity content of the solution are determined in combination with the multi-source data to perform scheduling and execution when the crystallization state or impurity content deviates from the normal range.

[0069] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the present application specifically comprises the following steps in step S130:

[0070] Step S131, the turbidity, crystallization particle characteristics and color change of the solution are monitored in real time through the optical sensor installed at the discharge port of the reaction kettle or the inlet of the crystallization workshop according to the set sampling frequency to obtain an optical data vector, and the optical data vector is fused with the multi-source data to construct a fusion vector.

[0071] Step S132, calling an abnormality detection based on a set normal working interval threshold or a machine learning algorithm on the fusion vector, triggering a control system to perform early warning and automatic adjustment to generate an abnormality index quantity when the fusion vector is abnormal.

[0072] Step S133, a control strategy function is obtained based on the abnormality index quantity and the correction amount of the original adjustment strategy, and when the abnormality index quantity is a first value, the production scheduling variable is revised through a digital twin model to obtain a revised scheduling scheme.

[0073] Step S140, the optical monitoring data and scheduling execution are fed back to the digital twin model based on the abnormality indication to correct the scheduling strategy and process parameters of the boron acid production line.

[0074] In some embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the present application specifically comprises the following steps in step S140:

[0075] Step S141, execution instructions and execution data logs are obtained based on the revised scheduling scheme, and in combination with the abnormality record and the optical monitoring data, a historical data matrix is constructed to predict the crystallization state or energy consumption at different times based on the historical data matrix by using a machine learning model.

[0076] Step S142, the scheduling strategy and optical threshold are optimized based on the prediction result of the machine learning model to construct an optimization objective function, obtain updated new model parameters, and map the new model parameters to the digital twin model.

[0077] The above boron acid production material scheduling method based on the Internet of Things obtains multi-source sensor data from the boron acid production line, pre-processes the multi-source sensor data, maps the pre-processed multi-source sensor data to a digital twin model, extracts scheduling requirements of the current workshop from the initial parameters of the digital twin model and the pre-processed multi-source data vector, and constructs a multi-objective scheduling model when the scheduling requirements meet the production requirements of the workshop. Then, the turbidity, crystalline particle characteristics and color change of the solution on the production line are monitored in real time by an optical system to obtain optical monitoring data, and the internal crystallization state and impurity content of the solution are determined in combination with the multi-source data to perform scheduling when the crystallization state or impurity content deviates from the normal range. Finally, the optical monitoring data and scheduling execution are fed back to the digital twin model based on the abnormal indication to correct the scheduling strategy and process parameters of the boron acid production line. This method corrects the scheduling strategy of the boron acid production line by fusing optical monitoring and multi-source sensor data. The sensor coverage is large and all real-time data are collected, effectively avoiding the loss caused by the original scheduling strategy when the boron acid production is abnormal. The assistance of the digital twin model also improves the intelligence and production efficiency of the boron acid production to some extent, ensuring the yield and production quality of the boron acid production.

[0078] In specific embodiments, the boron acid production material scheduling method based on the Internet of Things provided by the present application comprises steps 1-4:

[0079] Step 1, multi-source data acquisition and digital twin model initialization.

[0080] A complete boron acid production line data base is established to lay a data foundation for subsequent material scheduling and optical monitoring, overcome the defects of traditional single instrument data coverage and long collection period, and ensure full data coverage and high-frequency sampling.

[0081] Specifically, it comprises steps 1.1-1.4:

[0082] Step 1.1, sensor arrangement and original data acquisition.

[0083] Specifically, temperature, flow, pH, pressure and other sensors are installed at multiple key nodes such as raw material warehouse, conveying pipeline and reaction kettle to ensure coverage of all main links of boron acid production. Then, the original data vector is generated according to the collected sensor readings, which is composed of instantaneous readings (real-time update, such as updating the readings once a minute) of different sensors.

[0084] Step 1.2, data cleaning and preprocessing.

[0085] Specifically, the original data vector is pre-processed by missing value filling and denoising filtering to generate a clean and stable data vector composed of reliable readings after cleaning. For obviously abnormal sampling points, mean or median can be used for correction to avoid the influence of extreme values on subsequent analysis.

[0086] Step 1.3, parameter mapping and digital twin structure building.

[0087] Specifically, important process variables (such as temperature, boron ore impurity content, and flow rate, etc.) are determined according to the pre-processed data vector, and are mapped to the corresponding modules of the digital twin. The initial state of each sub-module in the digital twin can be recorded as the initial parameter vector, including the initial modeling values of reaction temperature, flow rate, and pH, which are used to establish the digital twin model.

[0088] Step 1.4, preliminary verification of the system.

[0089] Specifically, the established digital twin model is compared with part of the monitoring data on site to ensure that the fitting deviation is within an acceptable range. If the deviation exceeds the set threshold range, return to step 1.2 to adjust the data preprocessing strategy or correct the sensor bias, and obtain the verification result of the digital twin and the corrected parameters.

[0090] Step 2, material scheduling optimization and adaptive scheduling.

[0091] Under the premise of meeting the quality requirements of multiple production lines and different batches of boric acid products, the transportation route and feeding time sequence are dynamically optimized to solve the defects of manual scheduling and timely response to demand changes or material shortages.

[0092] Specifically, it includes steps 2.1-2.4:

[0093] Step 2.1, scheduling demand extraction and data input.

[0094] Specifically, the current actual demand, inventory status and production target of the workshop are extracted from the initial parameters of the digital twin model and the pre-processed data vector to form a scheduling demand list. According to the information about inventory, flow rate, temperature, etc. in the digital twin model, key parameters such as total feeding amount, feeding frequency, and expected yield are determined, and combined with the equipment capacity and personnel configuration of the workshop, a scheduling demand vector is obtained. The scheduling demand vector is composed of the demand for single batch of boron ore or borax, process time limit, and target yield. In addition, the data in the scheduling demand vector can also be updated in real time from the digital twin model, or mapped to the digital twin model from the target issued by the management layer.

[0095] Step 2.2, multi-objective scheduling model construction.

[0096] Specifically, under the premise of meeting production needs, transportation costs, energy consumption, and real-time monitoring needs of subsequent optical systems are taken into account to build a scheduling optimization model. First, a multi-objective function is set, such as minimizing total transportation time, minimizing energy consumption, and maximizing production capacity utilization rate, to balance the three aspects. A variable vector for decision-making is defined, including the feeding time of any batch of materials within the number of possible scheduling decisions or the route selection of the transportation vehicle.

[0097] Step 2.3, adaptive scheduling and optical system linkage.

[0098] Specifically, the scheduling decision is combined with the optical monitoring system, so that the scheduling can be dynamically adjusted according to the real-time optical signal feedback. First, a data interaction interface with the optical monitoring system is reserved in the scheduling optimization model: when the optical sensor detects abnormal solution turbidity or crystal particle size, a feedback signal can be triggered, and the scheduling optimization model can fine-tune the feeding time or batch order after capturing the feedback signal, to avoid continuing to feed in large quantities when an abnormal state occurs. If the current batch crystallization rate is too fast, the scheduling optimization model can delay the subsequent material delivery time or reduce the single feeding amount. An adaptive function is built based on optical monitoring feedback to calculate the new decision vector revised based on the original scheduling strategy. Through the above linkage, scheduling revision can be automatically triggered when the process is abnormal to reduce losses and stabilize production.

[0099] For example, when the scheduling system receives the optical feedback "abnormal precipitation is detected", the feeding time of the subsequent batch can be delayed by 10-20 minutes, or the feeding amount can be reduced, so that the staff can handle the crystallization situation.

[0100] Step 2.4, model solving and implementation.

[0101] Specifically, the above multi-objective scheduling model and adaptive strategy are implemented in actual production, and a suitable algorithm (improved particle swarm, genetic algorithm, or heuristic algorithm in edge computing) is used for solving, and the scheduling plan is made according to the solving result.

[0102] In this embodiment, different solving strategies are selected according to the calculation scale, network bandwidth, and real-time demand of the workshop, for example, if high precision is required, evolutionary algorithms can be used in the cloud, and if fast response is required, simplified heuristic algorithms can be deployed in the edge. Once the optimal solution or suboptimal solution is confirmed, the logistics equipment, feeding device, etc. are called according to the scheduling strategy before and after the revision, and if subsequent abnormal signals are received from the optical monitoring system, the current scheduling strategy is revised. Finally, during the scheduling execution, the execution process data needs to be recorded in real time and fed back to the digital twin model for model optimization.

[0103] Step 3, real-time control of production process combined with optical monitoring.

[0104] The optical system is introduced to finely detect the turbidity, refractive index, spectral characteristics, etc. of boric acid solution, and realize early abnormal identification.

[0105] Specifically, steps 3.1-3.4 are included:

[0106] Step 3.1, optical monitoring device arrangement and data acquisition.

[0107] Specifically, optical sensors (visible light, infrared light, laser scattering, etc.) are installed in the reaction kettle, crystallization tank or key pipeline to monitor the turbidity, crystallization particle characteristics and color change of the solution in real time. Among them, the arrangement position of the optical sensor is preferentially selected at the discharge port of the reaction kettle or the entrance of the crystallization workshop, etc. which is easy to observe the crystallization state. If the reaction kettle volume is large, multiple optical probes can be installed at different heights. In addition, the sampling frequency of the optical sensor can be set in the range of 1s-30s according to the production rhythm to ensure sufficient time resolution. Finally, based on the arrangement position and sampling frequency of the optical sensor, the corresponding optical data vector is constructed, including light transmittance, light scattering intensity and other optical characteristics (such as infrared absorption peak intensity). The above optical monitoring data is helpful to judge the particle size and impurity content in the solution.

[0108] Step 3.2, optical data fusion and abnormal identification.

[0109] Specifically, the optical data vector is fused with the existing multi-source sensor data to comprehensively judge the solution internal crystallization and impurity situation, and timely find the abnormal situation deviating from the normal range. First, based on the optical data vector and the foregoing sensor data, a fusion vector is constructed, so that all elements in the fusion vector are synchronously collected at the same time point for real-time analysis. Subsequently, set the normal working interval threshold, such as the optical transmittance in the range of 0.8-1 is the normal working range, use the algorithm based on the normal working interval threshold or machine learning to detect the abnormality of the fusion vector. Once the abnormality is identified, the control system can be triggered for early warning or automatic adjustment to avoid the decrease of yield or production quality caused by excessive crystallization or impurity aggregation, while outputting the corresponding abnormal indication quantity.

[0110] Step 3.3, control strategy linked with material scheduling.

[0111] Specifically, when an abnormality (such as early crystallization or impurity enrichment) is detected, the feeding rate or material transportation time can be dynamically adjusted to achieve adaptive linkage control. First, a control strategy function is defined based on the abnormality indicator and the revised amount of the original scheduling strategy (such as delaying the feeding time or reducing the feeding batch size), and when the abnormality indicator is not 0, the production impact is verified through the digital twin model, and then the scheduling variables are revised to obtain a revised new schedule to ensure that the production process does not lose control due to the occurrence of abnormal conditions. If the optical monitoring repeatedly prompts that the solution turbidity is high, the feeding system can be instructed to temporarily suspend the addition of the next batch of boron ore or sulfuric acid, and increase the stirring or increase the solution flow rate to speed up the impurity settlement.

[0112] Step 3.4, perform closed-loop feedback.

[0113] Specifically, the scheduling correction and control instructions are conveyed to the field execution layer (such as automatic valves, pumps, mixers, etc.), and the execution results are fed back to the digital twin model to form a complete closed-loop control. First, the revised new schedule and control strategy are sent to the corresponding equipment through the PLC system or DCS system, including specific instructions such as on-off valves, adjusted stirring, and feeding amount control. Then, the temperature, pH, and optical monitoring data after actual operation are detected again for abnormalities, and if there are still abnormalities, the current schedule is revised again, and if it has returned to normal, the next batch of feeding production is continued according to the current established schedule. Finally, all data of the current batch process control (such as the time when the abnormality occurs, the response measures) are recorded and synchronized to the digital twin model for model optimization.

[0114] Step 4, scheduling system iterative evolution.

[0115] The data of real-time monitoring and scheduling execution are returned to the digital twin and machine learning model to form a closed loop to continuously correct and upgrade the scheduling strategy and process parameters.

[0116] Specifically, it includes steps 4.1-4.3:

[0117] Step 4.1, data collection and historical record archiving.

[0118] Specifically, based on all the data (control data log, abnormality record, and optical monitoring data) recorded and returned in step 3.4 and the current schedule execution results, a historical data matrix is constructed for subsequent model learning or statistical analysis.

[0119] Step 4.2, algorithm training and parameter updating.

[0120] Specifically, a machine learning model (such as LSTM or random forest) is used to predict the crystallization state or energy consumption at different times, and the scheduling decision or optical threshold is optimized accordingly, and a corresponding target optimization function is defined, which is determined by the output of finished boric acid, total energy consumption, abnormal downtime and their respective adjustable weight coefficients. Finally, through training, the process and scheduling strategy that maximizes the optimization target function are found, and the updated model parameters (such as the optimal feeding frequency and improved threshold) are obtained.

[0121] Step 4.3, digital twin revision and simulation verification.

[0122] Specifically, the core equation in the digital twin model is optimized according to the updated parameters to ensure that the virtual factory closely corresponds to the real factory. A round of simulated production can be performed in the simulation environment to observe whether the product quality, energy consumption and optical monitoring value curve match the expected results. If the error between the expected results predicted by the revised digital twin and the real workshop does not exceed the set threshold, it means that the current scheduling revision is successful.

[0123] The boron acid production material scheduling system based on the Internet of Things provided by the application is described below. The boron acid production material scheduling system based on the Internet of Things described below can be mutually corresponding and referred to the boron acid production material scheduling method based on the Internet of Things described above.

[0124] In one embodiment, a boron acid production material scheduling system based on the Internet of Things includes a data preprocessing module, a model construction module, a scheduling execution module, and a scheduling strategy adjustment module.

[0125] The data preprocessing module is used to obtain multi-source data from the boron acid production line and preprocess the multi-source data to map the preprocessed multi-source data to the digital twin model.

[0126] The model construction module is used to extract the scheduling demand of the current workshop from the initial parameters of the digital twin model and the preprocessed multi-source data vector, and construct a multi-objective scheduling model when the scheduling demand meets the production demand of the workshop.

[0127] The scheduling execution module is used to monitor the turbidity, crystalline particle characteristics and color change of the solution on the production line in real time through the optical system to obtain optical monitoring data, and determine the crystallization state and impurity content inside the solution in combination with the multi-source data to perform scheduling execution when the crystallization state or impurity content deviates from the normal range.

[0128] The scheduling strategy adjustment module is used to feed back the optical monitoring data and scheduling execution to the digital twin model based on the abnormal indication to correct the scheduling strategy and process parameters of the boron acid production line.

[0129] The multi-source data includes sensor data of temperature, flow, pH and pressure of multiple key nodes installed on the boric acid production line, the preprocessing includes missing value filling and denoising filtering, and the multi-objective scheduling model is used to output scheduling decisions according to scheduling requirements and resource lists.

[0130] In the embodiment, the data preprocessing module of the boric acid production material scheduling system based on the Internet of Things is specifically used for:

[0131] The raw data vector is generated based on instantaneous readings from different sensors of the boric acid production line, and the raw data vector is subjected to missing value filling and denoising filtering to obtain the preprocessed data vector.

[0132] The key process variables are determined based on the preprocessed data vector, and the key process variables are mapped to the digital twin structure to build the digital twin model to obtain an initial parameter vector of the digital twin model.

[0133] The key process variables include temperature, boron ore impurity content and flow, and the initial parameter vector includes initial modeling values of reaction temperature, flow and pH of the boric acid production line.

[0134] In the embodiment, the data preprocessing module of the boric acid production material scheduling system based on the Internet of Things is specifically used for:

[0135] The established digital twin model is compared with the actual monitoring data to obtain an error between the model output and the actual monitoring data, and it is judged whether the error exceeds a first threshold value.

[0136] When the error exceeds the first threshold value, the initial parameters of the digital twin model are reprocessed or the bias of the sensor is corrected to obtain a verification result of the digital twin model and a corrected model parameter.

[0137] In the embodiment, the model construction module of the boric acid production material scheduling system based on the Internet of Things is specifically used for:

[0138] The key parameters of total feeding amount, feeding frequency and expected yield are determined according to the inventory data, flow, temperature and pH of the boric acid production line in the digital twin model, and the scheduling requirement vector is generated by combining the workshop equipment capacity and personnel configuration.

[0139] A multi-objective function for defining a decision variable vector is set, and a scheduling decision variable vector is obtained based on the multi-objective function to construct a multi-objective scheduling model.

[0140] The multi-objective function includes minimizing total transportation time, minimizing energy consumption and maximizing capacity utilization, and the variable vector for decision includes route selection of all batches of feeding time or transportation vehicles.

[0141] In the embodiment, the boron acid production material scheduling system based on Internet of Things provided by the application is characterized in that the model construction module is further used for:

[0142] When the optical sensor monitors that the solution turbidity or the crystallization particle size is abnormal, the first feedback signal is triggered and transmitted through the data interaction interface between the multi-objective scheduling model and the optical system, and when the multi-objective scheduling model captures the first feedback signal, the corresponding feeding time or batch sequence is fine-tuned.

[0143] When the crystallization rate of the current feeding batch exceeds the second threshold value, the second feedback signal is triggered and transmitted, and when the multi-objective scheduling model captures the second feedback signal, the subsequent material delivery time is delayed or the single feeding amount is reduced.

[0144] The adaptive function is constructed based on the first feedback signal and the second feedback signal, so that the revised scheduling decision is obtained on the basis of the original adjustment strategy according to the adaptive function, and the revised scheduling decision is used to generate the scheduling execution instruction.

[0145] In the embodiment, the boron acid production material scheduling system based on Internet of Things provided by the application is characterized in that the scheduling execution module is used for:

[0146] The optical sensor installed at the discharge port of the reaction kettle or the entrance of the crystallization workshop is used to monitor the turbidity, crystallization particle characteristics and color change of the solution in real time according to the set sampling frequency, so as to obtain the optical data vector, and the optical data vector is fused with the multi-source data to obtain the fusion vector.

[0147] The fusion vector is detected for abnormality based on the set normal working interval threshold value or the machine learning algorithm, and when the fusion vector is abnormal, the control system is triggered to generate the abnormal index quantity.

[0148] The control strategy function is obtained based on the abnormal index quantity and the correction amount of the original adjustment strategy, and when the abnormal index quantity is the first value, the production scheduling variable is revised through the digital twin model to obtain the revised scheduling scheme.

[0149] In the embodiment, the boron acid production material scheduling system based on Internet of Things provided by the application is characterized in that the scheduling strategy adjustment module is used for:

[0150] The execution instruction and the execution data log are obtained based on the revised scheduling scheme, and the historical data matrix is constructed by combining the abnormal record and the optical monitoring data, so that the machine learning model is used to predict the crystallization state or energy consumption at different times based on the historical data matrix.

[0151] The scheduling strategy and optical threshold are optimized based on the prediction result of the machine learning model to construct an optimization objective function, obtain updated new model parameters, and map the new model parameters to the digital twin model.

[0152] The present disclosure can be a system, a method, and / or a computer program product. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present disclosure.

[0153] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium can be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or punched tape, a

[0154] The computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0155] Computer readable program instructions for carrying out operations of the present disclosure can be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate array (FPGA), or programmable logic array (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0156] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0157] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other data storage device. When the computer readable program instructions are loaded into the computer and other programmable data processing apparatus, a series of operational steps are implemented that provide processes such that the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0158] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0159] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0160] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limiting the present application, and although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or equivalent replacements without departing from the spirit and scope of the present application, and any modifications or equivalent replacements without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A boron acid production material scheduling method based on the Internet of Things, characterized by, The method comprises: obtaining multi-source data from a boric acid production line, and preprocessing the multi-source data to map the preprocessed multi-source data to a digital twin model; extracting scheduling requirements of a current workshop from initial parameters of the digital twin model and a preprocessed multi-source data vector, and constructing a multi-objective scheduling model when the scheduling requirements meet production requirements of the workshop; real-time monitoring of turbidity, crystalline particle characteristics and color changes of a solution on the production line through an optical system to obtain optical monitoring data, and determining the internal crystallization state and impurity content of the solution in combination with the multi-source data to perform scheduling when the crystallization state or impurity content deviates from a normal range; feeding back the optical monitoring data and scheduling execution to the digital twin model based on an abnormality indication to correct scheduling strategies and process parameters of the boric acid production line; wherein the multi-source data comprises sensor data of temperature, flow rate, pH and pressure installed at multiple key nodes on the boric acid production line, the preprocessing comprises missing value filling and denoising filtering, and the multi-objective scheduling model is used to output scheduling decisions according to scheduling requirements and a resource list.

2. The IoT-based boron acid production material scheduling method according to claim 1, characterized in that, The method comprises: generating an original data vector based on instantaneous readings from different sensors of the boric acid production line, and performing missing value filling and denoising filtering on the original data vector to obtain a preprocessed data vector; determining key process variables based on the preprocessed data vector, and mapping the key process variables to a digital twin structure to build the digital twin model to obtain an initial parameter vector of the digital twin model; wherein the key process variables comprise temperature, boron ore impurity content and flow rate, and the initial parameter vector comprises initial modeling values of reaction temperature, flow rate and pH of the boric acid production line.

3. The IoT-based boron acid production material scheduling method according to claim 2, wherein, The method further comprises: comparing the established digital twin model with actual monitoring data to obtain an error between model output and actual monitoring data, and determining whether the error exceeds a first threshold value; when the error exceeds the first threshold value, reprocessing initial parameters of the digital twin model or correcting a bias of the sensor to obtain a verification result of the digital twin model and a corrected model parameter.

4. The IoT-based boron acid production material scheduling method according to claim 3, wherein, The method further comprises: determining key parameters of total feeding amount, feeding frequency and expected yield based on inventory data, flow rate, temperature and pH of the boric acid production line in the digital twin model, and combining workshop equipment capacity and personnel configuration to generate a scheduling requirement vector; Set a multi-objective function for defining a decision variable vector, and obtain a scheduling decision variable vector based on the multi-objective function, to construct the multi-objective scheduling model; Wherein, the multi-objective function includes minimizing total transportation time, minimizing energy consumption, and maximizing production capacity utilization, and the decision variable vector includes all batch feeding times or route selection of transport vehicles within the scheduling decision number.

5. The IoT-based boron acid production material scheduling method according to claim 4, wherein, The scheduling demand of the current workshop is extracted from the initial parameters of the digital twin model and the pre-processed multi-source data vector, and when the scheduling demand meets the production demand of the workshop, a multi-objective scheduling model is constructed, which further comprises: When the optical sensor monitors the abnormality of solution turbidity or crystalline particle size, a first feedback signal is triggered and transmitted through the data interaction interface between the multi-objective scheduling model and the optical system, and when the multi-objective scheduling model captures the first feedback signal, the corresponding feeding time or batch order is fine-tuned; When the crystallization rate of the current feeding batch exceeds a second threshold, a second feedback signal is triggered and transmitted, and when the multi-objective scheduling model captures the second feedback signal, the subsequent material delivery time is delayed or the single feeding amount is reduced; An adaptive function is constructed based on the first feedback signal and the second feedback signal to obtain a revised scheduling decision based on the adaptive function on the basis of the original adjustment strategy, and the revised scheduling decision is used to generate scheduling execution instructions.

6. The IoT-based boron acid production material scheduling method according to claim 5, wherein, The optical system is used to monitor the turbidity, crystalline particle characteristics and color change of the solution in real time, and the optical monitoring data is obtained, and the solution internal crystallization state and impurity content are determined in combination with the multi-source data, so that scheduling execution is performed when the crystallization state or impurity content deviates from the normal range, which comprises: According to the set sampling frequency, the optical sensor installed at the discharge port of the reaction kettle or the entrance of the crystallization workshop is used to monitor the turbidity, crystalline particle characteristics and color change of the solution in real time, and an optical data vector is obtained, and the optical data vector is fused with the multi-source data to obtain a fusion vector; Anomaly detection is performed on the fusion vector based on a set normal working interval threshold or a machine learning algorithm, and when the fusion vector is abnormal, the control system is triggered for early warning and automatic adjustment to generate an abnormal index quantity; Based on the abnormal index quantity and the modification amount of the original adjustment strategy, a defined control strategy function is obtained, and when the abnormal index quantity is a first value, the production scheduling variable is revised by the digital twin model to obtain a revised scheduling scheme.

7. The IoT-based boron acid production material scheduling method according to claim 6, wherein, The optical monitoring data and scheduling execution situation are fed back to the digital twin model based on the abnormal indication to modify the scheduling strategy and process parameters of the boric acid production line, which comprises: Based on the revised scheduling scheme, execution instructions and execution data logs are obtained, and in combination with abnormal records and optical monitoring data, a historical data matrix is constructed to predict the crystallization state or energy consumption at different times based on the historical data matrix by using a machine learning model; Optimize the scheduling strategy and optical threshold based on the prediction result of the machine learning model to construct an optimization objective function, obtain updated new model parameters, and map the new model parameters to the digital twin model.

8. A boron acid production material scheduling system based on the Internet of Things, characterized by, The system comprises: A data preprocessing module for obtaining multi-source data from a boric acid production line and preprocessing the multi-source data to map the preprocessed multi-source data to a digital twin model; A model construction module for extracting scheduling requirements of a current plant from initial parameters of the digital twin model and a preprocessed multi-source data vector, and constructing a multi-objective scheduling model when the scheduling requirements meet the production requirements of the plant; A scheduling execution module for real-time monitoring of the turbidity, crystalline particle characteristics, and color changes of the solution on the production line through an optical system to obtain optical monitoring data, and determining the internal crystallization state and impurity content of the solution in combination with the multi-source data to perform scheduling when the crystallization state or impurity content deviates from the normal range; A scheduling strategy adjustment module for feeding back the optical monitoring data and scheduling execution to the digital twin model based on abnormal indications to correct the scheduling strategy and process parameters of the boric acid production line; Wherein, the multi-source data includes sensor data of temperature, flow, pH, and pressure installed at multiple key nodes on the boric acid production line, the preprocessing includes missing value filling and denoising filtering, and the multi-objective scheduling model is used to output scheduling decisions according to scheduling requirements and resource lists.

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to the instructions to perform the steps of the method according to any one of claims 1-7.

10. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps of the method according to any one of claims 1-7.

Images