CN118261417A - Risk monitoring and early warning method and system for lithium hydroxide production process - Google Patents

Abstract

The embodiment of the disclosure discloses a risk monitoring and early warning method and system for a lithium hydroxide production process. The method comprises the following steps: preprocessing the acquired first data set to obtain a second data set; the first data set includes real-time data during lithium hydroxide production; training the constructed deep learning model based on the historical data set to obtain a first model; obtaining risk information based on a second dataset and the first model; determining an uncertainty factor to be quantified in the lithium hydroxide production process; uncertainty factors include raw material quality fluctuations and production environment changes; quantifying uncertainty factors based on a preset uncertainty quantification strategy to obtain evaluation information; and obtaining an early warning strategy based on the evaluation information and the risk information. The method can effectively improve the accuracy and the reliability of the early warning system.

Description

Risk monitoring and early warning method and system for lithium hydroxide production process

Technical Field

The disclosure relates to the technical field of environmental monitoring and early warning, in particular to a risk monitoring and early warning method and system for a lithium hydroxide production process.

Background

Lithium hydroxide production is a complex and multifactorial process, and conventional monitoring and risk early warning systems often rely on simple threshold decisions or empirical rules, which limit their effectiveness in handling complex scenarios and dynamic changes. Despite the significant progress in pattern recognition and prediction made in deep learning in recent years, these models still face challenges in dealing with highly uncertain and dynamically changing production environments, resulting in low accuracy and poor reliability of the actual early warning system.

Disclosure of Invention

In view of the above, the embodiments of the present disclosure provide a risk monitoring and early warning method and system for a lithium hydroxide production process, which can effectively improve the accuracy and reliability of the early warning system.

In a first aspect, an embodiment of the present disclosure provides a risk monitoring and early warning method for a lithium hydroxide production process, including:

Preprocessing the acquired first data set to obtain a second data set;

The first data set includes real-time data during lithium hydroxide production;

training the constructed deep learning model based on the historical data set to obtain a first model;

obtaining risk information based on the second dataset and the first model;

determining an uncertainty factor to be quantified in the lithium hydroxide production process; the uncertainty factors comprise raw material quality fluctuation and production environment change;

Quantifying the uncertainty factors based on a preset uncertainty quantification strategy to obtain evaluation information;

and acquiring an early warning strategy based on the evaluation information and the risk information.

Optionally, the first data set includes temperature, pressure, chemical concentration.

Optionally, the preprocessing includes one or more of data cleansing, missing value processing, time window partitioning, normalization, or normalization.

Optionally, training the constructed deep learning model based on the historical dataset to obtain a first model, including:

Dividing the historical data set into a training set and a verification set;

Constructing the deep learning model;

determining a loss function and an optimization algorithm matched with monitoring of the lithium hydroxide production process;

Training the deep learning model based on the training set until the preset iteration times are reached or the deep learning model converges;

and verifying the trained deep learning model by adopting the verification set, wherein the deep learning model meeting the verification condition is the first model.

Optionally, the loss function comprises a cross entropy loss function or a mean square error function;

The optimization algorithm comprises a gradient descent algorithm or an Adam algorithm.

Optionally, the quantifying the uncertainty factor based on a preset uncertainty quantifying policy to obtain evaluation information includes:

Constructing a neural network model;

Training the neural network model by using a preset strategy, and obtaining the distribution information of key parameters based on the trained neural network model;

And quantifying uncertainty factors in the production process based on the trained neural network model and the distribution information to obtain evaluation information.

Optionally, the neural network model includes any one of a markov chain model or a bayesian neural network model.

Optionally, when the neural network model is a bayesian neural network model, a lower variance bound is used as a loss function in training the neural network model.

Alternatively, when the neural network model is a markov chain model, a plurality of random samples are taken using a monte carlo method, and the production process is simulated by sampling the obtained samples.

In a second aspect, an embodiment of the present disclosure further provides a risk monitoring and early warning system for a lithium hydroxide production process, including:

The processing module is configured to preprocess the collected first data set to obtain a second data set;

The first data set includes real-time data during lithium hydroxide production;

the training module is configured to train the constructed deep learning model based on the historical data set to obtain a first model;

an analysis module configured to obtain risk information based on the second dataset and the first model;

The determining module is configured to determine uncertainty factors to be quantified in the lithium hydroxide production process; the uncertainty factors comprise raw material quality fluctuation and production environment change;

The quantization module is configured to quantize the uncertainty factors based on a preset uncertainty quantization strategy to obtain evaluation information;

And the acquisition module is configured to acquire an early warning strategy based on the evaluation information and the risk information.

In a third aspect, an embodiment of the present disclosure further provides an electronic device, which adopts the following technical scheme:

the electronic device includes:

At least one processor; and

A memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor to enable the at least one processor to perform any one of the above methods for risk monitoring and early warning of a lithium hydroxide production process.

In a fourth aspect, the embodiments of the present disclosure further provide a computer readable storage medium storing computer instructions for causing a computer to perform any one of the above methods for risk monitoring and early warning of a lithium hydroxide production process.

According to the method disclosed by the application, the latest state of the process can be timely obtained by collecting the real-time data in the lithium hydroxide production process and preprocessing the real-time data, so that the monitoring process is more accurate and timely, and the abnormal condition in the production process can be timely found; the model can be used for predicting and detecting potential risk conditions by constructing a deep learning model and training based on historical data, and possible problems can be found in advance and corresponding measures can be taken to avoid or reduce the influence of risk occurrence by acquiring risk information based on a second data set and the first model; the influence of the uncertainty factors can be quantized into specific evaluation information by determining the uncertainty factors to be quantized in the lithium hydroxide production process and quantizing the factors based on a preset quantization strategy, so that the importance degree of the uncertainty factors can be analyzed and understood, and more visual guidance is provided for decision making; the comprehensive analysis and evaluation of the lithium hydroxide production process can be performed by integrating the evaluation information and the risk information, and the corresponding early warning strategy can be formulated based on the evaluation information and the risk information, so that a decision maker in the production process can better understand the current situation and take appropriate action; the method can help monitor abnormal conditions and risks in the production process, guide decision makers to adjust and control timely through an early warning strategy, and is beneficial to improving the efficiency and quality of the production process, reducing possible loss and risk and ensuring stable production of products.

The foregoing description is only an overview of the disclosed technology, and may be implemented in accordance with the disclosure of the present disclosure, so that the above-mentioned and other objects, features and advantages of the present disclosure can be more clearly understood, and the following detailed description of the preferred embodiments is given with reference to the accompanying drawings.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present disclosure, and other drawings may be obtained according to these drawings without inventive effort to a person of ordinary skill in the art.

Fig. 1 is a schematic flow chart of a risk monitoring and early warning method in a lithium hydroxide production process according to an embodiment of the disclosure.

Fig. 2 is a flowchart of a method for obtaining a first model according to an embodiment of the disclosure.

Fig. 3 is a flowchart illustrating a method for obtaining evaluation information according to an embodiment of the present disclosure.

Fig. 4 is a schematic block diagram of a risk monitoring and early warning system in a lithium hydroxide production process according to an embodiment of the disclosure.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.

Detailed Description

Embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

It should be appreciated that the following specific embodiments of the disclosure are described in order to provide a better understanding of the present disclosure, and that other advantages and effects will be apparent to those skilled in the art from the present disclosure. It will be apparent that the described embodiments are merely some, but not all embodiments of the present disclosure. The disclosure may be embodied or practiced in other different specific embodiments, and details within the subject specification may be modified or changed from various points of view and applications without departing from the spirit of the disclosure. It should be noted that the following embodiments and features in the embodiments may be combined with each other without conflict. All other embodiments, which can be made by one of ordinary skill in the art without inventive effort, based on the embodiments in this disclosure are intended to be within the scope of this disclosure.

It is noted that various aspects of the embodiments are described below within the scope of the following claims. It should be apparent that the aspects described herein may be embodied in a wide variety of forms and that any specific structure and/or function described herein is merely illustrative. Based on the present disclosure, one skilled in the art will appreciate that one aspect described herein may be implemented independently of any other aspect, and that two or more of these aspects may be combined in various ways. For example, an apparatus may be implemented and/or a method practiced using any number of the aspects set forth herein. In addition, such apparatus may be implemented and/or such methods practiced using other structure and/or functionality in addition to one or more of the aspects set forth herein.

It should also be noted that the illustrations provided in the following embodiments merely illustrate the basic concepts of the disclosure by way of illustration, and only the components related to the disclosure are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, and the form, number and proportion of the components in actual implementation may be arbitrarily changed, and the layout of the components may be more complicated.

In addition, in the following description, specific details are provided in order to provide a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects may be practiced without these specific details.

Referring to fig. 1, the application discloses a method for monitoring and risk early warning in a lithium hydroxide production process, which comprises the following steps:

s100, preprocessing the collected first data set to obtain a second data set; the first data set includes real-time data during lithium hydroxide production.

In this embodiment, the first data set includes other data such as temperature, pressure, chemical concentration, etc.

And S200, training the constructed deep learning model based on the historical data set to obtain a first model.

And S300, acquiring risk information based on the second data set and the first model.

S400, determining uncertainty factors to be quantified in the lithium hydroxide production process; uncertainty factors include fluctuations in raw material quality and changes in production environment.

S500, quantifying uncertainty factors based on a preset uncertainty quantification strategy to obtain evaluation information.

S600, acquiring an early warning strategy based on the evaluation information and the risk information.

Wherein the preprocessing preferably includes one or more of data cleansing, missing value processing, time window partitioning, normalization or normalization to facilitate training and prediction of the deep learning model.

According to the method disclosed by the application, the latest state of the process can be timely obtained by collecting the real-time data in the lithium hydroxide production process and preprocessing the real-time data, so that the monitoring process is more accurate and timely, and the abnormal condition in the production process can be timely found; the model can be used for predicting and detecting potential risk conditions by constructing a deep learning model and training based on historical data, and possible problems can be found in advance and corresponding measures can be taken to avoid or reduce the influence of risk occurrence by acquiring risk information based on a second data set and the first model; the influence of the uncertainty factors can be quantized into specific evaluation information by determining the uncertainty factors to be quantized in the lithium hydroxide production process and quantizing the factors based on a preset quantization strategy, so that the importance degree of the uncertainty factors can be analyzed and understood, and more visual guidance is provided for decision making; the comprehensive analysis and evaluation of the lithium hydroxide production process can be performed by integrating the evaluation information and the risk information, and the corresponding early warning strategy can be formulated based on the evaluation information and the risk information, so that a decision maker in the production process can better understand the current situation and take appropriate action; the method can help monitor abnormal conditions and risks in the production process, guide decision makers to adjust and control timely through an early warning strategy, and is beneficial to improving the efficiency and quality of the production process, reducing possible loss and risk and ensuring stable production of products.

Therefore, the method can improve the monitoring capability of the production process, reduce the risk, optimize the decision and finally improve the production efficiency and quality through real-time monitoring, risk early warning, uncertainty factor quantification, comprehensive evaluation and decision support.

Referring to fig. 2, the method of obtaining the first model includes the steps of:

s210, dividing the historical data set into a training set and a verification set.

In this embodiment, 70% of the data in the historical data set is used for training, i.e., as a training set, and the remaining 30% of the data in the historical data set is used for verification, i.e., as a verification set.

S220, constructing a deep learning model.

In this embodiment, the deep learning model may be constructed as a convolutional neural network, a cyclic neural network, a long short term memory network (LSTM), a transducer, or the like.

In addition, traditional machine learning models, such as decision trees, support vector machines, random forests, and the like, can also be constructed.

S230, determining a loss function and an optimization algorithm matched with monitoring of the lithium hydroxide production process.

S240, training the deep learning model based on the training set until the preset iteration times or the deep learning model converges.

In a specific training, parameters (e.g., learning rate, batch size, number of iterations) are adjusted to optimize model performance.

S250, verifying the trained deep learning model by using a verification set, wherein the deep learning model meeting the verification conditions is the first model.

In the embodiment, the historical data set is divided into the training set and the verification set, and after the model is trained by the training set, the model is verified on the verification set, so that the first model has good generalization capability and can adapt to different production conditions and data distribution; when the deep learning model is constructed, a proper model type, such as a convolutional neural network, a cyclic neural network or a transducer, can be selected according to the complexity of tasks and the characteristics of data so as to improve the capturing capability of the data characteristics, and in addition, the traditional machine learning model can be considered, so that the first model has more selection and flexibility; the loss function and the optimization algorithm matched with the lithium hydroxide production process monitoring are determined, so that the first model can be better converged in the training process and can be better adapted to a specific production monitoring task; through training of a model based on a training set and a verification process of a verification set, the first model can be verified after being fully trained, and reliability and effectiveness of the model are guaranteed; after verification, the first model can better meet the requirements of monitoring the lithium hydroxide production process, improves the accuracy and generalization capability of the model, and can better adapt to the requirements of actual production scenes.

The first model obtained through the steps has better generalization capability, higher accuracy and better adaptability, and can better meet the monitoring requirement of the lithium hydroxide production process, for example, the abnormal mode and the potential risk in the production process can be better identified.

Furthermore, the accuracy and generalization capability of the model can be verified by methods such as cross verification and the like, so that the model can be ensured to work effectively under different production conditions. Specifically, it includes testing the model over different time periods to ensure that it maintains good performance under a variety of conditions.

In this embodiment, the deep learning model is preferably a long-short-term memory network (LSTM), specifically, according to the data characteristics (such as long-term dependency of time-series data) in the lithium hydroxide production process, the LSTM network architecture is designed, including determining the layer number, the number of neurons, the connection mode, and the like, specifically, including the settings of the forgetting gate, the input gate, and the output gate, so as to effectively process the time-series data.

In a specific training, suitable loss functions and optimizers may be set, such as using Mean Square Error (MSE) as the loss function, adam optimizers, etc.

Then, the network weights are adjusted by iterative training to achieve the best predicted performance.

Finally, the model can be optimized according to the verification result, such as adjusting the layer number, the neuron number, the learning rate and the like, so as to improve the model performance.

The method for verifying the trained model by using the verification set specifically comprises the following steps: by inputting the verification set into the model, calculating a prediction result, comparing the prediction result with a real label, and calculating an evaluation index (such as accuracy, precision, recall, and the like) to evaluate the performance of the model.

Further, the application also comprises the evaluation of the model performance, which specifically comprises the following steps: various performance metrics (e.g., accuracy, recall, F1 score) are used to evaluate the predictive capabilities of the model.

The analytical model performs in predicting key parameters (e.g., temperature, pressure, etc.) in the lithium hydroxide production process.

Analyzing the performance of the model under different conditions, and identifying the strong items and weak items of the model.

And according to the comparison of the prediction result of the model and the actual production condition, an improvement proposal and an optimization direction are provided.

Further optimization of the model based on the validation results may include adjusting architecture, adding regularization (e.g., dropout) to prevent overfitting, using different optimizers, etc.

Model tuning is implemented, particularly for conditions and data characteristics that are unique to the production environment. And deploying the trained model into a production environment, and integrating the trained model with the existing monitoring system.

Ensuring the stability and reliability of the model in actual operation includes continuously monitoring its performance and periodically performing maintenance and updating.

In this embodiment, the acquiring risk information specifically includes: and analyzing the second data set by using the first model, judging whether the data in the second data set is larger than a corresponding preset threshold value, if so, obtaining a corresponding risk level according to the exceeding degree, and if not, judging that the risk is absent.

For example, the corresponding risk levels can be classified into low risk, medium risk, high risk and high risk so as to be convenient for corresponding to specific measures, and quick response to different situations on site can be realized, so that the production on site is not delayed, and the production safety on site can be ensured.

Referring to fig. 3, the method for obtaining evaluation information specifically includes the steps of:

S510, constructing a neural network model.

Wherein the neural network model comprises any one of a Markov chain model or a Bayesian neural network model.

S520, training a neural network model by using a preset strategy, and obtaining the distribution information of the key parameters based on the trained neural network model simulation.

The key parameters include product quality, yield and the like.

And S530, quantifying uncertainty factors in the production process based on the trained neural network model and the distribution information to obtain evaluation information.

In the embodiment, by constructing the neural network model, a great amount of data can be trained and learned by utilizing the strong learning capability of the neural network model, so that the accuracy of information is improved, and compared with a traditional statistical method, the neural network can better identify the mode and rule in the data, so that the evaluation information is more accurately obtained; the neural network model can be adaptively adjusted according to different data and conditions, so that the neural network model is suitable for the change and uncertainty factors in different production processes, and the acquisition of the evaluation information is more flexible and adaptive; by training the neural network model by using a preset strategy, the influence of a plurality of factors on the evaluation information can be comprehensively considered, for example, the product quality and the yield can be influenced by the plurality of factors, and the neural network model can comprehensively measure the influence of the factors on the evaluation information by learning the complex relation among the factors; the neural network model can learn and evaluate the data in real time or near real time, so as to provide evaluation information in time, which is very important for the situation that in the production process, timely adjustment and decision making are needed, and can help the production process to be more efficient and optimized.

In conclusion, the method for obtaining the evaluation information by utilizing the neural network model has the advantages of high accuracy, self-adaptability, comprehensiveness and instantaneity, and can provide more reliable basis for decision and optimization in the production process.

When the neural network model is a bayesian neural network model, a suitable algorithm (e.g., variational inference) is used to train the bayesian neural network and estimate the posterior distribution of weights.

Through training of a large amount of data, the distribution of the weights is updated to be close to the real posterior distribution.

Specifically, a variance inference is employed, which is a method of estimating a posterior distribution of a probabilistic model. In bayesian neural networks, it is used to approximate complex posterior distributions of weights. Then, a simple distribution (e.g., gaussian) is found that is as close as possible to the true posterior distribution. Next, a standard neural network architecture is designed to determine the number of layers required and the number of neurons per layer. Then, introducing a Bayesian framework; a priori distribution, typically gaussian, is introduced for each weight.

A simple approximate posterior distribution (e.g., gaussian) is chosen for each weight, whose parameters (mean and variance) will be learned during the training process.

A lower variation bound (ELBO, evidence LowerBOund) is used as a loss function. ELBO is the difference between log likelihood and KL-divergence (Kullback-Leibler divergence) that aims to maximize the likelihood of the data while keeping the approximate posterior distribution as close as possible to the true posterior distribution.

Random gradient descent or variants thereof are used to optimize ELBO. In each iteration, the parameters approximating the posterior distribution are updated.

Evaluation of posterior distribution: the weights are sampled from the optimized approximate posterior distribution to evaluate the prediction uncertainty of the model.

Model verification: and predicting the verification data set by using the weight obtained by sampling, and evaluating the performance of the model.

The uncertainty of the model is interpreted from the resulting posterior distribution and applied to actual problems such as predictions, classifications or risk assessment.

Furthermore, the network architecture or the optimization method can be adjusted according to the feedback of practical application, so that the performance and accuracy of the model are further improved.

Through the steps, the Bayesian neural network can be effectively trained by using variation inference, and posterior distribution of weights is estimated; this approach provides a powerful tool to evaluate and handle neural network prediction uncertainty.

And predicting the lithium hydroxide production process by using the trained Bayesian neural network, and calculating the uncertainty of the prediction.

The uncertainty can be quantified by observing the change of weight distribution, confidence interval of output or variance of prediction result. The model is then cross-validated and its performance on different data sets is checked to assess its generalization ability and accuracy of uncertainty quantification. And adjusting the model according to the verification result, such as changing the network structure, optimizing the training algorithm and the like.

Finally, applying the prediction result and uncertainty analysis of the Bayesian neural network to monitoring and risk early warning of the production process; providing uncertainty-based decision support, such as taking more discreet measures in cases of higher uncertainty.

Through the steps, the Bayesian neural network can play an important role in monitoring and risk early warning of the lithium hydroxide production process, not only can an accurate prediction result be provided, but also the uncertainty of prediction can be quantified, so that important support is provided for decision making of the production process.

When the neural network model is a Markov chain model, this model is capable of describing the probability of transitions between states, i.e., the probability of transitioning from one production state to another.

The monte carlo method is used for carrying out a large number of random samples, and various conditions possibly occurring in the production process are simulated.

From these samples, the distribution of key parameters in the production process, such as product quality, yield, etc., is estimated.

Markov Chain Monte Carlo (MCMC) is a set of algorithms for obtaining a series of random samples by building a markov chain whose distribution will approximate the target probability distribution after a sufficiently long iteration.

Markov chains are a stochastic process in which the state of the next sample depends only on the current state, independent of the previous state (no memory properties).

For how to use the monte carlo method for a large number of random samplings, the following is specific:

step A1: a target distribution is defined.

Determining a target probability distribution to be sampled; in the context of production process simulation, this may be a distribution regarding product quality, production time, cost, etc.

Step A2: a markov chain is constructed.

A Markov chain is constructed to have a smooth distribution matching the target distribution. This typically involves defining a transition probability, i.e. the probability of moving from the current state to the next state.

Common MCMC algorithms include the Metropolis-Hastings algorithm and Gibbs Sampling (Gibbs Sampling).

Step A3: initialization and iteration.

An initial state is selected and an iterative process is started.

In each iteration, the next sample is generated based on the transition probability and a decision is made as to whether to accept the new sample based on certain acceptance criteria (e.g., the Metropolis criterion).

Step A4: convergence and sampling.

After a sufficient number of iterations, the Markov chain will reach its plateau, at which point sample collection may begin.

The collected samples will be used to estimate characteristics of the target distribution, such as mean, variance, confidence interval, etc.

Step A5: simulating the production process.

The collected samples were used to simulate the production process. For example, production variations, quality control results or cost fluctuations under different production conditions can be simulated.

These simulation results are analyzed to assess risk and uncertainty that may occur during the production process.

Through the steps, the MCMC method can be effectively used for risk assessment and decision support in a complex production process. This approach is particularly useful for complex systems where direct sampling or analytical computation is difficult.

Based on the Markov chain model and the Monte Carlo simulation result, uncertainty in the production process is quantified. And calculating probability distribution of the key parameters, and analyzing possible influences on production efficiency and product quality.

Uncertainty is quantified using Markov chain models and Monte Carlo simulations, including in particular: 1) Constructing a Markov chain model; 2) Defining states and transition probabilities; 3) Determining various conditions in the production process, such as different production phases or conditions; 4) Establishing a state transition probability matrix describing the probability of transitioning from one state to another; 5) Model parameterization: estimating parameters of the model, such as transition probabilities, by historical data or expert knowledge; 6) Monte Carlo simulation: generating random samples, specifically: a number of random samples based on markov chains are generated using the monte carlo method. These samples represent different sequences of states that may occur during the production process; 7) Simulating the production process: collecting data about system behavior, such as production time, cost, quality control index, etc., by simulating multiple production processes; 8) Uncertainty quantization: carrying out statistical analysis on the results of Monte Carlo simulation, and calculating the distribution of key indexes such as average value, standard deviation, confidence interval and the like; 9) Evaluating the change of the indexes under different production states, and quantifying the uncertainty of the production process; 10 Based on the results of the statistical analysis, assessing the probability of potential risks, such as delays, hyperbranched or quality problems, in the production process; 11 Sensitivity analysis to determine which factors have the greatest effect on the uncertainty of the production process; 12 Using quantized uncertainty results to support production management and decision making. For example, optimizing resource allocation, formulating risk mitigation strategies; 13 The model is applied to continuous monitoring of the actual production process, and model parameters are adjusted according to the real-time data so as to improve the prediction accuracy.

Further, the quantified results are analyzed to identify a primary source of uncertainty in the production process. The quantified results are applied to risk management and decision support, such as adjusting production strategies, optimizing raw material ratios, and the like.

Further, the application also comprises verification and adjustment, in particular: verifying the accuracy and the effectiveness of the quantization method through actual production data; and carrying out necessary adjustment and optimization on the model and the method according to the verification result.

Through the steps, the Markov chain Monte Carlo method can be effectively used for uncertainty quantification of the lithium hydroxide production process. The method not only can provide probability distribution of key parameters, but also can help managers to better understand risks and uncertainties in the production process, so that more scientific and reasonable decisions can be made.

Data uncertainty is also incorporated into the overall quantitative analysis, taking into account errors or imperfections that may be present in the data itself.

And providing a comprehensive uncertainty evaluation result by combining the data and the model uncertainty.

The method for evaluating the effectiveness comprises the following steps: the validity of the uncertainty quantification method is tested through actual production data, so that the accuracy and the reliability of the uncertainty quantification method in actual application are ensured.

The performance of the method under different conditions can be evaluated by means of simulation experiments or historical data return.

And carrying out necessary optimization and adjustment on the uncertainty quantization method according to the performance in practical application.

The uncertainty quantization model is periodically evaluated and updated to account for new production conditions and data characteristics, taking into account dynamic changes in the production process.

And according to the quantized uncertainty evaluation result, a corresponding early warning strategy is formulated.

In the application, for training a model, a training set can be used for training the model, specifically: the training set is divided into small batches of data for training, and the size of each batch can be adjusted according to actual conditions. During the training of each batch, the gradient of the loss function is calculated by a back propagation algorithm, and the parameters are updated according to a selected optimization algorithm. The process is repeated until a preset number of iterations or model convergence is reached.

In the present application, the obtained early warning information may specifically include:

1. Alarm notification: when the assessment information or risk information exceeds a given threshold or a given range, the system can trigger an alarm notification to timely remind related personnel to pay attention to and take corresponding actions. For example, when a certain critical parameter in the lithium hydroxide production process is out of a safe range, an alarm notification is sent for timely adjustment and control.

2. And (3) automatic shutdown: when the assessment information or risk information reaches a dangerous level, the system can automatically trigger a shutdown mechanism to ensure the safety of the production process. For example, when a significant change in the production environment occurs or a certain indicator reaches a dangerous level, an automatic shutdown is performed to prevent further risk.

3. Emergency treatment guidelines: based on the assessment information and the risk information, the system may provide corresponding emergency treatment guidelines to assist the decision maker in making proper decisions and countermeasures in the event of an emergency. For example, based on the assessment information and risk information, countermeasures advice for different uncertainty factors are given to minimize potential impact.

4. Real-time monitoring and reporting: the system can provide real-time monitoring and reporting functions, continuously track the change trend of the assessment information and the risk information, and generate a related report for a decision maker to refer to. Thus, the state of the production process can be known in time, the possible risk can be predicted, and corresponding measures can be taken in time.

The obtained early warning strategy can provide timely feedback and warning and help a decision maker to take appropriate action so as to ensure the safety and stability of the production process.

The risk monitoring and early warning method for the lithium hydroxide production process disclosed by the application can be applied to the following embodiments.

Example one: temperature anomaly detection

The temperature data is analyzed using a deep learning model to identify potential anomaly patterns.

The uncertainty quantization method is used to evaluate the confidence level of the predictions and to increase the frequency of manual checks at low confidence levels.

Example two: chemical reaction process monitoring

The model is used to predict risk points that may occur during chemical reactions.

In combination with the uncertainty evaluation, the production parameters are adjusted to prevent potential safety problems.

Example three: pressure fluctuation analysis

And monitoring the real-time fluctuation of the pressure through a deep learning model, and predicting an abnormal event.

And adjusting an early warning threshold according to the uncertainty quantification result, and optimizing a risk management strategy.

The scheme disclosed by the application aims to improve a monitoring and risk early warning system for a lithium hydroxide production process. By combining the deep learning technology and the uncertainty quantization method, the accuracy and the reliability of the early warning system are effectively improved; the deep learning model is used to analyze production data and predict potential risks, while the uncertainty quantization method is used to evaluate the credibility of these predictions. In addition, the scheme also introduces a set of qualitative standards based on uncertainty evaluation to guide strategic switching between model solutions and traditional algorithms.

Referring to fig. 4, a second aspect of the present application discloses a risk monitoring and early warning system for lithium hydroxide production process, comprising:

the processing module is configured to preprocess the collected first data set to obtain a second data set; the first data set includes real-time data during lithium hydroxide production;

the training module is configured to train the constructed deep learning model based on the historical data set to obtain a first model;

An analysis module configured to obtain risk information based on a second dataset and the first model;

the determining module is configured to determine uncertainty factors to be quantified in the lithium hydroxide production process; uncertainty factors include raw material quality fluctuations and production environment changes;

the quantization module is configured to quantize the uncertainty factors based on a preset uncertainty quantization strategy to obtain evaluation information;

And the acquisition module is configured to acquire an early warning strategy based on the evaluation information and the risk information.

An electronic device according to an embodiment of the present disclosure includes a memory and a processor. The memory is for storing non-transitory computer readable instructions. In particular, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory. The volatile memory may include, for example, random Access Memory (RAM) and/or cache memory (cache), and the like. The non-volatile memory may include, for example, read Only Memory (ROM), hard disk, flash memory, and the like.

The processor may be a Central Processing Unit (CPU) or other form of processing unit having data processing and/or instruction execution capabilities, and may control other components in the electronic device to perform the desired functions. In one embodiment of the disclosure, the processor is configured to execute the computer readable instructions stored in the memory, so that the electronic device performs all or part of the steps of the foregoing risk monitoring and early warning method for lithium hydroxide production process according to various embodiments of the disclosure.

It should be understood by those skilled in the art that, in order to solve the technical problem of how to obtain a good user experience effect, the present embodiment may also include well-known structures such as a communication bus, an interface, and the like, and these well-known structures are also included in the protection scope of the present disclosure.

Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure. A schematic diagram of an electronic device suitable for use in implementing embodiments of the present disclosure is shown. The electronic device shown in fig. 5 is merely an example and should not be construed to limit the functionality and scope of use of the disclosed embodiments.

As shown in fig. 5, the electronic device may include a processor (e.g., a central processing unit, a graphic processor, etc.) that may perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM) or a program loaded from a storage device into a Random Access Memory (RAM). In the RAM, various programs and data required for the operation of the electronic device are also stored. The processor, ROM and RAM are connected to each other by a bus. An input/output (I/O) interface is also connected to the bus.

In general, the following devices may be connected to the I/O interface: input means including, for example, sensors or visual information gathering devices; output devices including, for example, display screens and the like; storage devices including, for example, magnetic tape, hard disk, etc.; a communication device. The communication means may allow the electronic device to communicate wirelessly or by wire with other devices, such as edge computing devices, to exchange data. While fig. 5 shows an electronic device having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.

In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from ROM. When the computer program is executed by a processor, all or part of the steps of the risk monitoring and early warning method for the lithium hydroxide production process in the embodiment of the disclosure are executed.

The detailed description of the present embodiment may refer to the corresponding description in the foregoing embodiments, and will not be repeated herein.

A computer-readable storage medium according to an embodiment of the present disclosure has stored thereon non-transitory computer-readable instructions. When executed by a processor, the non-transitory computer readable instructions perform all or part of the steps of the aforementioned risk monitoring and pre-warning method for lithium hydroxide production process of embodiments of the disclosure.

The computer-readable storage medium described above includes, but is not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or removable hard disk), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

The detailed description of the present embodiment may refer to the corresponding description in the foregoing embodiments, and will not be repeated herein.

The basic principles of the present disclosure have been described above in connection with specific embodiments, but it should be noted that the advantages, benefits, effects, etc. mentioned in the present disclosure are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present disclosure. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, since the disclosure is not necessarily limited to practice with the specific details described.

In this disclosure, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions, and the block diagrams of devices, apparatuses, devices, systems involved in this disclosure are merely illustrative examples and are not intended to require or implicate that connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.

In addition, as used herein, the use of "or" in the recitation of items beginning with "at least one" indicates a separate recitation, such that recitation of "at least one of A, B or C" means a or B or C, or AB or AC or BC, or ABC (i.e., a and B and C), for example. Furthermore, the term "exemplary" does not mean that the described example is preferred or better than other examples.

It is also noted that in the systems and methods of the present disclosure, components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered equivalent to the present disclosure.

Various changes, substitutions, and alterations are possible to the techniques described herein without departing from the teachings of the techniques defined by the appended claims. Furthermore, the scope of the claims of the present disclosure is not limited to the particular aspects of the process, machine, manufacture, composition of matter, means, methods and acts described above. The processes, machines, manufacture, compositions of matter, means, methods, or acts, presently existing or later to be developed that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein may be utilized. Accordingly, the appended claims are intended to include within their scope such processes, machines, manufacture, compositions of matter, means, methods, or acts.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the disclosure. Thus, the present disclosure is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the disclosure to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.

Claims (10)

1. The risk monitoring and early warning method for the lithium hydroxide production process is characterized by comprising the following steps of:

Preprocessing the acquired first data set to obtain a second data set;

The first data set includes real-time data during lithium hydroxide production;

training the constructed deep learning model based on the historical data set to obtain a first model;

obtaining risk information based on the second dataset and the first model;

determining an uncertainty factor to be quantified in the lithium hydroxide production process; the uncertainty factors comprise raw material quality fluctuation and production environment change;

Quantifying the uncertainty factors based on a preset uncertainty quantification strategy to obtain evaluation information;

and acquiring an early warning strategy based on the evaluation information and the risk information.

2. The method of claim 1, wherein the first data set includes temperature, pressure, and chemical concentration.

3. The method of claim 1, wherein the preprocessing includes one or more of data cleaning, missing value processing, time window partitioning, normalization, or normalization.

4. The method for risk monitoring and early warning in a lithium hydroxide production process according to claim 1, wherein training the constructed deep learning model based on the historical dataset to obtain a first model comprises:

Dividing the historical data set into a training set and a verification set;

Constructing the deep learning model;

determining a loss function and an optimization algorithm matched with monitoring of the lithium hydroxide production process;

Training the deep learning model based on the training set until the preset iteration times are reached or the deep learning model converges;

and verifying the trained deep learning model by adopting the verification set, wherein the deep learning model meeting the verification condition is the first model.

5. The method for risk monitoring and early warning of a lithium hydroxide production process according to claim 4, wherein the loss function comprises a cross entropy loss function or a mean square error function;

The optimization algorithm comprises a gradient descent algorithm or an Adam algorithm.

6. The method for risk monitoring and early warning in a lithium hydroxide production process according to claim 4, wherein the quantifying the uncertainty factor based on a preset uncertainty quantification policy to obtain evaluation information comprises:

Constructing a neural network model;

Training the neural network model by using a preset strategy, and obtaining the distribution information of key parameters based on the trained neural network model;

And quantifying uncertainty factors in the production process based on the trained neural network model and the distribution information to obtain evaluation information.

7. The method of claim 6, wherein the neural network model comprises any one of a markov chain model or a bayesian neural network model.

8. The method of claim 7, wherein when the neural network model is a bayesian neural network model, a variance lower bound is used as a loss function in training the neural network model.

9. The method for risk monitoring and early warning of lithium hydroxide production process according to claim 7, wherein when the neural network model is a markov chain model, a plurality of random samplings are performed using a monte carlo method, and the production process is simulated by sampling the obtained samples.

10. The utility model provides a lithium hydroxide production process risk monitoring early warning system which characterized in that includes:

The processing module is configured to preprocess the collected first data set to obtain a second data set;

The first data set includes real-time data during lithium hydroxide production;

the training module is configured to train the constructed deep learning model based on the historical data set to obtain a first model;

an analysis module configured to obtain risk information based on the second dataset and the first model;

The determining module is configured to determine uncertainty factors to be quantified in the lithium hydroxide production process; the uncertainty factors comprise raw material quality fluctuation and production environment change;

The quantization module is configured to quantize the uncertainty factors based on a preset uncertainty quantization strategy to obtain evaluation information;

And the acquisition module is configured to acquire an early warning strategy based on the evaluation information and the risk information.