Fire extinguishing agent preparation method and system based on industrial intelligence

Abstract

The application discloses a fire extinguishing agent preparation method and system based on industrial intelligence, and relates to the technical field of artificial intelligence. The method comprises the following steps: in the fire extinguishing agent preparation process, real-time effective component concentration data of the fire extinguishing agent is acquired through an infrared spectrum analyzer; target feature variables corresponding to the real-time effective component concentration data are input into a fire extinguishing agent preparation planning model; feature variables with current target feature variables existing abnormities are screened out through a perception block; feature variables with abnormal risks existing in the future are calculated through a time block; the quality problem type existing in the current fire extinguishing agent is determined based on the feature variables through a decision block; the optimal fire extinguishing agent preparation planning strategy is determined based on the quality problem type through an action value function, and the optimal fire extinguishing agent preparation planning strategy is executed. The method is helpful to solve the problem that the prior art cannot quickly determine and effectively execute the optimal fire extinguishing agent preparation planning strategy when the fire extinguishing agent has quality problems.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method and system for preparing fire extinguishing agents based on industrial intelligence. Background Technology

[0002] In modern industrial production and fire safety, the quality of fire extinguishing agents is of paramount importance. With continuous industrial development, higher demands are placed on the performance and stability of fire extinguishing agents. Traditional methods of preparing fire extinguishing agents often rely on experience and simple quality control measures. During the preparation process, it is difficult to accurately control the concentration of the effective components of the fire extinguishing agent, resulting in inconsistent quality. If a poor-quality fire extinguishing agent is used during a fire, it may fail to effectively extinguish the fire, or even delay the firefighting effort, causing serious loss of life and property.

[0003] Although some companies have tried to introduce advanced testing equipment to monitor the concentration of fire extinguishing agent components, these devices can usually only provide data and cannot conduct in-depth analysis and processing of the data to guide production. At the same time, due to the lack of effective use of historical data, it is impossible to learn from past quality problems and it is difficult to optimize the preparation process.

[0004] In recent years, the development of industrial intelligence technology has provided new ideas and methods for solving the above problems. Through data collection, analysis, and modeling, precise control of the fire extinguishing agent preparation process can be achieved. However, research on applying industrial intelligence technology to the field of fire extinguishing agent preparation is still relatively limited, and a mature, complete, and efficient industrial intelligence-based fire extinguishing agent preparation system has not yet been formed.

[0005] Therefore, there is an urgent need for a method that can effectively improve the quality of fire extinguishing agent preparation by utilizing industrial intelligent technology. Summary of the Invention

[0006] In view of this, the present invention proposes a method and system for preparing fire extinguishing agents based on industrial intelligence, which can realize rapid and accurate monitoring of the quality of fire extinguishing agent preparation, timely resolution of quality problems, improvement of production safety and efficiency, and reduction of errors and lags in manual monitoring.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for preparing a fire extinguishing agent based on industrial intelligence, comprising:

[0009] Collect historical effective component concentration data of fire extinguishing agents with quality problems, and perform pre-screening processing on the historical effective component concentration data of fire extinguishing agents to obtain characteristic variables of quality problems;

[0010] A fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block is constructed. The quality problem characteristic variables are used as inputs to the fire extinguishing agent preparation planning model, and the fire extinguishing agent preparation planning model is trained to obtain the optimal fire extinguishing agent preparation planning model.

[0011] During the preparation of the extinguishing agent, real-time effective component concentration data of the extinguishing agent is acquired by an infrared spectroscopy analyzer. The target feature variables corresponding to the real-time effective component concentration data are input into the extinguishing agent preparation planning model. The perception block filters out the feature variables that are abnormal in the current target feature variables. The time block calculates the feature variables that will have abnormal risks in the future. The decision block determines the type of quality problem that exists in the current extinguishing agent based on the feature variables. The action value function determines the optimal extinguishing agent preparation planning strategy based on the type of quality problem and executes the optimal extinguishing agent preparation planning strategy.

[0012] Based on the above technical solution, the present invention can be further improved as follows:

[0013] Optionally, the pre-screening of the historical effective component concentration data of the fire extinguishing agent to obtain quality problem characteristic variables includes:

[0014] Multiple original feature variables are extracted from the historical effective component concentration data of the fire extinguishing agent. The correlation coefficient between the original feature variables and the quality problem of the fire extinguishing agent is calculated. The original feature variables with a correlation coefficient greater than a preset value are extracted as the quality problem feature variables.

[0015] Optionally, when the extinguishing agent is a water-based extinguishing agent, the quality problem characteristic variables include pH value, surfactant content, and additive concentration;

[0016] When the extinguishing agent is a foam extinguishing agent, the quality problem characteristic variables include expansion ratio, foam stability, and surface tension;

[0017] When the extinguishing agent is a dry powder extinguishing agent, the quality problem characteristic variables include the content of active ingredient, particle size, and moisture content;

[0018] When the extinguishing agent is carbon dioxide, the quality problem characteristic variables include purity, water content, and pressure.

[0019] Optionally, the step of filtering out anomalous feature variables of the current target feature variable through the perceptual block includes:

[0020] Calculate each feature variable x using formula (1). i Calculate an anomaly score S i ;

[0021]

[0022] In the formula, S i For outlier scores, w1 is the weight of standardized deviation, w2 is the weight of business-standardized deviation, w3 is the weight of trend score, and x... i Let μ be the characteristic variable, σ be the mean of the characteristic variable in the historical data, and σ be the standard deviation of the characteristic variable in the historical data. Let L be the standardized value of the i-th feature variable. i U is the lower limit of the business standard for the i-th feature variable. i Let trend be the upper limit of the business standard for the i-th feature variable. i The score is given for the trend of change of the i-th feature variable.

[0023] Optionally, the characteristic variables for calculating future anomaly risks through time blocks include:

[0024] The abnormal risk score of the feature variables at the h-th time step in the future is calculated using formula (2);

[0025]

[0026] In the formula, R t+h Let K1, K2, and K3 be the abnormal risk scores of the feature variables at the h-th time step in the future, where K1, K2, and K3 are the weights of different calculation parts, respectively. This is a value predicted by trend extrapolation. For reference only. μ is the predicted value using exponential smoothing. t σ is the historical mean. t C represents the historical standard deviation. t+h The confidence level of the prediction made by the machine learning model.

[0027] Optionally, the step of determining the type of quality problem existing in the current fire extinguishing agent based on the feature variables through the decision block includes:

[0028] The score for the j-th quality problem type is calculated using formula (3);

[0029]

[0030] In the formula, S j Let W be the score for the j-th type of quality problem, n be the abnormal characteristic variable, and W be the score for the j-th type of quality problem. ij T represents the weight of the influence of the i-th feature variable on the j-th type of quality problem. i Let D be the standard range of the i-th feature variable. i Let be the deviation of the i-th feature variable.

[0031] Optionally, determining the optimal fire extinguishing agent preparation planning strategy based on the quality problem type using the action value function includes:

[0032] The i-th strategy a is calculated using formula (4). i Action value function for quality problem type q;

[0033]

[0034] In the formula, V iq For strategy i, a i For the action value function of quality problem type q, λ is the cost-effectiveness balance coefficient, and E iq For strategy i, a i The evaluation value of the resolution effect of quality problem type q, C i To execute the i-th strategy a i The required costs.

[0035] An industrial intelligence-based fire extinguishing agent preparation system, comprising:

[0036] The module for determining the characteristic variables of quality problems is used to collect historical effective component concentration data of fire extinguishing agents with quality problems, and to pre-screen the historical effective component concentration data of fire extinguishing agents to obtain characteristic variables of quality problems.

[0037] The optimal fire extinguishing agent preparation planning model acquisition module is used to construct a fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block. The quality problem feature variables are used as inputs to the fire extinguishing agent preparation planning model, and the fire extinguishing agent preparation planning model is trained to obtain the optimal fire extinguishing agent preparation planning model.

[0038] The optimal fire extinguishing agent preparation planning strategy execution module is used to acquire real-time effective component concentration data of the fire extinguishing agent through an infrared spectroscopy analyzer during the fire extinguishing agent preparation process, input the target feature variables corresponding to the real-time effective component concentration data into the fire extinguishing agent preparation planning model, filter out feature variables with abnormal current target feature variables through a perception block, calculate feature variables with future abnormal risks through a time block, determine the type of quality problem of the current fire extinguishing agent based on the feature variables through a decision block, determine the optimal fire extinguishing agent preparation planning strategy based on the quality problem type through an action value function, and execute the optimal fire extinguishing agent preparation planning strategy.

[0039] An electronic device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method described herein.

[0040] A non-transitory computer-readable storage medium having a computer program stored thereon, the computer program implementing the steps of the method when executed by a processor.

[0041] The present invention has the following advantages:

[0042] The fire extinguishing agent preparation method based on industrial intelligence in this invention collects historical effective component concentration data of fire extinguishing agents with quality problems and performs pre-screening to obtain characteristic variables of quality problems. This allows for the accurate extraction of key information from massive historical data. In the subsequent preparation process, when real-time effective component concentration data is input into the fire extinguishing agent preparation planning model, the perception block, time block, and decision block work together to accurately screen out characteristic variables that are currently and in the future that are abnormal, and determine the type of quality problem. This enables quality problems to be detected in a timely and accurate manner during the preparation process, avoiding the production of substandard fire extinguishing agents and greatly improving the stability of product quality.

[0043] This invention presents an industrial intelligence-based fire extinguishing agent preparation method. It constructs and trains a fire extinguishing agent preparation planning model that includes a perception block, a time block, and a decision block. The optimal model trained using historical quality problem data provides scientific guidance for fire extinguishing agent preparation. During the production process, based on the optimal fire extinguishing agent preparation planning strategy output by the model, potential quality deviations are effectively corrected, ensuring that each batch of fire extinguishing agent meets high-quality standards and reducing problems such as product recalls or poor fire extinguishing effects caused by quality instability. Attached Figure Description

[0044] For illustrative and not limiting purposes, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein:

[0045] Figure 1 This is a schematic flowchart of the fire extinguishing agent preparation method based on industrial intelligence according to an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of the main components of the industrial intelligence-based fire extinguishing agent preparation system in an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of the physical structure of the electronic device provided by the present invention. Detailed Implementation

[0048] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0049] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0050] It should be noted that, where there is no conflict, the embodiments and features of the present invention can be combined with each other. The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0051] Figure 1 This is a schematic flowchart of the industrial intelligence-based fire extinguishing agent preparation method according to an embodiment of the present invention, as shown below. Figure 1 As shown, the industrial intelligence-based fire extinguishing agent preparation method provided in this embodiment of the invention includes the following steps S101 to S103.

[0052] S101. Collect historical effective component concentration data of fire extinguishing agents with quality problems, and perform pre-screening processing on the historical effective component concentration data of fire extinguishing agents to obtain characteristic variables of quality problems.

[0053] Multiple original feature variables were extracted from the historical effective component concentration data of fire extinguishing agents. The correlation coefficient between the original feature variables and the quality problems of fire extinguishing agents was calculated. Original feature variables with correlation coefficients greater than preset values ​​were extracted as feature variables of quality problems.

[0054] The original characteristic variables include, but are not limited to, the specific content of the active ingredient, the ratio between different components, the temperature and humidity of the production environment, and the operating parameters of the production equipment. Then, statistical methods are used to calculate the correlation coefficient between each original characteristic variable and the quality of the fire extinguishing agent. For example, the Pearson correlation coefficient can be used to measure linear correlation.

[0055] When the extinguishing agent is a water-based extinguishing agent, the characteristic variables of quality problems include pH value, surfactant content and additive concentration;

[0056] When the extinguishing agent is a foam extinguishing agent, the characteristic variables of quality problems include expansion ratio, foam stability and surface tension;

[0057] When the extinguishing agent is a dry powder extinguishing agent, the characteristic variables of quality problems include the content of active ingredient, particle size and moisture content;

[0058] When the extinguishing agent is carbon dioxide, the characteristic variables of quality issues include purity, water content, and pressure.

[0059] A preset value (e.g., 0.6) is set, and original characteristic variables with correlation coefficients greater than this preset value are selected as characteristic variables of quality problems. When the extinguishing agent is water-based, analysis shows that pH value, surfactant content, and additive concentration have a high correlation with quality problems, and these are identified as characteristic variables of quality problems. For foam extinguishing agents, expansion ratio, foam stability, and surface tension are important characteristic variables of quality problems. For dry powder extinguishing agents, the content of active ingredients, particle size, and moisture content have a significant impact on quality and are selected as characteristic variables. For carbon dioxide extinguishing agents, purity, water content, and pressure are key characteristic variables of quality problems.

[0060] S102, construct a fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block. Use the characteristic variables of the quality problem as inputs to the fire extinguishing agent preparation planning model, and train the fire extinguishing agent preparation planning model to obtain the optimal fire extinguishing agent preparation planning model.

[0061] After constructing the model, the previously obtained quality problem feature variables were used as input, and the fire extinguishing agent preparation planning model was trained using a large amount of historical data. During training, cross-validation was employed, dividing the dataset into training, validation, and test sets. The model was trained on the training set, and its parameters were continuously adjusted, such as the kernel function parameters of the SVM in the perception block, the number of hidden layer neurons in the LSTM in the time block, and the number of decision trees in the random forest in the decision block. The model's performance was evaluated on the validation set, and the model was optimized based on the evaluation results. Finally, the optimized model was tested on the test set to ensure that it had good generalization ability and accuracy. When the model reached the preset performance index on the test set, the optimal fire extinguishing agent preparation planning model was obtained.

[0062] S103, during the fire extinguishing agent preparation process, real-time effective component concentration data of the fire extinguishing agent is obtained through an infrared spectral analyzer. The target characteristic variables corresponding to the real-time effective component concentration data are input into the fire extinguishing agent preparation planning model. The perception block filters out the characteristic variables that are abnormal in the current target characteristic variables. The time block calculates the characteristic variables that will have abnormal risks in the future. The decision block determines the type of quality problem that exists in the current fire extinguishing agent based on the characteristic variables. The action value function determines the optimal fire extinguishing agent preparation planning strategy based on the type of quality problem and executes the optimal fire extinguishing agent preparation planning strategy.

[0063] The perceptual block filters out feature variables that exhibit anomalies in the current target feature variable, including:

[0064] Calculate each feature variable x using formula (1). iCalculate an anomaly score S i ;

[0065]

[0066] In the formula, S i For outlier scores, w1 is the weight of standardized deviation, w2 is the weight of business-standardized deviation, w3 is the weight of trend score, and x... i Let μ be the characteristic variable, σ be the mean of the characteristic variable in the historical data, and σ be the standard deviation of the characteristic variable in the historical data. Let L be the standardized value of the i-th feature variable. i U is the lower limit of the business standard for the i-th feature variable. i Let trend be the upper limit of the business standard for the i-th feature variable. i The score is given for the trend of change of the i-th feature variable.

[0067] Characteristic variables for calculating future anomaly risks using time blocks include:

[0068] The abnormal risk score of the feature variables at the h-th time step in the future is calculated using formula (2);

[0069]

[0070] In the formula, R t+h Let K1, K2, and K3 be the abnormal risk scores of the feature variables at the h-th time step in the future, where K1, K2, and K3 are the weights of different calculation parts, respectively. This is a value predicted by trend extrapolation. For reference only. μ is the predicted value using exponential smoothing. t σ is the historical mean. t C represents the historical standard deviation. t+h The confidence level of the prediction made by the machine learning model.

[0071] The decision block determines the type of quality problem present in the current fire extinguishing agent based on the aforementioned feature variables, including:

[0072] The score for the j-th quality problem type is calculated using formula (3);

[0073]

[0074] In the formula, S j Let W be the score for the j-th type of quality problem, n be the abnormal characteristic variable, and W be the score for the j-th type of quality problem. ij T represents the weight of the influence of the i-th feature variable on the j-th type of quality problem. i Let D be the standard range of the i-th feature variable. i Let be the deviation of the i-th feature variable.

[0075] The optimal fire extinguishing agent preparation planning strategy is determined based on the quality problem type using the action value function, including:

[0076] The i-th strategy a is calculated using formula (4). i Action value function for quality problem type q;

[0077]

[0078] In the formula, V iq For strategy i, a i For the action value function of quality problem type q, λ is the cost-effectiveness balance coefficient, and E iq For strategy i, a i The evaluation value of the resolution effect of quality problem type q, C i To execute the i-th strategy a i The required costs.

[0079] Taking the preparation of foam fire extinguishing agents as an example:

[0080] On the foam extinguishing agent production line, an infrared spectrometer collects real-time concentration data of the effective components of the extinguishing agent every 10 minutes, such as surfactant concentration and foam stabilizer concentration; in one instance, the surfactant concentration was 8% and the foam stabilizer concentration was 3%.

[0081] Historical data analysis shows that the historical average concentration of surfactants is 10% with a standard deviation of 1%; the historical average concentration of foam stabilizers is 5% with a standard deviation of 0.5%. Business standards specify a surfactant concentration range of 9%-11% and a foam stabilizer concentration range of 4%-6%; weights w1 = 0.3, w2 = 0.5, w3 = 0.2.

[0082] Standardized value Standardized deviation Deviation from business standards =2, assuming the trend score is 0.1; substituting into formula (1) yields the anomaly score. .

[0083] Standardized value (3% - 5%) / 0.5% = -4, Standardized deviation |-4| = 4, Business standard deviation Assuming the trend score is 0.2, substituting it into formula (1) yields the abnormal score S2 = 0.3 × 4 + 0.5 × 2 + 0.2 × 0.2 = 2.24.

[0084] Therefore, the concentration of foam stabilizer is an abnormal characteristic variable.

[0085] For the concentration of the foam stabilizer, the trend extrapolation method predicts a value of 2.5% for the second time step (20 minutes later), the exponential smoothing method predicts a value of 2.8%, the historical mean is 5%, the standard deviation is 0.5%, the reference value is the historical mean, and the machine learning model predicts a confidence level of 0.7. The weights are K1 = 0.4, K2 = 0.4, and K3 = 0.2.

[0086] Substituting into formula (2), R t+2 =1.98.

[0087] This indicates a high risk of abnormal concentrations of foam stabilizer in the future.

[0088] The weight W of the foam stabilizer concentration on the foam persistence problem is known. 11 =0.7, the weight W for the impact on foam fire extinguishing efficiency. 12 =0.3, foam stabilizer concentration deviation 2.

[0089] Substituting into formula (3), the score for foam persistence is S1 = 0.7 × 2 = 1.4; the score for foam extinguishing efficiency is S2 = 0.3 × 2 = 0.6. Therefore, the primary quality issue is determined to be foam persistence.

[0090] The strategy library contains two strategies for addressing foam persistence issues. Strategy 1: Increase the amount of foam stabilizer added, with an effectiveness evaluation value of 0.8 and a cost of 600 yuan; Strategy 2: Adjust the mixing time, with an effectiveness evaluation value of 0.6 and a cost of 300 yuan. The cost-effectiveness balance coefficient k = 0.5.

[0091] Substituting into formula (4), the action value function of strategy 1 Action value function of strategy 2

[0092] Choose strategy 2, which involves adjusting the mixing time to solve the problem.

[0093] Figure 2 This is a schematic diagram of the main components of the industrial intelligence-based fire extinguishing agent preparation system according to an embodiment of the present invention. Figure 2 As shown, the industrial intelligence-based fire extinguishing agent preparation system 1 provided in this embodiment of the invention includes a quality problem characteristic variable determination module 10, an optimal fire extinguishing agent preparation planning model acquisition module 20, and an optimal fire extinguishing agent preparation planning strategy execution module 30.

[0094] The module 10 for determining the characteristic variables of quality problems is used to collect historical effective component concentration data of fire extinguishing agents with quality problems, and to pre-screen the historical effective component concentration data of fire extinguishing agents to obtain the characteristic variables of quality problems.

[0095] The optimal fire extinguishing agent preparation planning model acquisition module 20 is used to construct a fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block. The quality problem characteristic variables are used as inputs to the fire extinguishing agent preparation planning model to train the fire extinguishing agent preparation planning model and obtain the optimal fire extinguishing agent preparation planning model.

[0096] The optimal fire extinguishing agent preparation planning strategy execution module 30 is used to acquire real-time effective component concentration data of the fire extinguishing agent through an infrared spectrometer during the fire extinguishing agent preparation process, input the target feature variables corresponding to the real-time effective component concentration data into the fire extinguishing agent preparation planning model, filter out the feature variables that are abnormal in the current target feature variables through the perception block, calculate the feature variables of future abnormal risks through the time block, determine the type of quality problem of the current fire extinguishing agent based on the feature variables through the decision block, determine the optimal fire extinguishing agent preparation planning strategy based on the quality problem type through the action value function, and execute the optimal fire extinguishing agent preparation planning strategy.

[0097] Figure 3 This is a schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the electronic device 40 includes: a processor 401, a memory 402, and a bus 403;

[0098] The processor 401 and the memory 402 communicate with each other via the bus 403.

[0099] The processor 401 is used to call program instructions in the memory 402 to execute the methods provided in the above-described method embodiments, and to execute the methods provided in the embodiments of the present invention.

[0100] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions, which cause a computer to execute the method provided in this embodiment of the invention.

[0101] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0102] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for preparing a fire extinguishing agent based on industrial intelligence, characterized in that, include: Collect historical effective component concentration data of fire extinguishing agents with quality problems, and perform pre-screening processing on the historical effective component concentration data of fire extinguishing agents to obtain characteristic variables of quality problems; A fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block is constructed. The quality problem characteristic variables are used as inputs to the fire extinguishing agent preparation planning model, and the fire extinguishing agent preparation planning model is trained to obtain the optimal fire extinguishing agent preparation planning model. During the preparation of the extinguishing agent, real-time effective component concentration data of the extinguishing agent is obtained by an infrared spectroscopy analyzer. The target feature variables corresponding to the real-time effective component concentration data are input into the extinguishing agent preparation planning model. The perception block filters out the feature variables that are abnormal in the current target feature variables. The time block calculates the feature variables that will have abnormal risks in the future. The decision block determines the type of quality problem that the current extinguishing agent has based on the feature variables. The action value function determines the optimal extinguishing agent preparation planning strategy based on the type of quality problem and executes the optimal extinguishing agent preparation planning strategy. The step of filtering out anomalous feature variables of the current target feature variable through perceptual blocks includes: Calculate each feature variable using formula (1) Calculate an anomaly score ; Formula (1); In the formula, For abnormal scores, The weights for the standardized deviations, The weight of the deviation from business standards. The weights for the score of the changing trend, As characteristic variables, This is the average value of this feature variable in historical data. The standard deviation of this feature variable in historical data. Let i be the lower limit of the business standard for the i-th feature variable. Let i be the upper limit value of the business standard for the i-th feature variable. Score the trend of change of the i-th feature variable; The characteristic variables for calculating future anomaly risks through time blocks include: The abnormal risk score of the feature variables at the h-th time step in the future is calculated using formula (2); Official (2); In the formula, The abnormal risk score for the feature variables at the h-th time step in the future. , and These are the weights of different calculation parts. This is a value predicted by trend extrapolation. For reference only. These are values ​​predicted using exponential smoothing. This is the historical average. For historical standard deviation, The confidence level of the prediction made by the machine learning model; The process of determining the type of quality problem existing in the current fire extinguishing agent based on the feature variables through the decision block includes: The score for the j-th quality problem type is calculated using formula (3); Official (3); In the formula, Let be the score for the j-th type of quality problem, and n be the abnormal characteristic variable. Let be the weight of the influence of the i-th feature variable on the j-th quality problem type. Let be the deviation of the i-th feature variable.

2. The method for preparing fire extinguishing agent based on industrial intelligence according to claim 1, characterized in that, The pre-screening of the historical effective component concentration data of the fire extinguishing agent yields characteristic variables of quality problems, including: Multiple original feature variables are extracted from the historical effective component concentration data of the fire extinguishing agent. The correlation coefficient between the original feature variables and the quality problem of the fire extinguishing agent is calculated. The original feature variables with a correlation coefficient greater than a preset value are extracted as the quality problem feature variables.

3. The method for preparing fire extinguishing agent based on industrial intelligence according to claim 2, characterized in that, When the extinguishing agent is a water-based extinguishing agent, the characteristic variables of quality problems include pH value, surfactant content and additive concentration; When the extinguishing agent is a foam extinguishing agent, the quality problem characteristic variables include expansion ratio, foam stability, and surface tension; When the extinguishing agent is a dry powder extinguishing agent, the quality problem characteristic variables include the content of active ingredient, particle size, and moisture content; When the extinguishing agent is carbon dioxide, the quality problem characteristic variables include purity, water content, and pressure.

4. The method for preparing fire extinguishing agent based on industrial intelligence according to claim 1, characterized in that, The process of determining the optimal fire extinguishing agent preparation planning strategy based on the quality problem type using the action value function includes: The i-th strategy is calculated using formula (4). The action value function for quality problem type q; Official (4); In the formula, For the i-th strategy For the action value function of quality problem type q, Cost-effectiveness balance coefficient For the i-th strategy Evaluation value of the effectiveness of resolving quality problem type q. To execute the i-th strategy The required costs.

5. A system for preparing fire extinguishing agents based on industrial intelligence, characterized in that, include: The module for determining the characteristic variables of quality problems is used to collect historical effective component concentration data of fire extinguishing agents with quality problems, and to pre-screen the historical effective component concentration data of fire extinguishing agents to obtain characteristic variables of quality problems. The optimal fire extinguishing agent preparation planning model acquisition module is used to construct a fire extinguishing agent preparation planning model including a perception block, a time block, and a decision block. The quality problem feature variables are used as inputs to the fire extinguishing agent preparation planning model, and the fire extinguishing agent preparation planning model is trained to obtain the optimal fire extinguishing agent preparation planning model. The optimal fire extinguishing agent preparation planning strategy execution module is used to acquire real-time effective component concentration data of the fire extinguishing agent through an infrared spectroscopy analyzer during the fire extinguishing agent preparation process, input the target feature variables corresponding to the real-time effective component concentration data into the fire extinguishing agent preparation planning model, filter out feature variables with abnormal current target feature variables through a perception block, calculate feature variables with future abnormal risks through a time block, determine the type of quality problem of the current fire extinguishing agent based on the feature variables through a decision block, determine the optimal fire extinguishing agent preparation planning strategy based on the quality problem type through an action value function, and execute the optimal fire extinguishing agent preparation planning strategy. The optimal fire extinguishing agent preparation planning strategy execution module is also used for: Calculate each feature variable using formula (1) Calculate an anomaly score ; Formula (1); In the formula, For abnormal scores, The weights for the standardized deviations, The weight of the deviation from business standards. The weights for the score of the changing trend, As characteristic variables, This is the average value of this feature variable in historical data. The standard deviation of this feature variable in historical data. Let i be the lower limit of the business standard for the i-th feature variable. Let i be the upper limit value of the business standard for the i-th feature variable. Score the trend of change of the i-th feature variable; The abnormal risk score of the feature variables at the h-th time step in the future is calculated using formula (2); Official (2); In the formula, The abnormal risk score for the feature variables at the h-th time step in the future. , and These are the weights of different calculation parts. This is a value predicted by trend extrapolation. For reference only. These are values ​​predicted using exponential smoothing. This is the historical average. For historical standard deviation, The confidence level of the prediction made by the machine learning model; The score for the j-th quality problem type is calculated using formula (3); Official (3); In the formula, Let be the score for the j-th type of quality problem, and n be the abnormal characteristic variable. Let be the weight of the influence of the i-th feature variable on the j-th quality problem type. Let be the deviation of the i-th feature variable.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 4.

7. A non-transitory computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 4.

Images