Intelligent optimization system for firefighters' firefighting protective clothing performance based on artificial intelligence

Abstract

The present invention belongs to the technical field of performance optimization of protective clothing, and discloses an intelligent optimization system for the performance of fire-fighting protective clothing for firefighters based on artificial intelligence; the system comprises: a data acquisition module, which is used to acquire m groups of fire-fighting characteristic data; a fire simulation model is constructed according to the m groups of fire-fighting characteristic data; an environment classification module, which is used to classify the m groups of fire-fighting characteristic data into environment categories; a layout optimization module, which optimizes the reflective layout parameters of each environment category based on the fire simulation model; a design optimization module, which optimizes the reflective design parameters of each environment category based on the fire simulation model and the reflective layout parameters; the present invention can significantly improve the reaction speed and visibility of firefighters in complex environments by effectively improving the reflective effect and comfort of protective clothing, thereby effectively protecting the lives of firefighters.

Description

Technical Field

[0001] The present invention relates to the technical field of protective clothing performance optimization, and more specifically, to an artificial intelligence-based intelligent optimization system for the performance of firefighters' firefighting protective clothing. Background Art

[0002] Firefighters often face various complex lighting conditions during firefighting and rescue operations, including strong sunlight, smoke, flames, and low-light environments at night; these factors have a significant impact on firefighters' visibility, reducing their safety and reaction speed when performing their tasks; therefore, optimizing the layout and design of reflective materials on firefighters' firefighting protective clothing has become an important task to improve their visibility under various lighting conditions; traditional firefighters' protective clothing often only follows fixed design standards in the use of reflective materials, and fails to fully consider the actual needs in different environments; with the development of intelligent technology, artificial intelligence is used to analyze visibility under different lighting conditions, thereby optimizing the layout and design of reflective materials to ensure that firefighters' recognition can be effectively improved under various lighting conditions.

[0003] The patent with announcement number CN110147566A discloses a research method for high-temperature protective clothing based on genetic algorithms and nonlinear programming; including: according to given experimental data, using MATLAB to draw images and perform polynomial fitting on the data, and after repeated experiments, it is found that the cubic fitting result is most in line with the actual situation, and further solving the problem based on this result; predicting the optimal thickness of layer II of the protective clothing; predicting the optimal thickness of layer II and layer IV of the protective clothing; this invention adopts the method of establishing a mathematical model to refine each layer, so that the temperature change situation is more specific in space; using traditional nonlinear programming and genetic algorithms for iterative solution to optimize the model, so that the thickness estimation value is more accurate.

[0004] The genetic algorithm and nonlinear programming used in the above technology can also be used to optimize the layout and design of reflective materials, but the genetic algorithm and nonlinear programming cannot fully capture the nonlinear relationship between complex fire environments and reflective material performance, and the model expression ability is weak; and the optimization results obtained by genetic algorithms and nonlinear programming are often limited to a single performance indicator, unable to take into account multiple key factors, and difficult to achieve multi-objective optimization, which affects the overall optimization effect and applicability of protective clothing.

[0005] In view of this, the present invention proposes an artificial intelligence-based intelligent optimization system for the performance of firefighters' firefighting protective clothing to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: an intelligent optimization system for the performance of firefighting protective clothing for firefighters based on artificial intelligence, comprising:

[0007] A data collection module, used for collecting m groups of fire characteristic data;

[0008] A model building module builds a fire simulation model based on m groups of fire characteristic data;

[0009] An environmental classification module is used to classify m groups of fire feature data into environmental categories;

[0010] Layout optimization module, based on the fire simulation model, optimizes the reflective layout parameters for each environment category;

[0011] Design optimization module, based on fire simulation model and reflective layout parameters, optimizes reflective design parameters for each environment category.

[0012] Further, the fire protection characteristic data includes light environment data, reflective characteristic data, reflective property data and flexibility;

[0013] The light environment data includes illumination, light source angle and smoke density; the illumination is the light flux received per unit area in the fire area; the light source angle is the incident angle of the light source relative to the ground in the fire area; the smoke density is the concentration of smoke particles per unit volume in the fire area;

[0014] The reflective characteristic data includes reflective material position, reflective material area and reflective material type; the reflective material position is the position of each reflective material on the protective clothing; the reflective material area is the surface area of ​​each reflective material; the reflective material type is the type of each reflective material;

[0015] The reflective characteristic data includes a reflectivity mean value and a luminous flux density mean value;

[0016] The reflectivity is the ratio of the luminous flux reflected by the incident light on the surface of the reflective material to the incident luminous flux; the method for obtaining the mean reflectivity is: obtaining the reflected luminous flux and the incident luminous flux corresponding to each reflective material, dividing the reflected luminous flux of the reflective material corresponding to each position by the corresponding incident luminous flux, and obtaining the reflectivity corresponding to each reflective material; counting the number of reflective materials and marking them as the number of materials; adding the reflectivity of each reflective material in turn, and then dividing by the number of materials to obtain the mean reflectivity;

[0017] The luminous flux density is the luminous flux received per unit area in the reflective material; the method for obtaining the average luminous flux density is: dividing the incident luminous flux of each reflective material by the corresponding reflective material area to obtain the luminous flux density corresponding to each material; adding each luminous flux density in turn and dividing it by the number of materials to obtain the average luminous flux density;

[0018] The flexibility refers to the ability of firefighters to stretch their joints while wearing protective clothing during firefighting.

[0019] Furthermore, the method for constructing a fire simulation model includes:

[0020] Select a deep learning framework and build a fire simulation model architecture, which includes an input layer, a hidden layer, and an output layer;

[0021] Different digital labels are set for different reflective material positions in m groups of fire feature data, and marked as position labels; different digital labels are set for different reflective material types in m groups of fire feature data, and marked as type labels; the reflective material positions in each group of reflective feature data are replaced with corresponding position labels, and the reflective material types are replaced with corresponding type labels, and the replaced reflective feature data are marked as reflective replacement data; the light environment data and reflective replacement data in each group of fire feature data are used as training data, and the reflective characteristic data and flexibility in each group of fire feature data are used as evaluation data; the training data and evaluation data in each group of fire feature data are converted into a corresponding set of feature vectors;

[0022] Each group of feature vectors is used as the input of the fire simulation model. The fire simulation model takes a group of predicted evaluation data corresponding to each group of training data as output, and takes the actual evaluation data corresponding to each group of training data as the prediction target. The actual evaluation data is the pre-set evaluation data corresponding to the training data. Minimizing the sum of prediction errors of all training data is used as the training goal. The fire simulation model is trained until the sum of prediction errors converges, and the training is stopped to obtain the fire simulation model, which is a deep neural network model.

[0023] Furthermore, the step of dividing the m groups of fire protection characteristic data into environmental categories includes:

[0024] Step A: Take m groups of light environment data as sample points, and the sample points correspond to the light environment data one by one; preset the number of categories , randomly selected sample points as the center point; The center points are marked in ascending order. , , …, , marking the center point as , ;

[0025] Step B: Mark the sample points that are not the center points as classification points, and mark the classification points in ascending order. , , …, , that is, marking the classification points as , ; Calculate the point distance from each classification point to each center point in turn;

[0026] Step C: According to The corresponding center point is established environmental categories;

[0027] Step D: Classify the points The distance to each center point is compared and the classified points are Assign to the environment category corresponding to the center point with the minimum distance to the point;

[0028] Step E: Let , jump back to step D;

[0029] Step F: Repeat steps D to E until When the loop ends, it goes to step G;

[0030] Step G: Recalculate the new center point corresponding to each environment category;

[0031] Step H: Repeat steps B to G until the new center point of each environment category recalculated in step G is consistent with the new center point of the corresponding environment category calculated in the previous cycle, the cycle ends, and the Environmental categories and corresponding classification points.

[0032] Furthermore, the expression of point distance is: ; In the formula, For classification points To center point The point distance, For classification points Middle The value of the dimension, For classification points Middle The value of the dimension, ; Different dimensions represent different types of data in the light environment data;

[0033] The calculation method for the new center point of each environmental category includes: ;

[0034] In the formula, For the The environment category corresponds to the new center point. For the In the environmental category classification points, , For the first In the environmental category The corresponding values ​​of classification points in different dimensions, For the The number of classification points in each environment category, .

[0035] Furthermore, the step of optimizing the reflective layout parameters for each environment category includes:

[0036] Step 1: construct a parameter set, set a different digital label for each set of reflective layout parameters in the parameter set, and mark them as parameter labels;

[0037] Step 2: Select environment categories and initialize the quantum group , quantum group There are n particles in the quantum group, and the position of each particle corresponds to the parameter label one by one. The number of iterations t is 0, ;

[0038] Step 3: Define the fitness function;

[0039] Step 4: Based on the phenomenon of quantum entanglement, Each particle in the process updates its position;

[0040] Step 5: Based on the quantum interference phenomenon, Each particle in the process updates its position;

[0041] Step 6: Based on the quantum tunneling phenomenon, Each particle in the process updates its position;

[0042] Step 7: Use quantum annealing mechanism to Each particle in the process updates its position;

[0043] Step 8: Compare the number of iterations t with the preset iteration threshold T; if , then go to step 9; if , then let , and return to step 4;

[0044] Step 9: Calculate the adaptability of each particle, obtain the parameter label corresponding to the particle with the largest adaptability, obtain the reflective layout parameters corresponding to the parameter label, and use it as the first The reflective layout parameters corresponding to the environment category; , and return to step 2;

[0045] Step 10: Repeat steps 2 to 9 until Each environment category obtains the corresponding reflective layout parameters, and the loop ends.

[0046] Furthermore, in step 1, the method for constructing the parameter set is: obtaining a parameter range, the parameter range including a reflective material position range and a reflective material area range; randomly selecting a value from each range within the parameter range to construct a set of reflective layout parameters, constructing a total of N sets of reflective layout parameters, and taking the N sets of reflective layout parameters as a parameter set, where N is an integer greater than 1;

[0047] In step 2, the quantum group The expression for the position of each particle in is: ; In the formula, is the position of the ith particle, is the random coefficient of the ith particle, , ;

[0048] In step 3, the expression of the adaptability function is: ; In the formula, For adaptability, is the total mean reflectivity, is the total mean value of luminous flux density, For total flexibility, , , , All are preset weight coefficients;

[0049] The method for obtaining the total mean value of reflectivity, total mean value of luminous flux density and total flexibility is as follows: The light environment data corresponding to the classification points in each environment category are obtained and marked as predicted environment data; the reflective layout parameters corresponding to the parameter labels corresponding to the particle positions are obtained, and the reflective layout parameters, a group of predicted environment data and the reflective material type corresponding to the predicted environment data are taken as a group of analysis data; each group of analysis data is input into the fire simulation model respectively to obtain the corresponding evaluation data and marked as the first data; the number of first data is counted and marked as the first number; the mean reflectivity values ​​in each group of first data are added in turn, and then divided by the first number to obtain the total mean reflectivity; the mean luminous flux density values ​​in each group of first data are added in turn, and then divided by the first number to obtain the total mean luminous flux density; the flexibility in each group of first data is added in turn, and then divided by the first number to obtain the total flexibility.

[0050] Furthermore, in step 4, the quantum group The methods for updating the position of each particle in include: ;

[0051] In the formula, is the position of the ith particle after updating based on the quantum entanglement phenomenon, To update the position of the previous i-th particle based on the quantum entanglement phenomenon, is the entanglement coefficient, , To update the position of the nearest particle based on the quantum entanglement phenomenon, is a random number in the standard normal distribution, and the nearest particle is the particle closest to the i-th particle;

[0052] In step 5, the quantum group The methods for updating the position of each particle in include: ;

[0053] In the formula, is the position of the ith particle after updating based on the quantum interference phenomenon, is the first fluctuation coefficient, , To update the position of the best particle before based on the quantum interference phenomenon, is the interference coefficient, and the best particle is the quantum group The particle with the greatest adaptability.

[0054] Furthermore, in step 6, the quantum group The methods for updating the position of each particle in include: ;

[0055] In the formula, is the position of the ith particle after updating based on the quantum tunneling phenomenon, is the intensity factor, To update the position of the previous best particle based on the quantum tunneling phenomenon, is the second fluctuation coefficient, ; The expression of intensity factor is: ; In the formula, is an exponential function, is the adaptability of the previous ith particle updated based on the quantum tunneling phenomenon, To update the adaptability of the previous best particle based on the quantum tunneling phenomenon, is the adjustment factor;

[0056] In step 7, the quantum group The methods for updating the position of each particle in include: ;

[0057] In the formula, The position of the i-th particle after updating using the quantum annealing mechanism, To update the position of the previous best particle using the quantum annealing mechanism, To update the position of the nearest particle using the quantum annealing mechanism, is the third volatility coefficient, , is the third volatility coefficient, .

[0058] Furthermore, the method for optimizing the reflective design parameters for each environment category includes:

[0059] Get the number of positions according to the reflective material position in the reflective layout parameters , the number of positions is the number of reflective materials on the protective clothing; get the type label range, the type label range is , is the number of type labels; randomly selected from the type label range type tags, and as a type set, a total of A collection of types, for factorial; sort the type labels in each type set, and take one sorting result as a sorted set, and obtain a total of a sorted set, in which the type labels in the sorted set correspond to the positions of the reflective materials one by one; the light environment data corresponding to the classification points in each environmental category are obtained and marked as analysis environment data; the reflective layout parameters, a set of analysis environment data and a sorted set are used as a set of test data; each set of test data is input into the fire simulation model respectively to obtain the corresponding evaluation data, and marked as second data; according to the second data, the adaptability corresponding to each set of test data is calculated; the adaptability of the test data corresponding to each environmental category is compared to obtain the maximum adaptability corresponding to each environmental category; the sorted set in the test data corresponding to the maximum adaptability is used as the reflective design parameter of the corresponding environmental category.

[0060] The technical effects and advantages of the firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence of the present invention are as follows:

[0061] It can comprehensively collect and analyze fire-fighting characteristic data, including light environment data, reflective characteristic data, reflective property data and flexibility, etc.; by constructing a deep neural network model and dividing different environmental categories, combined with the optimization algorithm, it can realize the automatic optimization of the layout and design of reflective materials under different environmental categories, which not only effectively improves the reflective effect and comfort of protective clothing, but also improves the reaction speed of firefighters in complex environments, thereby effectively protecting the lives of firefighters; and by improving the intelligence level of firefighting equipment, it can significantly improve the visibility and safety of firefighters in complex environments, and improve the performance of protective clothing in emergency firefighting and rescue missions. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic diagram of an intelligent optimization system for firefighters' firefighting protective clothing performance based on artificial intelligence according to Embodiment 1 of the present invention;

[0063] Figure 2 This is a flow chart of a method for optimizing reflective layout parameters corresponding to environmental categories according to Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Example 1

[0066] See also Figure 1 As shown, the firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence in this embodiment includes a data acquisition module, a model building module, an environment classification module, a layout optimization module and a design optimization module; each module is connected by wired and / or wireless means to realize data transmission between modules.

[0067] The data acquisition module is used to collect m groups of fire characteristic data.

[0068] Firefighting characteristic data include light environment data, reflective characteristic data, reflective property data and flexibility;

[0069] Lighting environment data includes illumination, light source angle, and smoke density;

[0070] Illuminance is the luminous flux received per unit area in the fire area. It is used to evaluate the lighting conditions in the fire area. The illuminance is obtained through illuminance sensors distributed and installed in the fire area. The light source angle is the incident angle of the light source relative to the ground in the fire area. It is used to understand the illumination effect of the light source and evaluate the distribution of light in the fire area. The light source angle is obtained through goniometers distributed and installed in the fire area. Smoke density is the concentration of smoke particles per unit volume in the fire area. It is used to evaluate the impact of smoke on light propagation. The smoke density is obtained through smoke sensors distributed and installed in the fire area.

[0071] The reflective characteristic data includes the reflective material position, reflective material area and reflective material type;

[0072] The reflective material position refers to the position of each reflective material on the protective clothing, such as shoulders, back, chest, etc.; the reflective material area refers to the surface area of ​​each reflective material, indicating the size of the reflective material; the reflective material type refers to the type of each reflective material, such as micro glass bead material, aluminum foil reflective material, etc., reflecting the reflective performance and composition of the reflective material; the reflective material position and reflective material area are obtained through on-site measurements by technical personnel in this field, and the reflective material type is obtained through the label or instruction manual of the protective clothing.

[0073] The reflective characteristic data include the mean reflectivity and the mean luminous flux density;

[0074] The reflectivity is the ratio of the luminous flux reflected by the incident light on the surface of the reflective material to the incident luminous flux, which reflects the reflective ability of the reflective material to light; the method for obtaining the mean reflectivity is as follows: obtain the reflected luminous flux and incident luminous flux corresponding to each reflective material, divide the reflected luminous flux of the reflective material corresponding to each position by the corresponding incident luminous flux, and obtain the reflectivity corresponding to each reflective material; count the number of reflective materials and mark them as the number of materials; add the reflectivity of each reflective material in turn, and then divide it by the number of materials to obtain the mean reflectivity; the reflected luminous flux and incident luminous flux are obtained by a photometer integrated in the protective clothing; the luminous flux density is the luminous flux received per unit area in the reflective material, which is used to evaluate the lighting effect of the reflective material in the fire area; the method for obtaining the mean luminous flux density is as follows: divide the incident luminous flux of each reflective material by the corresponding reflective material area to obtain the luminous flux density corresponding to each material; add each luminous flux density in turn, and then divide it by the number of materials to obtain the mean luminous flux density.

[0075] Flexibility refers to the ability of firefighters' joints to stretch during firefighting while wearing protective clothing. It is used to evaluate the rationality of the layout and design of reflective materials. Flexibility is obtained by technical personnel in this field through tests on firefighters wearing protective clothing, such as joint range of motion tests, squat tests, and forward and backward stride tests.

[0076] The model building module builds a fire simulation model based on m groups of fire characteristic data.

[0077] Methods for building fire simulation models include:

[0078] Select a deep learning framework, such as TensorFlow, Keras, PyTorch, etc.; build a fire simulation model architecture, which includes an input layer, a hidden layer, and an output layer; each hidden layer includes multiple neurons, each neuron is connected to the next layer of neurons, and the connection contains weights, which determine the importance and influence of data transmission in the neural network; an activation function is applied to each neuron between the hidden layer and the output layer. The activation function introduces nonlinearity, allowing the network to learn more complex patterns and features.

[0079] Different digital labels are set for different reflective material positions in m groups of fire feature data, and marked as position labels; different digital labels are set for different reflective material types in m groups of fire feature data, and marked as type labels; the reflective material positions in each group of reflective feature data are replaced with corresponding position labels, the reflective material types are replaced with corresponding type labels, and the replaced reflective feature data are marked as reflective replacement data; the light environment data and reflective replacement data in each group of fire feature data are used as training data, and the reflective characteristic data and flexibility in each group of fire feature data are used as evaluation data; the training data and evaluation data in each group of fire feature data are converted into a corresponding set of feature vectors.

[0080] Each set of feature vectors is used as the input of the fire simulation model. The fire simulation model uses a set of predicted evaluation data corresponding to each set of training data as output, and the actual evaluation data corresponding to each set of training data as the prediction target. The actual evaluation data is the pre-set evaluation data corresponding to the training data. The training target is to minimize the sum of the prediction errors of all training data. The calculation formula of the prediction error is: ; In the formula, is the prediction error, is the group number of the feature vector corresponding to the training data, For the The predicted evaluation data corresponding to the group training data, For the The actual evaluation data corresponding to the group training data is obtained; the fire simulation model is trained until the sum of the prediction errors reaches convergence and the training is stopped to obtain the fire simulation model, which is specifically a deep neural network model.

[0081] The environment classification module is used to classify m groups of fire feature data into environmental categories.

[0082] The steps of classifying m groups of fire characteristic data into environmental categories include:

[0083] Step A: Take m groups of light environment data as sample points, and the sample points correspond to the light environment data one by one; preset the number of categories , number of categories Pre-set by technicians in this field according to actual conditions; randomly selected sample points as the center point; The center points are marked in ascending order. , , …, , marking the center point as , ;

[0084] Step B: Mark the sample points that are not the center points as classification points, and mark the classification points in ascending order. , , …, , that is, marking the classification points as , ; Calculate the point distance from each classification point to each center point in turn; the expression of point distance is: ; In the formula, For classification points To center point The point distance, For classification points Middle The value of the dimension, For classification points Middle The value of the dimension, ; Different dimensions represent different types of data in the light environment data, for example, illumination is one dimension and light source angle is another dimension;

[0085] Step C: According to The corresponding center point is established environmental categories;

[0086] Step D: Classify the points The distance to each center point is compared and the classified points are Assign to the environment category corresponding to the center point with the minimum distance to the point;

[0087] Step E: Let , jump back to step D;

[0088] Step F: Repeat steps D to E until When the loop ends, it goes to step G;

[0089] Step G: Recalculate the new center point corresponding to each environment category;

[0090] The calculation method for the new center point of each environmental category includes: ;

[0091] In the formula, For the The environment category corresponds to the new center point. For the In the environmental category classification points, , For the In the environmental category The corresponding values ​​of classification points in different dimensions, For the The number of classification points in each environment category, ;

[0092] Step H: Repeat steps B to G until the new center point of each environment category recalculated in step G is consistent with the new center point of the corresponding environment category calculated in the previous cycle, the cycle ends, and the Environmental categories and corresponding classification points.

[0093] The layout optimization module optimizes the reflective layout parameters for each environment category based on the fire simulation model.

[0094] like Figure 2 As shown, the steps to optimize the reflective layout parameters for each environment category include:

[0095] Step 1: construct a parameter set, set a different digital label for each set of reflective layout parameters in the parameter set, and mark them as parameter labels;

[0096] Step 2: Select environment categories and initialize the quantum group , quantum group There are n particles in the quantum group, and the position of each particle corresponds to the parameter label one by one. The number of iterations t is 0, ;

[0097] Step 3: Define the fitness function;

[0098] Step 4: Based on the phenomenon of quantum entanglement, Each particle in the process updates its position;

[0099] Step 5: Based on the quantum interference phenomenon, Each particle in the process updates its position;

[0100] Step 6: Based on the quantum tunneling phenomenon, Each particle in the process updates its position;

[0101] Step 7: Use quantum annealing mechanism to Each particle in the process updates its position;

[0102] Step 8: Compare the number of iterations t with the preset iteration threshold T; if , then go to step 9; if , then let , and return to step 4;

[0103] Step 9: Calculate the adaptability of each particle, obtain the parameter label corresponding to the particle with the largest adaptability, obtain the reflective layout parameters corresponding to the parameter label, and use it as the first The reflective layout parameters corresponding to the environment category; , and return to step 2;

[0104] Step 10: Repeat steps 2 to 9 until Each environment category obtains the corresponding reflective layout parameters, and the loop ends.

[0105] In the above step 1, the method for constructing the parameter set is: obtaining a parameter range, the parameter range includes a reflective material position range and a reflective material area range; the parameter range is obtained by a person skilled in the art according to industry standards and specifications and in combination with actual conditions; a value is randomly selected from each range within the parameter range to construct a set of reflective layout parameters, and a total of N sets of reflective layout parameters are constructed, and the N sets of reflective layout parameters are used as a parameter set, where N is an integer greater than 1.

[0106] In step 2 above, the quantum group The expression for the position of each particle in is: ; In the formula, is the position of the ith particle, is the random coefficient of the ith particle, , .

[0107] In the above step 3, the expression of the fitness function is: ; In the formula, For adaptability, is the total mean reflectivity, is the total mean value of luminous flux density, For total flexibility, , , , are all preset weight coefficients; the specific values ​​of the weight coefficients can be set according to actual conditions, and the weight coefficients reflect the influence of the total mean reflectivity, the total mean luminous flux density and the total flexibility on adaptability. Technical personnel in this field can preset corresponding weight coefficients according to the actual influence of the total mean reflectivity, the total mean luminous flux density and the total flexibility on adaptability, so as to accurately evaluate the overall performance of protective clothing under different reflective layout parameters.

[0108] It should be understood that the higher the total mean reflectivity, the more light the reflective material can reflect, so the reflective material will appear brighter, thereby improving visibility; the larger the total mean luminous flux density, the more light energy the reflective material receives, so the reflective material will appear clearer, thereby improving its visibility; the higher the visibility, the easier it is to identify protective clothing in low light or complex backgrounds, thereby improving the safety of firefighters, so the greater the adaptability; protective clothing with higher total flexibility allows firefighters to move freely and perform various operations, which is crucial for firefighting, rescue and other work, so the greater the adaptability.

[0109] The method for obtaining the total mean value of reflectivity, total mean value of luminous flux density and total flexibility is as follows: The light environment data corresponding to the classification points in each environment category are obtained and marked as predicted environment data; the reflective layout parameters corresponding to the parameter labels corresponding to the particle positions are obtained, and the reflective layout parameters, a group of predicted environment data and the reflective material type corresponding to the predicted environment data are taken as a group of analysis data; each group of analysis data is input into the fire simulation model respectively to obtain the corresponding evaluation data and marked as the first data; the number of first data is counted and marked as the first number; the mean reflectivity values ​​in each group of first data are added in turn, and then divided by the first number to obtain the total mean reflectivity; the mean luminous flux density values ​​in each group of first data are added in turn, and then divided by the first number to obtain the total mean luminous flux density; the flexibility in each group of first data is added in turn, and then divided by the first number to obtain the total flexibility.

[0110] In the above step 4, for the quantum group The methods for updating the position of each particle in include: ;

[0111] In the formula, is the position of the ith particle after updating based on the quantum entanglement phenomenon, To update the position of the previous i-th particle based on the quantum entanglement phenomenon, is the entanglement coefficient, , To update the position of the nearest particle based on the quantum entanglement phenomenon, is a random number from the standard normal distribution, and the nearest particle is the particle closest to the i-th particle.

[0112] In the above step 5, the quantum group The methods for updating row positions include: ;

[0113] In the formula, is the position of the ith particle after updating based on the quantum interference phenomenon, is the first fluctuation coefficient, , To update the position of the previous best particle based on the quantum interference phenomenon, is the interference coefficient, and the best particle is the quantum group The interference coefficient is preset by those skilled in the art according to actual conditions for the particles with the greatest adaptability.

[0114] In step 6 above, for the quantum group The methods for updating the position of each particle in include: ;

[0115] In the formula, is the position of the ith particle after updating based on the quantum tunneling phenomenon, is the intensity factor, To update the position of the previous best particle based on the quantum tunneling phenomenon, is the second fluctuation coefficient, ;

[0116] The expression of intensity factor is: ; In the formula, is an exponential function, is the adaptability of the previous ith particle updated based on the quantum tunneling phenomenon, To update the adaptability of the previous best particle based on the quantum tunneling phenomenon, It is an adjustment factor, which is preset by those skilled in the art according to actual conditions.

[0117] In the above step 7, for the quantum group The methods for updating the position of each particle in include: ;

[0118] In the formula, The position of the i-th particle after updating using the quantum annealing mechanism, To update the position of the previous best particle using the quantum annealing mechanism, To update the position of the nearest particle using the quantum annealing mechanism, is the third volatility coefficient, , is the third volatility coefficient, .

[0119] Design optimization module, based on fire simulation model and reflective layout parameters, optimizes reflective design parameters for each environment category.

[0120] Methods for optimizing reflective design parameters for each environmental category include:

[0121] Get the number of positions according to the reflective material position in the reflective layout parameters , the number of positions is the number of reflective materials on the protective clothing; get the type label range, the type label range is obtained by setting different digital labels for different reflective material types in the m groups of fire feature data, the type label range is , is the number of type labels, which is obtained by technical personnel in the field of reflective materials by statistics of all types of reflective materials; randomly selected from the range of type labels type tags, and as a type set, a total of A collection of types, for factorial; sort the type labels in each type set, and take one sorting result as a sorted set, and obtain a total of a sorted set, in which the type labels in the sorted set correspond to the positions of the reflective materials one by one; the light environment data corresponding to the classification points in each environmental category are obtained and marked as analysis environment data; the reflective layout parameters, a set of analysis environment data and a sorted set are used as a set of test data; each set of test data is input into the fire simulation model respectively to obtain the corresponding evaluation data, and marked as second data; according to the second data, the adaptability corresponding to each set of test data is calculated; the adaptability of the test data corresponding to each environmental category is compared to obtain the maximum adaptability corresponding to each environmental category; the sorted set in the test data corresponding to the maximum adaptability is used as the reflective design parameter of the corresponding environmental category.

[0122] This embodiment can comprehensively collect and analyze fire feature data, including light environment data, reflective feature data, reflective property data and flexibility, etc.; by constructing a deep neural network model and dividing different environmental categories, and then combining the optimization algorithm, the layout and design of reflective materials in different environmental categories can be automatically optimized, which not only effectively improves the reflective effect and comfort of protective clothing, but also improves the reaction speed of firefighters in complex environments, thereby effectively protecting the lives of firefighters; and by improving the intelligence level of firefighting equipment, it can significantly improve the visibility and safety of firefighters in complex environments, and improve the performance of protective clothing in emergency firefighting and rescue missions.

[0123] Example 2

[0124] The present application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable codes, and when the computer-readable codes are executed by the one or more processors, the firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence as described above may be executed.

[0125] The method or system according to the implementation mode of the present application can also be implemented with the help of the architecture of the electronic device shown in the present application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. A storage device in the electronic device, such as a ROM or a hard disk, can store the firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence provided in the present application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in the present application is only exemplary. When implementing different devices, one or more components in the electronic device shown in the present application may be omitted according to actual needs.

[0126] Example 3

[0127] One embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, the firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0128] In addition, according to the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided by the present application, such as: an intelligent optimization system for the performance of firefighters' firefighting protective clothing based on artificial intelligence. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0129] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0130] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence is characterized by: include: A data collection module, used for collecting m groups of fire characteristic data; A model building module builds a fire simulation model based on m groups of fire characteristic data; The environment classification module is used to classify the m groups of fire feature data into environment categories; the step of classifying the m groups of fire feature data into environment categories comprises: Step A: Take m groups of light environment data as sample points, and the sample points correspond to the light environment data one by one; preset the number of categories a, randomly select a sample points as center points; mark the a center points in ascending order as ω1, ω2, ..., ω a , that is, mark the center point as ω b , b∈[1,a]; Step B: Mark the sample points that are not used as the center points as classification points, and mark the classification points in ascending order as ψ1, ψ2, …, ψ m-a , that is, marking the classification point as ψ c , c∈[1,ma]; calculate the point distance from each classification point to each center point in turn; Step C: Establish a corresponding a environment categories according to a center points; Step D: Classify the points ψ c The distance to each center point is compared and the classification point ψ c Assign to the environment category corresponding to the center point with the minimum distance to the point; Step E: Let c=c+1, and jump back to step D; Step F: loop through steps D to E until c=ma, and then proceed to step G; Step G: Recalculate the new center point corresponding to each environment category; Step H: Repeat steps B to G until the new center point of each environmental category recalculated in step G is consistent with the new center point of the corresponding environmental category calculated in the previous cycle, the cycle ends, and a environmental categories and corresponding classification points are obtained; The layout optimization module optimizes the reflective layout parameters of each environment category based on the fire simulation model; the step of optimizing the reflective layout parameters of each environment category includes: Step 1: construct a parameter set, set a different digital label for each set of reflective layout parameters in the parameter set, and mark them as parameter labels; Step 2: Select the a′th environment category and initialize the quantum group S. The quantum group S includes n particles. The position of each particle corresponds to the parameter label one by one. The number of iterations t of the initialization quantum group S is 0, a′∈[1,a]; Step 3: Define the fitness function; Step 4: Based on the quantum entanglement phenomenon, the position of each particle in the quantum group S is updated; Step 5: Based on the quantum interference phenomenon, the position of each particle in the quantum group S is updated; Step 6: Based on the quantum tunneling phenomenon, the position of each particle in the quantum group S is updated; Step 7: Using the quantum annealing mechanism, the position of each particle in the quantum group S is updated; Step 8: Compare the number of iterations t with the preset iteration threshold T; if t≥T, proceed to step 9; if t<T, set t=t+1 and return to step 4; Step 9: Calculate the adaptability corresponding to each particle, obtain the parameter label corresponding to the particle with the largest adaptability, obtain the reflective layout parameter corresponding to the parameter label, and use it as the reflective layout parameter corresponding to the a′th environmental category; let a′=a′+1, and return to step 2; Step 10: looping steps 2 to 9 until all a environment categories obtain corresponding reflective layout parameters, and then the loop ends; Design optimization module, based on fire simulation model and reflective layout parameters, optimizes reflective design parameters for each environment category.

2. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 1 is characterized in that: The fire-fighting characteristic data includes light environment data, reflective characteristic data, reflective property data and flexibility; The light environment data includes illumination, light source angle and smoke density; the illumination is the light flux received per unit area in the fire area; the light source angle is the incident angle of the light source relative to the ground in the fire area; the smoke density is the concentration of smoke particles per unit volume in the fire area; The reflective characteristic data includes reflective material position, reflective material area and reflective material type; the reflective material position is the position of each reflective material on the protective clothing; the reflective material area is the surface area of ​​each reflective material; the reflective material type is the type of each reflective material; The reflective characteristic data includes a reflectivity mean value and a luminous flux density mean value; The reflectivity is the ratio of the luminous flux reflected by the incident light on the surface of the reflective material to the incident luminous flux; the method for obtaining the mean reflectivity is: obtaining the reflected luminous flux and the incident luminous flux corresponding to each reflective material, dividing the reflected luminous flux of the reflective material corresponding to each position by the corresponding incident luminous flux, and obtaining the reflectivity corresponding to each reflective material; counting the number of reflective materials and marking them as the number of materials; adding the reflectivity of each reflective material in turn, and then dividing by the number of materials to obtain the mean reflectivity; The luminous flux density is the luminous flux received per unit area in the reflective material; the method for obtaining the average luminous flux density is: dividing the incident luminous flux of each reflective material by the corresponding reflective material area to obtain the luminous flux density corresponding to each material; adding each luminous flux density in turn and dividing it by the number of materials to obtain the average luminous flux density; The flexibility refers to the ability of firefighters to stretch their joints while wearing protective clothing during firefighting.

3. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 2 is characterized in that: The method for constructing a fire simulation model comprises: Select a deep learning framework and build a fire simulation model architecture, which includes an input layer, a hidden layer, and an output layer; Different digital labels are set for different reflective material positions in m groups of fire feature data, and marked as position labels; different digital labels are set for different reflective material types in m groups of fire feature data, and marked as type labels; the reflective material positions in each group of reflective feature data are replaced with corresponding position labels, and the reflective material types are replaced with corresponding type labels, and the replaced reflective feature data are marked as reflective replacement data; the light environment data and reflective replacement data in each group of fire feature data are used as training data, and the reflective characteristic data and flexibility in each group of fire feature data are used as evaluation data; the training data and evaluation data in each group of fire feature data are converted into a corresponding set of feature vectors; Each group of feature vectors is used as the input of the fire simulation model. The fire simulation model takes a group of predicted evaluation data corresponding to each group of training data as output, and takes the actual evaluation data corresponding to each group of training data as the prediction target. The actual evaluation data is the pre-set evaluation data corresponding to the training data. Minimizing the sum of prediction errors of all training data is used as the training goal. The fire simulation model is trained until the sum of prediction errors converges, and the training is stopped to obtain the fire simulation model, which is a deep neural network model.

4. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 3 is characterized in that: The expression for point distance is: Where D cb is the classification point ψ c To the center point ω b The point distance, ω bd is the classification point ω b The value of the dth dimension in , ψ cd is the classification point ψ c The value of the dth dimension in , d∈[1,3]; where different dimensions represent different types of data in the light environment data; The calculation method for the new center point of each environmental category includes: In the formula, ω′ a is the new center point corresponding to the a-th environment category, ψ ar is the rth classification point in the ath environmental category, ψ ar =(ψ ar1 ,ψ ar1 ,ψ ar3 ), (ψ ar1 ,ψ ar1 ,ψ ar3 ) is the value corresponding to the r-th classification point in the a-th environmental category in different dimensions, R a is the number of classification points in the ath environment category, r∈[1,R a ].

5. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 4 is characterized in that: In the step 1, the method for constructing the parameter set is: obtaining a parameter range, the parameter range including a reflective material position range and a reflective material area range; randomly selecting a value in each range within the parameter range to construct a set of reflective layout parameters, constructing a total of N sets of reflective layout parameters, and taking the N sets of reflective layout parameters as a parameter set, where N is an integer greater than 1; In step 2, the expression for the position of each particle in the quantum group S is: In the formula, is the position of the ith particle, P i is the random coefficient of the ith particle, P i ∈[0,1], i∈[1,n]; In step 3, the expression of the adaptability function is: Where f is adaptability, fs is the total mean reflectivity, gt is the total mean luminous flux density, lh is the total flexibility, ζ1, ζ2, ζ3, All are preset weight coefficients; The method for obtaining the total mean value of reflectivity, the total mean value of luminous flux density and the total flexibility is as follows: obtain the light environment data corresponding to the classification point in the a′th environmental category, and mark it as predicted environmental data; obtain the reflective layout parameters corresponding to the parameter label corresponding to the particle position, and use the reflective layout parameters, a group of predicted environmental data and the reflective material type corresponding to the predicted environmental data as a group of analysis data; input each group of analysis data into the fire simulation model respectively, obtain the corresponding evaluation data, and mark it as the first data; count the number of first data and mark it as the first number; add the reflectivity means in each group of first data in turn, and then divide it by the first number to obtain the total mean value of reflectivity; add the luminous flux density means in each group of first data in turn, and then divide it by the first number to obtain the total mean value of luminous flux density; add the flexibility in each group of first data in turn, and then divide it by the first number to obtain the total flexibility.

6. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 5 is characterized in that: In step 4, the method for updating the position of each particle in the quantum group S includes: In the formula, is the position of the ith particle after updating based on the quantum entanglement phenomenon, is the position of the i-th particle before updating based on the quantum entanglement phenomenon, α is the entanglement coefficient, α∈[0,1], is the position of the nearest particle before updating based on the quantum entanglement phenomenon, N(0,1) is a random number in the standard normal distribution, and the nearest particle is the particle closest to the i-th particle; In step 5, the method for updating the position of each particle in the quantum group S includes: In the formula, is the position of the ith particle after update based on the quantum interference phenomenon, C1 is the first fluctuation coefficient, C1∈[0,1], is the position of the best particle before updating based on the quantum interference phenomenon, δ is the interference coefficient, and the best particle is the particle with the greatest adaptability in the quantum group S.

7. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 6 is characterized in that: In step 6, the method for updating the position of each particle in the quantum group S includes: In the formula, is the position of the ith particle after updating based on the quantum tunneling phenomenon, γ is the intensity factor, is the position of the best particle before updating based on the quantum tunneling phenomenon, C2 is the second fluctuation coefficient, C2∈[0,1]; The expression of intensity factor is: Where exp is an exponential function, fi is the fitness of the ith particle before updating based on the quantum tunneling phenomenon, fb is the fitness of the best particle before updating based on the quantum tunneling phenomenon, and y is the adjustment factor; In step 7, the method for updating the position of each particle in the quantum group S includes: In the formula, The position of the i-th particle after updating using the quantum annealing mechanism, To update the position of the previous best particle using the quantum annealing mechanism, To update the position of the nearest particle before using the quantum annealing mechanism, C3 is the third fluctuation coefficient, C3∈[0,1], and C4 is the third fluctuation coefficient, C4∈[0,1].

8. The firefighter firefighting protective clothing performance intelligent optimization system based on artificial intelligence according to claim 7 is characterized in that: The method for optimizing the reflective design parameters for each environmental category includes: According to the reflective material position in the reflective layout parameters, obtain the number of positions v, which is the number of reflective materials on the protective clothing; obtain the type label range, the type label range is [1, V], and V is the number of type labels; randomly select v type labels from the type label range and take them as a type set, and obtain a total of type sets, V! is the factorial of V; sort the type labels in each type set, and take one sorting result as a sorted set, and get a total of a sorted set, in which the type labels in the sorted set correspond to the positions of the reflective materials one by one; the light environment data corresponding to the classification points in each environmental category are obtained and marked as analysis environment data; the reflective layout parameters, a set of analysis environment data and a sorted set are used as a set of test data; each set of test data is input into the fire simulation model respectively to obtain the corresponding evaluation data, and marked as second data; according to the second data, the adaptability corresponding to each set of test data is calculated; the adaptability of the test data corresponding to each environmental category is compared to obtain the maximum adaptability corresponding to each environmental category; the sorted set in the test data corresponding to the maximum adaptability is used as the reflective design parameter of the corresponding environmental category.

Images