Natural language based job behavior simulation generation method

Abstract

The application provides a natural language-based job behavior simulation generation method, and relates to the technical field of data simulation.The application establishes a fire combustion model in a simulation environment, then obtains ignition conditions through an LLM model, obtains a development weight in natural language, simulates the fire combustion model, obtains the volume, surface area and shape factor of a combustion field, forms an average fire index, obtains an overall excess temperature according to the temperature of the combustion field, further obtains a fire intensity index, adjusts the intensity threshold according to the development weight, compares the fire intensity index and the intensity threshold to judge whether the simulation result meets the expectation of the natural language and adjusts the simulation result, provides specific quantitative guidance for users, and ensures that the simulation result meets the expectation of the natural language.

Description

Technical Field

[0001] This invention relates to the field of data simulation technology, specifically to a method for generating simulations of job behaviors based on natural language. Background Technology

[0002] With the development of emerging technologies, more and more cutting-edge technologies are being applied to traditional fields. Natural language extraction technology can convert technical data that is normally required to be recognized by language conversion systems, lowering the barrier for users and facilitating data adjustment. It avoids complex and professional data creation processes. Through the powerful flame simulation capabilities of Computational Fluid Dynamics (CFD) software, the development process of a fire can be simulated in a computer. It has complete data construction and processing modules that are easy for users to access. Combined with scene modeling, fire scenes can be simulated in specific scenarios. By sending the data extracted from natural language into the CFD software, simulation can be directly achieved through language control, and then visualized through VR to intuitively show the situation of the fire scene.

[0003] In fields such as fire emergency teaching and training, traditional teaching methods are still limited to videos and on-site simulations. Videos cannot allow trainees to truly feel the harm and destructive power of fire, while on-site simulations require a lot of resources and can only simulate specific fire situations in a single session, failing to demonstrate the complex and ever-changing conditions of a fire scene. Simulating a fire scene using natural language control can greatly improve training outcomes.

[0004] However, due to the limited amount of precise information carried by natural language itself, simulation software cannot obtain accurate conditions and data. Furthermore, due to the complexity of the fire development process, it is even more difficult to ensure that the simulation results meet the expectations of natural language itself, thus hindering fire emergency training.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a method for generating simulations of job behaviors based on natural language, so as to solve the problems mentioned in the background art.

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

[0008] The method for generating job behavior simulation based on natural language includes the following steps:

[0009] Step 1: Construct a three-dimensional Cartesian coordinate system in the simulation environment and build a scene model based on the real environment at a 1:1 scale. Set the names of each component and establish the connection relationship between each independent component. Set the air data and establish the simulation relationship to form a fire combustion model.

[0010] Step 2: Obtain the ignition conditions of components in the fire combustion model through the LLM model based on natural language, and form biased development weights based on the corresponding words in the natural language.

[0011] Step 3: Obtain the combustion field at equal time intervals according to the simulation of the fire combustion model to form a combustion field sequence. Obtain the volume, surface area and shape factor of each combustion field in the combustion field sequence to form a fire index. Form a fire index sequence according to the order of the combustion fields. Filter the fire index sequence to obtain the average fire index.

[0012] Step 4: Obtain the trend of temperature change over time in each unit block of the combustion field, obtain the excess temperature of each unit block based on the difference between the unit block temperature and the ambient temperature, and combine the air data to obtain the overall excess temperature of all combustion fields in the combustion field sequence.

[0013] Step 5: Based on the average fire intensity index and the overall excess temperature, form the fire intensity index for this simulation, set the intensity threshold and correct it through development weights, obtain the deviation of the fire intensity index from the corrected intensity index, feed back the deviation, and execute the simulation based on the feedback results.

[0014] Furthermore, a three-dimensional Cartesian coordinate system is constructed in the CFD environment. The unit length of the three-dimensional Cartesian coordinate system is mapped 1:1 with respect to the real environment. Scene modeling simulating the real environment is carried out in the three-dimensional Cartesian coordinate system. The scene modeling uses three-dimensional modeling to describe the shape of each component and name it.

[0015] The logic for determining the connection relationship between independent components is as follows:

[0016] If two unit blocks belonging to two independent components are connected at these two unit blocks if the distance between the midpoints of their two nearest faces is less than 1 cm, then the two independent components are considered to be connected at these two unit blocks; otherwise, the two unit blocks are separate. This is used to determine the areas where all components are connected, and the heat transfer in the connected areas is carried out by heat conduction.

[0017] Set the ambient air data, which includes density and specific heat capacity of air.

[0018] Furthermore, heat conduction and heat radiation equations are set according to the connection relationship of each component, and boundary conditions are set to form a fire combustion model.

[0019] Furthermore, natural language is sent to the fire combustion model. The LLM model obtains the ignition conditions based on the natural language, where the ignition conditions are that a specified component is in a burning state. The ignition conditions are then input into the fire combustion model for simulation. The natural language is analyzed again to obtain the development weights. The logic is as follows:

[0020] A tendency dictionary is established, containing five tendency words: intense, slightly large, normal, slow, and weak. A corresponding word is assigned to each tendency word category, along with a weight. The corresponding words are selected from the natural language database. The LLM model extracts the corresponding words from the natural language and maps them to each tendency word category, obtaining the weight of the corresponding tendency word in the current natural language simulation. This weight is the development weight for this simulation. If no corresponding word is extracted, the development weight for this simulation is considered to be the normal type.

[0021] Furthermore, a full-process combustion simulation is performed in a CFD environment. This full-process combustion simulation refers to the intermediate process from the generation of ignition conditions to the determination by the fire combustion model that there is no open flame and the smoke concentration is less than 30%. The combustion field during the full-process combustion simulation is obtained, and the logic is as follows:

[0022] The unit block containing the point with a temperature exceeding 600 degrees Celsius is defined as the combustion zone, and the collection of all combustion zones at the same time is defined as the combustion field;

[0023] The combustion field during the entire combustion simulation process is acquired at equal time intervals and formed into a combustion field sequence according to time order;

[0024] The volume of each combustion field is obtained, and the closed surface of the combustion field is generated using the MarchingCubes algorithm and its area is calculated. The maximum distance of each combustion field on the X, Y, and Z axes is obtained, and the shape factor of each combustion field is obtained using the following formula:

[0025]

[0026] in, Indicates the first The shape factor of the combustion field Indicates the first The volume of the combustion field, Indicates the first The maximum distance of each combustion field on the X-axis Indicates the first The maximum distance of each combustion field on the Y-axis Indicates the first The maximum distance of the combustion field on the Z-axis Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0027] Furthermore, the fire intensity index for each combustion zone is obtained using the following formula:

[0028]

[0029] in, Indicates the first The fire intensity index of each burning area Indicates the first The volume of the combustion field, Indicates the first The shape factor of the combustion field For the first The surface area of ​​the combustion field For coefficients, Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0030] The fire intensity index of each combustion field is sorted according to the combustion field sequence to form a fire intensity index sequence. The average fire intensity index is then obtained by filtering the fire intensity index sequence. The logic is as follows:

[0031] Obtain the average value of the fire index in the fire index sequence, identify and remove combustion fields whose fire index is less than one-third of the average value, obtain the average value of the fire index of the remaining combustion fields, and label the obtained average value as the average fire index.

[0032] Furthermore, the trend of temperature change over time within each unit block of the combustion field is obtained, along with the ambient temperature, to form the excess temperature of each unit block in the combustion field. The formula used is as follows:

[0033]

[0034] in, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The excess temperature of the unit block, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The temperature of the unit block, For ambient temperature, Retrieve variables for X-axis coordinates. , , This represents the minimum value of the combustion field on the X-axis. This represents the maximum value of the combustion field on the X-axis. Retrieve variables for the Y-axis coordinate. , , This represents the minimum value of the combustion field on the Y-axis. This represents the maximum value of the combustion field on the Y-axis. Retrieve variables for Z-axis coordinates. , , This represents the minimum value of the combustion field on the Z-axis. This represents the maximum value of the combustion field on the Z-axis. Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0035] The overall excess temperature of the combustion field sequence is obtained using the following formula:

[0036]

[0037] in, This represents the total excess temperature of the combustion field sequence. air density, This is the specific heat capacity of air.

[0038] Furthermore, the fire intensity index for the entire combustion process is obtained using the following formula:

[0039]

[0040] in, This is an index indicating the intensity of the fire. This represents the total excess temperature of the combustion field sequence. This represents the average fire intensity index.

[0041] A severity threshold is set, and this threshold is then adjusted using development weights in natural language processing. The formula used is as follows:

[0042]

[0043] in, For the correction of the severity threshold, The threshold for severity, For development weight.

[0044] Furthermore, the deviation of the fire intensity index from the severity threshold is obtained using the following formula:

[0045]

[0046] in, For deviation, This is an index indicating the intensity of the fire. For the adjusted severity threshold;

[0047] The deviation is fed back to the natural language input user for confirmation. If the natural language input user modifies the input language, the deviation is obtained again for the modified full-process combustion simulation and fed back. When the natural language input user confirms that the simulation will be performed with the current deviation, the simulation will be displayed through VR.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] This invention simulates fire scenes by combining real-world scenario modeling in a CFD environment with natural language extraction from an LLM model. It identifies the combustion field in the entire combustion simulation, obtains the average fire intensity index from the flame shape dimension, and obtains the overall excess temperature from the combustion field temperature dimension. The two are combined to form a fire intensity index that reflects the development trend of the fire. The intensity threshold of the fire scene is adjusted based on the intention tendency in the natural language. The degree of deviation between the fire intensity index and the intensity threshold reflects the difference between the input natural language and the actual simulation results, providing users with specific quantitative guidance to ensure that the simulation results match the expectations of the natural language. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the overall method flow of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0052] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0053] Example:

[0054] Please see Figure 1 The present invention provides a technical solution:

[0055] The method for generating job behavior simulation based on natural language includes the following steps:

[0056] Step 1: Construct a three-dimensional Cartesian coordinate system in the simulation environment and build a scene model based on the real environment at a 1:1 scale. Set the names of each component and establish the connection relationship between each independent component. Set the air data and establish the simulation relationship to form a fire combustion model.

[0057] Step 1 includes the following:

[0058] A three-dimensional Cartesian coordinate system is constructed in the CFD environment. The unit length of the three-dimensional Cartesian coordinate system is mapped to the real environment at a 1:1 ratio. Scene modeling simulating the real environment is carried out in the three-dimensional Cartesian coordinate system. The scene modeling uses three-dimensional modeling to describe the shape of each component and name it.

[0059] In a preferred embodiment, a BIM model of the real environment is obtained. The BIM model simulates various components and their materials in the real environment. These components include various flammable and non-flammable items such as walls, tables, chairs, and electrical equipment. The BIM model is then imported into CFD environment software, such as Fire Dynamics Simulator (FDS, a fire dynamics simulation tool developed by the National Bureau of Standards and Technology). Simulating the real environment in CFD environment software is a common technical feature in this field and will not be elaborated here. A three-dimensional Cartesian coordinate system assigns precise coordinates to each position in the scene model. A solid block constructed from the unit length of the three coordinate axes in the three-dimensional Cartesian coordinate system is a unit block. The unit length is set according to the real environment; in this embodiment, 1 cm can be set as the unit length.

[0060] The logic for determining the connection relationship between independent components is as follows:

[0061] If two unit blocks belonging to two independent components are connected at these two unit blocks if the distance between the midpoints of their two nearest faces is less than 1 cm, then the two independent components are considered to be connected at these two unit blocks; otherwise, the two unit blocks are separate. This is used to determine the areas where all components are connected, and the heat transfer in the connected areas is carried out by heat conduction.

[0062] Set the ambient air data, which includes density and specific heat capacity of air.

[0063] The heat conduction and heat radiation equations are set according to the connection relationship of each component, and the boundary conditions are set to form a fire combustion model.

[0064] By constructing a 1:1 scale three-dimensional Cartesian coordinate system within a CFD environment and using this as the basis for detailed 3D scene modeling, a high-fidelity physical foundation was provided for the subsequent simulation system. Each component was explicitly named in the model, and geometric information such as distance and midpoint was used to determine the connection relationships between independent unit blocks, ensuring accurate identification of the actual connection interfaces between components. This judgment logic not only reflects the minute distance changes at the contact surfaces of actual objects but also provides a refined basis for the specific implementation of heat conduction and heat radiation transfer modes. The introduction of air density and specific heat capacity data ensures the reliability and accuracy of the physical model in heat transfer and energy conversion processes. This series of processing logics built a foundation for the fire combustion model that possesses both physical realism and the ability to capture subtle local differences, thus playing a crucial role in laying a precise and reliable foundation for simulation data within the overall simulation system, providing solid data support for subsequent fire behavior prediction and safety assessment.

[0065] Step 2: Obtain the ignition conditions of components in the fire combustion model through the LLM model based on natural language, and form biased development weights based on the corresponding words in the natural language.

[0066] Step 2 includes the following:

[0067] Natural language is sent to the fire combustion model. The LLM model obtains the ignition conditions based on the natural language, where the ignition conditions are that a specified component is in a burning state. The ignition conditions are then input into the fire combustion model for simulation. The natural language is analyzed again to obtain the development weights. The logic is as follows:

[0068] A tendency dictionary is established, containing five tendency words: intense, slightly large, normal, slow, and weak. A corresponding word is assigned to each tendency word category, along with a weight. The corresponding words are selected from the natural language database. The LLM model extracts the corresponding words from the natural language and maps them to each tendency word category, obtaining the weight of the corresponding tendency word in the current natural language simulation. This weight is the development weight for this simulation. If no corresponding word is extracted, the development weight for this simulation is considered to be the normal type.

[0069] A Large Language Model (LLM) is introduced to interpret natural language input and extract descriptive features of fire development based on a predefined sentiment dictionary. This step first transforms external natural language into ignition conditions within the fire combustion model, bridging the gap between phenomenological descriptions and physical state parameters. Specifically, by analyzing natural language to determine the combustion state of specified components, the simulation system can quickly respond to changes in on-site descriptions and ignition conditions. Furthermore, by defining five sentiment words (violent, moderate, normal, slow, weak) and their corresponding weights, the system maps the corresponding words in the natural language to the model parameters, providing a quantitative correction mechanism based on linguistic sentiment and descriptive intensity for the fire's development speed and energy transfer. This mechanism helps the simulation system better adapt to complex and changing fire scene information, guiding the matching of simulation results with the expected natural language input, and improving the simulation results' ability and accuracy in predicting actual conditions.

[0070] Step 3: Obtain the combustion field at equal time intervals according to the simulation of the fire combustion model to form a combustion field sequence. Obtain the volume, surface area and shape factor of each combustion field in the combustion field sequence to form a fire index. Form a fire index sequence according to the order of the combustion fields. Filter the fire index sequence to obtain the average fire index.

[0071] Step 3 includes the following:

[0072] Step 301: Perform a full-process combustion simulation in the CFD environment. The full-process combustion simulation refers to the intermediate process from the generation of ignition conditions to the determination by the fire combustion model that there is no open flame and the smoke concentration is less than 30%. Obtain the combustion field during the full-process combustion simulation. The logic is as follows:

[0073] The unit block containing the point with a temperature exceeding 600 degrees Celsius is defined as the combustion zone, and the collection of all combustion zones at the same time is defined as the combustion field;

[0074] The combustion field during the entire combustion simulation process is acquired at equal time intervals and formed into a combustion field sequence according to time order;

[0075] The volume of each combustion field is obtained, and the closed surface of the combustion field is generated using the MarchingCubes algorithm and its area is calculated. The maximum distance of each combustion field on the X, Y, and Z axes is obtained, and the shape factor of each combustion field is obtained using the following formula:

[0076]

[0077] in, Indicates the first The shape factor of the combustion field Indicates the first The volume of the combustion field, Indicates the first The maximum distance of each combustion field on the X-axis Indicates the first The maximum distance of each combustion field on the Y-axis Indicates the first The maximum distance of the combustion field on the Z-axis Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0078] in Indicates the first The volume of a combustion field reflects the amount of space occupied by heat release and material combustion in the combustion zone in a real environment; , and These represent the maximum distances of the combustion field in the three orthogonal directions X, Y, and Z, respectively, reflecting the extent of the combustion field's expansion in each spatial dimension. Combustion field volume. An increase in [amount] means a larger burning area, and the fire may be more intense; while when [amount] increases, it means a larger burning area, and the fire may be more intense; , and When any value increases, if the volume remains constant, it indicates that the shape of the combustion field tends to become elongated or flattened, thus reducing the shape factor. Conversely, if these lengths decrease while the volume remains constant, it indicates that the combustion field is more compact, the shape factor increases, and heat energy is released more concentratedly. By comparing the volume with the product of the three spatial dimensions, the compactness and spatial distribution of the combustion field are essentially measured, reflecting the geometric characteristics of the fire. This helps to quantify the three-dimensional morphology of the combustion field and incorporate it into the calculation of the fire intensity index, thus providing crucial geometric basis and data support for subsequent quantitative analysis of fire development, fire intensity assessment, and dynamic simulation.

[0079] In the full-process combustion simulation, by identifying the combustion zones in the temperature field and acquiring the combustion field at equal time intervals, and using methods to calculate volume, surface area, and shape factor, the combustion zones in the complex fire development process are quantitatively described, thus providing sufficient and objective geometric and physical parameter support for the generation of the fire intensity index. This process ensures the continuity and temporal sequence of the simulation data, providing detailed physical quantity indicators for assessing the fire scale and combustion characteristics.

[0080] Combustion zones are determined by a temperature threshold exceeding 600 degrees Celsius. A closed-surface extraction algorithm is used to generate the geometry of the combustion field, and detailed calculations are performed on the volume, surface area, and shape factor of the combustion field. This ensures both the accuracy of spatial region division and the ability to capture subtle changes in fire spread and contraction over time.

[0081] Step 302: Obtain the fire intensity index for each combustion zone, based on the following formula:

[0082]

[0083] in, Indicates the first The fire intensity index of each burning area Indicates the first The volume of the combustion field, Indicates the first The shape factor of the combustion field For the first The surface area of ​​the combustion field For coefficients, Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0084] Indicates the first The fire intensity index of each combustion field reflects the overall intensity of the fire in that combustion area; It represents the volume of the combustion field, indicating the amount of material actually participating in combustion and the overall spatial scale of the released heat energy in the combustion region; This represents the surface area of ​​the flame in the combustion zone, reflecting the degree of dispersion of the fire. This is a correction factor used to make necessary scale adjustments to the overall fire situation assessment; It is the shape factor of the combustion field, used to measure the spatial compactness of the combustion zone. When the volume of the combustion zone... An increase in the fire intensity index indicates an expansion in the scale of the fire. It tends to increase; if An increase indicates a greater degree of dispersion within a unit area, which is detrimental to combustion development and will... reduce; A scaling factor is provided to adjust the shape factor value; and the shape factor An increase in the fire intensity index usually indicates a more compact combustion zone, which to some extent exacerbates the concentrated release of fire and its harmful effects; conversely, a decrease in the fire intensity index will reduce the fire intensity index. This formula integrates the geometric characteristics of the combustion field, the scale of heat release, and the characteristics of local flames. By quantitatively describing the fire intensity of the combustion field, it provides a crucial quantitative indicator for fire assessment, dynamic trend analysis, and emergency control, thus serving as a bridge and link in the simulation process.

[0085] The fire intensity index of each combustion field is sorted according to the combustion field sequence to form a fire intensity index sequence. The average fire intensity index is then obtained by filtering the fire intensity index sequence. The logic is as follows:

[0086] Obtain the average value of the fire index in the fire index sequence, identify and remove combustion fields whose fire index is less than one-third of the average value, obtain the average value of the fire index of the remaining combustion fields, and label the obtained average value as the average fire index.

[0087] Further, a fire intensity index is calculated for each combustion field, and outliers below one-third of the average value are filtered out through sorting and selection, thereby extracting core indicators representing the overall fire development. The advantage of this step is that it not only provides a quantitative description of fire development but also utilizes geometric features such as shape factors to make the data more three-dimensional and multi-dimensional, enabling a more accurate description of the spatial spread and energy release characteristics of the fire. The data processing logic in step 3 achieves an organic unity of space, time, and physical quantities, providing strong data support for the quantitative analysis and trend prediction of fire evolution, making the entire simulation process more scientific, systematic, and operational.

[0088] Step 4: Obtain the trend of temperature change over time in each unit block of the combustion field, obtain the excess temperature of each unit block based on the difference between the unit block temperature and the ambient temperature, and combine the air data to obtain the overall excess temperature of all combustion fields in the combustion field sequence.

[0089] Step 4 includes the following:

[0090] The trend of temperature change over time within each unit block of the combustion field is obtained, along with the ambient temperature, to form the excess temperature of each unit block in the combustion field. The formula used is as follows:

[0091]

[0092] in, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The excess temperature of the unit block, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The temperature of the unit block, For ambient temperature, Retrieve variables for X-axis coordinates. , , This represents the minimum value of the combustion field on the X-axis. This represents the maximum value of the combustion field on the X-axis. Retrieve variables for the Y-axis coordinate. , , This represents the minimum value of the combustion field on the Y-axis. This represents the maximum value of the combustion field on the Y-axis. Retrieve variables for Z-axis coordinates. , , This represents the minimum value of the combustion field on the Z-axis. This represents the maximum value of the combustion field on the Z-axis. Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

[0093] The overall excess temperature of the combustion field sequence is obtained using the following formula:

[0094]

[0095] in, This represents the total excess temperature of the combustion field sequence. air density, This is the specific heat capacity of air.

[0096] It represents the total excess temperature energy accumulated by all unit blocks in the entire combustion field sequence, reflecting the energy enrichment caused by the temperature being higher than the ambient temperature during the fire. Density is the density of air, which in a real environment represents the mass of air per unit volume. It is the specific heat capacity of air, representing the amount of heat absorbed by a unit mass of air when the temperature increases by 1 degree Celsius; Then it means in the first The coordinates of the combustion field The unit block, the temperature difference between its temperature and the ambient temperature. When the air density Or specific heat capacity As the temperature increases, the energy contained in a unit temperature difference also increases, thus increasing the overall excess temperature. The corresponding increase; similarly, if the excess temperature within a certain unit block... An increase means that more heat energy is stored at that point, thus making... Furthermore, as temperature anomalies within the combustion field expand to more unit blocks (i.e., the spatial range expands), the accumulated thermal energy across the entire combustion zone increases due to the accumulation of excess temperature at each coordinate point. This formula converts the temperature anomalies of each tiny unit in the combustion field into energy magnitudes through physical parameters and integrates them spatially, thus providing a scientific and systematic evaluation index for further quantification of fire thermal energy distribution and fire intensity. This index accurately reflects the release and accumulation of thermal energy during the fire process, providing crucial evidence for subsequent fire intensity assessment and natural language input feedback adjustments, achieving macroscopic energy conversion and comprehensive analysis of microscopic temperature data.

[0097] By capturing the temperature change trajectories of each unit block in the combustion field and comparing them with the ambient temperature, excess temperature is calculated to reflect the local energy accumulation of the unit blocks during the fire. This not only visually displays the distribution of temperature anomaly areas but also reflects the overall thermal energy situation throughout the entire fire combustion process by summarizing the global excess temperature. By utilizing air data such as air density and specific heat capacity to convert the temperature differences of each unit block into energy magnitudes and finally summarizing them into an overall excess temperature index, this method integrates massive amounts of temperature data into a physically meaningful energy evaluation parameter. It not only comprehensively describes the energy release and dissipation of a fire but also provides a direct quantitative reference for assessing fire hazard and destructive power. Therefore, it acts as a bridge connecting microscopic changes and macroscopic manifestations in the overall scheme, making fire analysis more systematic and comprehensive.

[0098] Step 5: Based on the average fire intensity index and the overall excess temperature, form the fire intensity index for this simulation, set the intensity threshold and correct it through development weights, obtain the deviation of the fire intensity index from the corrected intensity index, feed back the deviation, and execute the simulation based on the feedback results.

[0099] Step 5 includes the following:

[0100] The formula used to obtain the fire intensity index throughout the entire combustion process is as follows:

[0101]

[0102] in, This is an index indicating the intensity of the fire. This represents the total excess temperature of the combustion field sequence. This represents the average fire intensity index.

[0103] The fire intensity index is used to comprehensively reflect the combined performance of heat energy accumulation and fire intensity during a fire. It represents the total excess temperature of the combustion field sequence. In a real environment, it represents the accumulated heat energy in the entire combustion area where the temperature is higher than the ambient temperature. It can quantify the total effect of fire heat release and accumulation. This is the average fire intensity index, representing the average fire intensity extracted from the fire intensity index calculations at various times. It reflects the combined level of the geometric characteristics of the combustion field and the local combustion intensity. When the accumulated heat energy in the combustion zone ( An increase in this number indicates a larger number of abnormal temperature areas or a greater temperature difference, which in turn will... Increase; similarly, if the average fire intensity index... An increase indicates that the overall fire intensity in the combustion field is more intense, which also promotes... An increase in the fire intensity index will lead to an increase in the fire intensity index; conversely, a decrease in any component will result in a corresponding decrease in the fire intensity index, thereby reducing the overall assessment of fire hazard. This formula extracts the energy integral result from the microscopic temperature distribution (…). ) and dynamic geometric characteristics of fire ( This effectively integrates various methods to achieve a quantitative description of the intensity of a fire. It also serves as a feedback control basis for subsequent correction of the intensity threshold using trend words extracted from natural language, providing a scientific and intuitive evaluation indicator for fire simulation and emergency decision-making.

[0104] A severity threshold is set, and this threshold is then adjusted using development weights in natural language processing. The formula used is as follows:

[0105]

[0106] in, For the correction of the severity threshold, The threshold for severity, For development weight.

[0107] This represents the revised intensity threshold, which is an important critical value used to determine the intensity of a fire. This represents the original intensity threshold, which is based on historical data, experimental results, or preset standards and is used to measure the critical limit of fire intensity during fire development. This is a biased weight extracted from natural language, reflecting a biased correction factor for the development trend of a fire in the current fire scenario. When biased words such as "fierce" or "slightly large" appear in the natural language description, the corresponding biased weight will be adjusted. The value increases or decreases, thereby adjusting the original threshold accordingly. ), making the corrected severity threshold ( It is closer to the actual situation on site. A value greater than 1 indicates a high expected fire intensity, requiring a higher threshold to reflect the difference; conversely, when... A value less than 1 indicates a relatively mild fire development and a more moderate fire expectation, thus lowering the critical value. By correcting the original intensity threshold, dynamic response and precise control of fire scenarios are achieved, ensuring close coupling between the fire intensity index calculation and real-time data feedback, and providing scientific and effective technical support for fire emergency response and decision-making.

[0108] The deviation of the fire intensity index from the severity threshold is determined by the following formula:

[0109]

[0110] in, For deviation, This is an index indicating the intensity of the fire. For the adjusted severity threshold;

[0111] This indicates the degree of deviation of the fire intensity index from the correction threshold, used to quantify the deviation between the current fire simulation results and the preset fire intensity critical standard; The fire intensity index integrates the overall excess temperature of the combustion field with the average fire intensity index, reflecting the degree of combustion activity at the fire scene; It is the severity threshold after being corrected by the bias weight extracted from natural language, representing the critical value adjusted according to the current fire development trend. Greater than At that time, then ( A positive value indicates the degree of deviation. As the value increases, it indicates that the actual development of the fire has exceeded the revised critical value, suggesting a relatively serious fire situation; conversely, when... Below At that time, A negative or low value indicates a relatively stable or less intense fire, quantifying the difference between the actual simulation results and the expected intensity. This mechanism enhances the sensitivity of fire simulation and provides a scientific basis for on-site emergency control and decision-making guidance. It achieves an organic connection between quantitative assessment and actual control strategies, helping to adjust prevention and control measures in a timely manner during a fire.

[0112] The deviation is fed back to the natural language input user for confirmation. If the natural language input user modifies the input language, the deviation is obtained again for the modified full-process combustion simulation and fed back. When the natural language input user confirms that the simulation will be performed with the current deviation, the simulation will be displayed through VR.

[0113] As a preferred embodiment, displaying the CFD simulation process through VR is a common technical feature. For example, TechViz, as a professional virtual reality (VR) software, can enhance the effect of design review and demonstration through immersive technology and display the CFD simulation process, which will not be elaborated here.

[0114] The aforementioned data indicators are integrated to form a fire intensity index. This index is then multiplied by the overall excess temperature and the average fire intensity index to arrive at an overall fire hazard assessment. Simultaneously, the intensity threshold is dynamically corrected using trend-based words extracted from natural language. This step establishes a feedback loop by calculating the deviation between the fire intensity index and the corrected threshold, feeding real-time fire evolution data back to the natural language input user. This feedback mechanism not only captures potential deviations during fire development but also provides a basis for timely adjustments to simulation parameters and control strategies for the control system. Through this real-time feedback and dynamic correction, the simulation system's adaptability to complex fire scenarios is improved, as well as the real-time nature and relevance of fire prevention and emergency decision-making. This achieves closed-loop control within the overall simulation system, enhancing its safety and reliability.

[0115] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0116] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0117] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0118] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for simulating and generating job behaviors based on natural language, characterized in that, The specific steps include: Step 1: Construct a three-dimensional Cartesian coordinate system in the simulation environment and build a scene model at a 1:1 scale based on the real environment. Set the names of each component, divide the system into unit blocks according to the unit length of each axis in the three-dimensional Cartesian coordinate system, establish the connection relationship between each independent component based on the unit blocks, set the air data, and establish the simulation relationship to form a fire combustion model. Step 2: Input natural language into the LLM model for analysis, obtain ignition conditions, and input the ignition conditions into the fire combustion model. At the same time, form biased development weights based on the reference words in the natural language. Natural language is sent to the fire combustion model. The LLM model obtains the ignition conditions based on the natural language, where the ignition conditions are that a specified component is in a burning state. The ignition conditions are then input into the fire combustion model for simulation. The natural language is analyzed again to obtain the development weights. The logic is as follows: A tendency dictionary is set up, which contains five tendency words: fierce, slightly large, normal, slow, and weak. A reference word is set for each tendency word, and a weight is assigned to each tendency word. The reference words are words selected from the natural language database. The reference words are extracted from the natural language database using an LLM model and mapped to each tendency word to obtain the weight of the corresponding tendency word in the current natural language. The weight is the development weight of this simulation. If no reference word is extracted, the development weight of this simulation is considered to be the normal type development weight. Step 3: Obtain the combustion field at equal time intervals according to the simulation of the fire combustion model to form a combustion field sequence. Obtain the volume and shape factor of each combustion field in the combustion field sequence to form the fire intensity index surface area and number. Form a fire intensity index sequence according to the order of the combustion fields. Filter the fire intensity index sequence to obtain the average fire intensity index. Step 4: Obtain the trend of temperature change over time in each unit block of the combustion field, obtain the excess temperature of each unit block based on the difference between the unit block temperature and the ambient temperature, and combine the air data to obtain the overall excess temperature of all combustion fields in the combustion field sequence. Step 5: Based on the average fire intensity index and the overall excess temperature, form the fire intensity index for this simulation, set the intensity threshold and correct it through development weights, obtain the deviation of the fire intensity index from the corrected intensity index, feed back the deviation, and execute the simulation based on the feedback results.

2. The method for generating simulation of job behavior based on natural language according to claim 1, characterized in that: A three-dimensional Cartesian coordinate system is constructed in the CFD environment. The unit length of the three-dimensional Cartesian coordinate system is mapped to the real environment at a 1:1 ratio. Scene modeling simulating the real environment is carried out in the three-dimensional Cartesian coordinate system. The scene modeling uses three-dimensional modeling to describe the shape of each component and name it. The logic for determining the connection relationship between independent components is as follows: If two unit blocks belonging to two independent components are connected at these two unit blocks if the distance between the midpoints of their two nearest faces is less than 1 cm, then the two independent components are considered to be connected at these two unit blocks; otherwise, the two unit blocks are separate. This is used to determine the areas where all components are connected, and the heat transfer in the connected areas is carried out by heat conduction. Set the ambient air data, which includes density and specific heat capacity of air.

3. The method for generating simulations of job behaviors based on natural language according to claim 2, characterized in that: The heat conduction and heat radiation equations are set according to the connection relationship of each component, and the boundary conditions are set to form a fire combustion model.

4. The method for generating simulation of job behavior based on natural language according to claim 3, characterized in that: A full-process combustion simulation is performed in a CFD environment. This full-process combustion simulation refers to the intermediate process from the generation of ignition conditions to the determination by the fire combustion model that there is no open flame and the smoke concentration is less than 30%. The combustion field during the full-process combustion simulation is obtained, and the logic is as follows: The unit block containing the point with a temperature exceeding 600 degrees Celsius is defined as the combustion zone, and the collection of all combustion zones at the same time is defined as the combustion field; The combustion field during the entire combustion simulation process is acquired at equal time intervals and formed into a combustion field sequence according to time order; The volume of each combustion field is obtained, and the closed surface of the combustion field is generated using the MarchingCubes algorithm and its area is calculated. The maximum distance of each combustion field on the X, Y, and Z axes is obtained, and the shape factor of each combustion field is obtained using the following formula: in, Indicates the first The shape factor of the combustion field Indicates the first The volume of the combustion field, Indicates the first The maximum distance of each combustion field on the X-axis Indicates the first The maximum distance of each combustion field on the Y-axis Indicates the first The maximum distance of the combustion field on the Z-axis Retrieve variables for combustion field number. , , This represents the total number of combustion zones.

5. The method for generating simulation of job behavior based on natural language according to claim 4, characterized in that: The fire intensity index for each combustion zone is obtained using the following formula: in, Indicates the first The fire intensity index of each burning area Indicates the first The volume of the combustion field, Indicates the first The shape factor of the combustion field For the first The surface area of ​​the combustion field For coefficients, Retrieve variables for combustion field number. , , This represents the total number of combustion zones. The fire intensity index of each combustion field is sorted according to the combustion field sequence to form a fire intensity index sequence. The average fire intensity index is then obtained by filtering the fire intensity index sequence. The logic is as follows: Obtain the average value of the fire index in the fire index sequence, identify and remove combustion fields whose fire index is less than one-third of the average value, obtain the average value of the fire index of the remaining combustion fields, and label the obtained average value as the average fire index.

6. The method for generating simulation of job behavior based on natural language according to claim 5, characterized in that: The trend of temperature change over time within each unit block of the combustion field is obtained, along with the ambient temperature, to form the excess temperature of each unit block in the combustion field. The formula used is as follows: in, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The excess temperature of the unit block, Indicates the first The X-axis coordinate in each combustion field is Y-axis coordinate is Z-axis coordinates are The temperature of the unit block, For ambient temperature, Retrieve variables for X-axis coordinates. , , This represents the minimum value of the combustion field on the X-axis. This represents the maximum value of the combustion field on the X-axis. Retrieve variables for the Y-axis coordinate. , , This represents the minimum value of the combustion field on the Y-axis. This represents the maximum value of the combustion field on the Y-axis. Retrieve variables for Z-axis coordinates. , , This represents the minimum value of the combustion field on the Z-axis. This represents the maximum value of the combustion field on the Z-axis. Retrieve variables for combustion field number. , , This represents the total number of combustion zones. The overall excess temperature of the combustion field sequence is obtained using the following formula: in, This represents the total excess temperature of the combustion field sequence. air density, This is the specific heat capacity of air.

7. The method for generating simulation of job behavior based on natural language according to claim 6, characterized in that: The formula used to obtain the fire intensity index throughout the entire combustion process is as follows: in, This is an index indicating the intensity of the fire. This represents the total excess temperature of the combustion field sequence. This represents the average fire intensity index. A severity threshold is set, and this threshold is then adjusted using development weights in natural language processing. The formula used is as follows: in, For the correction of the severity threshold, The threshold for severity, For development weight.

8. The method for generating simulation of job behavior based on natural language according to claim 7, characterized in that: The deviation of the fire intensity index from the severity threshold is determined by the following formula: in, For deviation, This is an index indicating the intensity of the fire. For the adjusted severity threshold; The deviation is fed back to the natural language input user for confirmation. If the natural language input user modifies the input language, the deviation is obtained again for the modified full-process combustion simulation and fed back. When the natural language input user confirms that the simulation will be performed with the current deviation, the simulation will be displayed through VR.

Images