Intelligent early warning method and system based on big data analysis

Through big data analysis and elevator dynamic safety floor selection strategy, early warning and automatic evacuation of elevator passengers is solved, and the problem of existing elevator early warning systems responding to fire lag is improved, improving the efficiency and safety of elevator safety management.

CN118894424AInactive Publication Date: 2024-11-05BEIJING SHICHUANG ELEVATOR ENG CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411188648.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-05
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When a fire occurs in an office building, the existing elevator warning system will be delayed when a fire occurs near the office building, causing the elevator to respond passively or suspend operations, causing greater losses and failing to ensure the safety of passengers in a timely manner.

Method used

Using intelligent early warning method based on big data analysis, the building and greening layout data, smoke sensor and temperature sensor data of the park management system are obtained, combined with geospatial technology and meteorological sensor data, the fire spread time is estimated, and dynamic safety is carried out through the elevator. Select a floor strategy, early warning and automatically transport passengers to evacuate.

Benefits of technology

Accurate and accurate evacuation of elevator passengers before the fire arrives, reducing the amount of follow-up rescue work, and improving the safety of park management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118894424A_ABST
    Figure CN118894424A_ABST
Patent Text Reader

Abstract

The invention provides an intelligent early warning method and system based on big data analysis, and the method comprises the steps: obtaining the building and greening layout data of a park management system, and combining the building combustion characteristic data to obtain the combustion load of a fire spreading region; according to the combustion load of the fire spreading area, the distance from the fire point to the elevator building, the wind speed and the wind direction data, the estimated remaining time of the fire reaching the elevator building is obtained; first time sequence data of a smoke sensor and first time sequence data of a temperature sensor at each floor monitoring point are monitored and loaded to a floor environment estimation model to obtain second time sequence data of the smoke sensor and the temperature sensor at each floor monitoring point, and a target floor where an elevator stops is obtained through an elevator dynamic safety floor selection strategy. According to dynamic stopping of the elevator stopping target floor, early warning of the elevator of the office building in advance when a fire occurs in the area close to the office building in the office park is achieved, people in the elevator are helped to evacuate before the fire arrives, and the park management safety is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of elevator safety management, and in particular to an intelligent early warning method and system based on big data analysis. Background Art

[0002] With the rapid development of urban construction, elevators have become an indispensable means of transportation in modern high-rise office buildings. The safety management of elevators in office parks is particularly important. Traditional elevator early warning methods and systems mainly provide early warnings based on passenger behavior and elevator failures. However, when a fire occurs in the vicinity of the office building where the elevator is located, the spread of the fire will pose a threat to the operational safety of elevators in surrounding buildings. Existing elevator early warning methods and systems mainly focus on early warning research when a fire occurs in the building itself. The early warning measures have a lag effect, and when the fire has reached the building itself, evacuation becomes more passive. A new intelligent elevator early warning method and system is urgently needed to achieve accurate and real-time monitoring and early warning of office building elevators in advance when a fire occurs in the vicinity of the office building in the office park, to ensure the safety of elevator passengers. Summary of the Invention

[0003] (1) Technical problems to be solved

[0004] The present invention provides an intelligent early warning method and system based on big data analysis. In the event of a fire in an area near an office building in an office park, a timely early warning is issued before the fire reaches the office building and elevator passengers are automatically transported to evacuate, avoiding the situation where the elevator passively responds or stops operating after the fire reaches the office building, causing greater losses.

[0005] (2) Technical solution

[0006] To achieve the above objectives, the present invention provides, on the one hand, an intelligent early warning method based on big data analysis, the method comprising:

[0007] Obtain building and greenery layout data from the park management system, match this building and greenery layout data with a preset combustible material information database to obtain building combustion characteristic data, calculate the combustion load of the fire spread area based on the building combustion characteristic data and the building and greenery layout data, obtain the distance from the fire point to the elevator building using geospatial technology, obtain wind speed and direction data from meteorological sensors in the park area, and calculate the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and direction data;

[0008] Obtain the first time series data of smoke sensors at monitoring points on each floor of the elevator shaft. When the data value in the first time series data of the smoke sensors at at least one monitoring point on the floor rises to a preset concentration threshold and shows an increasing trend, activate the first early warning mode of the elevator. Subtract the preset safe evacuation time threshold from the estimated remaining time for the fire to reach the elevator building to obtain the estimated remaining time until the elevator will no longer respond to regular floor call requests. Activate the elevator broadcast system to remind people in the elevator to pay attention to fire evacuation and announce the estimated remaining time until the elevator will no longer respond to regular floor call requests.

[0009] When the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the elevator's second early warning mode is activated, and the first time series data of the smoke sensor and the first time series data of the temperature sensor at each floor monitoring point are obtained and loaded into the floor environment prediction model to obtain the second time series data of the smoke sensor and the second time series data of the temperature sensor at each floor monitoring point;

[0010] According to the second time series data of the smoke sensors at the monitoring points on each floor, the second time series data of the temperature sensors at the monitoring points on each floor, and the estimated remaining time for the fire to reach the elevator building, the elevator dynamic safety floor selection strategy is used to obtain the elevator stop target floor. The elevator runs to the target floor and stops at the target floor.

[0011] Furthermore, the acquisition of building and greenery layout data from the park management system, matching the building and greenery layout data with a preset combustible material information database to obtain building combustion characteristic data, calculating the combustion load of the fire spread area based on the building combustion characteristic data and the building and greenery layout data, obtaining the distance from the fire point to the elevator building through geospatial technology, obtaining wind speed and direction data through meteorological sensors in the park area, and obtaining the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and direction data include:

[0012] Obtain the building and green layout data of the park management system to obtain the building type, building number, building area, and building height;

[0013] Match the building type with the preset combustible information database to obtain the building combustion calorific value;

[0014] According to the building area, number of buildings, building height, and building combustion heat value, the combustion load L of the fire spread area is:

[0015]

[0016] Among them A i is the area of ​​the i-th building, h iis the height of the i-th building, q i is the calorific value of the i-th building, n is the number of buildings, and L is the combustion load of the fire spread area;

[0017] The fire spread coefficient is obtained by matching the combustion load of the fire spread area with the preset fire spread coefficient list. The wind speed and direction data are obtained through the meteorological sensors in the park area. The fire spread rate R is obtained based on the fire spread coefficient, wind speed and wind direction data:

[0018] R = β·V·cosθ;

[0019] Where V is the wind speed, β is the fire spread coefficient, θ is the angle between the wind direction and the straight line from the fire point to the building, and R is the fire spread rate;

[0020] The distance from the fire point to the elevator building is obtained through geospatial technology. Based on the distance from the fire point to the elevator building and the fire spread rate, the estimated remaining time T for the fire to reach the elevator building is obtained as follows:

[0021]

[0022] Where D is the distance from the fire point to the elevator building, R is the fire spread rate, and T is the estimated remaining time for the fire to reach the elevator building.

[0023] Furthermore, the floor environment prediction model is a model established based on a sequence-to-sequence long short-term memory network to generate smoke sensor and temperature sensor prediction data based on the historical data of smoke sensors and temperature sensors at monitoring points on each floor;

[0024] The historical data of the smoke sensor and temperature sensor at each floor monitoring point is the first time series data of the smoke sensor at each floor monitoring point and the first time series data of the temperature sensor at each floor monitoring point; the predicted data of the smoke sensor and temperature sensor is the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point;

[0025] The method of obtaining the first time series data of the smoke sensor at each monitoring point on each floor and the first time series data of the temperature sensor at each monitoring point on each floor and loading them into the floor environment prediction model to obtain the second time series data of the smoke sensor at each monitoring point on each floor and the second time series data of the temperature sensor at each monitoring point on each floor includes:

[0026] Based on the time when the estimated remaining time for the fire to reach the elevator building is less than the preset safe evacuation time threshold, the first time series data of the smoke sensor and the first time series data of the temperature sensor at each floor monitoring point are obtained according to the preset historical time window and merged to obtain the input data of the floor environment estimation model;

[0027] The input data of the floor environment prediction model is processed by a sequence-to-sequence long short-term memory network encoder and passed to the decoder. The decoder uses the time when the elevator runs one floor as the timestamp interval and generates the second time series data of the smoke sensor and the second time series data of the temperature sensor at each floor monitoring point according to the preset prediction time window.

[0028] Furthermore, the method of obtaining the elevator stop target floor by using the elevator dynamic safety floor selection strategy based on the second time series data of the smoke sensor at the monitoring point on each floor, the second time series data of the temperature sensor at the monitoring point on each floor, and the estimated remaining time for the fire to reach the elevator building, and the elevator running to the target floor includes:

[0029] When the estimated remaining time for the fire to reach the elevator building is the preset safe evacuation time threshold, the elevator's current floor is obtained. The elevator arrival time t is calculated based on the elevator's current floor and the time it takes for the elevator to run one floor. n for:

[0030] t n =T+t(|An|);

[0031] Where T is the time when the fire reaches the building of the elevator and the estimated remaining time reaches the preset safe evacuation time threshold, t is the time for the elevator to run one floor, A is the floor where the elevator is located, n is the floor the elevator arrives at, t n The arrival time of the elevator from the current floor to the nth floor;

[0032] The elevator arrival time from the current floor to each floor is obtained based on the elevator arrival time from the current floor to the nth floor, and the smoke concentration data value and temperature data value of each floor at the elevator arrival time from the smoke sensor of the monitoring point of each floor and the second time series data of the temperature sensor of the monitoring point of each floor are obtained;

[0033] The floor environment index value is obtained by weighted summing the floor smoke concentration data value and the temperature data value of each floor at the elevator arrival time when the elevator arrives at the floor from the floor where the elevator is located according to the preset weight, and the floor whose environment index value is less than the preset environment index threshold is selected as the elevator stop candidate floor, and it is judged whether there is a floor among the elevator stop candidate floors. When there is a floor among the elevator stop candidate floors, a floor is selected as the elevator stop target floor. When there is no floor among the elevator stop candidate floors, the floor with the largest difference in floor number from the floor with the largest environment index value is selected as the elevator stop target floor, and the elevator runs according to the elevator stop target floor and stops at the target floor.

[0034] Furthermore, when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the elevator automatic control timing is started, and the starting time of the elevator automatic control timing is when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold;

[0035] After the elevator runs to the target floor and stops at the target floor, the method further includes:

[0036] When the elevator door is closed, obtain the elevator automatic control time, and subtract the elevator automatic control time from the safe evacuation time threshold to obtain the remaining automatic control time of the elevator;

[0037] The floors that are not included in the elevator stop candidate floors are obtained as the rescue floors. According to the elevator stop target floor, the time for the elevator to run one floor and the system preset elevator automatic door opening and closing time, the time t' for the elevator to run to each rescue floor and return to the elevator stop target floor is obtained. n for:

[0038] t′ n =2t(|nm|)+w;

[0039] Where t is the time it takes for the elevator to run one floor, m is the target floor the elevator stops at, n is the floor to be rescued, w is the system preset time for the elevator to automatically open and close the door, t' n It is the time it takes for the elevator to reach the nth rescue floor and return to the target floor where the elevator stops;

[0040] The time it takes for the elevator to run to each rescue floor and return to the target floor is compared with the remaining automatic control time of the elevator. The floors whose time it takes for the elevator to run to the rescue floor and return to the target floor is less than the remaining automatic control time are obtained as rescueable floors. The floor with the largest environmental index value among the rescueable floors is selected as the target rescue floor. After the elevator runs to the target rescue floor and picks up the passengers, it returns to the target floor.

[0041] When the elevator door is closed, the elevator automatic control time update data is obtained, and the elevator remaining automatic control time update data is obtained by subtracting the elevator automatic control time update data from the safe evacuation time threshold;

[0042] Get the remaining floors to be rescued, get the floors for which the time it takes for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is less than the elevator's remaining automatic control time update data, get the rescueable floors, run the elevator as described above, and update the elevator automatic control time again until the time it takes for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is greater than the elevator's remaining automatic control time update data, then stop the elevator and enter the fire control mode.

[0043] Furthermore, the sequence-to-sequence long short-term memory network training process includes:

[0044] Collect time series data from smoke and temperature sensors at building monitoring points in historical fire cases near the building. Use 70% of the data as a training set to train the model, and 30% of the data as a validation set to adjust parameters and monitor overfitting.

[0045] The first LSTM layer is used as an encoder to receive time series data from the smoke sensor and temperature sensor and generate a context vector to capture the key information in the input sequence data. The second LSTM layer is used as a decoder to receive the context vector generated by the encoder and gradually generate an output sequence. The output layer is a fully connected layer that converts the output of the second LSTM into predicted values ​​of smoke concentration and temperature.

[0046] The training set data is loaded into the model to calculate the predicted values ​​of smoke concentration and temperature through forward propagation. The loss between the predicted value and the actual value is calculated using the mean square error loss function. The model parameters are fitted by minimizing the loss function using the Adam optimizer.

[0047] The model was validated using validation set data, and the model parameters were adjusted based on the predicted value data performance of smoke concentration and temperature in the validation set to iteratively optimize the model.

[0048] Furthermore, after the second early warning mode of the elevator is activated when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the intelligent early warning method for the elevator based on big data analysis further includes:

[0049] Send fire warning information to designated elevator emergency management personnel through the communication system;

[0050] The smoke sensor concentration data in the elevator car is obtained, and the emergency lighting and ventilation systems are turned on when the smoke concentration in the elevator car exceeds the preset car smoke concentration threshold.

[0051] Based on the same inventive concept, the present invention also provides an intelligent early warning system based on big data analysis, the system comprising:

[0052] The fire sensing module is used to obtain building and green space layout data from the park management system, match the building and green space layout data with a preset combustible material information database to obtain building combustion characteristic data, calculate the combustion load of the fire spread area based on the building combustion characteristic data and the building and green space layout data, obtain the distance from the fire point to the elevator building through geospatial technology, obtain wind speed and direction data through meteorological sensors in the park area, and calculate the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and direction data;

[0053] A first warning module is configured to obtain first time series data from smoke sensors at monitoring points on each floor of the elevator shaft, and activate a first warning mode of the elevator when the data value in the first time series data of the smoke sensors at at least one monitoring point on the floor rises to a preset concentration threshold and shows an increasing trend. The module deducts a preset safe evacuation time threshold from the estimated remaining time for the fire to reach the elevator building to obtain an estimated remaining time until the elevator will no longer respond to regular floor call requests. The module activates the elevator broadcast system to remind people in the elevator to evacuate the fire and announces the estimated remaining time until the elevator will no longer respond to regular floor call requests.

[0054] The second early warning module is used to activate the second early warning mode of the elevator when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, obtain the first time series data of the smoke sensor and the first time series data of the temperature sensor at the monitoring point on each floor, and load them into the floor environment prediction model to obtain the second time series data of the smoke sensor and the second time series data of the temperature sensor at the monitoring point on each floor;

[0055] The dynamic adjustment module is used to obtain the elevator's target floor through the elevator's dynamic safety floor selection strategy based on the second time series data of the smoke sensor at the monitoring point on each floor, the second time series data of the temperature sensor at the monitoring point on each floor, and the estimated remaining time for the fire to reach the elevator building. The elevator runs to the target floor and stops at the target floor.

[0056] (3) Beneficial effects

[0057] The beneficial effects of the present invention are: by monitoring the fire situation in the area adjacent to the office building, the smoke concentration and temperature in the elevator are predicted based on big data analysis and combined with the floor environment prediction model, and the best floor is dynamically selected according to the prediction results combined with the elevator's dynamic safety floor selection strategy, helping people in the elevator to evacuate before the fire reaches, thereby reducing the subsequent rescue workload and promoting the improvement of park management safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0059] Figure 1 This is a flowchart of an intelligent early warning method based on big data analysis according to Example 1 of the present invention;

[0060] Figure 2This is a module block diagram of the intelligent early warning system based on big data analysis in Example 2 of the present invention. DETAILED DESCRIPTION

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0062] Example 1: Figure 1 As shown, this embodiment provides an intelligent early warning method based on big data analysis, the method comprising:

[0063] S1. Obtain the building and green layout data of the park management system, match the building and green layout data with the preset combustible material information database to obtain the building combustion characteristic data, calculate the combustion load of the fire spread area based on the building combustion characteristic data and the building and green layout data, obtain the distance from the fire point to the elevator building through geospatial technology, obtain wind speed and wind direction data through meteorological sensors in the area where the park is located, and obtain the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and wind direction data; the geospatial technology, such as the use of common geospatial technologies such as the global positioning system GPS, the geographic information system GIS, and the web map service WMS, can achieve positioning and measurement of the distance between target buildings.

[0064] S2. Obtain first time series data from smoke sensors at monitoring points on each floor of the elevator shaft. When the data value in the first time series data from the smoke sensors at at least one monitoring point on the floor rises to a preset concentration threshold and shows an increasing trend, activate the first early warning mode of the elevator. Subtract a preset safe evacuation time threshold from the estimated remaining time for the fire to reach the elevator building to obtain the estimated remaining time until the elevator will no longer respond to regular floor call requests. Activate the elevator broadcast system to alert occupants of the elevator to evacuate the fire and announce the estimated remaining time until the elevator will no longer respond to regular floor call requests.

[0065] S3. When the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the elevator's second early warning mode is activated, and the first time series data of the smoke sensor and the first time series data of the temperature sensor at each floor monitoring point are obtained and loaded into the floor environment estimation model to obtain the second time series data of the smoke sensor and the second time series data of the temperature sensor at each floor monitoring point;

[0066] S4. The elevator's target floor is obtained based on the second time series data of the smoke sensors at the monitoring points on each floor, the second time series data of the temperature sensors at the monitoring points on each floor, and the estimated remaining time for the fire to reach the elevator building through the elevator's dynamic safety floor selection strategy. The elevator runs to the target floor and stops at the target floor.

[0067] Furthermore, the acquisition of building and greenery layout data from the park management system, matching the building and greenery layout data with a preset combustible material information database to obtain building combustion characteristic data, calculating the combustion load of the fire spread area based on the building combustion characteristic data and the building and greenery layout data, obtaining the distance from the fire point to the elevator building through geospatial technology, obtaining wind speed and direction data through meteorological sensors in the park area, and obtaining the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and direction data include:

[0068] S11. Obtain building and greening layout data from the park management system to obtain building types, building numbers, building areas, and building heights;

[0069] S12, matching the building type with a preset combustible material information database to obtain the building combustion calorific value;

[0070] S13. Based on the building area, number of buildings, building height, and building combustion heat value, the fire spread area combustion load L is:

[0071]

[0072] Among them A i is the area of ​​the i-th building, h i is the height of the i-th building, q i is the calorific value of the i-th building, n is the number of buildings, and L is the combustion load of the fire spread area;

[0073] S14. Match the combustion load of the fire spread area with the preset fire spread coefficient list to obtain the fire spread coefficient. Obtain wind speed and wind direction data through the meteorological sensor in the area where the park is located. According to the fire spread coefficient, wind speed and wind direction data, the fire spread rate R is obtained as follows:

[0074] R = β·V·cosθ;

[0075] Where V is the wind speed, β is the fire spread coefficient, θ is the angle between the wind direction and the straight line from the fire point to the building, and R is the fire spread rate; for example, when the wind is facing the elevator building, θ = 0, cosθ = 1, and when the wind is facing away from the building, θ = 180, cosθ = -1;

[0076] S15. Obtain the distance from the fire point to the elevator building using geospatial technology. Based on the distance from the fire point to the elevator building and the fire spread rate, obtain the estimated remaining time T for the fire to reach the elevator building:

[0077]

[0078] Where D is the distance from the fire point to the elevator building, R is the fire spread rate, and T is the estimated remaining time for the fire to reach the elevator building.

[0079] For example, a fire breaks out in a restaurant near an office building where an elevator is located in the park. The building and greening layout data of the park management system are obtained, and the buildings in the area from the restaurant to the office building where the elevator is located include a restaurant, a lawn green belt, and a shrub green belt. The restaurant has an area of ​​200 m2 and a height of 3 m. The lawn green belt has an area of ​​500 m2 and a height of 0.1 m. The shrub green belt has an area of ​​300 m2 and a height of 1.2 m. The restaurant and green belt are matched with the preset combustible information database to obtain the restaurant's combustion calorific value of 18,000 kJ / m 3 The calorific value of lawn combustion is 9000kJ / m 3 The calorific value of shrub combustion is 15000kJ / m 3 , so the fire spread area combustion load L = 200 × 3 × 18000 + 500 × 0.1 × 9000 + 300 × 1.2 × 15000 = 16650000 kJ. Matching the fire spread area combustion load with the preset fire propagation coefficient list, the fire propagation coefficient is assumed to be 0.2. Assuming the current wind speed is 3m / s and the wind direction θ is 30 degrees, the fire spread rate R = 0.2 × 3 × cos 30 = 0.52m / s. Using geospatial technology, the distance from the restaurant to the elevator office building is 30m. The estimated remaining time for the fire to reach the elevator building is Second.

[0080] Furthermore, the floor environment prediction model is a model established based on a sequence-to-sequence long short-term memory network to generate smoke sensor and temperature sensor prediction data based on the historical data of smoke sensors and temperature sensors at monitoring points on each floor;

[0081] The historical data of the smoke sensor and temperature sensor at each floor monitoring point is the first time series data of the smoke sensor at each floor monitoring point and the first time series data of the temperature sensor at each floor monitoring point; the predicted data of the smoke sensor and temperature sensor is the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point;

[0082] The method of obtaining the first time series data of the smoke sensor at each monitoring point on each floor and the first time series data of the temperature sensor at each monitoring point on each floor and loading them into the floor environment prediction model to obtain the second time series data of the smoke sensor at each monitoring point on each floor and the second time series data of the temperature sensor at each monitoring point on each floor includes:

[0083] S31. Taking the time when the estimated remaining time for the fire to reach the elevator building is less than the preset safe evacuation time threshold as a reference, obtain the first time series data of the smoke sensor and the first time series data of the temperature sensor at each floor monitoring point according to the preset historical time window and merge them to obtain the input data of the floor environment estimation model;

[0084] For example, the first 30 seconds before the estimated remaining time for the fire to reach the elevator building is less than the preset safe evacuation time threshold is used as the historical time window to obtain the first time series data of the smoke sensor and the first time series data of the temperature sensor at each floor monitoring point and merge them to obtain the input data of the floor environment estimation model as follows:

[0085]

[0086] S32. Process the input data of the floor environment prediction model through a sequence-to-sequence long short-term memory network encoder and pass it to the decoder. The decoder uses the time when the elevator runs one floor as the timestamp interval and generates the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point according to the preset prediction time window.

[0087] For example, if the elevator takes 3 seconds to run one floor, and the remaining 30 seconds after the fire reaches the elevator building is less than the preset safe evacuation time threshold, the second time series data of the smoke sensor and the second time series data of the temperature sensor at each floor monitoring point are obtained as follows:

[0088]

[0089] Furthermore, the method of obtaining the elevator stop target floor by using the elevator dynamic safety floor selection strategy based on the second time series data of the smoke sensor at the monitoring point on each floor, the second time series data of the temperature sensor at the monitoring point on each floor, and the estimated remaining time for the fire to reach the elevator building, and the elevator running to the target floor includes:

[0090] S41. When the estimated remaining time for the fire to reach the elevator building reaches the preset safe evacuation time threshold, obtain the elevator's current floor, and calculate the elevator arrival time t from the current floor to the nth floor based on the elevator's current floor and the time it takes the elevator to run one floor. n for:

[0091] t n=T+t(|An|);

[0092] Where T is the time when the fire reaches the building of the elevator and the estimated remaining time reaches the preset safe evacuation time threshold, t is the time for the elevator to run one floor, A is the floor where the elevator is located, n is the floor the elevator arrives at, t n The arrival time of the elevator from the current floor to the nth floor;

[0093] For example, when the safe evacuation time threshold is set to 30 seconds, that is, when the estimated remaining time for the fire to reach the building of this elevator is 30 seconds, the elevator floor is obtained. Assuming that the elevator is on the eighth floor and the electromagnetic operation time for one floor is 3 seconds, the elevator arrival time at the seventh floor is t7 = T+3, the elevator arrival time at the sixth floor is t6 = T+6, and the time to reach the first floor is t1 = T+21.

[0094] S42. Obtain the elevator arrival time from the current floor to each floor based on the elevator arrival time from the current floor to the nth floor, and obtain the floor smoke concentration data value and temperature data value of each floor at the elevator arrival time from the second time series data of the smoke sensor at the monitoring point on each floor and the second time series data of the temperature sensor at the monitoring point on each floor;

[0095] For example, if the elevator arrives at the seventh floor at T+3, then the smoke concentration data value of the seventh floor at T+3 is obtained from the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point. The temperature data value of the seventh layer at T+3 is obtained The elevator arrives at the sixth floor at T+6. The smoke concentration data value of the sixth floor at T+6 is obtained from the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point. Get the sixth layer temperature data value at T+6 The elevator arrives at the first floor at T+21. The smoke concentration data value of the first floor at T+21 is obtained from the second time series data of the smoke sensor at each floor monitoring point and the second time series data of the temperature sensor at each floor monitoring point. The first layer temperature data value at T+21 is

[0096] S43. The floor environment index value is obtained by weighted summing the floor smoke concentration data value and the temperature data value of each floor at the time when the elevator arrives at the floor from which the elevator is currently located according to the preset weights. The floor whose environment index value is less than the preset environment index threshold is selected as the elevator stop candidate floor. It is determined whether there is a floor among the elevator stop candidate floors. When there is a floor among the elevator stop candidate floors, a floor is selected as the elevator stop target floor. When there is no floor among the elevator stop candidate floors, the floor with the largest difference in floor number from the floor with the largest environment index value is selected as the elevator stop target floor. The elevator runs according to the elevator stop target floor and stops at the target floor.

[0097] For example, if the smoke concentration data value for the seventh floor is 55 and the temperature data is 31 at the time the elevator arrives from the current floor, the weighted sum of the smoke concentration data and the temperature data using the preset weights of 0.6 and 0.4 yields an environmental index value of 55 × 0.6 + 31 × 0.4 for the seventh floor, i.e., an environmental index value of 45.4. If the preset environmental index threshold is 40, the seventh floor is determined to be a candidate floor for the elevator to stop. If there is one floor among the candidate floors for the elevator to stop, the first floor is selected as the target floor for the elevator to stop to facilitate evacuation. If there is no floor among the candidate floors for the elevator to stop, for example, if the third floor is predicted to be the floor most affected by the fire, the environmental index value of the first floor may be greater than the preset environmental index threshold due to smoke from the lower floor. At this time, transporting passengers to the first floor may result in limited vision and suffocation due to the thick smoke. In this case, the floor with the largest difference from the third floor should be selected as the target floor for the elevator to stop to increase the waiting time for rescue.

[0098] Furthermore, when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the elevator automatic control timing is started, and the starting time of the elevator automatic control timing is when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold;

[0099] After the elevator runs to the target floor and stops at the target floor, the method further includes:

[0100] S431. When the elevator door is closed, obtain the elevator automatic control time, and subtract the elevator automatic control time from the safe evacuation time threshold to obtain the remaining automatic control time of the elevator;

[0101] S432: Acquire floors that are not included in the elevator stop candidate floors as floors to be rescued, and obtain the time t' for the elevator to run to each floor to be rescued and return to the elevator stop target floor based on the elevator stop target floor, the time for the elevator to run one floor, and the system preset elevator automatic door opening and closing time. n for:

[0102] t' n =2t(|nm|)+w;

[0103] Where t is the time it takes for the elevator to run one floor, m is the target floor the elevator stops at, n is the floor to be rescued, w is the system preset time for the elevator to automatically open and close the door, t' n It is the time it takes for the elevator to reach the nth rescue floor and return to the target floor where the elevator stops;

[0104] S433: Compare the time it takes for the elevator to reach each rescue floor and return to the target floor with the remaining automatic control time of the elevator. Obtain the floors for which the time it takes for the elevator to reach the rescue floor and return to the target floor is less than the remaining automatic control time. Select the floor with the largest environmental index value among the rescueable floors as the target rescue floor. After the elevator reaches the target rescue floor and picks up passengers, it returns to the target floor.

[0105] S434. When the elevator door is closed, obtain the elevator automatic control time update data, and subtract the elevator automatic control time update data from the safe evacuation time threshold to obtain the elevator remaining automatic control time update data;

[0106] S435, obtain the remaining floors to be rescued, obtain the floors for which the time it takes for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is less than the updated data of the remaining automatic control time of the elevator, obtain the rescueable floors, and operate the elevator as described above and update the automatic control time of the elevator again until the time it takes for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is greater than the updated data of the remaining automatic control time of the elevator, then stop the elevator operation and enter the fire control mode.

[0107] When the fire reaches the elevator building and the estimated remaining time is the preset safe evacuation time threshold, the remaining time is very tight. However, there are still some cases where the remaining automatic control time of the elevator is relatively long after the passengers in the elevator are transported to the elevator's target floor. For example, when the fire reaches the elevator building and the estimated remaining time is the preset safe evacuation time threshold, the elevator is located on the second floor, and the elevator's target floor is the first floor obtained by the elevator dynamic safe floor selection strategy. Under the condition that the safe evacuation time threshold is 30 seconds, if the previous elevator runs one floor every 3 seconds, it only takes 3 seconds for the elevator to run from the second floor to the first floor. After the personnel get off the elevator, if the elevator door is closed, the automatic control time of the elevator is 20 seconds. At this time, if the third, fourth and fifth floors are the floors to be rescued, when the elevator door opening and closing time is 5 seconds, the time t3' of the elevator running to the third floor and returning to the first floor is 2× 3×(|3-1|)+5=17 seconds, the time it takes for the elevator to run to the fourth floor and return to the first floor is t4'=2×3×(|4-1|)+5=23 seconds, and the time it takes for the elevator to run to the fifth floor and return to the first floor is t'5=2×3×(|5-1|)+5=29 seconds. The elevator can use the remaining time to run to the third floor to transport the people on the third floor to the first floor, further helping people evacuate. After the elevator transports the people on the third floor to the first floor, the remaining automatic control time of the elevator is only a few seconds. At this time, the time it takes for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is greater than the remaining automatic control time of the elevator, that is, the elevator cannot transport the people on the remaining floors to be rescued to the first floor before the fire arrives. Since the circuits and other aspects of the building will be affected after a fire, the elevator may operate abnormally. Therefore, the elevator stops running and enters the fire control mode, waiting for firefighters to take over and cooperate with the subsequent arrangements of the fire.

[0108] Furthermore, the sequence-to-sequence long short-term memory network training process includes:

[0109] A1. Collect time series data from smoke and temperature sensors at building monitoring points in historical fire cases near the building. Use 70% of the data as a training set to train the model, and 30% of the data as a validation set to adjust parameters and monitor overfitting.

[0110] A2. Use the first LSTM layer as an encoder to receive time series data from the smoke sensor and temperature sensor and generate a context vector to capture key information in the input sequence data. Use the second LSTM layer as a decoder to receive the context vector generated by the encoder and gradually generate an output sequence. The output layer is a fully connected layer that converts the output of the second LSTM into predicted values ​​of smoke concentration and temperature.

[0111] A3. Load the training data into the model and calculate the predicted values ​​of smoke concentration and temperature through forward propagation. Calculate the loss between the predicted value and the actual value using the mean square error loss function. Fit the model parameters by minimizing the loss function using the Adam optimizer.

[0112] A4. Use the validation set data to verify the model. Based on the predicted smoke concentration and temperature data performance of the validation set, adjust the model parameters and iteratively optimize the model.

[0113] Furthermore, after the second early warning mode of the elevator is activated when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the intelligent early warning method for the elevator based on big data analysis further includes:

[0114] S31. Send fire warning information to designated elevator emergency management personnel through the communication system;

[0115] S32. Obtain smoke sensor concentration data in the elevator car, and turn on the emergency lighting and ventilation systems when the smoke concentration in the elevator car is greater than a preset car smoke concentration threshold.

[0116] Example 2: Based on the same inventive concept, Figure 2 As shown, this embodiment also provides an intelligent early warning system based on big data analysis, the system comprising:

[0117] The fire sensing module is used to obtain building and green space layout data from the park management system, match the building and green space layout data with a preset combustible material information database to obtain building combustion characteristic data, calculate the combustion load of the fire spread area based on the building combustion characteristic data and the building and green space layout data, obtain the distance from the fire point to the elevator building through geospatial technology, obtain wind speed and direction data through meteorological sensors in the park area, and calculate the estimated remaining time for the fire to reach the elevator building based on the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and direction data;

[0118] A first warning module is configured to obtain first time series data from smoke sensors at monitoring points on each floor of the elevator shaft, and activate a first warning mode of the elevator when the data value in the first time series data of the smoke sensors at at least one monitoring point on the floor rises to a preset concentration threshold and shows an increasing trend. The module deducts a preset safe evacuation time threshold from the estimated remaining time for the fire to reach the elevator building to obtain an estimated remaining time until the elevator will no longer respond to regular floor call requests. The module activates the elevator broadcast system to remind people in the elevator to evacuate the fire and announces the estimated remaining time until the elevator will no longer respond to regular floor call requests.

[0119] The second early warning module is used to activate the second early warning mode of the elevator when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, obtain the first time series data of the smoke sensor and the first time series data of the temperature sensor at the monitoring point on each floor, and load them into the floor environment prediction model to obtain the second time series data of the smoke sensor and the second time series data of the temperature sensor at the monitoring point on each floor;

[0120] The dynamic adjustment module is used to obtain the elevator's target floor through the elevator's dynamic safety floor selection strategy based on the second time series data of the smoke sensor at the monitoring point on each floor, the second time series data of the temperature sensor at the monitoring point on each floor, and the estimated remaining time for the fire to reach the elevator building. The elevator runs to the target floor and stops at the target floor.

[0121] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0122] Finally, it should be noted that although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent early warning method based on big data analysis, characterized in that: The method comprises: Obtain the building and green layout data of the park management system, match the building and green layout data with the preset combustible information database to obtain the building combustion characteristic data, calculate the fire spread area combustion load based on the building combustion characteristic data and the building and green layout data, obtain the distance from the fire point to the elevator building through geospatial technology, obtain the wind speed and wind direction data through the meteorological sensor in the area where the park is located, and obtain the estimated remaining time for the fire to reach the elevator building based on the fire spread area combustion load, the distance from the fire point to the elevator building, and the wind speed and wind direction data; The first time series data of smoke sensors arranged at monitoring points on each floor in the elevator shaft are obtained. When the data value in the first time series data of smoke sensors at at least one monitoring point on the floor rises to a preset concentration threshold and shows an increasing trend, the first early warning mode of the elevator is started. The estimated remaining time before the elevator will no longer respond to regular floor call requests is subtracted from the preset safe evacuation time threshold by the estimated remaining time before the fire reaches the elevator building. The elevator broadcast system is started to remind people in the elevator to pay attention to fire evacuation and broadcast the estimated remaining time before the elevator will no longer respond to regular floor call requests. When the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the second early warning mode of the elevator is started, the first time series data of the smoke sensor at the monitoring point of each floor and the first time series data of the temperature sensor at the monitoring point of each floor are obtained and loaded into the floor environment estimation model to obtain the second time series data of the smoke sensor at the monitoring point of each floor and the second time series data of the temperature sensor at the monitoring point of each floor; According to the second time series data of the smoke sensors at the monitoring points on each floor, the second time series data of the temperature sensors at the monitoring points on each floor and the estimated remaining time for the fire to reach the elevator building, the elevator dynamic safety floor selection strategy is used to obtain the elevator stop target floor, and the elevator runs to the target floor.

2. The intelligent early warning method based on big data analysis according to claim 1 is characterized in that: The acquisition of the building and greening layout data of the park management system, matching the building and greening layout data with the preset combustible information database to obtain the building combustion characteristic data, calculating the fire spread area combustion load according to the building combustion characteristic data and the building and greening layout data, obtaining the distance from the fire point to the elevator building through geographic space technology, obtaining the wind speed and wind direction data through the meteorological sensor in the area where the park is located, and obtaining the estimated remaining time for the fire to reach the elevator building according to the fire spread area combustion load, the distance from the fire point to the elevator building, the wind speed and wind direction data include: Obtain the building and green layout data of the park management system to obtain the building type, building quantity, building area, and building height; Match the building type with the preset combustible information database to obtain the building combustion calorific value; According to the building area, number of buildings, building height, and building combustion heat value, the combustion load L of the fire spread area is: Among them A i is the area of ​​the ith building, h i is the height of the i-th building, q i is the calorific value of the i-th building, n is the number of buildings, and L is the combustion load of the fire spread area; The fire spread coefficient is obtained by matching the combustion load in the fire spread area with the preset fire spread coefficient list. The wind speed and wind direction data are obtained through the meteorological sensors in the area where the park is located. The fire spread rate R is obtained based on the fire spread coefficient, wind speed and wind direction data: R = β·V·cosθ; Where V is the wind speed, β is the fire spread coefficient, θ is the angle between the wind direction and the straight line from the fire point to the building, and R is the fire spread rate; The distance from the fire point to the elevator building is obtained through geospatial technology. According to the distance from the fire point to the elevator building and the fire spread rate, the estimated remaining time T for the fire to reach the elevator building is obtained as follows: Where D is the distance from the fire point to the elevator building, R is the fire spread rate, and T is the estimated remaining time for the fire to reach the elevator building.

3. The intelligent early warning method based on big data analysis according to claim 1 is characterized in that: The floor environment prediction model is a model established according to a sequence-to-sequence long short-term memory network to generate smoke sensor and temperature sensor prediction data based on the historical data of smoke sensors and temperature sensors at monitoring points on each floor; The historical data of the smoke sensor and temperature sensor at the monitoring point of each floor are the first time series data of the smoke sensor at the monitoring point of each floor and the first time series data of the temperature sensor at the monitoring point of each floor, and the predicted data of the smoke sensor and temperature sensor are the second time series data of the smoke sensor at the monitoring point of each floor and the second time series data of the temperature sensor at the monitoring point of each floor; The method of obtaining the first time series data of the smoke sensor at each monitoring point on each floor and the first time series data of the temperature sensor at each monitoring point on each floor and loading them into the floor environment prediction model to obtain the second time series data of the smoke sensor at each monitoring point on each floor and the second time series data of the temperature sensor at each monitoring point on each floor includes: Taking the time when the estimated remaining time for the fire to reach the elevator building is less than the preset safe evacuation time threshold as the benchmark, the first time series data of the smoke sensor at each floor monitoring point and the first time series data of the temperature sensor at each floor monitoring point are obtained according to the preset historical time window and merged to obtain the input data of the floor environment estimation model; The input data of the floor environment prediction model is processed by a sequence-to-sequence long short-term memory network encoder and passed to the decoder. The decoder uses the time when the elevator runs one floor as the timestamp interval and generates the second time series data of the smoke sensors at the monitoring points on each floor and the second time series data of the temperature sensors at the monitoring points on each floor according to the preset prediction time window.

4. The intelligent early warning method based on big data analysis according to claim 1 is characterized in that: The elevator stops at the target floor by using the elevator dynamic safety floor selection strategy according to the second time series data of the smoke sensor at each floor monitoring point, the second time series data of the temperature sensor at each floor monitoring point and the estimated remaining time for the fire to reach the elevator building, and the elevator runs to the target floor according to the elevator stop target floor, including: When the estimated remaining time for the fire to reach the elevator building is the preset safe evacuation time threshold, the elevator floor is obtained, and the elevator arrival time t from the current elevator floor to the nth floor is calculated based on the elevator floor and the time the elevator runs for one floor. n for: t n =T+t(|A-n|); Where T is the time when the fire reaches the building of the elevator and the estimated remaining time reaches the preset safe evacuation time threshold, t is the time for the elevator to run one floor, A is the floor where the elevator is located, n is the floor where the elevator arrives, and t n is the arrival time of the elevator from the current floor to the nth floor; The elevator arrival time of the elevator from the current floor to each floor is obtained according to the elevator arrival time of the elevator from the current floor to the nth floor, and the floor smoke concentration data value and temperature data value of each floor at the elevator arrival time when the elevator arrives at the current floor from the second time series data of the smoke sensor at the monitoring point of each floor and the second time series data of the temperature sensor at the monitoring point of each floor are obtained; The floor environment index value is obtained by weighted summing the floor smoke concentration data value and the temperature data value according to the preset weights according to the floor smoke concentration data value and the temperature data value when the elevator arrives at the floor from the current floor. The floor whose environment index value is less than the preset environment index threshold value is selected as the elevator stop candidate floor. It is judged whether there is a floor among the elevator stop candidate floors. When there is a floor among the elevator stop candidate floors, a floor is selected as the elevator stop target floor. When there is no floor among the elevator stop candidate floors, the floor with the largest difference in floor number from the floor with the largest environment index value is selected as the elevator stop target floor. The elevator runs according to the elevator stop target floor and stops at the target floor.

5. The intelligent early warning method based on big data analysis according to claim 4 is characterized in that: When the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the elevator automatic control timing is started, and the starting time of the elevator automatic control timing is when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold; After the elevator stops at the target floor according to the elevator stop target floor, the method further includes: When the elevator door is closed, the elevator automatic control duration is obtained, and the remaining automatic control time of the elevator is obtained by subtracting the elevator automatic control duration from the safe evacuation time threshold; The floors that are not included in the elevator stop candidate floors are obtained as the rescue floors. According to the elevator stop target floor, the time for the elevator to run one floor and the system preset elevator automatic door opening and closing time, the time t' for the elevator to run to each rescue floor and return to the elevator stop target floor n for: t' n =2t(|n-m|)+w; Where t is the time for the elevator to run one floor, m is the target floor for the elevator to stop, n is the floor to be rescued, w is the system preset time for the elevator to automatically open and close the door, t' n It is the time it takes for the elevator to run to the nth rescue floor and return to the target floor where the elevator stops; The time it takes for the elevator to run to each rescue floor and return to the target floor is compared with the remaining automatic control time of the elevator. The floors whose time it takes for the elevator to run to the rescue floor and return to the target floor are less than the remaining automatic control time of the elevator are obtained as rescueable floors. The floor with the largest environmental index value among the rescueable floors is selected as the target rescue floor. After the elevator runs to the target rescue floor and picks up people, it returns to the target floor where the elevator stops. When the elevator door is closed, the elevator automatic control duration update data is obtained, and the elevator remaining automatic control time update data is obtained by subtracting the elevator automatic control duration update data from the safe evacuation time threshold; Get the remaining floors to be rescued, get the floors for which the time for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is less than the remaining automatic control time update data of the elevator, get the rescueable floors, and run the elevator as mentioned above and update the elevator automatic control duration again until the time for the elevator to run to the remaining floors to be rescued and return to the target floor to be stopped is greater than the remaining automatic control time update data of the elevator, then stop the elevator operation and enter the fire control mode.

6. The intelligent early warning method based on big data analysis according to claim 3 is characterized in that: The sequence-to-sequence long short-term memory network training process includes: Collect the time series data of smoke sensors and temperature sensors at the monitoring points of buildings in historical cases of fires in the vicinity of buildings, use 70% of the data as a training set to train the model, and use 30% of the data as a validation set to adjust parameters and monitor overfitting; The first LSTM layer is used as an encoder to receive the time series data of the smoke sensor and the temperature sensor and generate a context vector to capture the key information in the input sequence data. The second LSTM layer is used as a decoder to receive the context vector generated by the encoder and gradually generate an output sequence. The output layer is a fully connected layer used to convert the output of the second LSTM into the predicted values ​​of smoke concentration and temperature. The training set data is loaded into the model to calculate the predicted values ​​of smoke concentration and temperature through forward propagation, the loss between the predicted value and the actual value is calculated through the mean square error loss function, and the model parameters are fitted by minimizing the loss function through the Adam optimizer; The model was verified using the validation set data, and the model parameters were adjusted based on the predicted value data performance of the validation set smoke concentration and temperature to iteratively optimize the model.

7. The intelligent early warning method based on big data analysis according to claim 1 is characterized in that: After the second early warning mode of the elevator is started when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, the intelligent early warning method for the elevator by big data analysis also includes: Send fire warning information to designated elevator emergency management personnel through the communication system; The smoke sensor concentration data in the elevator car is obtained, and the emergency lighting and ventilation systems are turned on when the smoke concentration in the elevator car is greater than the preset car smoke concentration threshold.

8. Intelligent early warning system based on big data analysis, characterized by: The system comprises: The fire sensing module is used to obtain the building and green layout data of the park management system, match the building and green layout data with the preset combustible information database to obtain the building combustion characteristic data, calculate the combustion load of the fire spread area according to the building combustion characteristic data and the building and green layout data, obtain the distance from the fire point to the elevator building through geographic space technology, obtain the wind speed and wind direction data through the meteorological sensor in the area where the park is located, and obtain the estimated remaining time for the fire to reach the elevator building according to the combustion load of the fire spread area, the distance from the fire point to the elevator building, and the wind speed and wind direction data; The first warning module is used to obtain the first time series data of the smoke sensors arranged at the monitoring points on each floor in the elevator shaft, and start the first warning mode of the elevator when the data value in the first time series data of the smoke sensors at the monitoring points on at least one floor rises to the preset concentration threshold and shows an increasing trend, and the estimated remaining time before the elevator will no longer respond to the conventional floor call request is obtained by subtracting the preset safe evacuation time threshold from the estimated remaining time when the fire reaches the elevator building, and the elevator broadcast system is started to remind the people in the elevator to pay attention to the fire evacuation and broadcast the estimated remaining time before the elevator will no longer respond to the conventional floor call request; The second early warning module is used to start the second early warning mode of the elevator when the estimated remaining time for the fire to reach the elevator building is equal to the preset safe evacuation time threshold, obtain the first time series data of the smoke sensor at the monitoring point of each floor and the first time series data of the temperature sensor at the monitoring point of each floor and load them into the floor environment estimation model to obtain the second time series data of the smoke sensor at the monitoring point of each floor and the second time series data of the temperature sensor at the monitoring point of each floor; The dynamic adjustment module is used to obtain the elevator stop target floor through the elevator dynamic safety floor selection strategy according to the second time series data of the smoke sensor at the monitoring point of each floor, the second time series data of the temperature sensor at the monitoring point of each floor and the estimated remaining time for the fire to reach the elevator building. The elevator runs according to the elevator stop target floor and stops at the target floor.