Urban rail engineering building energy-saving design method and system based on deep learning

Abstract

According to the urban rail engineering building energy-saving design method and system based on deep learning, a plurality of entrances and exits corresponding to each urban rail station serve as analysis targets, and a plurality of first building portraits are generated according to a plurality of first buildings corresponding to each first entrance and exit; therefore, multiple first entrance and exit portraits are obtained according to the multiple first building portraits, a first illumination control strategy is obtained in combination with the historical crowd travel rule of the first entrance and exit, and in addition, a first air conditioner control strategy is obtained according to the multiple first entrance and exit portraits and the clothes comparison condition when passengers get on and off the bus; and integrating the first illumination control strategy and the first air conditioner control strategy in the dimension of the time interval to obtain the first urban rail energy-saving adjustment strategy. According to the technical scheme, the lighting strategy and the air conditioner operation strategy in the urban rail station can better meet the actual requirements of users, and therefore the energy consumption of the urban rail station is effectively reduced.

Description

Technical Field

[0001] The present invention belongs to the field of energy-saving control technology, and in particular to an energy-saving design method and system for urban rail engineering buildings based on deep learning. Background Art

[0002] In the field of urban rail engineering construction, lighting equipment and air conditioning equipment are the main energy consumption factors. At present, during the operation stage of urban rail, the on and off of lighting equipment and the operating status of air conditioning equipment are fixed, which is somewhat mismatched with the actual lighting needs and air conditioning on and off needs. If the operation of lighting and air conditioning equipment can be adaptively adjusted according to the characteristics and rules of people's travel, it can not only meet the actual needs but also achieve the technical effect of energy saving and consumption reduction. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for energy-saving design of urban rail engineering buildings based on deep learning, so as to solve the technical problem that the lighting and air-conditioning equipment in urban rail stations cannot be adaptively adjusted in the prior art.

[0004] The urban rail engineering building energy-saving design method based on deep learning is characterized in that the method includes: S1: acquiring a plurality of first entrance and exit portraits corresponding to the first urban rail station according to a first correspondence between buildings around the first urban rail station and a plurality of first entrances and exits; The first entrance and exit portrait is used to represent the characteristic information of the travel group corresponding to the first entrance and exit; S2: determining a first lighting control strategy corresponding to each of the first entrances and exits according to the first historical crowd travel information and the first entrance and exit portraits; S3: Integrate the first plurality of import and export portraits to obtain a first station travel portrait; S4: identifying first clothing status information and second clothing status information of a plurality of first passengers within a second preset time interval, and obtaining a plurality of first comparison information according to the plurality of first clothing status information and the plurality of second clothing status information; The first clothing status information refers to the clothing information of the first passenger when waiting for the bus, and the second clothing status information refers to the clothing information of the first passenger after getting on the bus; S5: determining a first air conditioning control strategy according to the plurality of first comparison information and the first site travel portrait; S6: Obtain a first urban rail energy-saving adjustment strategy according to the first lighting control strategy and the first air-conditioning control strategy.

[0005] Preferably, S1 comprises the following sub-steps: S11: Acquire multiple first entrances and exits corresponding to the first urban rail station; S12: Taking each of the first entrances and exits as a reference, acquiring a plurality of first buildings within a first preset distance range according to a plurality of the first corresponding relationships; S13: determining a plurality of first building portraits according to a plurality of first historical behavior data corresponding to each of the first buildings; S14: Generate the first entrance and exit portrait according to the plurality of first building portraits.

[0006] Preferably, the S13 comprises the following sub-steps: S131: For each of the first buildings, obtaining a plurality of first historical behavior data within a first preset time interval; S132: Determine a first building portrait according to the first historical behavior data and the first travel pattern data.

[0007] Preferably, S2 comprises the following sub-steps: S21: Acquire first historical crowd travel information of each of the first entrances and exits within a first preset time interval; S22: establishing a first mapping relationship between the first building portrait corresponding to each of the first entrances and exits and the first historical crowd travel information, thereby obtaining a plurality of first mapping relationships; S23: inputting each of the first mapping relationships into a first group travel pattern prediction model one by one to obtain a plurality of first group travel pattern information; S24: Determine the first lighting control strategy according to the travel pattern information of the first group of people.

[0008] Preferably, S3 includes the following sub-steps: S31: extracting a plurality of first import and export label information from a plurality of the first import and export portraits; S32: Obtain a first label union for the plurality of first import and export label information; S33: Obtain the first site travel portrait according to the first label union.

[0009] Preferably, S4 includes the following sub-steps: S41: acquiring first clothing status information and second clothing status information of each of the first passengers within a second preset time interval; S42: For each of the first passengers, input the first clothing state information and the second clothing state information into a first clothing comparison model to obtain a plurality of first comparison information.

[0010] This application also proposes a deep learning-based energy-saving design system for urban rail engineering buildings, which is used to implement the above-mentioned deep learning-based energy-saving design method for urban rail engineering buildings.

[0011] The urban rail engineering building energy-saving design method and system based on deep learning proposed in this application takes the multiple first entrances and exits corresponding to each urban rail station as analysis targets, generates multiple first building portraits according to the multiple first buildings corresponding to each first entrance and exit, and obtains multiple first entrance and exit portraits according to the multiple first building portraits, and obtains the first lighting control strategy in combination with the historical travel patterns of the first entrance and exit. In addition, the first air-conditioning control strategy is obtained based on the multiple first entrance and exit portraits and the comparison of the clothing of passengers when getting on and off the train, so as to integrate the first lighting control strategy and the first air-conditioning control strategy in the dimension of time interval to obtain the first urban rail energy-saving adjustment strategy. Through the technical solution of this application, the lighting strategy and air-conditioning operation strategy in the urban rail station can be made more in line with the actual needs of users, thereby effectively reducing the energy consumption of the urban rail station. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0013] Figure 1 It is an execution flow chart of the energy-saving design method for urban rail engineering buildings based on deep learning of the present invention.

[0014] Figure 2 It is a schematic diagram of the correspondence between the first urban rail station, multiple first entrances and exits, and multiple first buildings in the present invention. DETAILED DESCRIPTION

[0015] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments, wherein the illustrative embodiments and descriptions are only used to explain the present invention but are not intended to limit the present invention.

[0017] The energy-saving design method and system for urban rail engineering buildings based on deep learning of the present invention are described in detail below.

[0018] This embodiment proposes a deep learning-based urban rail engineering building energy-saving design method, the method flow is as follows: Figure 1 As shown, the specific steps include: S1: According to a first correspondence between buildings surrounding a first urban rail station and a plurality of first entrances and exits, a plurality of first entrance and exit portraits corresponding to the first urban rail station are obtained.

[0019] Urban rail stations usually correspond to multiple entrances and exits, and each entrance and exit corresponds to multiple surrounding buildings. The characteristics of the active crowd corresponding to each surrounding building may be different. For example, according to the building type, it can be divided into office buildings, supermarkets or residential buildings. For each type of building, other dimensions can be divided. For example, for residential buildings, multiple dimensions can be divided according to the newness and value. Therefore, according to the multiple surrounding buildings corresponding to each of the first entrances and exits, the first entrance and exit portrait used to characterize the characteristics of the travel crowd is determined.

[0020] The first urban rail station, the first entrances and exits, and the first buildings are in a corresponding relationship with each other. For details, see Figure 2 .

[0021] The S1 comprises the following sub-steps: S11: Obtain multiple first entrances and exits corresponding to the first urban rail station.

[0022] The first entrance and exit can be obtained from the planning map of the first urban rail station.

[0023] S12: Taking each of the first entrances and exits as a reference, acquiring a plurality of first buildings within a first preset distance range according to a plurality of the first corresponding relationships.

[0024] In order to obtain the travel characteristics of the crowd corresponding to each of the first entrances and exits, it is necessary to obtain a plurality of first buildings within a preset distance range based on each of the first entrances and exits in this step.

[0025] Preferably, each of the first entrances and exits may be stored in a one-to-one correspondence with a plurality of first buildings within the first preset distance range in advance to obtain a plurality of the first corresponding relationships.

[0026] For example, if there are four buildings within a 3-kilometer radius of the first entrance and exit, namely office building 1, office building 2, residence 1 and residence 2, the above four buildings can be regarded as the multiple first buildings.

[0027] S13: Determine a plurality of first building portraits according to a plurality of first historical behavior data corresponding to each of the first buildings.

[0028] In order to determine the travel characteristics of the people corresponding to each of the first buildings, the first historical behavior data including those related to shared transportation or receiving takeout can be obtained, so as to determine the first building portrait for each of the first buildings, and then the first entrance and exit portraits can be finally obtained by integrating multiple first buildings corresponding to each of the first entrances and exits.

[0029] The S13 may include the following sub-steps: S131: For each of the first buildings, obtain a plurality of first historical behavior data within a first preset time interval.

[0030] This step is used to obtain the number of shared transportation orders or takeout delivery addresses signed for with the first building as the departure or destination within the first preset time interval.

[0031] Since the common means of transportation for workplace personnel to get to their offices from urban rail stations are shared transportation such as shared bicycles, and orders for takeout and the like are usually completed while the corresponding personnel are in the first building, the travel time pattern of personnel in the first building can be obtained through the above-mentioned first historical behavior data.

[0032] S132: Determine a first building portrait according to the first historical behavior data and the first travel pattern data.

[0033] Each piece of the first historical behavior data indicates that the person corresponding to the first building has arrived or left.

[0034] In order to further clarify the entry and exit actions, it is necessary to perform similarity matching between the plurality of the first historical behavior data and the first travel pattern data. The first travel pattern data refers to a general travel pattern. For example, for office buildings, arrival actions are generally generated in the morning and travel actions are generated in the afternoon, while for residential buildings, travel actions are generally generated in the morning and arrival actions are generated in the afternoon.

[0035] The first building portrait can clearly represent the entry and exit time and the number of people corresponding to the first building. For example, going to work at 8 am and 100 people can be used as two labels to generate the first building portrait.

[0036] In addition, since the first historical behavior data can also reflect the age group information of the personnel, the age group information can also be used as a part of the first building portrait.

[0037] S14: Generate the first entrance and exit portrait according to the plurality of first building portraits.

[0038] Since one entrance and exit of a city rail station may correspond to a plurality of the first buildings, the plurality of first building images corresponding to the one entrance and exit obtained in S13 may be integrated to generate the first entrance and exit image.

[0039] Preferably, a plurality of first labels may be extracted from a plurality of the first building portraits, and a union of the plurality of the first labels may be taken as the first entrance / exit portrait.

[0040] Preferably, a corresponding weight may be assigned to each of the first tags in the first import and export portrait according to the frequency of occurrence of each of the first tags, so as to increase the influence of the first tags.

[0041] S2: Determine a first lighting control strategy corresponding to each of the first entrances and exits based on the first historical crowd travel information and the plurality of first entrance and exit portraits.

[0042] Each of the first entrance and exit portraits can reflect the travel patterns of the people corresponding to the first entrance and exit to a certain extent. However, due to the influence of other factors, it is also necessary to adjust the predicted results of the travel patterns of the people based on the historical travel information of the people at the first entrance and exit sites.

[0043] The S2 comprises the following sub-steps: S21: Acquire first historical crowd travel information of each of the first entrances and exits within a first preset time interval.

[0044] Since the plurality of first building portraits are all established based on the travel data of people within the first preset time interval, the first historical crowd travel information should also use the travel data of people within the first preset time interval.

[0045] S22: Establishing a first mapping relationship between the first building portrait corresponding to each of the first entrances and exits and the first historical crowd travel information, thereby obtaining a plurality of first mapping relationships.

[0046] Since it is necessary to predict the travel pattern of the crowd at the first entrance and exit based on the first building portrait corresponding to each first entrance and exit and the first historical crowd travel information, the two need to be associated and stored in this step.

[0047] S23: Input each of the first mapping relationships into the first group travel pattern prediction model one by one to obtain a plurality of first group travel pattern information.

[0048] The first crowd travel pattern prediction model is obtained by training a convolutional neural network model based on historical data. The specific training process is as follows: First, the real historical data of known urban rail station entrances and exits are used as training sample data. For example, for each historical urban rail station entrance and exit, the architectural image and historical crowd travel information of the historical urban rail station entrance and exit in a preset time interval are used as input data, and the actual passenger flow data of the historical urban rail station entrance and exit in a future preset time interval is used as output data to train the convolutional neural network model.

[0049] Secondly, the trained model is tested using test samples. If the similarity between the output data and the actual data exceeds a preset value, the training of the first group travel pattern prediction model is completed.

[0050] The first group travel pattern information reflects the flow of people at each of the first entrances and exits. For example, the first building portrait may have a peak flow of 500 people from 7:00 to 8:00, and the first historical group travel information may have a peak flow of 7:05 to 8:10, and a flow of 700 people. Then, the first group travel pattern information can be obtained according to the first building portrait and the first historical group travel information in accordance with the specified calculation strategy and fitting rule.

[0051] S24: Determine the first lighting control strategy according to the travel pattern information of the first group of people.

[0052] The first lighting control strategy is formulated according to the travel pattern information of the first group of people.

[0053] Preferably, a preset correspondence can be established in advance between the first group travel regularity information and the number of lighting devices to be turned on, so that according to the flow of people and travel time interval information reflected in the first group travel regularity information, the specified number of lighting devices can be controlled to be turned on and off according to the travel time interval information. This improves the matching degree between the flow of people and the lighting devices, so as to achieve the technical effect of saving and reducing consumption.

[0054] S3: Integrate the multiple first import and export portraits to obtain the first station travel portrait.

[0055] Since the passenger flow data of the urban rail within the preset time interval is aggregated from multiple entrances and exits corresponding to an urban rail station, in order to formulate an air conditioning adjustment strategy on the urban rail, it is necessary to integrate multiple first entrance and exit portraits to obtain the first station travel portrait of the first urban rail station.

[0056] The S3 may include the following sub-steps: S31: extracting a plurality of first import and export label information from a plurality of the first import and export portraits.

[0057] For example, the first import and export label information may be that the peak traffic flow is from 7 to 8 o'clock or the traffic flow is 500 people, and the main age group is 25 to 35 years old.

[0058] S32: Obtain a first label union set from the plurality of first import and export label information.

[0059] The first label union may be to obtain the passenger flow of the first urban rail station at different times based on the first import and export label information, thereby obtaining passenger flow sequences in multiple time periods.

[0060] For example, during the time period from 7 to 8 o'clock, the passenger flow at the first entrance and exit is 500 and 200 respectively. Then, during this time period, the passenger flow at the first urban rail station is 700 people. Therefore, 700 people from 7 to 8 o'clock can be used as the label data in the first label set.

[0061] S33: Obtain the first site travel portrait according to the first label union.

[0062] The first site travel portrait is obtained by describing the passenger flow sequence in the multiple time periods.

[0063] S4: identifying first clothing status information and second clothing status information of a plurality of first passengers within a second preset time interval, and obtaining a plurality of first comparison information according to the plurality of first clothing status information and the plurality of second clothing status information.

[0064] In order to verify whether the operating status of the air conditioner matches the actual needs of the passengers, pattern recognition technology is needed in this step to compare the status information of the passengers while waiting and after getting on the bus.

[0065] The S4 comprises the following sub-steps: S41: within a second preset time interval, obtaining first clothing status information and second clothing status information of each of the first passengers.

[0066] The second preset time interval refers to a preset time interval before the time node at which the air conditioner operation state is to be adjusted.

[0067] Among them, each of the first passengers corresponds to a first clothing status information and a second clothing status information. The first clothing status information refers to the clothing information of the first passenger when waiting for the bus, and the second clothing status information refers to the clothing information of the first passenger after getting on the bus.

[0068] For example, for a first passenger, the first clothing status information may be that the passenger is wearing shorts and shirt while waiting for the bus, and the second clothing status information may be that the passenger is wearing longs and shirt after getting on the bus.

[0069] S42: For each of the first passengers, input the first clothing state information and the second clothing state information into a first clothing comparison model to obtain a plurality of first comparison information.

[0070] The first clothing comparison model is used to compare the clothing status of each of the first passengers while waiting for the bus and after getting on the bus, so as to further analyze and obtain the matching degree between the operating status of the air conditioner and the passenger.

[0071] The first clothing comparison model can preferably also be obtained by training a convolutional neural network model using historical sample data.

[0072] The first comparison information can be divided into two types: adding clothes and removing clothes. Adding clothes indicates that the operating temperature of the air conditioner is too low, while removing clothes indicates that the operating temperature of the air conditioner is too high.

[0073] S5: Determine a first air conditioning control strategy according to the first comparison information and the first site travel portrait.

[0074] Since people of different age groups have different preferences for air-conditioning operating status, based on the multiple first comparison information determined in S4, it is also necessary to combine the age group information of the travelers reflected in the travel portrait of the first site to generate the final first air-conditioning control strategy.

[0075] Preferably, the convolutional neural network model can be trained based on the historical sample information composed of the comparison information and the site travel portrait, so as to obtain the first air conditioning control strategy determination model. In the training process, the comparison information and the site travel portrait are used as input data, and the air conditioning control strategy composed of parameters such as the temperature and wind speed of the air conditioner is used as output data.

[0076] Thus, a plurality of the first comparison information and the first site travel portraits are input into the first air-conditioning control strategy determination model to obtain the first air-conditioning control strategy.

[0077] S6: Obtain a first urban rail energy-saving adjustment strategy according to the first lighting control strategy and the first air-conditioning control strategy.

[0078] In this step, the first lighting control strategy and the first air-conditioning control strategy corresponding to each time interval are integrated according to a specified time sequence to obtain the first urban rail energy-saving adjustment strategy.

[0079] Preferably, between 7 and 8 o'clock, the first urban rail energy-saving adjustment strategy may be to turn on 50% of the lighting equipment and 50% of the air conditioning.

[0080] This application also proposes a deep learning-based urban rail engineering building energy-saving design system, which is used to execute the above-mentioned deep learning-based urban rail engineering building energy-saving design method.

[0081] The urban rail engineering building energy-saving design method and system based on deep learning proposed in this application takes the multiple first entrances and exits corresponding to each urban rail station as analysis targets, generates multiple first building portraits according to the multiple first buildings corresponding to each first entrance and exit, and obtains multiple first entrance and exit portraits according to the multiple first building portraits, and obtains the first lighting control strategy in combination with the historical travel patterns of the first entrance and exit. In addition, the first air-conditioning control strategy is obtained based on the multiple first entrance and exit portraits and the comparison of the clothing of passengers when getting on and off the train, so as to integrate the first lighting control strategy and the first air-conditioning control strategy in the dimension of time interval to obtain the first urban rail energy-saving adjustment strategy. Through the technical solution of this application, the lighting strategy and air-conditioning operation strategy in the urban rail station can be made more in line with the actual needs of users, thereby effectively reducing the energy consumption of the urban rail station.

[0082] The above description is only a preferred embodiment of the present invention, so all equivalent changes or modifications made according to the structure, characteristics and principles described in the scope of the patent application of the present invention are included in the scope of the patent application of the present invention.

Claims

1. The energy-saving design method of urban rail engineering buildings based on deep learning is characterized by: The method includes: S1: acquiring a plurality of first entrance and exit portraits corresponding to the first urban rail station according to a first correspondence between buildings around the first urban rail station and a plurality of first entrances and exits; The first entrance and exit portrait is used to represent the characteristic information of the travel group corresponding to the first entrance and exit; S2: determining a first lighting control strategy corresponding to each of the first entrances and exits according to the first historical crowd travel information and the first entrance and exit portraits; S3: Integrate the first plurality of import and export portraits to obtain a first station travel portrait; S4: identifying first clothing status information and second clothing status information of a plurality of first passengers within a second preset time interval, and obtaining a plurality of first comparison information according to the plurality of first clothing status information and the plurality of second clothing status information; The first clothing status information refers to the clothing information of the first passenger when waiting for the bus, and the second clothing status information refers to the clothing information of the first passenger after getting on the bus; S5: determining a first air conditioning control strategy according to the plurality of first comparison information and the first site travel portrait; S6: Obtain a first urban rail energy-saving adjustment strategy according to the first lighting control strategy and the first air-conditioning control strategy.

2. The urban rail engineering building energy-saving design method based on deep learning according to claim 1 is characterized in that: The S1 comprises the following sub-steps: S11: Acquire multiple first entrances and exits corresponding to the first urban rail station; S12: Taking each of the first entrances and exits as a reference, acquiring a plurality of first buildings within a first preset distance range according to a plurality of the first corresponding relationships; S13: determining a plurality of first building portraits according to a plurality of first historical behavior data corresponding to each of the first buildings; S14: Generate the first entrance and exit portrait according to the plurality of first building portraits.

3. The urban rail engineering building energy-saving design method based on deep learning according to claim 2 is characterized in that: The S13 comprises the following sub-steps: S131: For each of the first buildings, obtaining a plurality of first historical behavior data within a first preset time interval; S132: Determine a first building portrait according to the first historical behavior data and the first travel pattern data.

4. The urban rail engineering building energy-saving design method based on deep learning according to claim 3 is characterized in that: The S2 comprises the following sub-steps: S21: Acquire first historical crowd travel information of each of the first entrances and exits within a first preset time interval; S22: establishing a first mapping relationship between the first building portrait corresponding to each of the first entrances and exits and the first historical crowd travel information, thereby obtaining a plurality of first mapping relationships; S23: inputting each of the first mapping relationships into a first group travel pattern prediction model one by one to obtain a plurality of first group travel pattern information; S24: Determine the first lighting control strategy according to the travel pattern information of the first group of people.

5. The urban rail engineering building energy-saving design method based on deep learning according to claim 4 is characterized in that: The S3 comprises the following sub-steps: S31: extracting a plurality of first import and export label information from a plurality of the first import and export portraits; S32: Obtain a first label union for the plurality of first import and export label information; S33: Obtain the first site travel portrait according to the first label union.

6. The urban rail engineering building energy-saving design method based on deep learning according to claim 5 is characterized in that: The S4 comprises the following sub-steps: S41: acquiring first clothing status information and second clothing status information of each of the first passengers within a second preset time interval; S42: For each of the first passengers, input the first clothing state information and the second clothing state information into a first clothing comparison model to obtain a plurality of first comparison information.

7. An urban rail engineering building energy-saving design system based on deep learning, used to implement the urban rail engineering building energy-saving design method based on deep learning in claims 1-6.

Images